Commit Graph

5613 Commits

Author SHA1 Message Date
Matthias Koefferlein 2d5778a860 'tap': do not consider texts as detecting them requires capturing a large area - which is against performance 2025-07-21 20:40:01 +02:00
Matthias Koefferlein 57025a7f9f Added fill tests with differen origins and tiled mode 2025-07-19 22:37:38 +02:00
Matthias Koefferlein e48e030521 Fixing DRC fill in tiled mode - must not use overlapping tiles with shifting origin 2025-07-19 22:30:06 +02:00
Matthias Koefferlein 1a9c12ce18 Merge branch 'master' into devel 2025-07-19 21:27:52 +02:00
Matthias Koefferlein 9d4a578a5e Merge branch 'bugfix/issue-2088' 2025-07-19 19:00:33 +02:00
Matthias Köfferlein 71202ea9d3
Merge pull request #2098 from KLayout/bugfix/issue-2094
Fixing issue #2094 (slow insert into Shapes)
2025-07-19 18:58:04 +02:00
Matthias Köfferlein ff7d7f20ee
Merge pull request #2089 from KLayout/bugfix/issue-2087
Bugfix/issue 2087
2025-07-19 18:57:25 +02:00
Matthias Köfferlein 131f36a4e0
Merge pull request #2084 from KLayout/feature/issue-2044
Feature/issue 2044
2025-07-19 18:57:10 +02:00
Matthias Köfferlein 316a58eca5
Merge pull request #2079 from KLayout/bugfix/issue-2075
Fixing issue #2075 (strm2oas ignores def net routing wire extensions)
2025-07-19 18:56:35 +02:00
Matthias Koefferlein 4ab64ca5d3 Fixing issue #2094 (slow insert into Shapes)
Solution is to use a temporary container instead
of locking (only if needed).
2025-07-19 15:08:09 +02:00
Matthias Koefferlein e6fa072bee Fixing 'tap' feature
After introducing the "do not select shapes
or instances if they overlap the viewport entirely"
feature, 'tap' was not finding shapes which extended
beyond the view's borders.

This got fixed and for 'tap', such shapes are
considered again.
2025-07-18 23:00:23 +02:00
Matthias Koefferlein d0b935d9e5 Fixed issue #2088 in reader and writer
Reader fix: the reader will not error out duplicate cell
names, but rename the cells.

Writer fix: the writer will uniquify cell names *after*
illegal character substitution.
2025-07-15 22:52:49 +02:00
Matthias Koefferlein 1dc0a56633 Bugfix on layout query
The following query did not render the correct box
for top-level objects:

select path_dtrans*shape.dbbox.center,cell_name from shapes on layer 1/0 from instances of .*..

Reason was that "path_dtrans" returned a ICplxTrans for top
level instead of a DCplxTrans object.
2025-07-13 21:42:35 +02:00
Matthias Koefferlein 67994a2f8a Introducing convenience method ObjectInstPath#cell 2025-07-13 21:27:26 +02:00
Matthias Koefferlein 83913f07bd Updated documentation 2025-07-13 21:20:37 +02:00
Matthias Koefferlein 87bc1a60e1 Introducing macro attributes and Macro#current 2025-07-13 21:14:53 +02:00
Matthias Koefferlein cc796317a0 Safeguard against broken code table in OASIS CBLOCKs 2025-07-13 19:31:47 +02:00
Matthias Koefferlein a6a012beac Added DRC full scale tests, first is Greyhound with M1 IHP fill 2025-07-12 22:02:44 +02:00
Matthias Koefferlein e7bc997e9f Further compressing fill cell arrays in x direction 2025-07-12 14:28:42 +02:00
Matthias Koefferlein 76f7c620f6 [consider merging] Enabling Shapes-based RecursiveShapeIterators for MT-use
This fixes a crash when using DRC fill in tiled mode with multiple
threads.
2025-07-12 13:59:06 +02:00
Matthias Koefferlein 70a609e995 [consider merging] Bugfix in polygon rasterizer 2025-07-11 23:57:03 +02:00
Matthias Koefferlein 46f9fba0ab Generalizing fill tool code 2025-07-11 23:56:40 +02:00
Matthias Koefferlein 7a06bfd66a Bugfix 2025-07-10 23:32:19 +02:00
Matthias Koefferlein 9bc49dd5a3 Added fill_pattern's 'margin' feature. This allows keeping a distance to features outside of the fill region in a more consistent way. 2025-07-10 00:31:25 +02:00
Matthias Koefferlein d5f1017f83 More efficient computation of remaining fill area 2025-07-10 00:24:24 +02:00
Matthias Koefferlein 7dafdb2316 Fixing bug in first place. Needs testing. 2025-07-08 23:36:44 +02:00
Matthias Koefferlein 15120760fe Handling of ghost cells in strm2x merge
Ghost cells act as anchors for top cells of the other
layout in "," merges. Unlike other cells which are
renamed.
2025-07-07 21:52:57 +02:00
Matthias Koefferlein d382629e8e Proposal for a fix of issue #2044 and #2066
The solution is to separate the meaning of "," and "+" layout
file combination operators in strm2x tools:

