Commit Graph

1236 Commits

Author SHA1 Message Date
Matthias Koefferlein 4ffcaba5d1 Important bug fix for MacOS
Unlike Linux, RTTI does not work in MacOS/clang when the classes
originate from different compile units. I think that Linux's C++
runtime not only checks for identical vtable, but alternatively
for same name. On MacOS, dynamic_cast will fail instead. This fix
solves this issue by placing the important steam format option
specializations into a single specific shared object (the DB plugin).
2018-10-04 14:10:22 -07:00
Matthias Koefferlein d8b1808234 Enhanced the build script to properly set the library IDs
This fix assigns proper library IDs and load paths for the dependent
libraries even if they are distributed among several directories.

This is in particular important for the stream and tool plugins
which exist in a folder next to the application but have to refer
to libraries from the Frameworks folder.

The fix consists of extending the library analysis within build4mac.py
and supplying an additional directory listing the target directories.
2018-10-02 18:57:04 -07:00
Matthias Koefferlein f7c4aa0348 (Partial) attempt to fix the plugin detection issue on MacOS 2018-10-02 17:36:30 -07:00
Matthias Koefferlein c053e4f8ec Manual merge of remaining files from lightwave-lab:pymod-test to pymod
git merge failed, probably because there were two parallel attempts
to merge master into pymod leading to opposite resolutions.
2018-10-02 22:41:08 +02:00
Matthias Koefferlein 795db6e03a Merge branch 'lightwave-lab-pymod-test' into pymod 2018-09-30 23:22:56 +02:00
Matthias Koefferlein 51d0cd52cf Merge branch 'pymod-test' of https://github.com/lightwave-lab/klayout into lightwave-lab-pymod-test 2018-09-30 23:19:37 +02:00
Matthias Koefferlein 1ff34fc1e2 Fixed build.bat - wasn't properly exiting and did not really perform a build without install (the latter is important for producing the unit test runner binary 2018-09-27 21:58:16 +02:00
Matthias Koefferlein d03d4bec1d Fixed build after merge. 2018-09-26 21:27:31 +02:00
Matthias Koefferlein 6795de1dbb Merge remote-tracking branch 'remotes/origin/master' into pymod
Provides fixes for issue #166 and #172.
2018-09-26 21:15:08 +02:00
Matthias Koefferlein 02c9661e25 Updated Changelog. 2018-09-26 20:57:50 +02:00
Matthias Köfferlein d676e4ec8a
Merge pull request #173 from klayoutmatthias/issue-172
Fixed #172 (DEF reader does not pull vias from LEF)
2018-09-26 20:50:08 +02:00
Matthias Koefferlein d4adb194e3 Fixed #172 (DEF reader does not pull vias from LEF) 2018-09-26 20:49:04 +02:00
Matthias Köfferlein 808f91223a
Merge pull request #171 from lightwave-lab/testfix/pya_dbTransTest
Changing pya.CplxTrans.angle test from assertEqual to assertAlmostEqual
2018-09-25 20:44:28 +02:00
Matthias Köfferlein ff9a47ca15
Merge pull request #170 from lightwave-lab/testfix/test_42
fixing test test_42 in qtbinding.py and adding note to qtbinding.rb
2018-09-25 20:43:45 +02:00
Thomas Ferreira de Lima 6ddecb1020
travis bugfix: cp -a behaves differently in macs and linux 2018-09-22 19:38:25 -04:00
Thomas Ferreira de Lima 186d7c3ddc
reorganizing dropbox folder to deploy/$(MACOS_VERSION)/$(PYTHON_VERSION)/$(KLAYOUT_VERSION) 2018-09-22 15:42:55 -04:00
Thomas Ferreira de Lima 214662997c
Changing pya.CplxTrans.angle test from assertEqual to assertAlmostEqual
Rationale:
pya.CplxTrans.angle is computed by doing atan2(m_sin, m_cos), but the algorithm used both in C and python for sin, cos and atan2 are not as precise as IEEE's float.

