luke
/
iverilog
mirror of https://github.com/steveicarus/iverilog.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
iverilog/vpi/sys_display.c

2719 lines
88 KiB
C
Raw Normal View History

Convert vvm to implement system tasks with vpi.
1999-08-15 03:23:56 +02:00
/*
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
* Copyright (c) 1999-2025 Stephen Williams (steve@icarus.com)
Convert vvm to implement system tasks with vpi.
1999-08-15 03:23:56 +02:00
*
* This source code is free software; you can redistribute it
* and/or modify it in source code form under the terms of the GNU
* General Public License as published by the Free Software
* Foundation; either version 2 of the License, or (at your option)
* any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
updated FSF-address
2012-08-29 03:41:23 +02:00
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
Convert vvm to implement system tasks with vpi.
1999-08-15 03:23:56 +02:00
*/
scanf support for real values.
2006-08-12 05:38:12 +02:00
# include "sys_priv.h"
Convert vvm to implement system tasks with vpi.
1999-08-15 03:23:56 +02:00
# include <assert.h>
# include <string.h>
Add full $ferror() support, bug fixes, etc. This patch adds $ferror() support. This is enhanced over what the standard says needs to be supported. All file I/O routines set/clear the errno flag, This allows $ferror() to check when an invalid fd/MCD is passed to a system task. The $ferror() fd is only checked if errno is not set since all the other routines will set EBADF when they are called. This allows us to test the output from $fopen() which sets the fd to zero (an invalid fd) when it has a problem. The following bugs (changes) are also fixed: Fix $fdisplay*, $fwrite*, $fstrobe*, $fflush, $fclose to not accept 0 as a valid MCD. For the $fdisplay*, $fwrite* and $fstrobe tasks an invalid fd/MCD is now a warning instead of an error like the other file I/O routines. Modify $fputc() to be a function and return a value like $ungetc(). Modify $ungetc() to return a value like the standard dictates (-1 on error else 0) and pass the value as a full int to get EOF as an error. Remove some extra checks that could not happen in $fseek() and to be safe convert the numeric codes to the system symbolic values. Skip $fstrobe callback when the fd/MCD is closed. The standard says we need to cancel the callback, but we currently do not have the ability to do that so just skipping this will work for now. Free the argument handle when $fscanf() is called with an invalid fd (memory leak).
2009-07-31 22:06:49 +02:00
# include <errno.h>
Convert vvm to implement system tasks with vpi.
1999-08-15 03:23:56 +02:00
# include <ctype.h>
Add vpi_fopen and vpi_get_file.
2003-05-23 06:04:02 +02:00
# include <stdio.h>
Convert vvm to implement system tasks with vpi.
1999-08-15 03:23:56 +02:00
# include <stdlib.h>
Format real values as time.
2003-02-06 18:40:02 +01:00
# include <math.h>
Make ivl_alloc.h the last include so it doesn't effect any system includes. In ivl_alloc.h we redefine malloc(), realloc() and calloc() to have standard error checking. We don't want to do this for anything that comes from the standard headers. This specifically doesn't work if a C++ header files does std::malloc, etc. Also change to -W instead of -Wextra since that is more portable. I plan to add a check from -Wextra and use it when available since it is more descriptive.
2010-10-24 00:52:56 +02:00
# include "ivl_alloc.h"
Convert vvm to implement system tasks with vpi.
1999-08-15 03:23:56 +02:00
Add -compatible command line option to vvp. This option is intended to make it easier to compare results from Icarus with results from other simulators. For now, the only effect it has is to change the default format for displaying real numbers when no format string is supplied.
2013-10-20 23:59:53 +02:00
// Flag to enable better compatibility with other simulators
static unsigned compatible_flag = 0;
static void check_command_line_args(void)
{
struct t_vpi_vlog_info vlog_info;
vpi_get_vlog_info(&vlog_info);
Clean up some more sign-compare issues The only known problems left are in files imported from gtkwave, if not for them you could turn on -Wsign-compare. Assumes c99 for c code, so the scope of for-loop indexes can be made sane.
2014-06-10 20:24:03 +02:00
for (int idx = 0 ; idx < vlog_info.argc ; idx += 1) {
Add -compatible command line option to vvp. This option is intended to make it easier to compare results from Icarus with results from other simulators. For now, the only effect it has is to change the default format for displaying real numbers when no format string is supplied.
2013-10-20 23:59:53 +02:00
if (strcmp(vlog_info.argv[idx],"-compatible") == 0) {
compatible_flag = 1;
}
}
}
Add vpi_fopen and vpi_get_file.
2003-05-23 06:04:02 +02:00
/* Printf wrapper to handle both MCD/FD */
static PLI_INT32 my_mcd_printf(PLI_UINT32 mcd, const char *fmt, ...)
{
int r = 0;
va_list ap;
va_start(ap, fmt);
if (IS_MCD(mcd)) {
r = vpi_mcd_vprintf(mcd, fmt, ap);
} else {
FILE *fp = vpi_get_file(mcd);
if (fp) r = vfprintf(fp, fmt, ap);
}
va_end(ap);
return r;
}
Use fwrite to write $display output, instead of fprintf This change makes it safe to write non-ascii characters, which is an issue when the %u format is used.
2015-06-03 19:00:35 +02:00
static void my_mcd_rawwrite(PLI_UINT32 mcd, const char*buf, size_t count)
{
if (IS_MCD(mcd)) {
vpip_mcd_rawwrite(mcd, buf, count);
} else {
FILE*fp = vpi_get_file(mcd);
if (fp) {
while (count > 0) {
size_t rc = fwrite(buf, 1, count, fp);
Fix some cppcheck warnings
2015-10-02 18:42:36 +02:00
if (rc == 0) break;
Use fwrite to write $display output, instead of fprintf This change makes it safe to write non-ascii characters, which is an issue when the %u format is used.
2015-06-03 19:00:35 +02:00
count -= rc;
buf += rc;
}
}
}
}
Add support for $timeformat.
2002-05-31 06:26:54 +02:00
struct timeformat_info_s timeformat_info = { 0, 0, 0, 20 };
Unify display and strobe format routines.
1999-11-07 00:32:14 +01:00
struct strobe_cb_info {
Add possibility of const-correctness Follow-up to "Brainless start to const-correct changes" Still actually does nothing, but now if the #define ICARUS_VPI_CONST in vpi_user.h is changed to const, Icarus is almost const-correct, as checked with gcc flags -Wcast-qual -Wwrite-strings. Choosing when to #define ICARUS_VPI_CONST const is left as an exercise for the reader. With these two patches applied, and the const define, there are still about a dozen const problems left.
2010-10-04 18:53:55 +02:00
const char*name;
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
char*filename;
int lineno;
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
int default_format;
Scope information is needed by all types of display tasks.
2000-11-04 02:52:57 +01:00
vpiHandle scope;
Unify display and strobe format routines.
1999-11-07 00:32:14 +01:00
vpiHandle*items;
unsigned nitems;
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
unsigned fd_mcd;
Unify display and strobe format routines.
1999-11-07 00:32:14 +01:00
};
Convert vvm to implement system tasks with vpi.
1999-08-15 03:23:56 +02:00
Content-free portability fixes. These are patches from ITOH Yasufumi and Larry Doolittle of pr2120948. They change nothing, but make the code compile with older compilers.
2008-09-30 03:06:47 +02:00
/*
* The number of decimal digits needed to represent a
* nr_bits binary number is floor(nr_bits*log_10(2))+1,
* where log_10(2) = 0.30102999566398.... and I approximate
* this transcendental number as 146/485, to avoid the vagaries
* of floating-point. The smallest nr_bits for which this
* approximation fails is 2621,
* 2621*log_10(2)=789.9996, but (2621*146+484)/485=790 (exactly).
* In cases like this, all that happens is we allocate one
* unneeded char for the output. I add a "L" suffix to 146
* to make sure the computation is done as long ints, otherwise
* on a 16-bit int machine (allowed by ISO C) we would mangle
* this computation for bit-length of 224. I'd like to put
* in a test for nr_bits < LONG_MAX/146, but don't know how
* to fail, other than crashing.
*
* In an April 2000 thread in comp.unix.programmer, with subject
* "integer -> string", I <LRDoolittle@lbl.gov> give the 28/93
* approximation, but overstate its accuracy: that version first
* fails when the number of bits is 289, not 671.
*
* This result does not include space for a trailing '\0', if any.
*/
__inline__ static int calc_dec_size(int nr_bits, int is_signed)
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
{
Better calcuation of dec string width (Larry Doolittle)
2002-01-22 01:18:10 +01:00
int r;
if (is_signed) --nr_bits;
r = (nr_bits * 146L + 484) / 485;
if (is_signed) ++r;
return r;
}
static int vpi_get_dec_size(vpiHandle item)
{
return calc_dec_size(
vpi_get(vpiSize, item),
vpi_get(vpiSigned, item)==1
);
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
}
Improved automatic variable error reporting. This patch moves the checks for automatic variables being passed to the $monitor or $strobe system tasks from the calltf routine to a new compiletf routine. It also adds the file name and line number of an offending call to the error message.
2008-11-15 21:33:23 +01:00
static void array_from_iterator(struct strobe_cb_info*info, vpiHandle argv)
Convert vvm to implement system tasks with vpi.
1999-08-15 03:23:56 +02:00
{
$display can take 0 arguments.
2001-03-18 01:31:32 +01:00
if (argv) {
vpiHandle item;
unsigned nitems = 1;
vpiHandle*items = malloc(sizeof(vpiHandle));
items[0] = vpi_scan(argv);
Fix display of no arguments.
2002-07-23 04:41:15 +02:00
if (items[0] == 0) {
free(items);
info->nitems = 0;
info->items = 0;
Improved automatic variable error reporting. This patch moves the checks for automatic variables being passed to the $monitor or $strobe system tasks from the calltf routine to a new compiletf routine. It also adds the file name and line number of an offending call to the error message.
2008-11-15 21:33:23 +01:00
return;
Fix display of no arguments.
2002-07-23 04:41:15 +02:00
}
$display can take 0 arguments.
2001-03-18 01:31:32 +01:00
for (item = vpi_scan(argv) ; item ; item = vpi_scan(argv)) {
items = realloc(items, (nitems+1)*sizeof(vpiHandle));
items[nitems] = item;
nitems += 1;
}
info->nitems = nitems;
info->items = items;
Convert vvm to implement system tasks with vpi.
1999-08-15 03:23:56 +02:00
$display can take 0 arguments.
2001-03-18 01:31:32 +01:00
} else {
info->nitems = 0;
info->items = 0;
}
Convert vvm to implement system tasks with vpi.
1999-08-15 03:23:56 +02:00
}
Add possibility of const-correctness Follow-up to "Brainless start to const-correct changes" Still actually does nothing, but now if the #define ICARUS_VPI_CONST in vpi_user.h is changed to const, Icarus is almost const-correct, as checked with gcc flags -Wcast-qual -Wwrite-strings. Choosing when to #define ICARUS_VPI_CONST const is left as an exercise for the reader. With these two patches applied, and the const define, there are still about a dozen const problems left.
2010-10-04 18:53:55 +02:00
static int get_default_format(const char *name)
Add monitoron and monitoroff system tasks.
2002-07-25 05:35:51 +02:00
{
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
int default_format;
Add monitoron and monitoroff system tasks.
2002-07-25 05:35:51 +02:00
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
switch(name[ strlen(name)-1 ]){
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
/* (f)writE/(f)strobE or (f)monitoR or (f)displaY or sformaT/sformatF */
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
case 'e':
case 'r':
case 't':
vpi: $sformatf() function.
2015-12-11 12:58:25 +01:00
case 'f':
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
case 'y': default_format = vpiDecStrVal; break;
case 'h': default_format = vpiHexStrVal; break;
case 'o': default_format = vpiOctStrVal; break;
case 'b': default_format = vpiBinStrVal; break;
default:
Add some casts in vpi to remove warnings. The Cygwin compiler is a bit picky. This patch adds some casts to remove compilation warnings. In the past I have had warnings off because of problems with the STL, but for C directories like this it makes sense to enable the warnings. It also does not recognize that an assert(0) or assert(false) ends a routine so it complains about no return at end of function or variables not being defined.
2009-12-10 21:49:39 +01:00
default_format = -1;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
assert(0);
}
return default_format;
Support vpiValueChance callbacks.
1999-10-29 05:37:22 +02:00
}
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
/* Build the format using the variables that control how the item will
* be printed. This is used in error messages and directly by the e/f/g
* format codes (minus the enclosing <>). The user needs to free the
* returned string. */
static char * format_as_string(int ljust, int plus, int ld_zero, int width,
int prec, char fmt)
{
char buf[256];
unsigned int size = 0;
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
/* Do not remove/change the "<" without also changing the e/f/g format
* code below! */
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
buf[size++] = '<';
buf[size++] = '%';
if (ljust == 1) buf[size++] = '-';
if (plus == 1) buf[size++] = '+';
if (ld_zero == 1) buf[size++] = '0';
if (width != -1)
size += sprintf(&buf[size], "%d", width);
if (prec != -1)
size += sprintf(&buf[size], ".%d", prec);
Convert a single % at the end of a format string to %%. If there is a single % at the end of a format string then automatically convert this to %% so that we get a single % in the output.
2011-04-20 00:31:45 +02:00
if (fmt) buf[size++] = fmt;
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
/* The same goes here ">"! */
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
buf[size++] = '>';
buf[size] = '\0';
return strdup(buf);
}
static void get_time(char *rtn, const char *value, int prec,
PLI_INT32 time_units)
{
int shift = time_units - timeformat_info.units;
/* Strip any leading zeros, but leave a single zero. */
while (value[0] == '0' && value[1] != '\0') value += 1;
/* We need to scale the number up. */
if (shift >= 0) {
strcpy(rtn, value);
/* Shift only non-zero values. */
while (shift > 0 && value[0] != '0') {
strcat(rtn, "0");
shift -= 1;
}
if (prec > 0) strcat(rtn, ".");
Factor out the common code for checking if a FD/MCD is valid
2020-05-17 03:02:46 +02:00
while (prec > 0) {
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
strcat(rtn, "0");
prec -= 1;
}
/* We need to scale the number down. */
} else {
Fix the formatting of time values that are less than 1 This patch is a modification of a user contributed patch that fixes the time formatting when the time is less than 1 and $timeformat() is used.
2012-01-24 03:42:02 +01:00
int len = strlen(value);
int head = len + shift;
int tail;
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
/* We have digits to the left of the decimal point. */
if (head > 0) {
strncpy(rtn, value, head);
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
*(rtn+head) = '\0';
if (prec > 0) {
strcat(rtn, ".");
strncat(rtn, &value[head], prec);
tail = prec + shift;
while (tail > 0) {
strcat(rtn, "0");
tail -= 1;
}
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
}
/* All digits are to the right of the decimal point. */
} else {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
strcpy(rtn, "0");
if (prec > 0) strcat(rtn, ".");
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
/* Add leading zeros as needed. */
Fix the formatting of time values that are less than 1 This patch is a modification of a user contributed patch that fixes the time formatting when the time is less than 1 and $timeformat() is used.
2012-01-24 03:42:02 +01:00
head = -head;
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
if (head > prec) head = prec;
while (head > 0) {
strcat(rtn, "0");
head -= 1;
}
/* Add digits from the value if they fit. */
Fix the formatting of time values that are less than 1 This patch is a modification of a user contributed patch that fixes the time formatting when the time is less than 1 and $timeformat() is used.
2012-01-24 03:42:02 +01:00
tail = prec + len + shift;
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
if (tail > 0) {
strncat(rtn, value, tail);
/* Add trailing zeros to fill out the precision. */
Fix the formatting of time values that are less than 1 This patch is a modification of a user contributed patch that fixes the time formatting when the time is less than 1 and $timeformat() is used.
2012-01-24 03:42:02 +01:00
tail = prec + shift + 1 - len;
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
while (tail > 0) {
strcat(rtn, "0");
tail -= 1;
}
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
}
}
}
strcat(rtn, timeformat_info.suff);
}
static void get_time_real(char *rtn, double value, int prec,
PLI_INT32 time_units)
{
Reduce the error when scaling a real time value. When scaling a time value we would often use the power operator to create constants 10**-N that were then multiplied with the original value. The problem with this is that there is some errors in the representation of the fractional number. It is better to create a integer value 10**N and then divide the original value by this exact constant. You still have the calculation error, but the scale value is now an exactly real value.
2011-10-31 07:29:02 +01:00
/* Scale the value from its time units to the format time units. */
if (time_units >= timeformat_info.units) {
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
value *= pow(10.0, time_units - timeformat_info.units);
Reduce the error when scaling a real time value. When scaling a time value we would often use the power operator to create constants 10**-N that were then multiplied with the original value. The problem with this is that there is some errors in the representation of the fractional number. It is better to create a integer value 10**N and then divide the original value by this exact constant. You still have the calculation error, but the scale value is now an exactly real value.
2011-10-31 07:29:02 +01:00
} else {
value /= pow(10.0, timeformat_info.units - time_units);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
}
sprintf(rtn, "%0.*f%s", prec, value, timeformat_info.suff);
}
static unsigned int get_format_char(char **rtn, int ljust, int plus,
int ld_zero, int width, int prec,
Add possibility of const-correctness Follow-up to "Brainless start to const-correct changes" Still actually does nothing, but now if the #define ICARUS_VPI_CONST in vpi_user.h is changed to const, Icarus is almost const-correct, as checked with gcc flags -Wcast-qual -Wwrite-strings. Choosing when to #define ICARUS_VPI_CONST const is left as an exercise for the reader. With these two patches applied, and the const define, there are still about a dozen const problems left.
2010-10-04 18:53:55 +02:00
char fmt, const struct strobe_cb_info *info,
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
unsigned int *idx)
{
s_vpi_value value;
char *result, *fmtb;
unsigned int size;
Use a default buffer size of 512 vs 256 for displaying. This patch changes the initial buffer size when displaying results to 512 from 256. This initial buffer is used as the default for the floating point results and failed when %f is given a large, but valid value. To make this even more safe we add the precision to the size to make sure we do not overflow the buffer.
2009-06-15 20:11:38 +02:00
unsigned int ini_size = 512; /* The initial size of the buffer. */
Make %v print all the bits of a vector. This patch reworks the %v code to print the strength information for all the bits of a vector. The code previously only printed the LSB information.
2008-04-10 03:16:21 +02:00
/* Make sure the width fits in the initial buffer. */
Revert bad merge from vhdl branch
2010-10-02 20:02:27 +02:00
assert(width >= -1);
if ((unsigned int)(width+1) > ini_size) ini_size = width + 1;
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
/* The default return value is the full format. */
Make %v print all the bits of a vector. This patch reworks the %v code to print the strength information for all the bits of a vector. The code previously only printed the LSB information.
2008-04-10 03:16:21 +02:00
result = malloc(ini_size*sizeof(char));
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
fmtb = format_as_string(ljust, plus, ld_zero, width, prec, fmt);
strcpy(result, fmtb);
Make %v print all the bits of a vector. This patch reworks the %v code to print the strength information for all the bits of a vector. The code previously only printed the LSB information.
2008-04-10 03:16:21 +02:00
size = strlen(result) + 1; /* fallback value if errors */
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
switch (fmt) {
Convert a single % at the end of a format string to %%. If there is a single % at the end of a format string then automatically convert this to %% so that we get a single % in the output.
2011-04-20 00:31:45 +02:00
case '\0':
Add a warning when a single % is at the end of the format string
2021-01-26 06:48:00 +01:00
vpi_printf("WARNING: %s:%d: a single %% at the end of format string "
"%s%s will be displayed as '%%'.\n",
info->filename, info->lineno, info->name, fmtb);
Some clean up and add initial support for elaboration system tasks
2021-02-01 09:20:09 +01:00
// fallthrough
Add a warning when a single % is at the end of the format string
2021-01-26 06:48:00 +01:00
case '%':
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
if (ljust != 0 || plus != 0 || ld_zero != 0 || width != -1 ||
prec != -1) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: invalid format %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
}
strcpy(result, "%");
size = strlen(result) + 1;
break;
case 'b':
case 'B':
case 'o':
case 'O':
case 'h':
case 'H':
case 'x':
case 'X':
*idx += 1;
if (plus != 0 || prec != -1) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: invalid format %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
}
if (*idx >= info->nitems) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: missing argument for %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
} else {
switch (fmt) {
case 'b':
case 'B':
value.format = vpiBinStrVal;
break;
case 'o':
case 'O':
value.format = vpiOctStrVal;
break;
case 'h':
case 'H':
case 'x':
case 'X':
value.format = vpiHexStrVal;
break;
}
vpi_get_value(info->items[*idx], &value);
if (value.format == vpiSuppressVal) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: incompatible value for %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
} else {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
unsigned swidth = strlen(value.value.str), free_flag = 0;;
char *cp = value.value.str;
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
if (ld_zero == 1) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
/* Strip the leading zeros if a width is not given. */
fixing: %0b format prints nothing for a 0 input the following example returns ||. This patch fixes it to return |0|. module t(); wire [3:0] b; initial $monitor("|%0b|", b); assign b = 0; endmodule
2009-03-19 04:27:48 +01:00
if (width == -1) while (*cp == '0' && *(cp+1) != '\0') cp++;
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
/* Pad with leading zeros. */
else if (ljust == 0 && (signed)swidth < width) {
unsigned pad = (unsigned)width - swidth;
cp = malloc((width+1)*sizeof(char));
memset(cp, '0', pad);
strcpy(cp+pad, value.value.str);
free_flag = 1;
/* For a left aligned value also strip the leading zeros. */
fixing: %0b format prints nothing for a 0 input the following example returns ||. This patch fixes it to return |0|. module t(); wire [3:0] b; initial $monitor("|%0b|", b); assign b = 0; endmodule
2009-03-19 04:27:48 +01:00
} else if (ljust != 0) while (*cp == '0' && *(cp+1) != '\0') cp++;
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
}
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
/* If a width was not given, use a width of zero. */
if (width == -1) width = 0;
Fix typos in vpi/sys_display.c Fallout from me trying to understand the origin of pr1780480 and pr1830834. Too bad I don't understand c++ and vvp as well as I understand English! (Spelling and grammer fixes in comments. -ed.)