* "+" does "blending"
* "," does merge

"+" has priority over "," ('a+b,c' acts as '(a+b),c').
NOTE: LEF caching only happens across "+".
2025-07-05 22:48:05 +02:00
Matthias Koefferlein 0ba6064507 Also supporting DBU check for PCB 2025-07-05 21:17:46 +02:00
Matthias Koefferlein 9e9344fdfa Introducing a check whether DBU is compatible with a given one in most readers (except PCB) 2025-07-05 21:13:19 +02:00
Matthias Koefferlein fbca4e6c18 Fixing layout text writer for 'long long' property names on Windows 2025-06-29 22:16:46 +02:00
Matthias Koefferlein 733a0ec4cf Proper solution for bug fix - wasn't done in the right order before. 2025-06-29 20:23:47 +02:00
Matthias Koefferlein abeab280a3 [consider merging] fixed a crash happening while blending two OASIS files with ghost cells 2025-06-28 15:50:27 +02:00
Matthias Koefferlein 5a401127a2 Trying to fix MacOS and Windows builds 2025-06-28 15:31:46 +02:00
Matthias Koefferlein a42e639d83 Trying to fix builds on MacOS 2025-06-27 17:56:45 +02:00
Matthias Koefferlein 7acd11abac Added test 2025-06-27 17:37:19 +02:00
Matthias Koefferlein aba1b30c8d Fixing issue #2075 (strm2oas ignores def net routing wire extensions) 2025-06-27 01:04:19 +02:00
Matthias Koefferlein 6ce61c8654 Fixing type of modal variables for box width and height 2025-06-26 22:54:48 +02:00
Matthias Koefferlein a08d36b5f5 Consistently replacing 'long' and 'long long' by 64 bit int in OASIS reader/writer, so there should not be a difference between Windows and Linux in reader and writer behavior. Database limitations still apply. 2025-06-26 21:59:28 +02:00
Matthias Koefferlein 5c7ac018a8 Not asking for copy mode for leaf cells isn't a bad idea in general, but it creates confusing behavior when pasting instances into a different layout where the target cell does not exist 2025-06-26 20:54:16 +02:00
Matthias Koefferlein f3cc7cadc0 Enabling Qt-less DB plugins for Qt-enabled applications 2025-06-26 20:35:39 +02:00
Matthias Koefferlein eb3be68030 Removing an unused declaration 2025-06-26 20:34:53 +02:00
Matthias Koefferlein 5cd1c187ab Modified display of devices in LVS xref view, so the primary and secondary parameters are separated 2025-05-31 18:30:13 +02:00
Matthias Koefferlein f2172d8e2a Extension of the L2N/LVSDB format to support deleted terminals and parameters
Previously, when deleting parameters or terminals from device
definitions, these parameters or terminals re-appeared in the
netlist browser, because they were generated from the
template class (e.g. "A" and "B" from "RES").

An additional token is added that indicates whether to
remove all entries before adding new ones.