In python, for example, this happens:

``` python
>>> from math import pi, cos, sin, atan2
>>> atan2(sin(45*pi/180), cos(45*pi/180)) * 180 / pi == 45
False
>>> atan2(sin(45*pi/180), cos(45*pi/180)) * 180 / pi, 45
(44.99999999999999, 45)
```
2018-09-22 15:15:33 -04:00
Thomas Ferreira de Lima e217e6a891
[skip ci] fixing test test_42 in qtbinding.py and adding note to qtbinding.rb
Sorry, don't know ruby
2018-09-22 14:29:21 -04:00
Thomas Ferreira de Lima ff6e818da8
[skip ci] minor change in deploy folder name 2018-09-22 13:49:18 -04:00
Thomas Ferreira de Lima 456c062ac0
Tested pymod python build with linux OS 2018-09-22 13:21:40 -04:00
Thomas Ferreira de Lima c6f1cfbae8
adding -std=c++0x flag to linux pymod builds 2018-09-22 12:45:43 -04:00
Thomas Ferreira de Lima 6e0dc93b55
Merge master into pymod (attempt by @tlima) 2018-09-21 16:06:43 -04:00
Matthias Koefferlein 2236a7662d Preparations for 0.25.5 release (updated Changelogs, version) 2018-09-19 22:05:39 +02:00
Matthias Köfferlein 5b34701f8f
Merge pull request #168 from lightwave-lab/travis-ci/v0.25.4
Travis automation formulas for macOS builds
2018-09-19 21:58:34 +02:00
Thomas Ferreira de Lima ef1a8d7f2b
[skip ci] Tested B37 in HighSierra, Sierra and El Capitan 2018-09-18 19:49:11 -04:00
Thomas Ferreira de Lima df301fac06
resizing image to fit python and numpy 2018-09-18 16:41:15 -04:00
Thomas Ferreira de Lima 30193cd7e9
bugfix python embedding (try 2) 2018-09-18 14:34:29 -04:00
Thomas Ferreira de Lima f0086aa5d4
bugfix python embedding (try 1) 2018-09-18 12:19:12 -04:00
Thomas Ferreira de Lima 1ddcfdf1f9
deleting obsolete file filter-clang.sh 2018-09-18 10:44:10 -04:00
Thomas Ferreira de Lima 9a24d32d9f
making DMG instead of tarballs 2018-09-17 21:46:32 -04:00
Matthias Koefferlein 9ccc0a86b1 Fixed a merge issue (build was broken) 2018-09-17 22:34:01 +02:00
Matthias Koefferlein 3b59d97195 Ported Ruby initialization issue bugfix from pymod to master.
On OpenSuSE 15 (gcc 7.3.1, Ruby 2.5.0) crashes have been
observed on starting the application. The main reason seems
to be the attempt to disable $0 redirection in Ruby.

This is not required for Ruby >= 1.9, so this is better dropped.
2018-09-17 21:43:38 +02:00
Matthias Köfferlein 11d5819c77 Fix of a small bug found by MSVC iterator assertions 2018-09-17 21:31:52 +02:00
Matthias Köfferlein cb15132029 Fixed an issue with destruction callbacks in the QApplication destructor - in this case the interpreter may be nulled already and callbacks must not happen into Ruby/Python. 2018-09-17 21:30:54 +02:00
Matthias Köfferlein 8c4099b77d Bugfix in the annotation area
The annotation ID was taken from an object already deleted - on MSVC 
this becomes an issue.
2018-09-17 21:25:28 +02:00
Matthias Köfferlein 88e0bd77a2 Fix in Image object handling
Id was taken from an object destructed already.
2018-09-17 21:24:47 +02:00
Matthias Köfferlein 1d7101bced
Merge pull request #167 from klayoutmatthias/issue-166
Issue 166
2018-09-17 21:21:13 +02:00
Thomas Ferreira de Lima d68e0bbe76
writing unique names for tarballs
KLAYOUT_VERSION incorporated in makefile
2018-09-17 14:50:27 -04:00
Matthias Koefferlein 6efdb3a9e1 Added testcase for #166. 2018-09-17 01:51:23 +02:00
Matthias Koefferlein f72fda7bae Added test case for #166. 2018-09-17 01:48:31 +02:00
Matthias Koefferlein e9de4252fb Fixed #166 (internal error when writing GDS file)
This fix consists of computing the intersection
points in the split procedure with higher resolution
to avoid topology changes due to snapping on the cut
line.
2018-09-17 01:29:31 +02:00
Matthias Köfferlein fe26785203 Fixed install target for MSVC - copy needs to use Windows paths. 2018-09-13 22:31:45 +02:00
Matthias Köfferlein f7664c09d5 Windows build script and setup.py for Anaconda (install needs testing). 2018-09-13 01:11:18 +02:00
Matthias Koefferlein 06c0b8f6e1 And back to Windows ... re-inserting the ruby_sysinit workaround. 2018-09-12 00:06:48 +02:00
Matthias Koefferlein 0a2574cd3a Now it breaks the ruby 1.8.x tests ... grrr 2018-09-11 23:19:52 +02:00
Matthias Koefferlein 121e1301b0 Hopefully this fixes the crashes on OpenSuSE 15. 2018-09-11 22:35:57 +02:00
Matthias Koefferlein 372823efcb grrr ... Debian package still complains. 2018-09-11 20:41:46 +02:00
Matthias Koefferlein 6536e9a8dc Modified command line prep for Ruby integration - maybe that fixes the OpenSuSE 15 crashes. 2018-09-11 08:08:18 +02:00
Matthias Koefferlein e16a37d3af Debian's packager check is still super-picky about Changelog. 2018-09-11 07:36:24 +02:00
Matthias Köfferlein 6a12a6a4cd Normalize CRLF for OASIS reader test's golden data on Windows. 2018-09-11 07:35:01 +02:00