2008-02-01 21:15:40 +01:00
/* If the default buffer is too small, make it big enough. */
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
size = strlen(cp) + 1;
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
if ((signed)size < (width+1)) size = width+1;
Make %v print all the bits of a vector. This patch reworks the %v code to print the strength information for all the bits of a vector. The code previously only printed the LSB information.
2008-04-10 03:16:21 +02:00
if (size > ini_size) result = realloc(result, size*sizeof(char));
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
if (ljust == 0) sprintf(result, "%*s", width, cp);
else sprintf(result, "%-*s", width, cp);
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
if (free_flag) free(cp);
Make %v print all the bits of a vector. This patch reworks the %v code to print the strength information for all the bits of a vector. The code previously only printed the LSB information.
2008-04-10 03:16:21 +02:00
size = strlen(result) + 1;
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
}
}
break;
case 'c':
case 'C':
*idx += 1;
if (plus != 0 || prec != -1) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: invalid format %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
}
if (*idx >= info->nitems) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: missing argument for %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
} else {
Fix bug in output of null character with %c format.
2018-10-07 19:59:53 +02:00
value.format = vpiIntVal;
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
vpi_get_value(info->items[*idx], &value);
if (value.format == vpiSuppressVal) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: incompatible value for %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
} else {
Fix bug in output of null character with %c format.
2018-10-07 19:59:53 +02:00
char ch = value.value.integer;
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
/* If the width is less than one then use a width of one. */
if (width < 1) width = 1;
Allow %c format to output null characters (GitHub issue #209) Currently $display et al. output nothing when the expression corresponding to a %c format specification has the value 0. As Verilog provides no other way to write raw bytes to a file, we should allow 0 values to be written. Other simulators allow this.
2018-09-29 22:22:17 +02:00
size = width + 1;
assert(size <= ini_size);
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
if (ljust == 0) {
if (width > 1) {
Allow %c format to output null characters (GitHub issue #209) Currently $display et al. output nothing when the expression corresponding to a %c format specification has the value 0. As Verilog provides no other way to write raw bytes to a file, we should allow 0 values to be written. Other simulators allow this.
2018-09-29 22:22:17 +02:00
memset(result, (ld_zero == 1 ? '0': ' '), width-1);
}
result[width-1] = ch;
} else {
result[0] = ch;
if (width > 1) {
memset(result+1, ' ', width-1);
}
}
result[width] = '\0';
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
}
}
break;
case 'd':
case 'D':
*idx += 1;
if (prec != -1) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: invalid format %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
}
if (*idx >= info->nitems) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: missing argument for %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
} else {
value.format = vpiDecStrVal;
vpi_get_value(info->items[*idx], &value);
if (value.format == vpiSuppressVal) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: incompatible value for %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
} else {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
unsigned pad = 0;
Fix space issues in VPI directory
2010-05-23 23:30:48 +02:00
unsigned swidth = strlen(value.value.str) +
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
(value.value.str[0] == '-' ? 0 : (unsigned)plus);
char *tbuf, *cpb, *cp = value.value.str;
/* Allocate storage and calculate the pad if needed. */
if (ljust == 0 && ld_zero == 1 && (signed)swidth < width) {
tbuf = malloc((width+1)*sizeof(char));
pad = (unsigned)width - swidth;
} else {
tbuf = malloc((swidth+1)*sizeof(char));
}
cpb = tbuf;
/* Insert the sign if needed. */
if (plus == 1 && *cp != '-') {
*cpb = '+';
cpb += 1;
} else if (*cp == '-') {
*cpb = '-';
cpb += 1;
cp += 1;
}
/* Now add padding if it is needed and then add the value. */
memset(cpb, '0', pad);
strcpy(cpb+pad, cp);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
Fix typos in vpi/sys_display.c Fallout from me trying to understand the origin of pr1780480 and pr1830834. Too bad I don't understand c++ and vvp as well as I understand English! (Spelling and grammer fixes in comments. -ed.)
2008-02-01 21:15:40 +01:00
/* If a width was not given, use the default, unless we have a
When printing a Dec numeric value use the string size and the minimum size. The vpiDecStrVal case for the get_numeric() function needs to use the existing string width as the minimum result size. -1 can be represented as a signed value with a width of 1. This gives a display width of -1 which is too small for the -1 string value. This was creating valgrind issues. In general the string value should be the minimum so this is a safe change.
2011-10-10 07:12:10 +02:00
* leading zero (width of zero). Because the width of a real in
* Icarus is 1 the string length will set the width of a real
* displayed using %d. */
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
if (width == -1) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
width = (ld_zero == 1) ? 0 : vpi_get_dec_size(info->items[*idx]);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
}
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
Fix typos in vpi/sys_display.c Fallout from me trying to understand the origin of pr1780480 and pr1830834. Too bad I don't understand c++ and vvp as well as I understand English! (Spelling and grammer fixes in comments. -ed.)
2008-02-01 21:15:40 +01:00
/* If the default buffer is too small make it big enough. */
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
size = strlen(tbuf) + 1;
if ((signed)size < (width+1)) size = width+1;
Make %v print all the bits of a vector. This patch reworks the %v code to print the strength information for all the bits of a vector. The code previously only printed the LSB information.
2008-04-10 03:16:21 +02:00
if (size > ini_size) result = realloc(result, size*sizeof(char));
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
if (ljust == 0) sprintf(result, "%*s", width, tbuf);
else sprintf(result, "%-*s", width, tbuf);
free(tbuf);
Make %v print all the bits of a vector. This patch reworks the %v code to print the strength information for all the bits of a vector. The code previously only printed the LSB information.
2008-04-10 03:16:21 +02:00
size = strlen(result) + 1;
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
}
}
break;
case 'e':
case 'E':
case 'f':
case 'F':
case 'g':
case 'G':
*idx += 1;
if (*idx >= info->nitems) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: missing argument for %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
} else {
value.format = vpiRealVal;
vpi_get_value(info->items[*idx], &value);
if (value.format == vpiSuppressVal) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: incompatible value for %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
} else {
char *cp = fmtb;
if (fmt == 'F') {
while (*cp != 'F') cp++;
*cp = 'f';
}
while (*cp != '>') cp++;
*cp = '\0';
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
/* If the default buffer is too small make it big enough.
*
Use a default buffer size of 512 vs 256 for displaying. This patch changes the initial buffer size when displaying results to 512 from 256. This initial buffer is used as the default for the floating point results and failed when %f is given a large, but valid value. To make this even more safe we add the precision to the size to make sure we do not overflow the buffer.
2009-06-15 20:11:38 +02:00
* This should always give enough space. The maximum double
* is approximately 1.8*10^308 this means we could need 310
* characters plus the precision. We'll use 320 to give some
* extra buffer space. The initial buffers size should work
* for most cases, but to be safe we add the precision to
* the maximum size (think %6.300f when passed 1.2*10^308). */
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
size = width + 1;
Use a default buffer size of 512 vs 256 for displaying. This patch changes the initial buffer size when displaying results to 512 from 256. This initial buffer is used as the default for the floating point results and failed when %f is given a large, but valid value. To make this even more safe we add the precision to the size to make sure we do not overflow the buffer.
2009-06-15 20:11:38 +02:00
if (size < 320) size = 320;
size += prec;
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
if (size > ini_size) result = realloc(result, size*sizeof(char));
Changes to vpi/sys_display.c First and second areas were changed for "Microsoft Visual Studio Express 2015 RC Web" so that gold files would match. For the third area of change, "Microsoft Visual Studio Express 2015 RC Web" wanted to print 6 zeros after the decimals for the values of 0.0 and -0.0, and then the gold files wouldn't match.
2015-07-10 18:50:39 +02:00
#if !defined(__GNUC__)
if (isnan(value.value.real))
sprintf(result, "%s", "nan");
else
sprintf(result, fmtb+1, value.value.real);
#else
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
sprintf(result, fmtb+1, value.value.real);
Changes to vpi/sys_display.c First and second areas were changed for "Microsoft Visual Studio Express 2015 RC Web" so that gold files would match. For the third area of change, "Microsoft Visual Studio Express 2015 RC Web" wanted to print 6 zeros after the decimals for the values of 0.0 and -0.0, and then the gold files wouldn't match.
2015-07-10 18:50:39 +02:00
#endif
Make %v print all the bits of a vector. This patch reworks the %v code to print the strength information for all the bits of a vector. The code previously only printed the LSB information.
2008-04-10 03:16:21 +02:00
size = strlen(result) + 1;
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
}
}
break;
/* This Verilog format specifier is not currently supported!
* vpiCell and vpiLibrary need to be implemented first. */
case 'l':
case 'L':
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: %%%c currently unsupported %s%s.\n",
info->filename, info->lineno, fmt, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
break;
case 'm':
case 'M':
if (plus != 0 || prec != -1) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: invalid format %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
}
Fix typos in vpi/sys_display.c Fallout from me trying to understand the origin of pr1780480 and pr1830834. Too bad I don't understand c++ and vvp as well as I understand English! (Spelling and grammer fixes in comments. -ed.)
2008-02-01 21:15:40 +01:00
/* If a width was not given, use a width of zero. */
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
if (width == -1) width = 0;
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
{
Some cppcheck cleanup
2025-10-13 02:35:15 +02:00
const char *cp = vpi_get_str(vpiFullName, info->scope);
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
/* If the default buffer is too small, make it big enough. */
size = strlen(cp) + 1;
if ((signed)size < (width+1)) size = width+1;
if (size > ini_size) result = realloc(result, size*sizeof(char));
if (ljust == 0) sprintf(result, "%*s", width, cp);
else sprintf(result, "%-*s", width, cp);
}
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
size = strlen(result) + 1;
break;
case 's':
case 'S':
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
/* Strings are not numeric and are not zero filled, so %08s => %8s. */
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
*idx += 1;
if (plus != 0 || prec != -1) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: invalid format %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
}
if (*idx >= info->nitems) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: missing argument for %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
} else {
value.format = vpiStringVal;
vpi_get_value(info->items[*idx], &value);
if (value.format == vpiSuppressVal) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: incompatible value for %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
} else {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
if (width == -1) {
/* If all we have is a leading zero then we want a zero width. */
if (ld_zero == 1) width = 0;
/* Otherwise if a width was not given, use the value width. */
else width = (vpi_get(vpiSize, info->items[*idx])+7) / 8;
}
Fix typos in vpi/sys_display.c Fallout from me trying to understand the origin of pr1780480 and pr1830834. Too bad I don't understand c++ and vvp as well as I understand English! (Spelling and grammer fixes in comments. -ed.)
2008-02-01 21:15:40 +01:00
/* If the default buffer is too small make it big enough. */
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
size = strlen(value.value.str) + 1;
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
if ((signed)size < (width+1)) size = width+1;
Make %v print all the bits of a vector. This patch reworks the %v code to print the strength information for all the bits of a vector. The code previously only printed the LSB information.
2008-04-10 03:16:21 +02:00
if (size > ini_size) result = realloc(result, size*sizeof(char));
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
if (ljust == 0) sprintf(result, "%*s", width, value.value.str);
else sprintf(result, "%-*s", width, value.value.str);
Make %v print all the bits of a vector. This patch reworks the %v code to print the strength information for all the bits of a vector. The code previously only printed the LSB information.
2008-04-10 03:16:21 +02:00
size = strlen(result) + 1;
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
}
}
break;
case 't':
case 'T':
*idx += 1;
if (*idx >= info->nitems) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: missing argument for %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
} else {
PLI_INT32 type;
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
/* Get the argument type and value. */
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
type = vpi_get(vpiType, info->items[*idx]);
if (((type == vpiConstant || type == vpiParameter) &&
vpi_get(vpiConstType, info->items[*idx]) == vpiRealConst) ||
Fix space issues in VPI directory
2010-05-23 23:30:48 +02:00
type == vpiRealVar || (type == vpiSysFuncCall &&
Fix %t to work with real system functions. Fix %t to get the real value when the argument comes from a real valued system function.
2009-05-06 01:13:15 +02:00
vpi_get(vpiFuncType, info->items[*idx]) == vpiRealFunc)) {
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
value.format = vpiRealVal;
} else {
value.format = vpiDecStrVal;
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
}
vpi_get_value(info->items[*idx], &value);
if (value.format == vpiSuppressVal) {
vpi_printf("WARNING: %s:%d: incompatible value for %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
} else {
Fix compile warning
2016-02-11 07:32:31 +01:00
char *tbuf, *prev_suff = 0;
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
PLI_INT32 time_units = vpi_get(vpiTimeUnit, info->scope);
vhdlpp: Added 'image attribute for time type.
2016-01-06 14:08:31 +01:00
if (plus != 0) {
/* Icarus-specific extension to print out time units.
* It is needed by vhdlpp to correctly implement time'image(). */
PLI_INT32 time_prec = vpi_get(vpiTimePrecision, info->scope);
Save the time format suffix string. In case of a vhdlpp-specific () call, the timeformat suffix string was overridden, therefore it has to be restored to the previous value after the call.
2016-02-09 16:17:16 +01:00
/* We are about to override the suffix string set with $timeformat(),
* therefore we need to restore it after the call. */
prev_suff = timeformat_info.suff;
Factor out the common code for checking if a FD/MCD is valid
2020-05-17 03:02:46 +02:00
if (time_units < -12)
vhdlpp: Added 'image attribute for time type.
2016-01-06 14:08:31 +01:00
timeformat_info.suff = strdup(" fs");
Factor out the common code for checking if a FD/MCD is valid
2020-05-17 03:02:46 +02:00
else if (time_units < -9)
vhdlpp: Added 'image attribute for time type.
2016-01-06 14:08:31 +01:00
timeformat_info.suff = strdup(" ps");
Factor out the common code for checking if a FD/MCD is valid
2020-05-17 03:02:46 +02:00
else if (time_units < -6)
vhdlpp: Added 'image attribute for time type.
2016-01-06 14:08:31 +01:00
timeformat_info.suff = strdup(" ns");
Factor out the common code for checking if a FD/MCD is valid
2020-05-17 03:02:46 +02:00
else if (time_units < -3)
vhdlpp: Added 'image attribute for time type.
2016-01-06 14:08:31 +01:00
timeformat_info.suff = strdup(" us");
Factor out the common code for checking if a FD/MCD is valid
2020-05-17 03:02:46 +02:00
else if (time_units < 0)
vhdlpp: Added 'image attribute for time type.
2016-01-06 14:08:31 +01:00
timeformat_info.suff = strdup(" ms");
else
timeformat_info.suff = strdup(" s");
/* Adjust shift for get_time(), so the number indeed matches the unit */
time_units += (3 + (time_units % 3)) % 3 + (time_prec - time_units);
}
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
unsigned swidth, free_flag = 0;
unsigned suff_len = strlen(timeformat_info.suff);
char *cp;
/* The 512 (513-1 for EOL) is more than enough for any double
* value (309 digits plus a decimal point maximum). Because of
* scaling this could be larger. For decimal values you can
Spelling and related fixes No effect on executable code.
2011-04-25 21:21:39 +02:00
* have an arbitrary value so you can overflow the buffer, but
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
* for now we will assume the user will use this as intended
* (pass a time variable or the result of a time function). */
tbuf = malloc((513+suff_len)*sizeof(char));
if (prec == -1) prec = timeformat_info.prec;
if (value.format == vpiRealVal) {
get_time_real(tbuf, value.value.real, prec, time_units);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
} else {
get_time(tbuf, value.value.str, prec, time_units);
}
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
cp = tbuf;
swidth = strlen(tbuf);
Factor out the common code for checking if a FD/MCD is valid
2020-05-17 03:02:46 +02:00
if (plus != 0) {
Save the time format suffix string. In case of a vhdlpp-specific () call, the timeformat suffix string was overridden, therefore it has to be restored to the previous value after the call.
2016-02-09 16:17:16 +01:00
/* Restore the previous suffix string, overridden by
* a vhdlpp-specific $sformatf() call. */
free(timeformat_info.suff);
timeformat_info.suff = prev_suff;
}
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
if (ld_zero == 1) {
/* No leading zeros are created by this conversion so just make
* the width 0 for this case. */
if (width == -1) width = 0;
/* Pad with leading zeros. */
else if (ljust == 0 && (signed)swidth < width) {
unsigned pad = (unsigned)width - swidth;
cp = malloc((width+1)*sizeof(char));
memset(cp, '0', pad);
strcpy(cp+pad, tbuf);
free_flag = 1;
}
}
if (width == -1) width = timeformat_info.width;
/* If the default buffer is too small make it big enough. */
size = strlen(tbuf) + 1;
if ((signed)size < (width+1)) size = width+1;
if (size > ini_size) result = realloc(result, size*sizeof(char));
if (ljust == 0) sprintf(result, "%*s", width, cp);
else sprintf(result, "%-*s", width, cp);
if (free_flag) free(cp);
free(tbuf);
size = strlen(result) + 1;
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
}
}
break;
case 'u':
case 'U':
*idx += 1;
if (ljust != 0 || plus != 0 || ld_zero != 0 || width != -1 ||
prec != -1) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: invalid format %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
}
if (*idx >= info->nitems) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: missing argument for %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
} else {
value.format = vpiVectorVal;
vpi_get_value(info->items[*idx], &value);
if (value.format == vpiSuppressVal) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: incompatible value for %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
} else {
Fix a few cppcheck warnings.
2014-06-28 03:39:06 +02:00
PLI_INT32 veclen, word, byte;
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
char *cp;
veclen = (vpi_get(vpiSize, info->items[*idx])+31)/32;
size = veclen * 4 + 1;
Fix typos in vpi/sys_display.c Fallout from me trying to understand the origin of pr1780480 and pr1830834. Too bad I don't understand c++ and vvp as well as I understand English! (Spelling and grammer fixes in comments. -ed.)
2008-02-01 21:15:40 +01:00
/* If the default buffer is too small, make it big enough. */
Make %v print all the bits of a vector. This patch reworks the %v code to print the strength information for all the bits of a vector. The code previously only printed the LSB information.
2008-04-10 03:16:21 +02:00
if (size > ini_size) result = realloc(result, size*sizeof(char));
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
cp = result;
for (word = 0; word < veclen; word += 1) {
Fix a few cppcheck warnings.
2014-06-28 03:39:06 +02:00
PLI_INT32 bits = value.value.vector[word].aval &
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
~value.value.vector[word].bval;
#ifdef WORDS_BIGENDIAN
for (byte = 3; byte >= 0; byte -= 1) {
#else
for (byte = 0; byte <= 3; byte += 1) {
#endif
*cp = (bits >> byte*8) & 0xff;
cp += 1;
}
}
*cp = '\0';
}
}
/* size is defined above! We can't use strlen here since this can
* be a binary string (can contain NULLs). */
break;
case 'v':
case 'V':
*idx += 1;
if (plus != 0 || prec != -1) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: invalid format %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
}
if (*idx >= info->nitems) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: missing argument for %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
} else {
value.format = vpiStrengthVal;
vpi_get_value(info->items[*idx], &value);
if (value.format == vpiSuppressVal) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: incompatible value for %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
} else {
Make %v print all the bits of a vector. This patch reworks the %v code to print the strength information for all the bits of a vector. The code previously only printed the LSB information.
2008-04-10 03:16:21 +02:00
char tbuf[4], *rbuf;
PLI_INT32 nbits;
int bit;
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
/* If a width was not given use a width of zero. */
if (width == -1) width = 0;
Make %v print all the bits of a vector. This patch reworks the %v code to print the strength information for all the bits of a vector. The code previously only printed the LSB information.
2008-04-10 03:16:21 +02:00
nbits = vpi_get(vpiSize, info->items[*idx]);
Spelling fixes Mostly comments, but includes quite a few user-visible error, debug, and help messages.
2008-06-12 19:04:29 +02:00
/* This is 4 chars for all but the last bit (strength + "_")
Add an underscore between multiple strength values. When printing the strength information for a multi bit net this patch adds an underscore between the individual bit strength values. This makes it easier to see the individual bit values.
2008-04-15 00:41:39 +02:00
* which only needs three chars (strength), but then you need
* space for the EOS '\0', so it is just number of bits * 4. */
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
size = nbits*4;
rbuf = malloc(size*sizeof(char));
if ((signed)size < (width+1)) size = width+1;
if (size > ini_size) result = realloc(result, size*sizeof(char));
Make %v print all the bits of a vector. This patch reworks the %v code to print the strength information for all the bits of a vector. The code previously only printed the LSB information.
2008-04-10 03:16:21 +02:00
strcpy(rbuf, "");
for (bit = nbits-1; bit >= 0; bit -= 1) {
vpip_format_strength(tbuf, &value, bit);
strcat(rbuf, tbuf);
Add an underscore between multiple strength values. When printing the strength information for a multi bit net this patch adds an underscore between the individual bit strength values. This makes it easier to see the individual bit values.
2008-04-15 00:41:39 +02:00
if (bit > 0) strcat(rbuf, "_");
Make %v print all the bits of a vector. This patch reworks the %v code to print the strength information for all the bits of a vector. The code previously only printed the LSB information.