This feature is backward-compatible and LVSDB/L2N files have to
be generated from new KLayout versions to make use of that
feature.
2025-05-31 17:33:44 +02:00
Matthias Koefferlein 769b7e9170 Typo fixed 2025-05-29 13:31:36 +02:00
Matthias Koefferlein 98834d8e3b Reference new golden testdata for LEFDEF/strm2oas test 2025-05-29 12:00:44 +02:00
Matthias Koefferlein f1cc048a4c Refreshing stubs 2025-05-29 09:56:11 +02:00
Matthias Köfferlein 45fdaf0a6d
Merge pull request #2049 from KLayout/devel
Devel
2025-05-29 09:43:23 +02:00
Matthias Köfferlein 3ce50679fa
Merge pull request #2050 from KLayout/dev-polygon-graph
Dev polygon graph
2025-05-29 09:43:06 +02:00
Matthias Köfferlein 6ae3737cd5
Merge pull request #2065 from KLayout/feature/issue-2041
Feature/issue 2041
2025-05-29 09:42:46 +02:00
Matthias Koefferlein 9917978566 Merge branch 'master' into dev-polygon-graph 2025-05-28 19:01:42 +02:00
Matthias Koefferlein ad444c5552 Don't error out on missing MACRO in LEF, but create a dummy macro with size 0,0 instead and issue a warning. This will place nicely with N, but not with other orientations. 2025-05-28 18:58:28 +02:00
Matthias Koefferlein 0cb5f35659 Allowing the polygon neighborhood visitor to specify a variant type for deep mode 2025-05-27 21:07:28 +02:00
Matthias Köfferlein dfd7d8ec1f
Merge pull request #2061 from KLayout/bugfix/issue-2060
Bugfix/issue 2060
2025-05-27 19:17:45 +02:00
Matthias Köfferlein 836c50d6a0
Merge pull request #2062 from KLayout/feature/issue-2055
Feature/issue 2055
2025-05-27 19:17:33 +02:00
Matthias Koefferlein 010fa2d5ba Avoiding a segfault when trying to run Ruby from a non-Ruby thread 2025-05-27 00:48:10 +02:00
Matthias Koefferlein 13bc723831 Fix for MSVC builds 2025-05-25 23:24:38 +02:00
Matthias Koefferlein 6cccb81293 Implemented solution for #2057 (nanometer scalebar)
The scale bar switches to nm below 0.1µm. Correspondingly
it will switch to mm above 100µm.
2025-05-25 21:28:11 +02:00
Matthias Koefferlein 534b33be1c MSVC does not assume UTF-8 encoding by default, hence use explicit bytes 2025-05-25 20:40:45 +02:00
Matthias Koefferlein bfe49c22e3 Properly using 'long' and 'unsigned long' in tl::sprintf to fit Windows definition of 'long' 2025-05-25 19:42:14 +02:00
Matthias Koefferlein 329ac6c5b7 Merge branch 'master' into devel 2025-05-25 19:37:57 +02:00
Matthias Koefferlein fb16c8c6f6 Trying to fix a linker issue 2025-05-25 19:01:04 +02:00
Matthias Koefferlein 4a20a30888 Hope to fix dependency on Ruby version rgarding string encoding. 2025-05-25 18:55:52 +02:00
Matthias Koefferlein 125e06bd49 Added doc, added a warning about no pins found at top level. 2025-05-25 17:53:11 +02:00
Matthias Koefferlein ad80019b12 Adding to_s (aka str(...)) methods to RNetExtractor tech objects 2025-05-25 16:31:20 +02:00
Matthias Koefferlein dec7ad9da1 [consider merging] properly conveying string encoding for Ruby, so that UTF-8 encoding is maintained when returning such strings from C++ 2025-05-25 16:28:46 +02:00
Matthias Koefferlein 094e11897a Added the new feature: 'flag_missing_ports' 2025-05-24 22:33:18 +02:00
Matthias Koefferlein aba912107b Refined solution, so it would allow a larger coordinate range. 2025-05-24 00:09:52 +02:00
Matthias Koefferlein 6d28c78eeb Fixed issue #2060 - the precision of the check was too low. 2025-05-22 22:48:48 +02:00
Matthias Koefferlein e96f20a106 Introducing case sensitivity in circuit and netlist queries, makeing 'same_nets' more case sensitivity-aware in case of mixed sensitivity 2025-05-22 22:25:49 +02:00
Matthias Koefferlein 57c2add00a API enhancement for NetlistCrossReference: obtaining information from one object only (net, circuit), does not need a pair anymore 2025-05-22 20:54:21 +02:00
Matthias Koefferlein 90c4f654b6 API for adding log entries to LVS and L2N databases 2025-05-22 19:57:06 +02:00
Matthias Koefferlein 4dd4524da9 Fixed typo 2025-05-22 19:47:39 +02:00
Matthias Köfferlein 23b049627f
Merge pull request #2040 from KLayout/feature/maly
Feature/maly
2025-05-22 19:22:52 +02:00
Matthias Köfferlein ba640b6d04
Merge pull request #2039 from KLayout/feature/issue-2016
First attempt to implement a solution for issue #2016
2025-05-22 19:22:32 +02:00
Matthias Koefferlein 6b8c79c488 Fixed a segfault (thanks, Martin\!) 2025-05-22 18:50:25 +02:00
Henner Zeller f9e2a9257f Fix logic error in hex2int.
This was probably never noticed as it was the last branch and would
behave benign on valid input.