2008-04-10 03:16:21 +02:00
}
if (ljust == 0) sprintf(result, "%*s", width, rbuf);
else sprintf(result, "%-*s", width, rbuf);
free(rbuf);
size = strlen(result) + 1;
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
}
}
break;
case 'z':
case 'Z':
*idx += 1;
Make %v print all the bits of a vector. This patch reworks the %v code to print the strength information for all the bits of a vector. The code previously only printed the LSB information.
2008-04-10 03:16:21 +02:00
size = strlen(result) + 1; /* fallback value if errors */
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
if (ljust != 0 || plus != 0 || ld_zero != 0 || width != -1 ||
prec != -1) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: invalid format %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
}
if (*idx >= info->nitems) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: missing argument for %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
} else {
value.format = vpiVectorVal;
vpi_get_value(info->items[*idx], &value);
if (value.format == vpiSuppressVal) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: incompatible value for %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
} else {
PLI_INT32 veclen, word, elem, bits, byte;
char *cp;
veclen = (vpi_get(vpiSize, info->items[*idx])+31)/32;
size = 2 * veclen * 4 + 1;
Fix typos in vpi/sys_display.c Fallout from me trying to understand the origin of pr1780480 and pr1830834. Too bad I don't understand c++ and vvp as well as I understand English! (Spelling and grammer fixes in comments. -ed.)
2008-02-01 21:15:40 +01:00
/* If the default buffer is too small, make it big enough. */
Make %v print all the bits of a vector. This patch reworks the %v code to print the strength information for all the bits of a vector. The code previously only printed the LSB information.
2008-04-10 03:16:21 +02:00
if (size > ini_size) result = realloc(result, size*sizeof(char));
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
cp = result;
for (word = 0; word < veclen; word += 1) {
/* Write the aval followed by the bval in endian order. */
for (elem = 0; elem < 2; elem += 1) {
bits = *(&value.value.vector[word].aval+elem);
#ifdef WORDS_BIGENDIAN
for (byte = 3; byte >= 0; byte -= 1) {
#else
for (byte = 0; byte <= 3; byte += 1) {
#endif
*cp = (bits >> byte*8) & 0xff;
cp += 1;
}
}
}
*cp = '\0';
}
}
/* size is defined above! We can't use strlen here since this can
* be a binary string (can contain NULLs). */
break;
default:
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: unknown format %s%s.\n",
info->filename, info->lineno, info->name, fmtb);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
size = strlen(result) + 1;
break;
}
free(fmtb);
Warnings and spelling fixes. Correct a variety of spelling errors, and eliminate a variety of compile time warnings.
2007-12-18 02:42:09 +01:00
/* We can't use strdup here since %u and %z can insert NULL
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
* characters into the stream. */
*rtn = malloc(size*sizeof(char));
memcpy(*rtn, result, size);
Make the $fdisplay/$fwrite family use the string formatting. This patch makes the $fdisplay() and $fwrite() family of task use the string formatting code. It also adds a compiletf routine for the tasks and a run time check that the FD or MCD is correct. If not it fails instead of just dropping the output. It also cleans up a couple of memory leaks and adds no argument compiletf routines for the $monitoron/$monitoroff tasks.
2009-01-10 02:27:27 +01:00
free(result);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
return size - 1;
}
/* We can't use the normal str functions on the return value since
* %u and %z can insert NULL characters into the stream. */
static unsigned int get_format(char **rtn, char *fmt,
Add possibility of const-correctness Follow-up to "Brainless start to const-correct changes" Still actually does nothing, but now if the #define ICARUS_VPI_CONST in vpi_user.h is changed to const, Icarus is almost const-correct, as checked with gcc flags -Wcast-qual -Wwrite-strings. Choosing when to #define ICARUS_VPI_CONST const is left as an exercise for the reader. With these two patches applied, and the const define, there are still about a dozen const problems left.
2010-10-04 18:53:55 +02:00
const struct strobe_cb_info *info, unsigned int *idx)
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
{
char *cp = fmt;
unsigned int size;
*rtn = strdup("");
size = 1;
while (*cp) {
size_t cnt = strcspn(cp, "%");
if (cnt > 0) {
*rtn = realloc(*rtn, (size+cnt)*sizeof(char));
memcpy(*rtn+size-1, cp, cnt);
size += cnt;
cp += cnt;
} else {
Shadow reduction part 5 Continue cleaning up shadowed variables, flagged by turning on -Wshadow. No intended change in functionality. Patch looks right, and is tested to compile and run on my machine. No regressions in test suite. This patch set covers C code, with the notable exception of vpi/lxt_write{,2}.c.
2008-12-18 17:45:47 +01:00
int ljust = 0, plus = 0, ld_zero = 0, width = -1, prec = -1;
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
char *result;
cp += 1;
while ((*cp == '-') || (*cp == '+')) {
if (*cp == '-') ljust = 1;
else plus = 1;
cp += 1;
}
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
if (*cp == '0') {
ld_zero = 1;
cp += 1;
}
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
if (isdigit((int)*cp)) width = strtoul(cp, &cp, 10);
if (*cp == '.') {
cp += 1;
prec = strtoul(cp, &cp, 10);
}
cnt = get_format_char(&result, ljust, plus, ld_zero, width, prec, *cp,
info, idx);
*rtn = realloc(*rtn, (size+cnt)*sizeof(char));
memcpy(*rtn+size-1, result, cnt);
free(result);
size += cnt;
Convert a single % at the end of a format string to %%. If there is a single % at the end of a format string then automatically convert this to %% so that we get a single % in the output.
2011-04-20 00:31:45 +02:00
if (*cp) cp += 1;
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
}
}
*(*rtn+size-1) = '\0';
return size - 1;
}
Add possibility of const-correctness Follow-up to "Brainless start to const-correct changes" Still actually does nothing, but now if the #define ICARUS_VPI_CONST in vpi_user.h is changed to const, Icarus is almost const-correct, as checked with gcc flags -Wcast-qual -Wwrite-strings. Choosing when to #define ICARUS_VPI_CONST const is left as an exercise for the reader. With these two patches applied, and the const define, there are still about a dozen const problems left.
2010-10-04 18:53:55 +02:00
static unsigned int get_numeric(char **rtn, const struct strobe_cb_info *info,
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
vpiHandle item)
{
When printing a Dec numeric value use the string size and the minimum size. The vpiDecStrVal case for the get_numeric() function needs to use the existing string width as the minimum result size. -1 can be represented as a signed value with a width of 1. This gives a display width of -1 which is too small for the -1 string value. This was creating valgrind issues. In general the string value should be the minimum so this is a safe change.
2011-10-10 07:12:10 +02:00
int size, min;
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
s_vpi_value val;
val.format = info->default_format;
vpi_get_value(item, &val);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
switch(info->default_format){
case vpiDecStrVal:
size = vpi_get_dec_size(item);
Fix space issues. This patch removes space before EOL, etc.
2011-11-03 02:10:41 +01:00
/* -1 can be represented as a one bit signed value. This returns
When printing a Dec numeric value use the string size and the minimum size. The vpiDecStrVal case for the get_numeric() function needs to use the existing string width as the minimum result size. -1 can be represented as a signed value with a width of 1. This gives a display width of -1 which is too small for the -1 string value. This was creating valgrind issues. In general the string value should be the minimum so this is a safe change.
2011-10-10 07:12:10 +02:00
* a size of 1 which is too small for the -1 string value so make
* the string width the minimum display width. */
min = strlen(val.value.str);
if (size < min) size = min;
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
*rtn = malloc((size+1)*sizeof(char));
sprintf(*rtn, "%*s", size, val.value.str);
break;
default:
*rtn = strdup(val.value.str);
}
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
return strlen(*rtn);
}
/* In many places we can't use the normal str functions since %u and %z
* can insert NULL characters into the stream. */
Add possibility of const-correctness Follow-up to "Brainless start to const-correct changes" Still actually does nothing, but now if the #define ICARUS_VPI_CONST in vpi_user.h is changed to const, Icarus is almost const-correct, as checked with gcc flags -Wcast-qual -Wwrite-strings. Choosing when to #define ICARUS_VPI_CONST const is left as an exercise for the reader. With these two patches applied, and the const define, there are still about a dozen const problems left.
2010-10-04 18:53:55 +02:00
static char *get_display(unsigned int *rtnsz, const struct strobe_cb_info *info)
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
{
char *result, *fmt, *rtn, *func_name;
Down payment on const-correctness Clears out the unambiguous and easy-to-fix const faults, so we can focus better on the fundamental ones.
2010-09-27 20:03:43 +02:00
const char *cresult;
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
s_vpi_value value;
unsigned int idx, size, width;
char buf[256];
rtn = strdup("");
size = 1;
for (idx = 0; idx < info->nitems; idx += 1) {
vpiHandle item = info->items[idx];
switch (vpi_get(vpiType, item)) {
case vpiConstant:
case vpiParameter:
if (vpi_get(vpiConstType, item) == vpiStringConst) {
value.format = vpiStringVal;
vpi_get_value(item, &value);
fmt = strdup(value.value.str);
width = get_format(&result, fmt, info, &idx);
free(fmt);
Fix some valgrind warnings and display a real parameter as a real. This patch fixes some technically benign problems found by valgrind, but we may as well give the values a default to make sure things work correctly. I believe that the default I choose is the appropriate value for the context. The other problem is that real valued (local)parameters should be printed as a real value like a real variable, etc. Trying to do this as a decimal with the string formatting was causing a memory problem. I could have fixed the routine to handle this, but the better solution was to just display things the right way.
2009-01-08 04:01:09 +01:00
} else if (vpi_get(vpiConstType, item) == vpiRealConst) {
value.format = vpiRealVal;
vpi_get_value(item, &value);
Changes to vpi/sys_display.c First and second areas were changed for "Microsoft Visual Studio Express 2015 RC Web" so that gold files would match. For the third area of change, "Microsoft Visual Studio Express 2015 RC Web" wanted to print 6 zeros after the decimals for the values of 0.0 and -0.0, and then the gold files wouldn't match.
2015-07-10 18:50:39 +02:00
#if !defined(__GNUC__)
Factor out the common code for checking if a FD/MCD is valid
2020-05-17 03:02:46 +02:00
if (compatible_flag)
Changes to vpi/sys_display.c First and second areas were changed for "Microsoft Visual Studio Express 2015 RC Web" so that gold files would match. For the third area of change, "Microsoft Visual Studio Express 2015 RC Web" wanted to print 6 zeros after the decimals for the values of 0.0 and -0.0, and then the gold files wouldn't match.
2015-07-10 18:50:39 +02:00
sprintf(buf, "%g", value.value.real);
else {
Factor out the common code for checking if a FD/MCD is valid
2020-05-17 03:02:46 +02:00
if (value.value.real == 0.0 || value.value.real == -0.0)
Changes to vpi/sys_display.c First and second areas were changed for "Microsoft Visual Studio Express 2015 RC Web" so that gold files would match. For the third area of change, "Microsoft Visual Studio Express 2015 RC Web" wanted to print 6 zeros after the decimals for the values of 0.0 and -0.0, and then the gold files wouldn't match.
2015-07-10 18:50:39 +02:00
sprintf(buf, "%.05f", value.value.real);
else
sprintf(buf, "%#g", value.value.real);
}
#else
Add -compatible command line option to vvp. This option is intended to make it easier to compare results from Icarus with results from other simulators. For now, the only effect it has is to change the default format for displaying real numbers when no format string is supplied.
2013-10-20 23:59:53 +02:00
sprintf(buf, compatible_flag ? "%g" : "%#g", value.value.real);
Changes to vpi/sys_display.c First and second areas were changed for "Microsoft Visual Studio Express 2015 RC Web" so that gold files would match. For the third area of change, "Microsoft Visual Studio Express 2015 RC Web" wanted to print 6 zeros after the decimals for the values of 0.0 and -0.0, and then the gold files wouldn't match.
2015-07-10 18:50:39 +02:00
#endif
Fix some valgrind warnings and display a real parameter as a real. This patch fixes some technically benign problems found by valgrind, but we may as well give the values a default to make sure things work correctly. I believe that the default I choose is the appropriate value for the context. The other problem is that real valued (local)parameters should be printed as a real value like a real variable, etc. Trying to do this as a decimal with the string formatting was causing a memory problem. I could have fixed the routine to handle this, but the better solution was to just display things the right way.
2009-01-08 04:01:09 +01:00
result = strdup(buf);
width = strlen(result);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
} else {
width = get_numeric(&result, info, item);
}
rtn = realloc(rtn, (size+width)*sizeof(char));
memcpy(rtn+size-1, result, width);
free(result);
break;
case vpiNet:
case vpiReg:
Add the ability to $display and $monitor a bit, byte, short, int or long var. Add code so that we can $display and $monitor the two-state variables.
2011-09-06 01:25:05 +02:00
case vpiBitVar:
case vpiByteVar:
case vpiShortIntVar:
case vpiIntVar:
case vpiLongIntVar:
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
case vpiIntegerVar:
case vpiMemoryWord:
Add smart part select for system functions &PV<>. This patch adds a smart part select that allows system functions to have full access to the real bits of the part select.
2008-05-27 23:29:08 +02:00
case vpiPartSelect:
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
width = get_numeric(&result, info, item);
rtn = realloc(rtn, (size+width)*sizeof(char));
memcpy(rtn+size-1, result, width);
free(result);
break;
/* It appears that this is not currently used! A time variable is
passed as an integer and processed above. Hence this code has
only been visually checked. */
case vpiTimeVar:
value.format = vpiDecStrVal;
vpi_get_value(item, &value);
get_time(buf, value.value.str, timeformat_info.prec,
vpi_get(vpiTimeUnit, info->scope));
width = strlen(buf);
if (width < timeformat_info.width) width = timeformat_info.width;
rtn = realloc(rtn, (size+width)*sizeof(char));
sprintf(rtn+size-1, "%*s", width, buf);
break;
/* Realtime variables are also processed here. */
case vpiRealVar:
value.format = vpiRealVal;
vpi_get_value(item, &value);
Changes to vpi/sys_display.c First and second areas were changed for "Microsoft Visual Studio Express 2015 RC Web" so that gold files would match. For the third area of change, "Microsoft Visual Studio Express 2015 RC Web" wanted to print 6 zeros after the decimals for the values of 0.0 and -0.0, and then the gold files wouldn't match.
2015-07-10 18:50:39 +02:00
#if !defined(__GNUC__)
if (compatible_flag)
sprintf(buf, "%g", value.value.real);
else {
if (value.value.real == 0.0 || value.value.real == -0.0)
sprintf(buf, "%.05f", value.value.real);
else
sprintf(buf, "%#g", value.value.real);
}
#else
Add -compatible command line option to vvp. This option is intended to make it easier to compare results from Icarus with results from other simulators. For now, the only effect it has is to change the default format for displaying real numbers when no format string is supplied.
2013-10-20 23:59:53 +02:00
sprintf(buf, compatible_flag ? "%g" : "%#g", value.value.real);
Changes to vpi/sys_display.c First and second areas were changed for "Microsoft Visual Studio Express 2015 RC Web" so that gold files would match. For the third area of change, "Microsoft Visual Studio Express 2015 RC Web" wanted to print 6 zeros after the decimals for the values of 0.0 and -0.0, and then the gold files wouldn't match.
2015-07-10 18:50:39 +02:00
#endif
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
width = strlen(buf);
rtn = realloc(rtn, (size+width)*sizeof(char));
memcpy(rtn+size-1, buf, width);
break;
First pass at getting strings to work. In vvp, create the .var/str variable for representing strings, and handle strings in the $display system task. Add to vvp threads the concept of a stack of strings. This is going to be how complex objects are to me handled in the future: forth-like operation stacks. Also add the first two instructions to minimally get strings to work. In the parser, handle the variable declaration and make it available to the ivl_target.h code generator. The vvp code generator can use this information to generate the code for new vvp support.
2012-06-18 03:22:50 +02:00
/* Process string variables like string constants: interpret
the contained strings like format strings. */
case vpiStringVar:
value.format = vpiStringVal;
vpi_get_value(item, &value);
Allow strings to be constructed formats in $display et al.
2012-06-23 17:56:38 +02:00
fmt = strdup(value.value.str);
width = get_format(&result, fmt, info, &idx);
free(fmt);
First pass at getting strings to work. In vvp, create the .var/str variable for representing strings, and handle strings in the $display system task. Add to vvp threads the concept of a stack of strings. This is going to be how complex objects are to me handled in the future: forth-like operation stacks. Also add the first two instructions to minimally get strings to work. In the parser, handle the variable declaration and make it available to the ivl_target.h code generator. The vvp code generator can use this information to generate the code for new vvp support.
2012-06-18 03:22:50 +02:00
rtn = realloc(rtn, (size+width)*sizeof(char));
Allow strings to be constructed formats in $display et al.
2012-06-23 17:56:38 +02:00
memcpy(rtn+size-1, result, width);
Fix memory leak when displaying a SV string variable.
2013-01-03 03:30:59 +01:00
free(result);
First pass at getting strings to work. In vvp, create the .var/str variable for representing strings, and handle strings in the $display system task. Add to vvp threads the concept of a stack of strings. This is going to be how complex objects are to me handled in the future: forth-like operation stacks. Also add the first two instructions to minimally get strings to work. In the parser, handle the variable declaration and make it available to the ivl_target.h code generator. The vvp code generator can use this information to generate the code for new vvp support.
2012-06-18 03:22:50 +02:00
break;
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
case vpiSysFuncCall:
func_name = vpi_get_str(vpiName, item);
if (strcmp(func_name, "$time") == 0) {
value.format = vpiDecStrVal;
vpi_get_value(item, &value);
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
width = strlen(value.value.str);
if (width < 20) width = 20;
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
rtn = realloc(rtn, (size+width)*sizeof(char));
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
sprintf(rtn+size-1, "%*s", width, value.value.str);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
Make a plain $stime print in 10 digits (32 bits). This patch makes $stime a unique call in display and returns the lower 32 bits in a 10 digit field when it is called.
2009-01-09 04:10:15 +01:00
} else if (strcmp(func_name, "$stime") == 0) {
value.format = vpiDecStrVal;
vpi_get_value(item, &value);
width = strlen(value.value.str);
if (width < 10) width = 10;
rtn = realloc(rtn, (size+width)*sizeof(char));
sprintf(rtn+size-1, "%*s", width, value.value.str);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
} else if (strcmp(func_name, "$simtime") == 0) {
value.format = vpiDecStrVal;
vpi_get_value(item, &value);
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
width = strlen(value.value.str);
if (width < 20) width = 20;
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
rtn = realloc(rtn, (size+width)*sizeof(char));
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
sprintf(rtn+size-1, "%*s", width, value.value.str);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
} else if (strcmp(func_name, "$realtime") == 0) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
/* Use the local scope precision. */
int use_prec = vpi_get(vpiTimeUnit, info->scope) -
vpi_get(vpiTimePrecision, info->scope);
assert(use_prec >= 0);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
value.format = vpiRealVal;
vpi_get_value(item, &value);
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
sprintf(buf, "%.*f", use_prec, value.value.real);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
width = strlen(buf);
rtn = realloc(rtn, (size+width)*sizeof(char));
sprintf(rtn+size-1, "%*s", width, buf);
} else {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: %s does not support %s as an argument!\n",
info->filename, info->lineno, info->name, func_name);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
strcpy(buf, "<?>");
width = strlen(buf);
rtn = realloc(rtn, (size+width)*sizeof(char));
memcpy(rtn+size-1, buf, width);
}
break;
default:
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
vpi_printf("WARNING: %s:%d: unknown argument type (%s) given to %s!\n",
info->filename, info->lineno, vpi_get_str(vpiType, item),
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
info->name);
Down payment on const-correctness Clears out the unambiguous and easy-to-fix const faults, so we can focus better on the fundamental ones.
2010-09-27 20:03:43 +02:00
cresult = "<?>";
width = strlen(cresult);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
rtn = realloc(rtn, (size+width)*sizeof(char));
Down payment on const-correctness Clears out the unambiguous and easy-to-fix const faults, so we can focus better on the fundamental ones.
2010-09-27 20:03:43 +02:00
memcpy(rtn+size-1, cresult, width);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
break;
}
size += width;
}
rtn[size-1] = '\0';
*rtnsz = size - 1;
return rtn;
}
Stopgap fix for br916. Currently, when a variable expression is passed to a system task, the expression value is stored in thread memory. Values stored in thread memory cannot safely be passed to $strobe or $monitor, because the thread memory may get reused or deallocated before the $strobe or $monitor task actually executes. As a temporary measure, we just trap this case and terminate with a "sorry" message. A proper fix would require the expression value to be calculated at the time the $strobe or $monitor executes, not at the time it is called.
2013-02-02 19:09:38 +01:00
#ifdef BR916_STOPGAP_FIX
static char br916_hint_issued = 0;
#endif
Add possibility of const-correctness Follow-up to "Brainless start to const-correct changes" Still actually does nothing, but now if the #define ICARUS_VPI_CONST in vpi_user.h is changed to const, Icarus is almost const-correct, as checked with gcc flags -Wcast-qual -Wwrite-strings. Choosing when to #define ICARUS_VPI_CONST const is left as an exercise for the reader. With these two patches applied, and the const define, there are still about a dozen const problems left.
2010-10-04 18:53:55 +02:00
static int sys_check_args(vpiHandle callh, vpiHandle argv, const PLI_BYTE8*name,
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
int no_auto, int is_monitor)
{
vpiHandle arg;
int ret = 0;
/* If there are no arguments, just return. */
if (argv == 0) return ret;
for (arg = vpi_scan(argv); arg; arg = vpi_scan(argv)) {
if (no_auto && vpi_get(vpiAutomatic, arg)) {
vpi_printf("ERROR: %s:%d: ", vpi_get_str(vpiFile, callh),
(int)vpi_get(vpiLineNo, callh));
vpi_printf("%s argument \"%s\" is an automatic variable.\n",
name, vpi_get_str(vpiName, arg));
ret = 1;
}
switch (vpi_get(vpiType, arg)) {
case vpiMemoryWord:
case vpiPartSelect:
if (is_monitor && vpi_get(vpiConstantSelect, arg) == 0) {
vpi_printf("SORRY: %s:%d: ",
vpi_get_str(vpiFile, callh),
(int)vpi_get(vpiLineNo, callh));
vpi_printf("%s must have a constant %s select.\n",
name, vpi_get_str(vpiType, arg));
ret = 1;
}
Fix implicit fallthrough warnings when building with recent GCC.
2018-10-06 21:13:31 +02:00
// fallthrough
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
case vpiConstant:
case vpiParameter:
case vpiNet:
case vpiReg:
case vpiIntegerVar:
Add vpiBitVar support is $display and other vpi functions.
2011-04-04 02:22:12 +02:00
case vpiBitVar:
VPI access to atom2 types. Create the .var/2u and .var/2s variable records and give them basic implementations. Make available to VPI the proper types for the SystemVerilog types that these variables represent.
2010-10-08 03:01:11 +02:00
case vpiByteVar:
case vpiShortIntVar:
case vpiIntVar:
case vpiLongIntVar:
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
case vpiTimeVar:
case vpiRealVar:
First pass at getting strings to work. In vvp, create the .var/str variable for representing strings, and handle strings in the $display system task. Add to vvp threads the concept of a stack of strings. This is going to be how complex objects are to me handled in the future: forth-like operation stacks. Also add the first two instructions to minimally get strings to work. In the parser, handle the variable declaration and make it available to the ivl_target.h code generator. The vvp code generator can use this information to generate the code for new vvp support.
2012-06-18 03:22:50 +02:00
case vpiStringVar:
Stopgap fix for br916. Currently, when a variable expression is passed to a system task, the expression value is stored in thread memory. Values stored in thread memory cannot safely be passed to $strobe or $monitor, because the thread memory may get reused or deallocated before the $strobe or $monitor task actually executes. As a temporary measure, we just trap this case and terminate with a "sorry" message. A proper fix would require the expression value to be calculated at the time the $strobe or $monitor executes, not at the time it is called.
2013-02-02 19:09:38 +01:00
#ifdef BR916_STOPGAP_FIX
// no_auto implies either $strobe or $monitor
if (no_auto) {
Minor fixes to br916 patch.
2013-02-04 20:41:22 +01:00
switch (vpi_get(_vpiFromThr, arg)) {
Stopgap fix for br916. Currently, when a variable expression is passed to a system task, the expression value is stored in thread memory. Values stored in thread memory cannot safely be passed to $strobe or $monitor, because the thread memory may get reused or deallocated before the $strobe or $monitor task actually executes. As a temporary measure, we just trap this case and terminate with a "sorry" message. A proper fix would require the expression value to be calculated at the time the $strobe or $monitor executes, not at the time it is called.
2013-02-02 19:09:38 +01:00
case _vpiVThr:
case _vpiWord:
case _vpiString:
vpi_printf("SORRY: %s:%d: ",
vpi_get_str(vpiFile, callh),
(int)vpi_get(vpiLineNo, callh));
vpi_printf("currently only simple signals or constant "
"expressions may be passed to %s.\n", name);
if (!br916_hint_issued) {
Minor fixes to br916 patch.
2013-02-04 20:41:22 +01:00
vpi_printf("NOTE: You can work around this by "
"assigning the desired expression "
"to an\n"
" intermediate net (using a "
"continuous assignment) and passing "
"that net\n"
" to %s.\n", name);
br916_hint_issued = 1;
Stopgap fix for br916. Currently, when a variable expression is passed to a system task, the expression value is stored in thread memory. Values stored in thread memory cannot safely be passed to $strobe or $monitor, because the thread memory may get reused or deallocated before the $strobe or $monitor task actually executes. As a temporary measure, we just trap this case and terminate with a "sorry" message. A proper fix would require the expression value to be calculated at the time the $strobe or $monitor executes, not at the time it is called.
2013-02-02 19:09:38 +01:00
}
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
ret = 2;
Stopgap fix for br916. Currently, when a variable expression is passed to a system task, the expression value is stored in thread memory. Values stored in thread memory cannot safely be passed to $strobe or $monitor, because the thread memory may get reused or deallocated before the $strobe or $monitor task actually executes. As a temporary measure, we just trap this case and terminate with a "sorry" message. A proper fix would require the expression value to be calculated at the time the $strobe or $monitor executes, not at the time it is called.
2013-02-02 19:09:38 +01:00
default:
break;
}
}
#endif
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
case vpiClassVar:
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
case vpiSysFuncCall:
break;
default:
vpi_printf("ERROR: %s:%d: ", vpi_get_str(vpiFile, callh),
(int)vpi_get(vpiLineNo, callh));
vpi_printf("%s does not support argument type (%s).\n", name,
vpi_get_str(vpiType, arg));
ret = 1;
break;
}
}
return ret;
}
Factor out some common compiletf code in vpi/sys_display.c This patch just factors out the common code for two of the compiletf routines.
2009-01-17 00:40:18 +01:00
/* Common compiletf routine. */
Brainless start to const-correct changes Results of running cd vpi for f in *.c *.h; do sed -i \ -e "s/_calltf(PLI_BYTE8/_calltf(ICARUS_VPI_CONST PLI_BYTE8/" \ -e "s/_compiletf(PLI_BYTE8/_compiletf(ICARUS_VPI_CONST PLI_BYTE8/" $f; done and a trivial patch to vpi_user.h to, among other things, make ICARUS_VPI_CONST blank. Thus, this patch does absolutely nothing. Will be followed by a (much shorter) patch that makes it do something. :-)
2010-10-04 18:31:01 +02:00
static PLI_INT32 sys_common_compiletf(ICARUS_VPI_CONST PLI_BYTE8*name, int no_auto,
Factor out some common compiletf code in vpi/sys_display.c This patch just factors out the common code for two of the compiletf routines.
2009-01-17 00:40:18 +01:00
int is_monitor)
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
{
vpiHandle callh, argv;
callh = vpi_handle(vpiSysTfCall, 0);
argv = vpi_iterate(vpiArgument, callh);
Factor out the common code for checking if a FD/MCD is valid
2020-05-17 03:02:46 +02:00
if (name[1] == 'f') {
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
/* Check that there is a fd/mcd and that it is numeric. */
if (argv == 0) {
vpi_printf("ERROR: %s:%d: ", vpi_get_str(vpiFile, callh),
(int)vpi_get(vpiLineNo, callh));
vpi_printf("%s requires at least a file descriptor/MCD.\n",
name);
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
vpip_set_return_value(1);
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
vpi_control(vpiFinish, 1);
return 0;
}
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
if (! is_numeric_obj(vpi_scan(argv))) {
vpi_printf("ERROR: %s:%d: ", vpi_get_str(vpiFile, callh),
(int)vpi_get(vpiLineNo, callh));
vpi_printf("%s's file descriptor/MCD must be numeric.\n",
name);
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
vpip_set_return_value(1);
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
vpi_control(vpiFinish, 1);
}
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
}
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
int rtn = sys_check_args(callh, argv, name, no_auto, is_monitor);
if (rtn) {
if (rtn==1) vpip_set_return_value(1);
Factor out some common compiletf code in vpi/sys_display.c This patch just factors out the common code for two of the compiletf routines.
2009-01-17 00:40:18 +01:00
vpi_control(vpiFinish, 1);
}
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
return 0;
}
Factor out some common compiletf code in vpi/sys_display.c This patch just factors out the common code for two of the compiletf routines.
2009-01-17 00:40:18 +01:00
/* Check the $display, $write, $fdisplay and $fwrite based tasks. */
Brainless start to const-correct changes Results of running cd vpi for f in *.c *.h; do sed -i \ -e "s/_calltf(PLI_BYTE8/_calltf(ICARUS_VPI_CONST PLI_BYTE8/" \ -e "s/_compiletf(PLI_BYTE8/_compiletf(ICARUS_VPI_CONST PLI_BYTE8/" $f; done and a trivial patch to vpi_user.h to, among other things, make ICARUS_VPI_CONST blank. Thus, this patch does absolutely nothing. Will be followed by a (much shorter) patch that makes it do something. :-)
2010-10-04 18:31:01 +02:00
static PLI_INT32 sys_display_compiletf(ICARUS_VPI_CONST PLI_BYTE8*name)
Factor out some common compiletf code in vpi/sys_display.c This patch just factors out the common code for two of the compiletf routines.
2009-01-17 00:40:18 +01:00
{
/* These tasks can have automatic variables and are not monitor. */
return sys_common_compiletf(name, 0, 0);
}
vpi: $sformatf() function.
2015-12-11 12:58:25 +01:00
/* This implements the $sformatf, $display/$fdisplay
* and the $write/$fwrite based tasks. */
Brainless start to const-correct changes Results of running cd vpi for f in *.c *.h; do sed -i \ -e "s/_calltf(PLI_BYTE8/_calltf(ICARUS_VPI_CONST PLI_BYTE8/" \ -e "s/_compiletf(PLI_BYTE8/_compiletf(ICARUS_VPI_CONST PLI_BYTE8/" $f; done and a trivial patch to vpi_user.h to, among other things, make ICARUS_VPI_CONST blank. Thus, this patch does absolutely nothing. Will be followed by a (much shorter) patch that makes it do something. :-)
2010-10-04 18:31:01 +02:00
static PLI_INT32 sys_display_calltf(ICARUS_VPI_CONST PLI_BYTE8 *name)
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
{
Make the $fdisplay/$fwrite family use the string formatting. This patch makes the $fdisplay() and $fwrite() family of task use the string formatting code. It also adds a compiletf routine for the tasks and a run time check that the FD or MCD is correct. If not it fails instead of just dropping the output. It also cleans up a couple of memory leaks and adds no argument compiletf routines for the $monitoron/$monitoroff tasks.
2009-01-10 02:27:27 +01:00
vpiHandle callh, argv, scope;
struct strobe_cb_info info;
char* result;
Use fwrite to write $display output, instead of fprintf This change makes it safe to write non-ascii characters, which is an issue when the %u format is used.
2015-06-03 19:00:35 +02:00
unsigned int size;
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
PLI_UINT32 fd_mcd;
vpi: $sformatf() function.
2015-12-11 12:58:25 +01:00
s_vpi_value val;
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
Make the $fdisplay/$fwrite family use the string formatting. This patch makes the $fdisplay() and $fwrite() family of task use the string formatting code. It also adds a compiletf routine for the tasks and a run time check that the FD or MCD is correct. If not it fails instead of just dropping the output. It also cleans up a couple of memory leaks and adds no argument compiletf routines for the $monitoron/$monitoroff tasks.
2009-01-10 02:27:27 +01:00
callh = vpi_handle(vpiSysTfCall, 0);
argv = vpi_iterate(vpiArgument, callh);
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
/* Get the file/MC descriptor and verify it is valid. */
Factor out the common code for checking if a FD/MCD is valid
2020-05-17 03:02:46 +02:00
if (name[1] == 'f') {
Use a common routine to get and check the FD/MCD
2020-05-17 06:05:44 +02:00
vpiHandle fd = vpi_scan(argv);
if (get_fd_mcd_from_arg(&fd_mcd, fd, callh, name)) {
Factor out the common code for checking if a FD/MCD is valid
2020-05-17 03:02:46 +02:00
vpi_free_object(argv);
return 0;
}
Use a common routine to get and check the FD/MCD
2020-05-17 06:05:44 +02:00
} else if (strncmp(name, "$sformatf", 9) == 0) {
vpi: $sformatf() function.
2015-12-11 12:58:25 +01:00
/* return as a string */
Factor out the common code for checking if a FD/MCD is valid
2020-05-17 03:02:46 +02:00
fd_mcd = 0;
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
} else {
vpi: $sformatf() function.
2015-12-11 12:58:25 +01:00
/* stdout */
Factor out the common code for checking if a FD/MCD is valid
2020-05-17 03:02:46 +02:00
fd_mcd = 1;
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
}
Make the $fdisplay/$fwrite family use the string formatting. This patch makes the $fdisplay() and $fwrite() family of task use the string formatting code. It also adds a compiletf routine for the tasks and a run time check that the FD or MCD is correct. If not it fails instead of just dropping the output. It also cleans up a couple of memory leaks and adds no argument compiletf routines for the $monitoron/$monitoroff tasks.
2009-01-10 02:27:27 +01:00
scope = vpi_handle(vpiScope, callh);
assert(scope);
/* We could use vpi_get_str(vpiName, callh) to get the task name,
* but name is already defined. */
info.name = name;
info.filename = strdup(vpi_get_str(vpiFile, callh));
info.lineno = (int)vpi_get(vpiLineNo, callh);
info.default_format = get_default_format(name);
info.scope = scope;
array_from_iterator(&info, argv);
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
Make the $fdisplay/$fwrite family use the string formatting. This patch makes the $fdisplay() and $fwrite() family of task use the string formatting code. It also adds a compiletf routine for the tasks and a run time check that the FD or MCD is correct. If not it fails instead of just dropping the output. It also cleans up a couple of memory leaks and adds no argument compiletf routines for the $monitoron/$monitoroff tasks.
2009-01-10 02:27:27 +01:00
/* Because %u and %z may put embedded NULL characters into the
* returned string strlen() may not match the real size! */
result = get_display(&size, &info);
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
Factor out the common code for checking if a FD/MCD is valid
2020-05-17 03:02:46 +02:00
if (fd_mcd > 0) {
my_mcd_rawwrite(fd_mcd, result, size);
if ((strncmp(name,"$display",8) == 0) ||
(strncmp(name,"$fdisplay",9) == 0)) my_mcd_rawwrite(fd_mcd, "\n", 1);
vpi: $sformatf() function.
2015-12-11 12:58:25 +01:00
} else {
Factor out the common code for checking if a FD/MCD is valid
2020-05-17 03:02:46 +02:00
/* Return as a string ($sformatf) */
val.format = vpiStringVal;
val.value.str = result;
vpi_put_value(callh, &val, 0, vpiNoDelay);
vpi: $sformatf() function.
2015-12-11 12:58:25 +01:00
}
free(result);
Make the $fdisplay/$fwrite family use the string formatting. This patch makes the $fdisplay() and $fwrite() family of task use the string formatting code. It also adds a compiletf routine for the tasks and a run time check that the FD or MCD is correct. If not it fails instead of just dropping the output. It also cleans up a couple of memory leaks and adds no argument compiletf routines for the $monitoron/$monitoroff tasks.
2009-01-10 02:27:27 +01:00
free(info.filename);
free(info.items);
return 0;
}
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
/*
* The strobe implementation takes the parameter handles that are
* passed to the calltf and puts them in to an array for safe
* keeping. That array (and other bookkeeping) is passed, via the
* struct_cb_info object, to the REadOnlySych function strobe_cb,
* where it is used to perform the actual formatting and printing.
*/
static PLI_INT32 strobe_cb(p_cb_data cb)
Make the $fdisplay/$fwrite family use the string formatting. This patch makes the $fdisplay() and $fwrite() family of task use the string formatting code. It also adds a compiletf routine for the tasks and a run time check that the FD or MCD is correct. If not it fails instead of just dropping the output. It also cleans up a couple of memory leaks and adds no argument compiletf routines for the $monitoron/$monitoroff tasks.
2009-01-10 02:27:27 +01:00
{
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
struct strobe_cb_info*info = (struct strobe_cb_info*)cb->user_data;
Add full $ferror() support, bug fixes, etc. This patch adds $ferror() support. This is enhanced over what the standard says needs to be supported. All file I/O routines set/clear the errno flag, This allows $ferror() to check when an invalid fd/MCD is passed to a system task. The $ferror() fd is only checked if errno is not set since all the other routines will set EBADF when they are called. This allows us to test the output from $fopen() which sets the fd to zero (an invalid fd) when it has a problem. The following bugs (changes) are also fixed: Fix $fdisplay*, $fwrite*, $fstrobe*, $fflush, $fclose to not accept 0 as a valid MCD. For the $fdisplay*, $fwrite* and $fstrobe tasks an invalid fd/MCD is now a warning instead of an error like the other file I/O routines. Modify $fputc() to be a function and return a value like $ungetc(). Modify $ungetc() to return a value like the standard dictates (-1 on error else 0) and pass the value as a full int to get EOF as an error. Remove some extra checks that could not happen in $fseek() and to be safe convert the numeric codes to the system symbolic values. Skip $fstrobe callback when the fd/MCD is closed. The standard says we need to cancel the callback, but we currently do not have the ability to do that so just skipping this will work for now. Free the argument handle when $fscanf() is called with an invalid fd (memory leak).
2009-07-31 22:06:49 +02:00
/* We really need to cancel any $fstrobe() calls for a file when it
* is closed, but for now we will just skip processing the result.
* Which has the same basic effect. */
Factor out the common code for checking if a FD/MCD is valid
2020-05-17 03:02:46 +02:00
if (is_valid_fd_mcd(info->fd_mcd)) {
Remove some cppcheck warnings, etc.
2012-11-13 03:13:41 +01:00
char* result;
Use fwrite to write $display output, instead of fprintf This change makes it safe to write non-ascii characters, which is an issue when the %u format is used.
2015-06-03 19:00:35 +02:00
unsigned int size;
Add full $ferror() support, bug fixes, etc. This patch adds $ferror() support. This is enhanced over what the standard says needs to be supported. All file I/O routines set/clear the errno flag, This allows $ferror() to check when an invalid fd/MCD is passed to a system task. The $ferror() fd is only checked if errno is not set since all the other routines will set EBADF when they are called. This allows us to test the output from $fopen() which sets the fd to zero (an invalid fd) when it has a problem. The following bugs (changes) are also fixed: Fix $fdisplay*, $fwrite*, $fstrobe*, $fflush, $fclose to not accept 0 as a valid MCD. For the $fdisplay*, $fwrite* and $fstrobe tasks an invalid fd/MCD is now a warning instead of an error like the other file I/O routines. Modify $fputc() to be a function and return a value like $ungetc(). Modify $ungetc() to return a value like the standard dictates (-1 on error else 0) and pass the value as a full int to get EOF as an error. Remove some extra checks that could not happen in $fseek() and to be safe convert the numeric codes to the system symbolic values. Skip $fstrobe callback when the fd/MCD is closed. The standard says we need to cancel the callback, but we currently do not have the ability to do that so just skipping this will work for now. Free the argument handle when $fscanf() is called with an invalid fd (memory leak).
2009-07-31 22:06:49 +02:00
/* Because %u and %z may put embedded NULL characters into the
* returned string strlen() may not match the real size! */
result = get_display(&size, info);
Use fwrite to write $display output, instead of fprintf This change makes it safe to write non-ascii characters, which is an issue when the %u format is used.
2015-06-03 19:00:35 +02:00
my_mcd_rawwrite(info->fd_mcd, result, size);
my_mcd_rawwrite(info->fd_mcd, "\n", 1);
Fix a few more cppcheck issues. Here are some more cleanups recommended by cppcheck. Mostly fixing the variable scope and removing some unneeded code.
2010-04-19 19:25:42 +02:00
free(result);
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
}
free(info->filename);
free(info->items);
free(info);
return 0;
}
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
/* Check both the $strobe and $fstrobe based tasks. */
Brainless start to const-correct changes Results of running cd vpi for f in *.c *.h; do sed -i \ -e "s/_calltf(PLI_BYTE8/_calltf(ICARUS_VPI_CONST PLI_BYTE8/" \ -e "s/_compiletf(PLI_BYTE8/_compiletf(ICARUS_VPI_CONST PLI_BYTE8/" $f; done and a trivial patch to vpi_user.h to, among other things, make ICARUS_VPI_CONST blank. Thus, this patch does absolutely nothing. Will be followed by a (much shorter) patch that makes it do something. :-)
2010-10-04 18:31:01 +02:00
static PLI_INT32 sys_strobe_compiletf(ICARUS_VPI_CONST PLI_BYTE8 *name)
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
{
Factor out some common compiletf code in vpi/sys_display.c This patch just factors out the common code for two of the compiletf routines.
2009-01-17 00:40:18 +01:00
/* These tasks can not have automatic variables and are not monitor. */
return sys_common_compiletf(name, 1, 0);
Make the $fdisplay/$fwrite family use the string formatting. This patch makes the $fdisplay() and $fwrite() family of task use the string formatting code. It also adds a compiletf routine for the tasks and a run time check that the FD or MCD is correct. If not it fails instead of just dropping the output. It also cleans up a couple of memory leaks and adds no argument compiletf routines for the $monitoron/$monitoroff tasks.