Signed-off-by: Henner Zeller <hzeller@google.com>
2025-05-19 14:16:18 +02:00
Matthias Koefferlein 9b03a1ba64 More robust tests 2025-05-18 22:41:46 +02:00
Matthias Koefferlein a2ac8d45de Postpone decision about "invalid vias are errors." 2025-05-18 22:28:11 +02:00
Matthias Koefferlein e27e24ff4f Bugfixing multi-DEF reader with LEF cache, adding a testcase. Making unknown vias an error. 2025-05-18 22:00:39 +02:00
Matthias Koefferlein 4306b24b4a Normalizing test case for RNetExtractor to reduce jitter 2025-05-18 09:27:10 +02:00
Matthias Koefferlein a07d742bee Fixed a problem with via uniquification - via definitions may be DEF local, so we need to clean them between different DEF reads. 2025-05-18 00:48:10 +02:00
Matthias Koefferlein 1fb0f318dc Added pex lib to bd tools. 2025-05-17 18:49:32 +02:00
Matthias Koefferlein 91cb8826c7 Including more dependencies in build 2025-05-17 18:41:22 +02:00
Matthias Koefferlein 5dd189d413 Including pex dependencies in build 2025-05-17 18:39:16 +02:00
Matthias Koefferlein ca53d8718b Fixed some build problems 2025-05-17 17:26:59 +02:00
Matthias Koefferlein 8221923ccb Added a large (private) testcase for the 'read LEF once' enhancement 2025-05-17 16:50:47 +02:00
Matthias Koefferlein 4206d7ee63 Tried a performance improvement experiment, don't dare to actually use it 2025-05-17 01:00:57 +02:00
Matthias Koefferlein baed5bd822 Small refactoring 2025-05-17 00:57:36 +02:00
Matthias Koefferlein 5c27fedfe5 Debugging and bug fixes. The 'big example' from issue 2014 now works 2025-05-17 00:44:15 +02:00
Matthias Koefferlein 113c701345 Trying explicit member template instantiation to solve linker issue 2025-05-16 23:23:11 +02:00
Matthias Koefferlein 1c4077449b Fixed build issue 2025-05-16 18:55:55 +02:00
Matthias Koefferlein 1b98efd7a8 First bug fixes 2025-05-15 23:41:23 +02:00
Matthias Koefferlein 271800ab94 WIP: first implementation. Builds, but needs testing. 2025-05-15 23:27:03 +02:00
Matthias Koefferlein 6cd4b97cd3 Including internal net names in LVS DB
Problem: so far, these names have not been conveyed.
But after the net IDs changed in the LVS DB (renumbering),
the names should be retained. Otherwise it is difficult to
relate log level entries or Spice netlists to them.
2025-05-15 21:17:15 +02:00
Matthias Koefferlein 7bee9ebcec Fixing a linker issue on Windows. 2025-05-11 21:48:01 +02:00
Charlie Lin e279888f52
Use Py_IncRef in pyaCallables.cc 2025-05-11 09:07:55 -04:00
Matthias Koefferlein 91005d5cb6 Tests for RBA+pex, bug fixes 2025-05-04 20:42:12 +02:00
Matthias Koefferlein 77aa729b06 GSI binding of RNetExtractor, introducing layers for nodes 2025-05-04 19:45:30 +02:00
Matthias Koefferlein fc25590dd7 Include a simplification step in the net extraction 2025-05-04 17:15:47 +02:00