2009-01-10 02:27:27 +01:00
}
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
/* This implements both the $strobe and $fstrobe based tasks. */
Brainless start to const-correct changes Results of running cd vpi for f in *.c *.h; do sed -i \ -e "s/_calltf(PLI_BYTE8/_calltf(ICARUS_VPI_CONST PLI_BYTE8/" \ -e "s/_compiletf(PLI_BYTE8/_compiletf(ICARUS_VPI_CONST PLI_BYTE8/" $f; done and a trivial patch to vpi_user.h to, among other things, make ICARUS_VPI_CONST blank. Thus, this patch does absolutely nothing. Will be followed by a (much shorter) patch that makes it do something. :-)
2010-10-04 18:31:01 +02:00
static PLI_INT32 sys_strobe_calltf(ICARUS_VPI_CONST PLI_BYTE8*name)
Make the $fdisplay/$fwrite family use the string formatting. This patch makes the $fdisplay() and $fwrite() family of task use the string formatting code. It also adds a compiletf routine for the tasks and a run time check that the FD or MCD is correct. If not it fails instead of just dropping the output. It also cleans up a couple of memory leaks and adds no argument compiletf routines for the $monitoron/$monitoroff tasks.
2009-01-10 02:27:27 +01:00
{
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
vpiHandle callh, argv, scope;
struct t_cb_data cb;
Fix shadow warnings found on OpenBSD. gcc on OpenBSD reported shadow warnings for variables, arguments named log, time and exp. This patch renanes those variables to logic, timerec and expr.
2010-05-28 05:09:32 +02:00
struct t_vpi_time timerec;
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
struct strobe_cb_info*info;
Make the $fdisplay/$fwrite family use the string formatting. This patch makes the $fdisplay() and $fwrite() family of task use the string formatting code. It also adds a compiletf routine for the tasks and a run time check that the FD or MCD is correct. If not it fails instead of just dropping the output. It also cleans up a couple of memory leaks and adds no argument compiletf routines for the $monitoron/$monitoroff tasks.
2009-01-10 02:27:27 +01:00
PLI_UINT32 fd_mcd;
callh = vpi_handle(vpiSysTfCall, 0);
argv = vpi_iterate(vpiArgument, callh);
/* Get the file/MC descriptor and verify it is valid. */
Factor out the common code for checking if a FD/MCD is valid
2020-05-17 03:02:46 +02:00
if (name[1] == 'f') {
Use a common routine to get and check the FD/MCD
2020-05-17 06:05:44 +02:00
vpiHandle fd = vpi_scan(argv);
if (get_fd_mcd_from_arg(&fd_mcd, fd, callh, name)) {
Factor out the common code for checking if a FD/MCD is valid
2020-05-17 03:02:46 +02:00
vpi_free_object(argv);
return 0;
}
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
} else {
Factor out the common code for checking if a FD/MCD is valid
2020-05-17 03:02:46 +02:00
fd_mcd = 1;
Make the $fdisplay/$fwrite family use the string formatting. This patch makes the $fdisplay() and $fwrite() family of task use the string formatting code. It also adds a compiletf routine for the tasks and a run time check that the FD or MCD is correct. If not it fails instead of just dropping the output. It also cleans up a couple of memory leaks and adds no argument compiletf routines for the $monitoron/$monitoroff tasks.
2009-01-10 02:27:27 +01:00
}
scope = vpi_handle(vpiScope, callh);
assert(scope);
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
info = calloc(1, sizeof(struct strobe_cb_info));
info->fd_mcd = fd_mcd;
Make the $fdisplay/$fwrite family use the string formatting. This patch makes the $fdisplay() and $fwrite() family of task use the string formatting code. It also adds a compiletf routine for the tasks and a run time check that the FD or MCD is correct. If not it fails instead of just dropping the output. It also cleans up a couple of memory leaks and adds no argument compiletf routines for the $monitoron/$monitoroff tasks.
2009-01-10 02:27:27 +01:00
/* We could use vpi_get_str(vpiName, callh) to get the task name,
* but name is already defined. */
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
info->name = name;
info->filename = strdup(vpi_get_str(vpiFile, callh));
info->lineno = (int)vpi_get(vpiLineNo, callh);
info->default_format = get_default_format(name);
info->scope= scope;
array_from_iterator(info, argv);
Make the $fdisplay/$fwrite family use the string formatting. This patch makes the $fdisplay() and $fwrite() family of task use the string formatting code. It also adds a compiletf routine for the tasks and a run time check that the FD or MCD is correct. If not it fails instead of just dropping the output. It also cleans up a couple of memory leaks and adds no argument compiletf routines for the $monitoron/$monitoroff tasks.
2009-01-10 02:27:27 +01:00
Fix shadow warnings found on OpenBSD. gcc on OpenBSD reported shadow warnings for variables, arguments named log, time and exp. This patch renanes those variables to logic, timerec and expr.
2010-05-28 05:09:32 +02:00
timerec.type = vpiSimTime;
timerec.low = 0;
timerec.high = 0;
Make the $fdisplay/$fwrite family use the string formatting. This patch makes the $fdisplay() and $fwrite() family of task use the string formatting code. It also adds a compiletf routine for the tasks and a run time check that the FD or MCD is correct. If not it fails instead of just dropping the output. It also cleans up a couple of memory leaks and adds no argument compiletf routines for the $monitoron/$monitoroff tasks.
2009-01-10 02:27:27 +01:00
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
cb.reason = cbReadOnlySynch;
cb.cb_rtn = strobe_cb;
Fix shadow warnings found on OpenBSD. gcc on OpenBSD reported shadow warnings for variables, arguments named log, time and exp. This patch renanes those variables to logic, timerec and expr.
2010-05-28 05:09:32 +02:00
cb.time = &timerec;
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
cb.obj = 0;
cb.value = 0;
cb.user_data = (char*)info;
vpi_register_cb(&cb);
Factor out the common code for checking if a FD/MCD is valid
2020-05-17 03:02:46 +02:00
Make the $fdisplay/$fwrite family use the string formatting. This patch makes the $fdisplay() and $fwrite() family of task use the string formatting code. It also adds a compiletf routine for the tasks and a run time check that the FD or MCD is correct. If not it fails instead of just dropping the output. It also cleans up a couple of memory leaks and adds no argument compiletf routines for the $monitoron/$monitoroff tasks.
2009-01-10 02:27:27 +01:00
return 0;
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
}
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
/*
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
* The monitor_info array records info for all the active monitor tasks.
* monitor_info[0] points to the info for the most recently called $monitor
* task (any preceding call to $monitor gets cancelled by that call) and is
* allocated before we register the VPI tasks. Further entries in monitor_info
* are allocated on demand by calls to $fmonitor.
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
*/
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
struct monitor_cb_info {
struct strobe_cb_info strobe;
vpiHandle *callbacks;
vpiHandle scheduled;
int disabled;
};
static struct monitor_cb_info**monitor_info = 0;
static unsigned monitor_info_len = 0;
static struct monitor_cb_info*allocate_monitor_info(void)
{
unsigned idx;
for (idx = 0; idx < monitor_info_len; idx += 1) {
if (monitor_info[idx] == 0)
goto found_empty_slot;
}
monitor_info = realloc(monitor_info, ++monitor_info_len * sizeof(struct monitor_cb_info*));
found_empty_slot:
monitor_info[idx] = calloc(1, sizeof(struct monitor_cb_info));
return monitor_info[idx];
}
static void cancel_monitor(struct monitor_cb_info*monitor)
{
unsigned idx;
if (monitor == 0) return;
if (monitor->scheduled) {
vpi_remove_cb(monitor->scheduled);
monitor->scheduled = 0;
}
for (idx = 0; idx < monitor->strobe.nitems; idx += 1) {
if (monitor->callbacks[idx])
vpi_remove_cb(monitor->callbacks[idx]);
}
free(monitor->callbacks);
monitor->callbacks = 0;
free(monitor->strobe.filename);
free(monitor->strobe.items);
monitor->strobe.items = 0;
monitor->strobe.nitems = 0;
monitor->strobe.name = 0;
}
static void cancel_fmonitor(unsigned idx)
{
cancel_monitor(monitor_info[idx]);
free(monitor_info[idx]);
monitor_info[idx] = 0;
}
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
static PLI_INT32 monitor_cb_2(p_cb_data cb)
{
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
struct monitor_cb_info*monitor = (struct monitor_cb_info*)cb->user_data;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
char* result;
Use fwrite to write $display output, instead of fprintf This change makes it safe to write non-ascii characters, which is an issue when the %u format is used.
2015-06-03 19:00:35 +02:00
unsigned int size;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
/* Because %u and %z may put embedded NULL characters into the
* returned string strlen() may not match the real size! */
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
result = get_display(&size, &(monitor->strobe));
my_mcd_rawwrite(monitor->strobe.fd_mcd, result, size);
my_mcd_rawwrite(monitor->strobe.fd_mcd, "\n", 1);
monitor->scheduled = 0;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
free(result);
return 0;
}
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
static void schedule_monitor_display(struct monitor_cb_info*monitor)
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
{
struct t_cb_data cb;
Fix shadow warnings found on OpenBSD. gcc on OpenBSD reported shadow warnings for variables, arguments named log, time and exp. This patch renanes those variables to logic, timerec and expr.
2010-05-28 05:09:32 +02:00
struct t_vpi_time timerec;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
Fix shadow warnings found on OpenBSD. gcc on OpenBSD reported shadow warnings for variables, arguments named log, time and exp. This patch renanes those variables to logic, timerec and expr.
2010-05-28 05:09:32 +02:00
timerec.type = vpiSimTime;
timerec.low = 0;
timerec.high = 0;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
cb.reason = cbReadOnlySynch;
cb.cb_rtn = monitor_cb_2;
Fix shadow warnings found on OpenBSD. gcc on OpenBSD reported shadow warnings for variables, arguments named log, time and exp. This patch renanes those variables to logic, timerec and expr.
2010-05-28 05:09:32 +02:00
cb.time = &timerec;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
cb.obj = 0;
cb.value = 0;
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
cb.user_data = (char*)monitor;
monitor->scheduled = vpi_register_cb(&cb);
}
/*
* The monitor_cb_1 callback is called when a monitored value changes.
* All this function does is schedule the actual display to occur in a
* ReadOnlySync callback. The scheduled event handle in the monitor info
* is used to ensure only one display is scheduled in each time slot,
* and can also be used to cancel the display event if the monitor gets
* cancelled.
*/
static PLI_INT32 monitor_cb_1(p_cb_data cb)
{
struct monitor_cb_info*monitor = (struct monitor_cb_info*)cb->user_data;
if (monitor->disabled) return 0;
if (monitor->scheduled) return 0;
schedule_monitor_display(monitor);
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
return 0;
}
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
/* Check both the $monitor and $fmonitor based tasks. */
Brainless start to const-correct changes Results of running cd vpi for f in *.c *.h; do sed -i \ -e "s/_calltf(PLI_BYTE8/_calltf(ICARUS_VPI_CONST PLI_BYTE8/" \ -e "s/_compiletf(PLI_BYTE8/_compiletf(ICARUS_VPI_CONST PLI_BYTE8/" $f; done and a trivial patch to vpi_user.h to, among other things, make ICARUS_VPI_CONST blank. Thus, this patch does absolutely nothing. Will be followed by a (much shorter) patch that makes it do something. :-)
2010-10-04 18:31:01 +02:00
static PLI_INT32 sys_monitor_compiletf(ICARUS_VPI_CONST PLI_BYTE8 *name)
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
{
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
/* These tasks can not have automatic variables and are monitor. */
return sys_common_compiletf(name, 1, 1);
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
}
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
/* This implements both the $monitor and $fmonitor based tasks. */
Brainless start to const-correct changes Results of running cd vpi for f in *.c *.h; do sed -i \ -e "s/_calltf(PLI_BYTE8/_calltf(ICARUS_VPI_CONST PLI_BYTE8/" \ -e "s/_compiletf(PLI_BYTE8/_compiletf(ICARUS_VPI_CONST PLI_BYTE8/" $f; done and a trivial patch to vpi_user.h to, among other things, make ICARUS_VPI_CONST blank. Thus, this patch does absolutely nothing. Will be followed by a (much shorter) patch that makes it do something. :-)
2010-10-04 18:31:01 +02:00
static PLI_INT32 sys_monitor_calltf(ICARUS_VPI_CONST PLI_BYTE8*name)
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
{
vpiHandle callh, argv, scope;
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
struct monitor_cb_info*monitor;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
struct t_cb_data cb;
Fix shadow warnings found on OpenBSD. gcc on OpenBSD reported shadow warnings for variables, arguments named log, time and exp. This patch renanes those variables to logic, timerec and expr.
2010-05-28 05:09:32 +02:00
struct t_vpi_time timerec;
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
PLI_UINT32 fd_mcd;
unsigned idx;
Remove missing parameter warnings
2014-07-09 23:16:57 +02:00
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
callh = vpi_handle(vpiSysTfCall, 0);
argv = vpi_iterate(vpiArgument, callh);
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
if (name[1] == 'f') {
vpiHandle fd = vpi_scan(argv);
if (get_fd_mcd_from_arg(&fd_mcd, fd, callh, name)) {
vpi_free_object(argv);
return 0;
}
monitor = allocate_monitor_info();
} else {
fd_mcd = 1;
assert(monitor_info);
monitor = monitor_info[0];
/* If there was a previous $monitor, cancel it. */
if (monitor->strobe.name)
cancel_monitor(monitor);
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
}
scope = vpi_handle(vpiScope, callh);
assert(scope);
/* Make an array of handles from the argument list. */
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
array_from_iterator(&(monitor->strobe), argv);
monitor->strobe.name = name;
monitor->strobe.filename = strdup(vpi_get_str(vpiFile, callh));
monitor->strobe.lineno = (int)vpi_get(vpiLineNo, callh);
monitor->strobe.default_format = get_default_format(name);
monitor->strobe.scope = scope;
monitor->strobe.fd_mcd = fd_mcd;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
/* Attach callbacks to all the parameters that might change. */
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
monitor->callbacks = calloc(monitor->strobe.nitems, sizeof(vpiHandle));
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
Fix shadow warnings found on OpenBSD. gcc on OpenBSD reported shadow warnings for variables, arguments named log, time and exp. This patch renanes those variables to logic, timerec and expr.
2010-05-28 05:09:32 +02:00
timerec.type = vpiSuppressTime;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
cb.reason = cbValueChange;
cb.cb_rtn = monitor_cb_1;
Fix shadow warnings found on OpenBSD. gcc on OpenBSD reported shadow warnings for variables, arguments named log, time and exp. This patch renanes those variables to logic, timerec and expr.
2010-05-28 05:09:32 +02:00
cb.time = &timerec;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
cb.value = NULL;
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
cb.user_data = (char*)monitor;
for (idx = 0 ; idx < monitor->strobe.nitems ; idx += 1) {
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
switch (vpi_get(vpiType, monitor->strobe.items[idx])) {
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
case vpiMemoryWord:
/*
* We only support constant selections. Make this
* better when we add a real compiletf routine.
*/
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
assert(vpi_get(vpiConstantSelect, monitor->strobe.items[idx]));
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
case vpiNet:
case vpiReg:
case vpiIntegerVar:
Add the ability to $display and $monitor a bit, byte, short, int or long var. Add code so that we can $display and $monitor the two-state variables.
2011-09-06 01:25:05 +02:00
case vpiBitVar:
case vpiByteVar:
case vpiShortIntVar:
case vpiIntVar:
case vpiLongIntVar:
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
case vpiRealVar:
case vpiPartSelect:
/* Monitoring reg and net values involves setting
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
a callback for value changes. */
cb.obj = monitor->strobe.items[idx];
monitor->callbacks[idx] = vpi_register_cb(&cb);
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
break;
}
}
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
/* When the monitor task is called, it schedules a first display
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
for the end of the current time, like a $strobe. */
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
schedule_monitor_display(monitor);
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
return 0;
}
Brainless start to const-correct changes Results of running cd vpi for f in *.c *.h; do sed -i \ -e "s/_calltf(PLI_BYTE8/_calltf(ICARUS_VPI_CONST PLI_BYTE8/" \ -e "s/_compiletf(PLI_BYTE8/_compiletf(ICARUS_VPI_CONST PLI_BYTE8/" $f; done and a trivial patch to vpi_user.h to, among other things, make ICARUS_VPI_CONST blank. Thus, this patch does absolutely nothing. Will be followed by a (much shorter) patch that makes it do something. :-)
2010-10-04 18:31:01 +02:00
static PLI_INT32 sys_monitoron_calltf(ICARUS_VPI_CONST PLI_BYTE8*name)
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
{
Remove missing parameter warnings
2014-07-09 23:16:57 +02:00
(void)name; /* Parameter is not used. */
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
assert(monitor_info);
monitor_info[0]->disabled = 0;
schedule_monitor_display(monitor_info[0]);
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
return 0;
}
Brainless start to const-correct changes Results of running cd vpi for f in *.c *.h; do sed -i \ -e "s/_calltf(PLI_BYTE8/_calltf(ICARUS_VPI_CONST PLI_BYTE8/" \ -e "s/_compiletf(PLI_BYTE8/_compiletf(ICARUS_VPI_CONST PLI_BYTE8/" $f; done and a trivial patch to vpi_user.h to, among other things, make ICARUS_VPI_CONST blank. Thus, this patch does absolutely nothing. Will be followed by a (much shorter) patch that makes it do something. :-)
2010-10-04 18:31:01 +02:00
static PLI_INT32 sys_monitoroff_calltf(ICARUS_VPI_CONST PLI_BYTE8*name)
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
{
Remove missing parameter warnings
2014-07-09 23:16:57 +02:00
(void)name; /* Parameter is not used. */
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
assert(monitor_info);
monitor_info[0]->disabled = 1;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
return 0;
}
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
void sys_monitor_fclose(PLI_UINT32 fd_mcd)
{
unsigned idx;
if (IS_MCD(fd_mcd)) {
// The LRM doesn't specify the exact behaviour in this case,
// but we want to cancel any active fmonitors where there is
// an intersection between the MCD being closed and the MCD
// being monitored. Otherwise a subsequent call to $fopen could
// reuse the bit(s) in the MCD and cause the monitor output to
// be written to the newly opened file(s).
for (idx = 1; idx < monitor_info_len; idx += 1) {
if (monitor_info[idx] && (monitor_info[idx]->strobe.fd_mcd & fd_mcd))
cancel_fmonitor(idx);
}
} else {
// Cancel any active fmonitors with a matching FD.
for (idx = 1; idx < monitor_info_len; idx += 1) {
if (monitor_info[idx] && (monitor_info[idx]->strobe.fd_mcd == fd_mcd))
cancel_fmonitor(idx);
}
}
}
Brainless start to const-correct changes Results of running cd vpi for f in *.c *.h; do sed -i \ -e "s/_calltf(PLI_BYTE8/_calltf(ICARUS_VPI_CONST PLI_BYTE8/" \ -e "s/_compiletf(PLI_BYTE8/_compiletf(ICARUS_VPI_CONST PLI_BYTE8/" $f; done and a trivial patch to vpi_user.h to, among other things, make ICARUS_VPI_CONST blank. Thus, this patch does absolutely nothing. Will be followed by a (much shorter) patch that makes it do something. :-)
2010-10-04 18:31:01 +02:00
static PLI_INT32 sys_swrite_compiletf(ICARUS_VPI_CONST PLI_BYTE8 *name)
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
{
vpiHandle callh = vpi_handle(vpiSysTfCall, 0);
vpiHandle argv = vpi_iterate(vpiArgument, callh);
Add support for a SV string to $swrite and $sformat
2015-02-28 20:45:44 +01:00
vpiHandle arg;
PLI_INT32 type;
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
/* Check that there are arguments. */
if (argv == 0) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("ERROR:%s:%d: ", vpi_get_str(vpiFile, callh),
(int)vpi_get(vpiLineNo, callh));
vpi_printf("%s requires at least one argument.\n", name);
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
vpip_set_return_value(1);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
vpi_control(vpiFinish, 1);
return 0;
}
Add support for a SV string to $swrite and $sformat
2015-02-28 20:45:44 +01:00
/* The first argument must be a register or a SV string. */
arg = vpi_scan(argv); /* This should never be zero. */
type = vpi_get(vpiType, arg);
if (type != vpiReg && type != vpiStringVar) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("ERROR:%s:%d: ", vpi_get_str(vpiFile, callh),
(int)vpi_get(vpiLineNo, callh));
Add support for a SV string to $swrite and $sformat
2015-02-28 20:45:44 +01:00
vpi_printf("%s's first argument must be a register or SV string.\n", name);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
vpi_control(vpiFinish, 1);
return 0;
}
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
int rtn = sys_check_args(callh, argv, name, 0, 0);
if (rtn) {
if (rtn == 1) vpip_set_return_value(1);
vpi_control(vpiFinish, 1);
}
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
return 0;
}
Brainless start to const-correct changes Results of running cd vpi for f in *.c *.h; do sed -i \ -e "s/_calltf(PLI_BYTE8/_calltf(ICARUS_VPI_CONST PLI_BYTE8/" \ -e "s/_compiletf(PLI_BYTE8/_compiletf(ICARUS_VPI_CONST PLI_BYTE8/" $f; done and a trivial patch to vpi_user.h to, among other things, make ICARUS_VPI_CONST blank. Thus, this patch does absolutely nothing. Will be followed by a (much shorter) patch that makes it do something. :-)
2010-10-04 18:31:01 +02:00
static PLI_INT32 sys_swrite_calltf(ICARUS_VPI_CONST PLI_BYTE8 *name)
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
{
vpiHandle callh, argv, reg, scope;
struct strobe_cb_info info;
s_vpi_value val;
unsigned int size;
callh = vpi_handle(vpiSysTfCall, 0);
argv = vpi_iterate(vpiArgument, callh);
reg = vpi_scan(argv);
scope = vpi_handle(vpiScope, callh);
assert(scope);
/* We could use vpi_get_str(vpiName, callh) to get the task name, but
* name is already defined. */
info.name = name;
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
info.filename = strdup(vpi_get_str(vpiFile, callh));
info.lineno = (int)vpi_get(vpiLineNo, callh);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
info.default_format = get_default_format(name);
info.scope = scope;
Improved automatic variable error reporting. This patch moves the checks for automatic variables being passed to the $monitor or $strobe system tasks from the calltf routine to a new compiletf routine. It also adds the file name and line number of an offending call to the error message.
2008-11-15 21:33:23 +01:00
array_from_iterator(&info, argv);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
/* Because %u and %z may put embedded NULL characters into the returned
* string strlen() may not match the real size! */
val.value.str = get_display(&size, &info);
val.format = vpiStringVal;
vpi_put_value(reg, &val, 0, vpiNoDelay);
if (size != strlen(val.value.str)) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: %s returned a value with an embedded NULL "
"(see %%u/%%z).\n", info.filename, info.lineno, name);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
}
Fix memory leaks in vvp and change vvp T_STRING token to be new based. This patch fixes a bunch of memory leaks in vvp and converts the T_STRING lexical token to be new based. There are still two known leaks that I need to find a way to fix and likely some unknown leaks that still need to be found and fixed.
2009-01-13 18:57:12 +01:00
free(val.value.str);
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
free(info.filename);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
free(info.items);
return 0;
}
Brainless start to const-correct changes Results of running cd vpi for f in *.c *.h; do sed -i \ -e "s/_calltf(PLI_BYTE8/_calltf(ICARUS_VPI_CONST PLI_BYTE8/" \ -e "s/_compiletf(PLI_BYTE8/_compiletf(ICARUS_VPI_CONST PLI_BYTE8/" $f; done and a trivial patch to vpi_user.h to, among other things, make ICARUS_VPI_CONST blank. Thus, this patch does absolutely nothing. Will be followed by a (much shorter) patch that makes it do something. :-)
2010-10-04 18:31:01 +02:00
static PLI_INT32 sys_sformat_compiletf(ICARUS_VPI_CONST PLI_BYTE8 *name)
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
{
vpiHandle callh = vpi_handle(vpiSysTfCall, 0);
vpiHandle argv = vpi_iterate(vpiArgument, callh);
vpiHandle arg;
PLI_INT32 type;
/* Check that there are arguments. */
if (argv == 0) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("ERROR:%s:%d: ", vpi_get_str(vpiFile, callh),
(int)vpi_get(vpiLineNo, callh));
vpi: $sformatf() function.
2015-12-11 12:58:25 +01:00
vpi_printf("%s requires at least two arguments.\n", name);
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
vpip_set_return_value(1);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
vpi_control(vpiFinish, 1);
return 0;
}
Add support for a SV string to $swrite and $sformat
2015-02-28 20:45:44 +01:00
/* The first argument must be a register or a SV string. */
Content-free portability fixes. These are patches from ITOH Yasufumi and Larry Doolittle of pr2120948. They change nothing, but make the code compile with older compilers.
2008-09-30 03:06:47 +02:00
arg = vpi_scan(argv); /* This should never be zero. */
Add support for a SV string to $swrite and $sformat
2015-02-28 20:45:44 +01:00
type = vpi_get(vpiType, arg);
if (type != vpiReg && type != vpiStringVar) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("ERROR:%s:%d: ", vpi_get_str(vpiFile, callh),
(int)vpi_get(vpiLineNo, callh));
Add support for a SV string to $swrite and $sformat
2015-02-28 20:45:44 +01:00
vpi_printf("%s's first argument must be a register or SV string.\n", name);
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
vpip_set_return_value(1);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
vpi_control(vpiFinish, 1);
return 0;
}
Add support for a SV string to $swrite and $sformat
2015-02-28 20:45:44 +01:00
/* The second argument must be a string, a register or a SV string. */
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
arg = vpi_scan(argv);
if (arg == 0) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("ERROR:%s:%d: ", vpi_get_str(vpiFile, callh),
(int)vpi_get(vpiLineNo, callh));
vpi: $sformatf() function.
2015-12-11 12:58:25 +01:00
vpi_printf("%s requires at least two arguments.\n", name);
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
vpip_set_return_value(1);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
vpi_control(vpiFinish, 1);
return 0;
}
type = vpi_get(vpiType, arg);
if (((type != vpiConstant && type != vpiParameter) ||
Add support for a SV string to $swrite and $sformat
2015-02-28 20:45:44 +01:00
vpi_get(vpiConstType, arg) != vpiStringConst) &&
type != vpiReg && type != vpiStringVar) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("ERROR:%s:%d: ", vpi_get_str(vpiFile, callh),
(int)vpi_get(vpiLineNo, callh));
vpi_printf("%s's second argument must be a string or a register.\n", name);
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
vpip_set_return_value(1);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
vpi_control(vpiFinish, 1);
return 0;
}
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
int rtn = sys_check_args(callh, argv, name, 0, 0);
if (rtn) {
if (rtn == 1) vpip_set_return_value(1);
vpi_control(vpiFinish, 1);
}
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
return 0;
}
Brainless start to const-correct changes Results of running cd vpi for f in *.c *.h; do sed -i \ -e "s/_calltf(PLI_BYTE8/_calltf(ICARUS_VPI_CONST PLI_BYTE8/" \ -e "s/_compiletf(PLI_BYTE8/_compiletf(ICARUS_VPI_CONST PLI_BYTE8/" $f; done and a trivial patch to vpi_user.h to, among other things, make ICARUS_VPI_CONST blank. Thus, this patch does absolutely nothing. Will be followed by a (much shorter) patch that makes it do something. :-)
2010-10-04 18:31:01 +02:00
static PLI_INT32 sys_sformat_calltf(ICARUS_VPI_CONST PLI_BYTE8 *name)
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
{
vpiHandle callh, argv, reg, scope;
struct strobe_cb_info info;
s_vpi_value val;
char *result, *fmt;
unsigned int idx, size;
callh = vpi_handle(vpiSysTfCall, 0);
argv = vpi_iterate(vpiArgument, callh);
reg = vpi_scan(argv);
val.format = vpiStringVal;
vpi_get_value(vpi_scan(argv), &val);
fmt = strdup(val.value.str);
scope = vpi_handle(vpiScope, callh);
assert(scope);
/* We could use vpi_get_str(vpiName, callh) to get the task name, but
* name is already defined. */
info.name = name;
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
info.filename = strdup(vpi_get_str(vpiFile, callh));
info.lineno = (int)vpi_get(vpiLineNo, callh);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
info.default_format = get_default_format(name);
info.scope = scope;
Improved automatic variable error reporting. This patch moves the checks for automatic variables being passed to the $monitor or $strobe system tasks from the calltf routine to a new compiletf routine. It also adds the file name and line number of an offending call to the error message.
2008-11-15 21:33:23 +01:00
array_from_iterator(&info, argv);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
idx = -1;
size = get_format(&result, fmt, &info, &idx);
free(fmt);
if (idx+1< info.nitems) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: %s has %d extra argument(s).\n",
info.filename, info.lineno, name,
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
info.nitems-idx-1);
}
val.value.str = result;
val.format = vpiStringVal;
vpi_put_value(reg, &val, 0, vpiNoDelay);
if (size != strlen(val.value.str)) {
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpi_printf("WARNING: %s:%d: %s returned a value with an embedded NULL "
"(see %%u/%%z).\n", info.filename, info.lineno, name);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
}
Fix memory leaks in vvp and change vvp T_STRING token to be new based. This patch fixes a bunch of memory leaks in vvp and converts the T_STRING lexical token to be new based. There are still two known leaks that I need to find a way to fix and likely some unknown leaks that still need to be found and fixed.
2009-01-13 18:57:12 +01:00
free(val.value.str);
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
free(info.filename);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
free(info.items);
return 0;
}
vpi: $sformatf() function.
2015-12-11 12:58:25 +01:00
static PLI_INT32 sys_sformatf_compiletf(ICARUS_VPI_CONST PLI_BYTE8 *name)
{
vpiHandle callh = vpi_handle(vpiSysTfCall, 0);
vpiHandle argv = vpi_iterate(vpiArgument, callh);
vpiHandle arg;
PLI_INT32 type;
/* Check that there are arguments. */
if (argv == 0) {
vpi_printf("ERROR:%s:%d: ", vpi_get_str(vpiFile, callh),
(int)vpi_get(vpiLineNo, callh));
vpi_printf("%s requires at least one argument.\n", name);
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
vpip_set_return_value(1);
vpi: $sformatf() function.
2015-12-11 12:58:25 +01:00
vpi_control(vpiFinish, 1);
return 0;
}
/* The first argument must be a string, a register or a SV string. */
arg = vpi_scan(argv);
if (arg == 0) {
vpi_printf("ERROR:%s:%d: ", vpi_get_str(vpiFile, callh),
(int)vpi_get(vpiLineNo, callh));
vpi_printf("%s requires at least one argument.\n", name);
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
vpip_set_return_value(1);
vpi: $sformatf() function.
2015-12-11 12:58:25 +01:00
vpi_control(vpiFinish, 1);
return 0;
}
type = vpi_get(vpiType, arg);
if (((type != vpiConstant && type != vpiParameter) ||
vpi_get(vpiConstType, arg) != vpiStringConst) &&
type != vpiReg && type != vpiStringVar) {
vpi_printf("ERROR:%s:%d: ", vpi_get_str(vpiFile, callh),
(int)vpi_get(vpiLineNo, callh));
vpi_printf("%s's first argument must be a string or a register.\n", name);
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
vpip_set_return_value(1);
vpi: $sformatf() function.
2015-12-11 12:58:25 +01:00
vpi_control(vpiFinish, 1);
return 0;
}
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
int rtn = sys_check_args(callh, argv, name, 0, 0);
if (rtn) {
if (rtn == 1) vpip_set_return_value(1);
vpi_control(vpiFinish, 1);
}
vpi: $sformatf() function.
2015-12-11 12:58:25 +01:00
return 0;
}
Cleanup timeformat argument checking.
2007-04-10 06:56:26 +02:00
static PLI_INT32 sys_end_of_compile(p_cb_data cb_data)
{
Remove missing parameter warnings
2014-07-09 23:16:57 +02:00
(void)cb_data; /* Parameter is not used. */
Cleanup timeformat argument checking.
2007-04-10 06:56:26 +02:00
/* The default timeformat prints times in unit of simulation
precision. */
Fix memory leaks in vvp and change vvp T_STRING token to be new based. This patch fixes a bunch of memory leaks in vvp and converts the T_STRING lexical token to be new based. There are still two known leaks that I need to find a way to fix and likely some unknown leaks that still need to be found and fixed.
2009-01-13 18:57:12 +01:00
free(timeformat_info.suff);
Cleanup timeformat argument checking.
2007-04-10 06:56:26 +02:00
timeformat_info.suff = strdup("");
timeformat_info.units = vpi_get(vpiTimePrecision, 0);
timeformat_info.prec = 0;
timeformat_info.width = 20;
return 0;
}
Brainless start to const-correct changes Results of running cd vpi for f in *.c *.h; do sed -i \ -e "s/_calltf(PLI_BYTE8/_calltf(ICARUS_VPI_CONST PLI_BYTE8/" \ -e "s/_compiletf(PLI_BYTE8/_compiletf(ICARUS_VPI_CONST PLI_BYTE8/" $f; done and a trivial patch to vpi_user.h to, among other things, make ICARUS_VPI_CONST blank. Thus, this patch does absolutely nothing. Will be followed by a (much shorter) patch that makes it do something. :-)
2010-10-04 18:31:01 +02:00
static PLI_INT32 sys_timeformat_compiletf(ICARUS_VPI_CONST PLI_BYTE8*name)
Add support for $timeformat.
2002-05-31 06:26:54 +02:00
{
Add leading 0 support to $swrite/$sformat and $display uses this code. This patch does a number of things. It adds file and line number information to the strobe_cb_info structure. This allows the $swrite and $sformat calls to reference their location in warning/error messages. Their compiletf and calltf routines also display file and line number information in messages. The $display routine was converted to use the string formatting routine. This can be reverted by defining USE_OLD_DISPLAY. The old code will be removed once there has been more testing. A number of bugs were fixed that were discovered with the $display code. The output of some conversions, variables and system functions were changed to match the original display code or to better match expectation. Leading zero support was added to the %b/%b, %o/%O, %h/%H, %x/%X, %c/%C and %d/%D. The floating point conversions already supported adding a leading zero. This is the start of converting all the various display tasks to use the string formatting code and a cleanup/update of the sys_display.c file in general.
2009-01-06 19:41:39 +01:00
vpiHandle callh = vpi_handle(vpiSysTfCall, 0);
vpiHandle argv = vpi_iterate(vpiArgument, callh);
Add support for $timeformat.
2002-05-31 06:26:54 +02:00
Cleanup timeformat argument checking.
2007-04-10 06:56:26 +02:00
if (argv) {
Add/modify compiletf routines for $timeformat and $printtimescale. Update/add compiletf routines for these two tasks. The $printtimescale task was updated to work with more types of objects.
2009-01-16 20:10:51 +01:00
vpiHandle arg;
/* Check that the unit argument is numeric. */
if (! is_numeric_obj(vpi_scan(argv))) {
vpi_printf("ERROR: %s:%d: ", vpi_get_str(vpiFile, callh),
(int)vpi_get(vpiLineNo, callh));
vpi_printf("%s's units argument must be numeric.\n", name);
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
vpip_set_return_value(1);
Add/modify compiletf routines for $timeformat and $printtimescale. Update/add compiletf routines for these two tasks. The $printtimescale task was updated to work with more types of objects.
2009-01-16 20:10:51 +01:00
vpi_control(vpiFinish, 1);
}
/* Check that the precision argument is given and is numeric. */
arg = vpi_scan(argv);
if (! arg) {
vpi_printf("ERROR: %s:%d: ", vpi_get_str(vpiFile, callh),
(int)vpi_get(vpiLineNo, callh));
Fix space issues in VPI directory
2010-05-23 23:30:48 +02:00
vpi_printf("%s requires zero or four arguments.\n", name);
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
vpip_set_return_value(1);
Add/modify compiletf routines for $timeformat and $printtimescale. Update/add compiletf routines for these two tasks. The $printtimescale task was updated to work with more types of objects.
2009-01-16 20:10:51 +01:00
vpi_control(vpiFinish, 1);
return 0;
}
if (! is_numeric_obj(arg)) {
vpi_printf("ERROR: %s:%d: ", vpi_get_str(vpiFile, callh),
(int)vpi_get(vpiLineNo, callh));
vpi_printf("%s's precision argument must be numeric.\n",
name);
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
vpip_set_return_value(1);
Add/modify compiletf routines for $timeformat and $printtimescale. Update/add compiletf routines for these two tasks. The $printtimescale task was updated to work with more types of objects.
2009-01-16 20:10:51 +01:00
vpi_control(vpiFinish, 1);
}
/* Check that the suffix argument is given and is a string. */
arg = vpi_scan(argv);
if (! arg) {
vpi_printf("ERROR: %s:%d: ", vpi_get_str(vpiFile, callh),
(int)vpi_get(vpiLineNo, callh));
Fix space issues in VPI directory
2010-05-23 23:30:48 +02:00
vpi_printf("%s requires zero or four arguments.\n", name);
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
vpip_set_return_value(1);
Add/modify compiletf routines for $timeformat and $printtimescale. Update/add compiletf routines for these two tasks. The $printtimescale task was updated to work with more types of objects.
2009-01-16 20:10:51 +01:00
vpi_control(vpiFinish, 1);
return 0;
}
if (! is_string_obj(arg)) {
vpi_printf("ERROR: %s:%d: ", vpi_get_str(vpiFile, callh),
(int)vpi_get(vpiLineNo, callh));
vpi_printf("%s's suffix argument must be a string.\n", name);
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
vpip_set_return_value(1);
Add/modify compiletf routines for $timeformat and $printtimescale. Update/add compiletf routines for these two tasks. The $printtimescale task was updated to work with more types of objects.
2009-01-16 20:10:51 +01:00
vpi_control(vpiFinish, 1);
}
/* Check that the min. width argument is given and is numeric. */
arg = vpi_scan(argv);
if (! arg) {
vpi_printf("ERROR: %s:%d: ", vpi_get_str(vpiFile, callh),
(int)vpi_get(vpiLineNo, callh));
Fix space issues in VPI directory
2010-05-23 23:30:48 +02:00
vpi_printf("%s requires zero or four arguments.\n", name);
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
vpip_set_return_value(1);
Add/modify compiletf routines for $timeformat and $printtimescale. Update/add compiletf routines for these two tasks. The $printtimescale task was updated to work with more types of objects.
2009-01-16 20:10:51 +01:00
vpi_control(vpiFinish, 1);
return 0;
}
if (! is_numeric_obj(arg)) {
vpi_printf("ERROR: %s:%d: ", vpi_get_str(vpiFile, callh),
(int)vpi_get(vpiLineNo, callh));
vpi_printf("%s's minimum width argument must be numeric.\n",
name);
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
vpip_set_return_value(1);
Add/modify compiletf routines for $timeformat and $printtimescale. Update/add compiletf routines for these two tasks. The $printtimescale task was updated to work with more types of objects.
2009-01-16 20:10:51 +01:00
vpi_control(vpiFinish, 1);
}
/* Make sure there are no extra arguments. */
More compiletf cleanup and refactoring. This patch adds compiletf routines for the mti random functions and adds a common routine to check for extra arguments. It also adds file and line information to the rest of the compiletf routines that were missing them.
2009-02-27 00:22:21 +01:00
check_for_extra_args(argv, callh, name, "four arguments", 0);
Cleanup timeformat argument checking.
2007-04-10 06:56:26 +02:00
}
Add support for $timeformat.
2002-05-31 06:26:54 +02:00
return 0;
}
Remove missing parameter warnings
2014-07-09 23:16:57 +02:00
static PLI_INT32 sys_timeformat_calltf(ICARUS_VPI_CONST PLI_BYTE8*name)
Add support for $timeformat.
2002-05-31 06:26:54 +02:00
{
s_vpi_value value;
vpiHandle sys = vpi_handle(vpiSysTfCall, 0);
vpiHandle argv = vpi_iterate(vpiArgument, sys);
Remove missing parameter warnings
2014-07-09 23:16:57 +02:00
(void)name; /* Parameter is not used. */
Cleanup timeformat argument checking.
2007-04-10 06:56:26 +02:00
if (argv) {
vpiHandle units = vpi_scan(argv);
vpiHandle prec = vpi_scan(argv);
vpiHandle suff = vpi_scan(argv);
vpiHandle wid = vpi_scan(argv);
Add support for $timeformat.
2002-05-31 06:26:54 +02:00
Cleanup timeformat argument checking.
2007-04-10 06:56:26 +02:00
vpi_free_object(argv);
Add support for $timeformat.
2002-05-31 06:26:54 +02:00
Cleanup timeformat argument checking.
2007-04-10 06:56:26 +02:00
value.format = vpiIntVal;
vpi_get_value(units, &value);
timeformat_info.units = value.value.integer;
Add support for $timeformat.
2002-05-31 06:26:54 +02:00
Cleanup timeformat argument checking.
2007-04-10 06:56:26 +02:00
value.format = vpiIntVal;
vpi_get_value(prec, &value);
timeformat_info.prec = value.value.integer;
Add support for $timeformat.
2002-05-31 06:26:54 +02:00
Cleanup timeformat argument checking.
2007-04-10 06:56:26 +02:00
value.format = vpiStringVal;
vpi_get_value(suff, &value);
Fix memory leaks in vvp and change vvp T_STRING token to be new based. This patch fixes a bunch of memory leaks in vvp and converts the T_STRING lexical token to be new based. There are still two known leaks that I need to find a way to fix and likely some unknown leaks that still need to be found and fixed.
2009-01-13 18:57:12 +01:00
free(timeformat_info.suff);
Cleanup timeformat argument checking.
2007-04-10 06:56:26 +02:00
timeformat_info.suff = strdup(value.value.str);
Add support for $timeformat.
2002-05-31 06:26:54 +02:00
Cleanup timeformat argument checking.
2007-04-10 06:56:26 +02:00
value.format = vpiIntVal;
vpi_get_value(wid, &value);
timeformat_info.width = value.value.integer;
} else {
/* If no arguments are given then use the default values. */
sys_end_of_compile(NULL);
}
Add support for $timeformat.
2002-05-31 06:26:54 +02:00
return 0;
}
Down payment on const-correctness Clears out the unambiguous and easy-to-fix const faults, so we can focus better on the fundamental ones.
2010-09-27 20:03:43 +02:00
static const char *pts_convert(int value)
Add printtimescale (caryr)
2007-04-12 04:50:51 +02:00
{
Down payment on const-correctness Clears out the unambiguous and easy-to-fix const faults, so we can focus better on the fundamental ones.
2010-09-27 20:03:43 +02:00
const char *string;
Add printtimescale (caryr)
2007-04-12 04:50:51 +02:00
switch (value) {
Fix $printtimescale to support 10s and 100s values.
2017-11-05 10:26:33 +01:00
case 2: string = "100s"; break;
case 1: string = "10s"; break;
Pass local scope precision from compiler to vvp/etc. This patch adds ivl_scope_time_precision() to the compiler which can be used to extract the local scope precision. tgt-stub and tgt-vvp have been modified to use this new function and output a value that is appropriate. The vvp runtime has been altered to use this new data which is accessed with the vpip_time_precision_from_handle() function. vpiTimePrecision uses this function to return the correct precision.
2007-09-28 20:21:08 +02:00
case 0: string = "1s"; break;
Add printtimescale (caryr)
2007-04-12 04:50:51 +02:00
case -1: string = "100ms"; break;
case -2: string = "10ms"; break;
case -3: string = "1ms"; break;
case -4: string = "100us"; break;
case -5: string = "10us"; break;
case -6: string = "1us"; break;
case -7: string = "100ns"; break;
case -8: string = "10ns"; break;
case -9: string = "1ns"; break;
case -10: string = "100ps"; break;
case -11: string = "10ps"; break;
case -12: string = "1ps"; break;
case -13: string = "100fs"; break;
case -14: string = "10fs"; break;
case -15: string = "1fs"; break;
Add some casts in vpi to remove warnings. The Cygwin compiler is a bit picky. This patch adds some casts to remove compilation warnings. In the past I have had warnings off because of problems with the STL, but for C directories like this it makes sense to enable the warnings. It also does not recognize that an assert(0) or assert(false) ends a routine so it complains about no return at end of function or variables not being defined.
2009-12-10 21:49:39 +01:00
default: string = "invalid"; assert(0);
Add printtimescale (caryr)
2007-04-12 04:50:51 +02:00
}
return string;
}
Brainless start to const-correct changes Results of running cd vpi for f in *.c *.h; do sed -i \ -e "s/_calltf(PLI_BYTE8/_calltf(ICARUS_VPI_CONST PLI_BYTE8/" \ -e "s/_compiletf(PLI_BYTE8/_compiletf(ICARUS_VPI_CONST PLI_BYTE8/" $f; done and a trivial patch to vpi_user.h to, among other things, make ICARUS_VPI_CONST blank. Thus, this patch does absolutely nothing. Will be followed by a (much shorter) patch that makes it do something. :-)
2010-10-04 18:31:01 +02:00
static PLI_INT32 sys_printtimescale_compiletf(ICARUS_VPI_CONST PLI_BYTE8*name)
Add/modify compiletf routines for $timeformat and $printtimescale. Update/add compiletf routines for these two tasks. The $printtimescale task was updated to work with more types of objects.
2009-01-16 20:10:51 +01:00
{
vpiHandle callh = vpi_handle(vpiSysTfCall, 0);
vpiHandle argv = vpi_iterate(vpiArgument, callh);
if (argv) {
vpiHandle arg = vpi_scan(argv);
switch (vpi_get(vpiType, arg)) {
case vpiFunction:
Add support for a generate scope to vvp and the vpi routines. Instead of just translating a generate scope to a named begin/end scope this patch creates a generate specific scope (vpiScopeGenerate) that is of the vpiGenScope type. This may not match the standard 100%, but does allow the FST dumper to denote generate scopes differently than the other scope types. Most of the VPI code treats a vpiGenScope just like a named block so only the FST dumper should have different behavior.
2013-07-17 19:55:46 +02:00
case vpiGenScope:
Add/modify compiletf routines for $timeformat and $printtimescale. Update/add compiletf routines for these two tasks. The $printtimescale task was updated to work with more types of objects.
2009-01-16 20:10:51 +01:00
case vpiIntegerVar:
Update $scanf, $printtimescale and the general is_numeric check for 2-state This patch updates the $scanf and $printtimescale routines to work with two-state variables. It also updates the general is_numeric check to recognize two-state variables as numeric.
2011-09-06 02:35:26 +02:00
case vpiBitVar:
case vpiByteVar:
case vpiShortIntVar:
case vpiIntVar:
case vpiLongIntVar:
Add/modify compiletf routines for $timeformat and $printtimescale. Update/add compiletf routines for these two tasks. The $printtimescale task was updated to work with more types of objects.
2009-01-16 20:10:51 +01:00
case vpiMemory:
case vpiMemoryWord:
case vpiModule:
case vpiNamedBegin:
case vpiNamedEvent:
case vpiNamedFork:
case vpiNet:
Add compiletf routines for the scanf functions. This patch adds compiletf routines for the $sscanf/$fscanf functions. It does not enhance the conversion routines it only adds better error/warning messages. It also removed the redundant IS_MCD in some of the fileio calltf routines. This is already checked in vpi_get_file(). I also added the vpiNetArray type to allow us to distinguish arrays of nets from other arrays (memories). Much more is needed here, but this solved the problem of the day. The problem was that we needed to make sure we could assign to the word of a variable array, but not the word of a net array. It's actually probably not needed since Icarus converts the net array word reference to just the net, but I thought it prudent to make sure things didn't slip by.
2009-02-27 21:34:48 +01:00
case vpiNetArray:
Add/modify compiletf routines for $timeformat and $printtimescale. Update/add compiletf routines for these two tasks. The $printtimescale task was updated to work with more types of objects.
2009-01-16 20:10:51 +01:00
// case vpiNetBit: // Unused and unavailable in Icarus
case vpiParameter:
case vpiPartSelect:
case vpiRealVar:
case vpiReg:
// case vpiRegBit: // Unused and unavailable in Icarus
case vpiTask:
case vpiTimeVar: // Unused in Icarus
break;
default:
vpi_printf("ERROR: %s:%d: ", vpi_get_str(vpiFile, callh),
(int)vpi_get(vpiLineNo, callh));
vpi_printf("%s's argument must have a module, given a %s.\n",
name, vpi_get_str(vpiType, arg));
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
vpip_set_return_value(1);
Add/modify compiletf routines for $timeformat and $printtimescale. Update/add compiletf routines for these two tasks. The $printtimescale task was updated to work with more types of objects.
2009-01-16 20:10:51 +01:00
vpi_control(vpiFinish, 1);
}
/* Make sure there are no extra arguments. */
More compiletf cleanup and refactoring. This patch adds compiletf routines for the mti random functions and adds a common routine to check for extra arguments. It also adds file and line information to the rest of the compiletf routines that were missing them.
2009-02-27 00:22:21 +01:00
check_for_extra_args(argv, callh, name, "one argument", 1);
Add/modify compiletf routines for $timeformat and $printtimescale. Update/add compiletf routines for these two tasks. The $printtimescale task was updated to work with more types of objects.
2009-01-16 20:10:51 +01:00
}
return 0;
}
Remove missing parameter warnings
2014-07-09 23:16:57 +02:00
static PLI_INT32 sys_printtimescale_calltf(ICARUS_VPI_CONST PLI_BYTE8*name)
Add printtimescale (caryr)
2007-04-12 04:50:51 +02:00
{
Add/modify compiletf routines for $timeformat and $printtimescale. Update/add compiletf routines for these two tasks. The $printtimescale task was updated to work with more types of objects.
2009-01-16 20:10:51 +01:00
vpiHandle callh = vpi_handle(vpiSysTfCall, 0);
vpiHandle argv = vpi_iterate(vpiArgument, callh);
vpiHandle item, scope;
Remove missing parameter warnings
2014-07-09 23:16:57 +02:00
(void)name; /* Parameter is not used. */
Add printtimescale (caryr)
2007-04-12 04:50:51 +02:00
if (!argv) {
Add/modify compiletf routines for $timeformat and $printtimescale. Update/add compiletf routines for these two tasks. The $printtimescale task was updated to work with more types of objects.
2009-01-16 20:10:51 +01:00
item = sys_func_module(callh);
Add printtimescale (caryr)
2007-04-12 04:50:51 +02:00
} else {
Add/modify compiletf routines for $timeformat and $printtimescale. Update/add compiletf routines for these two tasks. The $printtimescale task was updated to work with more types of objects.
2009-01-16 20:10:51 +01:00
item = vpi_scan(argv);
Add printtimescale (caryr)
2007-04-12 04:50:51 +02:00
vpi_free_object(argv);
}
Support timescales in design units that aren't inside a module. SystemVerilog allows tasks, functions, and classes to be defined at the root level or inside packages, so we can't rely on an enclosing module being present to provide the timescale.
2016-07-22 23:48:20 +02:00
scope = sys_func_module(item);
Add/modify compiletf routines for $timeformat and $printtimescale. Update/add compiletf routines for these two tasks. The $printtimescale task was updated to work with more types of objects.
2009-01-16 20:10:51 +01:00
vpi_printf("Time scale of (%s) is ", vpi_get_str(vpiFullName, item));
Add printtimescale (caryr)
2007-04-12 04:50:51 +02:00
vpi_printf("%s / ", pts_convert(vpi_get(vpiTimeUnit, scope)));
vpi_printf("%s\n", pts_convert(vpi_get(vpiTimePrecision, scope)));
return 0;
}
Brainless start to const-correct changes Results of running cd vpi for f in *.c *.h; do sed -i \ -e "s/_calltf(PLI_BYTE8/_calltf(ICARUS_VPI_CONST PLI_BYTE8/" \ -e "s/_compiletf(PLI_BYTE8/_compiletf(ICARUS_VPI_CONST PLI_BYTE8/" $f; done and a trivial patch to vpi_user.h to, among other things, make ICARUS_VPI_CONST blank. Thus, this patch does absolutely nothing. Will be followed by a (much shorter) patch that makes it do something. :-)
2010-10-04 18:31:01 +02:00
static PLI_INT32 sys_fatal_compiletf(ICARUS_VPI_CONST PLI_BYTE8*name)
Implement severity system tasks $fatal, $error, $warning, and $info.
2009-11-26 01:22:16 +01:00
{
Clean up the $fatal compiletf and calltf routines Some cleanup of the compiletf routine for $fatal and the same for the calltf routine that is used by $fatal, $error, $warning and $info.
2010-03-13 20:19:00 +01:00
vpiHandle callh = vpi_handle(vpiSysTfCall, 0);
vpiHandle argv = vpi_iterate(vpiArgument, callh);
Implement severity system tasks $fatal, $error, $warning, and $info.
2009-11-26 01:22:16 +01:00
if (argv) {
Clean up the $fatal compiletf and calltf routines Some cleanup of the compiletf routine for $fatal and the same for the calltf routine that is used by $fatal, $error, $warning and $info.
2010-03-13 20:19:00 +01:00
vpiHandle arg = vpi_scan(argv);
Implement severity system tasks $fatal, $error, $warning, and $info.
2009-11-26 01:22:16 +01:00
Clean up the $fatal compiletf and calltf routines Some cleanup of the compiletf routine for $fatal and the same for the calltf routine that is used by $fatal, $error, $warning and $info.
2010-03-13 20:19:00 +01:00
/* Check that finish_number (1st argument) is numeric */
Implement severity system tasks $fatal, $error, $warning, and $info.
2009-11-26 01:22:16 +01:00
if (! is_numeric_obj(arg)) {
Clean up the $fatal compiletf and calltf routines Some cleanup of the compiletf routine for $fatal and the same for the calltf routine that is used by $fatal, $error, $warning and $info.
2010-03-13 20:19:00 +01:00
vpi_printf("ERROR: %s:%d: ", vpi_get_str(vpiFile, callh),
Implement severity system tasks $fatal, $error, $warning, and $info.
2009-11-26 01:22:16 +01:00
(int)vpi_get(vpiLineNo, callh));
Clean up the $fatal compiletf and calltf routines Some cleanup of the compiletf routine for $fatal and the same for the calltf routine that is used by $fatal, $error, $warning and $info.
2010-03-13 20:19:00 +01:00
vpi_printf("%s's finish number must be numeric.\n", name);
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
vpip_set_return_value(1);
Clean up the $fatal compiletf and calltf routines Some cleanup of the compiletf routine for $fatal and the same for the calltf routine that is used by $fatal, $error, $warning and $info.
2010-03-13 20:19:00 +01:00
vpi_control(vpiFinish, 1);
Implement severity system tasks $fatal, $error, $warning, and $info.
2009-11-26 01:22:16 +01:00
}
Clean up the $fatal compiletf and calltf routines Some cleanup of the compiletf routine for $fatal and the same for the calltf routine that is used by $fatal, $error, $warning and $info.
2010-03-13 20:19:00 +01:00
Update VPI routines to on error signal VVP to return non-zero
2021-01-21 08:50:06 +01:00
int rtn = sys_check_args(callh, argv, name, 0, 0);
if (rtn) {
if (rtn == 1) vpip_set_return_value(1);
Clean up the $fatal compiletf and calltf routines Some cleanup of the compiletf routine for $fatal and the same for the calltf routine that is used by $fatal, $error, $warning and $info.
2010-03-13 20:19:00 +01:00
vpi_control(vpiFinish, 1);
}
Implement severity system tasks $fatal, $error, $warning, and $info.
2009-11-26 01:22:16 +01:00
}
Clean up the $fatal compiletf and calltf routines Some cleanup of the compiletf routine for $fatal and the same for the calltf routine that is used by $fatal, $error, $warning and $info.
2010-03-13 20:19:00 +01:00
Implement severity system tasks $fatal, $error, $warning, and $info.
2009-11-26 01:22:16 +01:00
return 0;
}
Brainless start to const-correct changes Results of running cd vpi for f in *.c *.h; do sed -i \ -e "s/_calltf(PLI_BYTE8/_calltf(ICARUS_VPI_CONST PLI_BYTE8/" \ -e "s/_compiletf(PLI_BYTE8/_compiletf(ICARUS_VPI_CONST PLI_BYTE8/" $f; done and a trivial patch to vpi_user.h to, among other things, make ICARUS_VPI_CONST blank. Thus, this patch does absolutely nothing. Will be followed by a (much shorter) patch that makes it do something. :-)
2010-10-04 18:31:01 +02:00
static PLI_INT32 sys_severity_calltf(ICARUS_VPI_CONST PLI_BYTE8*name)
Implement severity system tasks $fatal, $error, $warning, and $info.
2009-11-26 01:22:16 +01:00
{
Clean up the $fatal compiletf and calltf routines Some cleanup of the compiletf routine for $fatal and the same for the calltf routine that is used by $fatal, $error, $warning and $info.
2010-03-13 20:19:00 +01:00
vpiHandle callh = vpi_handle(vpiSysTfCall, 0);
vpiHandle argv = vpi_iterate(vpiArgument, callh);
vpiHandle scope;
Implement severity system tasks $fatal, $error, $warning, and $info.
2009-11-26 01:22:16 +01:00
struct strobe_cb_info info;
struct t_vpi_time now;
PLI_UINT64 now64;
char *sstr, *t, *dstr;
unsigned int size, location=0;
Clean up the $fatal compiletf and calltf routines Some cleanup of the compiletf routine for $fatal and the same for the calltf routine that is used by $fatal, $error, $warning and $info.
2010-03-13 20:19:00 +01:00
s_vpi_value finish_number;
Implement severity system tasks $fatal, $error, $warning, and $info.
2009-11-26 01:22:16 +01:00
Clean up the $fatal compiletf and calltf routines Some cleanup of the compiletf routine for $fatal and the same for the calltf routine that is used by $fatal, $error, $warning and $info.
2010-03-13 20:19:00 +01:00
/* Set the default finish number for $fatal. */
Move $fatal argument value check from compiletf to calltf. Move check of the value $fatal is called with from the compiletf to the calltf routine, since the value may not be known at compiletf time. If the value is invalid, it just prints a warning that $fatal was called with a bad finish_number, and resets it to the default 1. Changes the compile time check for a numeric argument to a warning. Also, fixes bug where $fatal called without an argument causes problems.
2010-03-09 07:25:12 +01:00
finish_number.value.integer = 1;
Implement severity system tasks $fatal, $error, $warning, and $info.
2009-11-26 01:22:16 +01:00
Clean up the $fatal compiletf and calltf routines Some cleanup of the compiletf routine for $fatal and the same for the calltf routine that is used by $fatal, $error, $warning and $info.
2010-03-13 20:19:00 +01:00
/* Check that the finish number is in range. */
if (strncmp(name,"$fatal", 6) == 0 && argv) {
Implement severity system tasks $fatal, $error, $warning, and $info.
2009-11-26 01:22:16 +01:00
vpiHandle arg = vpi_scan(argv);
finish_number.format = vpiIntVal;
vpi_get_value(arg, &finish_number);
Move $fatal argument value check from compiletf to calltf. Move check of the value $fatal is called with from the compiletf to the calltf routine, since the value may not be known at compiletf time. If the value is invalid, it just prints a warning that $fatal was called with a bad finish_number, and resets it to the default 1. Changes the compile time check for a numeric argument to a warning. Also, fixes bug where $fatal called without an argument causes problems.
2010-03-09 07:25:12 +01:00
if ((finish_number.value.integer < 0) ||
(finish_number.value.integer > 2)) {
vpi_printf("WARNING: %s:%d: ", vpi_get_str(vpiFile, callh),
(int)vpi_get(vpiLineNo, callh));
vpi_printf("$fatal called with finish_number of %d, "
Clean up the $fatal compiletf and calltf routines Some cleanup of the compiletf routine for $fatal and the same for the calltf routine that is used by $fatal, $error, $warning and $info.
2010-03-13 20:19:00 +01:00
"but it must be 0, 1, or 2.\n",
(int)finish_number.value.integer);
Move $fatal argument value check from compiletf to calltf. Move check of the value $fatal is called with from the compiletf to the calltf routine, since the value may not be known at compiletf time. If the value is invalid, it just prints a warning that $fatal was called with a bad finish_number, and resets it to the default 1. Changes the compile time check for a numeric argument to a warning. Also, fixes bug where $fatal called without an argument causes problems.
2010-03-09 07:25:12 +01:00
finish_number.value.integer = 1;
}
Implement severity system tasks $fatal, $error, $warning, and $info.
2009-11-26 01:22:16 +01:00
}
/* convert name to upper and drop $ to get severity string */
sstr = strdup(name) + 1;
Remove some gcc-4.3.4 warnings (Cygwin) The latest gcc with the latest Cygwin complains when passing a char to the toupper, tolower, isspace, isalnum, isprint, isdigit or isalpha functions/macros. These functions are defined to take an integer. This patch adds cast to int as needed to remove the warnings. After this there are still two warnings related to signed/unsigned comparison in yy_get_next_buffer() (part of flex).
2010-04-11 20:00:39 +02:00
for (t=sstr; *t; t+=1) *t = toupper((int)*t);
Implement severity system tasks $fatal, $error, $warning, and $info.
2009-11-26 01:22:16 +01:00
scope = vpi_handle(vpiScope, callh);
assert(scope);
info.name = name;
info.filename = strdup(vpi_get_str(vpiFile, callh));
info.lineno = (int)vpi_get(vpiLineNo, callh);
info.default_format = vpiDecStrVal;
info.scope = scope;
array_from_iterator(&info, argv);
vpi_printf("%s: %s:%d: ", sstr, info.filename, info.lineno);
dstr = get_display(&size, &info);
while (location < size) {
if (dstr[location] == '\0') {
my_mcd_printf(1, "%c", '\0');
location += 1;
} else {
my_mcd_printf(1, "%s", &dstr[location]);
location += strlen(&dstr[location]);
}
}
now.type = vpiSimTime;
vpi_get_time(0, &now);
now64 = timerec_to_time64(&now);
Some clean up and add initial support for elaboration system tasks
2021-02-01 09:20:09 +01:00
vpi_printf("\n%*s Time: %" PLI_UINT64_FMT " Scope: %s\n",
Clean up the $fatal compiletf and calltf routines Some cleanup of the compiletf routine for $fatal and the same for the calltf routine that is used by $fatal, $error, $warning and $info.
2010-03-13 20:19:00 +01:00
(int)strlen(sstr), " ", now64,
vpi_get_str(vpiFullName, scope));
Implement severity system tasks $fatal, $error, $warning, and $info.
2009-11-26 01:22:16 +01:00
Clean up the $fatal compiletf and calltf routines Some cleanup of the compiletf routine for $fatal and the same for the calltf routine that is used by $fatal, $error, $warning and $info.
2010-03-13 20:19:00 +01:00
free(--sstr); /* Get the $ back. */
Implement severity system tasks $fatal, $error, $warning, and $info.
2009-11-26 01:22:16 +01:00
free(info.filename);
free(info.items);
free(dstr);
Clean up the $fatal compiletf and calltf routines Some cleanup of the compiletf routine for $fatal and the same for the calltf routine that is used by $fatal, $error, $warning and $info.
2010-03-13 20:19:00 +01:00
if (strncmp(name,"$fatal",6) == 0) {
Make vvp exit with FAILURE if $fatal is used to exit the simulation.
2019-10-14 23:11:50 +02:00
/* Set the exit code from vvp as an error code. */
vpip_set_return_value(1);
/* Now tell the simulator to finish. */
Implement severity system tasks $fatal, $error, $warning, and $info.
2009-11-26 01:22:16 +01:00
vpi_control(vpiFinish, finish_number.value.integer);
Clean up the $fatal compiletf and calltf routines Some cleanup of the compiletf routine for $fatal and the same for the calltf routine that is used by $fatal, $error, $warning and $info.
2010-03-13 20:19:00 +01:00
}
Implement severity system tasks $fatal, $error, $warning, and $info.
2009-11-26 01:22:16 +01:00
return 0;
}
More end of simulation memory cleanup. This patch adds some more memory freeing routines to vvp and the vpi library. Much more to do before this is finished.
2009-01-24 01:04:44 +01:00
static PLI_INT32 sys_end_of_simulation(p_cb_data cb_data)
{
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
unsigned idx;
Remove missing parameter warnings
2014-07-09 23:16:57 +02:00
(void)cb_data; /* Parameter is not used. */
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
assert(monitor_info);
for (idx = 0; idx < monitor_info_len; idx +=1) {
cancel_monitor(monitor_info[idx]);
free(monitor_info[idx]);
}
free(monitor_info);
monitor_info_len = 0;
monitor_info = 0;
More end of simulation memory cleanup. This patch adds some more memory freeing routines to vvp and the vpi library. Much more to do before this is finished.
2009-01-24 01:04:44 +01:00
free(timeformat_info.suff);
timeformat_info.suff = 0;
return 0;
}
Fuss with C function prototypes 119 formal void parameters added to keep -Wstrict-prototypes happy. Process found one real missing prototype in vpi/vcd_priv.h: EXTERN void vcd_names_delete(struct vcd_names_list_s*tab); 8 such warnings left, all in Tony's code
2014-07-08 21:12:54 +02:00
void sys_display_register(void)
Convert vvm to implement system tasks with vpi.
1999-08-15 03:23:56 +02:00
{
Add support for $timeformat.
2002-05-31 06:26:54 +02:00
s_cb_data cb_data;
Convert vvm to implement system tasks with vpi.
1999-08-15 03:23:56 +02:00
s_vpi_systf_data tf_data;
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
vpiHandle res;
Convert vvm to implement system tasks with vpi.
1999-08-15 03:23:56 +02:00
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
// Allocate the first entry in monitor_info for use by $monitor
// based tasks.
(void)allocate_monitor_info();
Add -compatible command line option to vvp. This option is intended to make it easier to compare results from Icarus with results from other simulators. For now, the only effect it has is to change the default format for displaying real numbers when no format string is supplied.
2013-10-20 23:59:53 +02:00
check_command_line_args();
Content-free portability fixes. These are patches from ITOH Yasufumi and Larry Doolittle of pr2120948. They change nothing, but make the code compile with older compilers.
2008-09-30 03:06:47 +02:00
/*============================== display */
Convert vvm to implement system tasks with vpi.
1999-08-15 03:23:56 +02:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$display";
tf_data.calltf = sys_display_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_display_compiletf;
Convert vvm to implement system tasks with vpi.
1999-08-15 03:23:56 +02:00
tf_data.sizetf = 0;
Support the $write system task, and have the vpi_scan function free iterators as needed.
1999-09-29 03:41:18 +02:00
tf_data.user_data = "$display";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Support the $write system task, and have the vpi_scan function free iterators as needed.
1999-09-29 03:41:18 +02:00
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$displayh";
tf_data.calltf = sys_display_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_display_compiletf;
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.sizetf = 0;
tf_data.user_data = "$displayh";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$displayo";
tf_data.calltf = sys_display_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_display_compiletf;
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.sizetf = 0;
tf_data.user_data = "$displayo";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$displayb";
tf_data.calltf = sys_display_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_display_compiletf;
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.sizetf = 0;
tf_data.user_data = "$displayb";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
Content-free portability fixes. These are patches from ITOH Yasufumi and Larry Doolittle of pr2120948. They change nothing, but make the code compile with older compilers.
2008-09-30 03:06:47 +02:00
/*============================== write */
Support the $write system task, and have the vpi_scan function free iterators as needed.
1999-09-29 03:41:18 +02:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$write";
tf_data.calltf = sys_display_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_display_compiletf;
Support the $write system task, and have the vpi_scan function free iterators as needed.
1999-09-29 03:41:18 +02:00
tf_data.sizetf = 0;
tf_data.user_data = "$write";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Rewrite vvm VPI support to make objects more persistent, rewrite the simulation scheduler in C (to interface with VPI) and add VPI support for callbacks.
1999-10-28 02:47:24 +02:00
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$writeh";
tf_data.calltf = sys_display_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_display_compiletf;
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.sizetf = 0;
tf_data.user_data = "$writeh";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$writeo";
tf_data.calltf = sys_display_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_display_compiletf;
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.sizetf = 0;
tf_data.user_data = "$writeo";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$writeb";
tf_data.calltf = sys_display_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_display_compiletf;
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.sizetf = 0;
tf_data.user_data = "$writeb";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
Content-free portability fixes. These are patches from ITOH Yasufumi and Larry Doolittle of pr2120948. They change nothing, but make the code compile with older compilers.
2008-09-30 03:06:47 +02:00
/*============================== strobe */
Rewrite vvm VPI support to make objects more persistent, rewrite the simulation scheduler in C (to interface with VPI) and add VPI support for callbacks.
1999-10-28 02:47:24 +02:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$strobe";
tf_data.calltf = sys_strobe_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_strobe_compiletf;
Rewrite vvm VPI support to make objects more persistent, rewrite the simulation scheduler in C (to interface with VPI) and add VPI support for callbacks.
1999-10-28 02:47:24 +02:00
tf_data.sizetf = 0;
tf_data.user_data = "$strobe";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Support vpiValueChance callbacks.
1999-10-29 05:37:22 +02:00
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$strobeh";
tf_data.calltf = sys_strobe_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_strobe_compiletf;
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.sizetf = 0;
tf_data.user_data = "$strobeh";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$strobeo";
tf_data.calltf = sys_strobe_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_strobe_compiletf;
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.sizetf = 0;
tf_data.user_data = "$strobeo";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$strobeb";
tf_data.calltf = sys_strobe_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_strobe_compiletf;
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.sizetf = 0;
tf_data.user_data = "$strobeb";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
Content-free portability fixes. These are patches from ITOH Yasufumi and Larry Doolittle of pr2120948. They change nothing, but make the code compile with older compilers.
2008-09-30 03:06:47 +02:00
/*============================== fstrobe */
Add support for fstrobe system tasks.
2003-08-26 05:51:05 +02:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$fstrobe";
tf_data.calltf = sys_strobe_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_strobe_compiletf;
Add support for fstrobe system tasks.
2003-08-26 05:51:05 +02:00
tf_data.sizetf = 0;
tf_data.user_data = "$fstrobe";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Add support for fstrobe system tasks.
2003-08-26 05:51:05 +02:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$fstrobeh";
tf_data.calltf = sys_strobe_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_strobe_compiletf;
Add support for fstrobe system tasks.
2003-08-26 05:51:05 +02:00
tf_data.sizetf = 0;
tf_data.user_data = "$fstrobeh";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Add support for fstrobe system tasks.
2003-08-26 05:51:05 +02:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$fstrobeo";
tf_data.calltf = sys_strobe_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_strobe_compiletf;
Add support for fstrobe system tasks.
2003-08-26 05:51:05 +02:00
tf_data.sizetf = 0;
tf_data.user_data = "$fstrobeo";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Add support for fstrobe system tasks.
2003-08-26 05:51:05 +02:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$fstrobeb";
tf_data.calltf = sys_strobe_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_strobe_compiletf;
Add support for fstrobe system tasks.
2003-08-26 05:51:05 +02:00
tf_data.sizetf = 0;
tf_data.user_data = "$fstrobeb";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Add support for fstrobe system tasks.
2003-08-26 05:51:05 +02:00
Content-free portability fixes. These are patches from ITOH Yasufumi and Larry Doolittle of pr2120948. They change nothing, but make the code compile with older compilers.
2008-09-30 03:06:47 +02:00
/*============================== monitor */
Support vpiValueChance callbacks.
1999-10-29 05:37:22 +02:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$monitor";
tf_data.calltf = sys_monitor_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_monitor_compiletf;
Support vpiValueChance callbacks.
1999-10-29 05:37:22 +02:00
tf_data.sizetf = 0;
tf_data.user_data = "$monitor";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Import MCD support from Stephen Tell, and add system function parameter support to the IVL core.
2000-05-07 20:20:07 +02:00
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$monitorh";
tf_data.calltf = sys_monitor_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_monitor_compiletf;
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.sizetf = 0;
tf_data.user_data = "$monitorh";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$monitoro";
tf_data.calltf = sys_monitor_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_monitor_compiletf;
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.sizetf = 0;
tf_data.user_data = "$monitoro";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$monitorb";
tf_data.calltf = sys_monitor_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_monitor_compiletf;
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.sizetf = 0;
tf_data.user_data = "$monitorb";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
Add monitoron and monitoroff system tasks.
2002-07-25 05:35:51 +02:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$monitoron";
tf_data.calltf = sys_monitoron_calltf;
Make the $fdisplay/$fwrite family use the string formatting. This patch makes the $fdisplay() and $fwrite() family of task use the string formatting code. It also adds a compiletf routine for the tasks and a run time check that the FD or MCD is correct. If not it fails instead of just dropping the output. It also cleans up a couple of memory leaks and adds no argument compiletf routines for the $monitoron/$monitoroff tasks.
2009-01-10 02:27:27 +01:00
tf_data.compiletf = sys_no_arg_compiletf;
Add monitoron and monitoroff system tasks.
2002-07-25 05:35:51 +02:00
tf_data.sizetf = 0;
tf_data.user_data = "$monitoron";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Add monitoron and monitoroff system tasks.
2002-07-25 05:35:51 +02:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$monitoroff";
tf_data.calltf = sys_monitoroff_calltf;
Make the $fdisplay/$fwrite family use the string formatting. This patch makes the $fdisplay() and $fwrite() family of task use the string formatting code. It also adds a compiletf routine for the tasks and a run time check that the FD or MCD is correct. If not it fails instead of just dropping the output. It also cleans up a couple of memory leaks and adds no argument compiletf routines for the $monitoron/$monitoroff tasks.
2009-01-10 02:27:27 +01:00
tf_data.compiletf = sys_no_arg_compiletf;
Add monitoron and monitoroff system tasks.
2002-07-25 05:35:51 +02:00
tf_data.sizetf = 0;
tf_data.user_data = "$monitoroff";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Add support for $fmonitor tasks (issue #1280)
2025-10-05 13:34:25 +02:00
/*============================== fmonitor */
tf_data.type = vpiSysTask;
tf_data.tfname = "$fmonitor";
tf_data.calltf = sys_monitor_calltf;
tf_data.compiletf = sys_monitor_compiletf;
tf_data.sizetf = 0;
tf_data.user_data = "$fmonitor";
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
tf_data.type = vpiSysTask;
tf_data.tfname = "$fmonitorh";
tf_data.calltf = sys_monitor_calltf;
tf_data.compiletf = sys_monitor_compiletf;
tf_data.sizetf = 0;
tf_data.user_data = "$fmonitorh";
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
tf_data.type = vpiSysTask;
tf_data.tfname = "$fmonitoro";
tf_data.calltf = sys_monitor_calltf;
tf_data.compiletf = sys_monitor_compiletf;
tf_data.sizetf = 0;
tf_data.user_data = "$fmonitoro";
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
tf_data.type = vpiSysTask;
tf_data.tfname = "$fmonitorb";
tf_data.calltf = sys_monitor_calltf;
tf_data.compiletf = sys_monitor_compiletf;
tf_data.sizetf = 0;
tf_data.user_data = "$fmonitorb";
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Add monitoron and monitoroff system tasks.
2002-07-25 05:35:51 +02:00
Content-free portability fixes. These are patches from ITOH Yasufumi and Larry Doolittle of pr2120948. They change nothing, but make the code compile with older compilers.
2008-09-30 03:06:47 +02:00
/*============================== fdisplay */
Import MCD support from Stephen Tell, and add system function parameter support to the IVL core.
2000-05-07 20:20:07 +02:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$fdisplay";
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
tf_data.calltf = sys_display_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_display_compiletf;
Import MCD support from Stephen Tell, and add system function parameter support to the IVL core.
2000-05-07 20:20:07 +02:00
tf_data.sizetf = 0;
tf_data.user_data = "$fdisplay";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Import MCD support from Stephen Tell, and add system function parameter support to the IVL core.
2000-05-07 20:20:07 +02:00
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$fdisplayh";
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
tf_data.calltf = sys_display_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_display_compiletf;
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.sizetf = 0;
tf_data.user_data = "$fdisplayh";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$fdisplayo";
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
tf_data.calltf = sys_display_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_display_compiletf;
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.sizetf = 0;
tf_data.user_data = "$fdisplayo";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$fdisplayb";
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
tf_data.calltf = sys_display_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_display_compiletf;
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
tf_data.sizetf = 0;
tf_data.user_data = "$fdisplayb";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Default widths pad out as per the standard, add $displayb/o/h et al., and some better error messages for incorrect formats.
2002-01-15 04:23:34 +01:00
Content-free portability fixes. These are patches from ITOH Yasufumi and Larry Doolittle of pr2120948. They change nothing, but make the code compile with older compilers.
2008-09-30 03:06:47 +02:00
/*============================== fwrite */
Import MCD support from Stephen Tell, and add system function parameter support to the IVL core.
2000-05-07 20:20:07 +02:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$fwrite";
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
tf_data.calltf = sys_display_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_display_compiletf;
Import MCD support from Stephen Tell, and add system function parameter support to the IVL core.
2000-05-07 20:20:07 +02:00
tf_data.sizetf = 0;
tf_data.user_data = "$fwrite";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Attach line number information to task calls.
2007-04-15 22:45:40 +02:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$fwriteh";
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
tf_data.calltf = sys_display_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_display_compiletf;
Attach line number information to task calls.
2007-04-15 22:45:40 +02:00
tf_data.sizetf = 0;
tf_data.user_data = "$fwriteh";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Attach line number information to task calls.
2007-04-15 22:45:40 +02:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$fwriteo";
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
tf_data.calltf = sys_display_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_display_compiletf;
Attach line number information to task calls.
2007-04-15 22:45:40 +02:00
tf_data.sizetf = 0;
tf_data.user_data = "$fwriteo";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Attach line number information to task calls.
2007-04-15 22:45:40 +02:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$fwriteb";
Make $strobe and $fstrobe use the string based formatting. This patch modifies $strobe and $fstrobe to use the string formatting routines. It also combines the $display/$write and $fdisplay/$fwrite routines into one and removes the old routines that are no longer needed.
2009-01-15 01:18:36 +01:00
tf_data.calltf = sys_display_calltf;
Convert $monitor to use string based formatting. This completes the transition to the new string based formatting. All the tasks now use the string formatting routines. Better compile time checking and better messages were also added. Also a couple of types were added to vpi_get_str(vpiType, ...) and the calculation for vpiConstantSelect was fixed for both the &A<> and &PV<> constructs. If the value is a plain variable or if it is calculated in thread space we assume it is not a constant. This may not be true because of limitations/bugs in the compiler (constant user functions are one known problem).
2009-01-15 19:43:46 +01:00
tf_data.compiletf = sys_display_compiletf;
Attach line number information to task calls.
2007-04-15 22:45:40 +02:00
tf_data.sizetf = 0;
tf_data.user_data = "$fwriteb";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Add support for $timeformat.
2002-05-31 06:26:54 +02:00
Content-free portability fixes. These are patches from ITOH Yasufumi and Larry Doolittle of pr2120948. They change nothing, but make the code compile with older compilers.
2008-09-30 03:06:47 +02:00
/*============================== swrite */
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$swrite";
tf_data.calltf = sys_swrite_calltf;
tf_data.compiletf = sys_swrite_compiletf;
tf_data.sizetf = 0;
tf_data.user_data = "$swrite";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$swriteh";
tf_data.calltf = sys_swrite_calltf;
tf_data.compiletf = sys_swrite_compiletf;
tf_data.sizetf = 0;
tf_data.user_data = "$swriteh";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$swriteo";
tf_data.calltf = sys_swrite_calltf;
tf_data.compiletf = sys_swrite_compiletf;
tf_data.sizetf = 0;
tf_data.user_data = "$swriteo";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$swriteb";
tf_data.calltf = sys_swrite_calltf;
tf_data.compiletf = sys_swrite_compiletf;
tf_data.sizetf = 0;
tf_data.user_data = "$swriteb";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$sformat";
tf_data.calltf = sys_sformat_calltf;
tf_data.compiletf = sys_sformat_compiletf;
tf_data.sizetf = 0;
tf_data.user_data = "$sformat";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Implement the swrite* and sformat system tasks plus a few other fixes. This patch implements the swrite* and sformat system tasks. It also makes $simtime distinguishable from the other integer time tasks. This was needed to get the correct time units when $simtime was given as an argument to $swrite*. The string constant code was also modified to allow a string to be returned as a vector (0/1 bit pattern). Here are some more specifics about the swrite* changes. 1. They do not share formatting code with the other display functions, so they may/will produce different results. 2. All %{alpha} codes allow a width and justification. Others have been enhanced (%t allows the default width and precision to be overridden, time functions print with time formatting, better error checking and messages, etc.). 3. %u and %z formatting codes have been added. It is important to note that these two formats can produce embedded NULLs, since these functions are returning a string anything after the first NULL will not be reachable! memcpy is used instead of regular string processing where needed so that the original string will contain the total result. The size returned when the string is created is the true length. A warning will be printed if a string with embedded NULLs is produced (strlen() does not match the true length). 4. Real numbers are printed with %g instead of %f. Once this new formatting code has been evaluated we should incorporate the changes/fixes into the formatting code for the other functions or the other functions could be modified to use this new code. The true string length is available so we should be able to work around the embedded NULL problem.
2007-08-14 00:56:02 +02:00
vpi: $sformatf() function.
2015-12-11 12:58:25 +01:00
tf_data.type = vpiSysFunc;
tf_data.sysfunctype = vpiStringFunc;
tf_data.tfname = "$sformatf";
tf_data.calltf = sys_display_calltf;
tf_data.compiletf = sys_sformatf_compiletf;
tf_data.sizetf = 0;
tf_data.user_data = "$sformatf";
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Content-free portability fixes. These are patches from ITOH Yasufumi and Larry Doolittle of pr2120948. They change nothing, but make the code compile with older compilers.
2008-09-30 03:06:47 +02:00
/*============================ timeformat */
Add support for $timeformat.
2002-05-31 06:26:54 +02:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$timeformat";
tf_data.calltf = sys_timeformat_calltf;
tf_data.compiletf = sys_timeformat_compiletf;
tf_data.sizetf = 0;
Add/modify compiletf routines for $timeformat and $printtimescale. Update/add compiletf routines for these two tasks. The $printtimescale task was updated to work with more types of objects.
2009-01-16 20:10:51 +01:00
tf_data.user_data = "$timeformat";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Add support for $timeformat.
2002-05-31 06:26:54 +02:00
Add printtimescale (caryr)
2007-04-12 04:50:51 +02:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$printtimescale";
tf_data.calltf = sys_printtimescale_calltf;
Add/modify compiletf routines for $timeformat and $printtimescale. Update/add compiletf routines for these two tasks. The $printtimescale task was updated to work with more types of objects.
2009-01-16 20:10:51 +01:00
tf_data.compiletf = sys_printtimescale_compiletf;
Add printtimescale (caryr)
2007-04-12 04:50:51 +02:00
tf_data.sizetf = 0;
Add/modify compiletf routines for $timeformat and $printtimescale. Update/add compiletf routines for these two tasks. The $printtimescale task was updated to work with more types of objects.
2009-01-16 20:10:51 +01:00
tf_data.user_data = "$printtimescale";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Add printtimescale (caryr)
2007-04-12 04:50:51 +02:00
Implement severity system tasks $fatal, $error, $warning, and $info.
2009-11-26 01:22:16 +01:00
/*============================ severity tasks */
tf_data.type = vpiSysTask;
tf_data.tfname = "$fatal";
tf_data.calltf = sys_severity_calltf;
tf_data.compiletf = sys_fatal_compiletf;
tf_data.sizetf = 0;
tf_data.user_data = "$fatal";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Implement severity system tasks $fatal, $error, $warning, and $info.
2009-11-26 01:22:16 +01:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$error";
tf_data.calltf = sys_severity_calltf;
tf_data.compiletf = sys_display_compiletf;
tf_data.sizetf = 0;
tf_data.user_data = "$error";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Implement severity system tasks $fatal, $error, $warning, and $info.
2009-11-26 01:22:16 +01:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$warning";
tf_data.calltf = sys_severity_calltf;
tf_data.compiletf = sys_display_compiletf;
tf_data.sizetf = 0;
tf_data.user_data = "$warning";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Implement severity system tasks $fatal, $error, $warning, and $info.
2009-11-26 01:22:16 +01:00
tf_data.type = vpiSysTask;
tf_data.tfname = "$info";
tf_data.calltf = sys_severity_calltf;
tf_data.compiletf = sys_display_compiletf;
tf_data.sizetf = 0;
tf_data.user_data = "$info";
Add support for vpiUserSystf iteration, etc. This patch adds support for iterating over the list of vpiUserSystf tasks/functions that have vpiUserDefn set. The vpiUserDefn property is true by default, but you can call vpip_make_systf_system_defined() to set this property false (will hide the vpiUserSystf object). All the normal system tasks/functions have been modified to call this procedure to remove them from the list of vpiUserSystf objects. Only user defined system tasks or functions should appear in the list. vpi_compare_objects() is just a simple are the two pointers the same. This works correctly for the vpiUserSystf objects, but the other handle objects have not been checked.
2010-04-12 08:39:08 +02:00
res = vpi_register_systf(&tf_data);
vpip_make_systf_system_defined(res);
Implement severity system tasks $fatal, $error, $warning, and $info.
2009-11-26 01:22:16 +01:00
Add support for $timeformat.
2002-05-31 06:26:54 +02:00
cb_data.reason = cbEndOfCompile;
Fix a few memory leaks in vvp. This patch fixes two memory leaks in the compile stage and one easy fix in vpi/sys_display.c.
2009-01-22 01:16:49 +01:00
cb_data.time = 0;
Add support for $timeformat.
2002-05-31 06:26:54 +02:00
cb_data.cb_rtn = sys_end_of_compile;
cb_data.user_data = "system";
vpi_register_cb(&cb_data);
Fix a few memory leaks in vvp. This patch fixes two memory leaks in the compile stage and one easy fix in vpi/sys_display.c.
2009-01-22 01:16:49 +01:00
cb_data.reason = cbEndOfSimulation;
cb_data.cb_rtn = sys_end_of_simulation;
cb_data.user_data = "system";
vpi_register_cb(&cb_data);
Convert vvm to implement system tasks with vpi.
1999-08-15 03:23:56 +02:00
}