`PYBIND11_NAMESPACE` consistency fixes. #4043

rwgk · 2022-07-05T18:51:12Z

Description

This PR fixes a couple long-standing PYBIND11_NAMESPACE __attribute__((visibility("hidden"))) inconsistencies. The fixes only affect builds with __GNUG__ defined.

Based on experiments under PR #4044, this (currently) only makes a difference in an unusual situation: macOS, NOT using -fvisibility=hidden. However, if someone ventures there, surprises due to inconsistencies can be extremely time-consuming to troubleshoot.

(Discovered while experimenting related to PR #4022.)

Suggested changelog entry:

A couple long-standing ``PYBIND11_NAMESPACE`` ``__attribute__((visibility("hidden")))`` inconsistencies are now fixed (affects only unusual environments).

rwgk · 2022-07-06T21:29:12Z

Thanks!

rwgk · 2022-07-06T22:18:14Z

Just to note, the one post-merge CI failure is a flake I saw several times before already.

Run cmake --build build2 --target pytest
Microsoft (R) Build Engine version [1](https://github.com/pybind/pybind11/runs/7223230463?check_suite_focus=true#step:18:1)7.2.1+52cd2da31 for .NET Framework
Copyright (C) Microsoft Corporation. All rights reserved.
  cross_module_gil_utils.vcxproj -> D:\a\pybind11\pybind11\build2\tests\cross_module_gil_utils.pypy37-pp73-win_amd64.pyd
  cross_module_interleaved_error_already_set.vcxproj -> D:\a\pybind11\pybind11\build2\tests\cross_module_interleaved_error_already_set.pypy37-pp73-win_amd64.pyd
  pybind11_cross_module_tests.vcxproj -> D:\a\pybind11\pybind11\build2\tests\pybind11_cross_module_tests.pypy37-pp73-win_amd64.pyd
  pybind11_tests.vcxproj -> D:\a\pybind11\pybind11\build2\tests\pybind11_tests.pypy37-pp73-win_amd64.pyd
  ------ pybind11_tests.pypy37-pp73-win_amd64.pyd file size: 22369792 (no change)
C:\Program Files\Microsoft Visual Studio\2022\Enterprise\MSBuild\Microsoft\VC\v170\Microsoft.CppCommon.targets(245,5): error MSB8066: Custom build for 'D:\a\pybind11\pybind11\build2\CMakeFiles\c74e89d6f5f1ea[10](https://github.com/pybind/pybind11/runs/7223230463?check_suite_focus=true#step:18:11)8a7628c93f4b5083\pytest.rule;D:\a\pybind[11](https://github.com/pybind/pybind11/runs/7223230463?check_suite_focus=true#step:18:12)\pybind11\tests\CMakeLists.txt' exited with code -107374[18](https://github.com/pybind/pybind11/runs/7223230463?check_suite_focus=true#step:18:19)19. [D:\a\pybind11\pybind11\build2\tests\pytest.vcxproj]

To probe environment/toolchain/platform-specific behavior under the exact same conditions as normal tests. (An earlier version of this code was used to inform PR pybind#4043.)

rwgk · 2022-07-08T17:28:19Z

Pure record keeping, JIC it's useful for something later:

(For reasons that are difficult to explain) I compared XPASS before and after this PR was merged:

master CI PR docs: update changelog #4042 merged: https://github.com/pybind/pybind11/actions/runs/2625465392
master CI PR PYBIND11_NAMESPACE consistency fixes. #4043 merged: https://github.com/pybind/pybind11/actions/runs/2625732463

Due to the flake mentioned in my previous comment, one result is missing. Apart from that, the XPASS are identical between the two CI workflow logs.

Dumping the current XPASS here while I have them handy (complete list from runs/2625465392):

5______pypy-3.7_____ubuntu-latest_____x64.txt:XPASS test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
5______pypy-3.7_____ubuntu-latest_____x64.txt:XPASS test_exceptions.py::test_python_alreadyset_in_destructor Failure on PyPy 3.8 (7.3.7)
5______pypy-3.7_____ubuntu-latest_____x64.txt:XPASS ../../tests/test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
5______pypy-3.7_____ubuntu-latest_____x64.txt:XPASS ../../tests/test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
5______pypy-3.7_____ubuntu-latest_____x64.txt:XPASS ../../tests/test_exceptions.py::test_python_alreadyset_in_destructor Failure on PyPy 3.8 (7.3.7)
6______pypy-3.8_____ubuntu-latest_____x64_-DPYBIND11_FINDPYTHON=ON.txt:XPASS test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
6______pypy-3.8_____ubuntu-latest_____x64_-DPYBIND11_FINDPYTHON=ON.txt:XPASS ../../tests/test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
6______pypy-3.8_____ubuntu-latest_____x64_-DPYBIND11_FINDPYTHON=ON.txt:XPASS ../../tests/test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
7______pypy-3.9_____ubuntu-latest_____x64.txt:XPASS test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
7______pypy-3.9_____ubuntu-latest_____x64.txt:XPASS ../../tests/test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
7______pypy-3.9_____ubuntu-latest_____x64.txt:XPASS ../../tests/test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
12______pypy-3.7_____windows-2022_____x64.txt:  XPASS test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
12______pypy-3.7_____windows-2022_____x64.txt:  XPASS test_exceptions.py::test_python_alreadyset_in_destructor Failure on PyPy 3.8 (7.3.7)
12______pypy-3.7_____windows-2022_____x64.txt:  XPASS ..\..\tests\test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
12______pypy-3.7_____windows-2022_____x64.txt:  XPASS ..\..\tests\test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
12______pypy-3.7_____windows-2022_____x64.txt:  XPASS ..\..\tests\test_exceptions.py::test_python_alreadyset_in_destructor Failure on PyPy 3.8 (7.3.7)
13______pypy-3.8_____windows-2022_____x64.txt:  XPASS test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
13______pypy-3.8_____windows-2022_____x64.txt:  XPASS ..\..\tests\test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
13______pypy-3.8_____windows-2022_____x64.txt:  XPASS ..\..\tests\test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
14______pypy-3.9_____windows-2022_____x64.txt:  XPASS test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
14______pypy-3.9_____windows-2022_____x64.txt:  XPASS ..\..\tests\test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
14______pypy-3.9_____windows-2022_____x64.txt:  XPASS ..\..\tests\test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
19______pypy-3.7_____macos-latest_____x64.txt:XPASS test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
19______pypy-3.7_____macos-latest_____x64.txt:XPASS test_exceptions.py::test_python_alreadyset_in_destructor Failure on PyPy 3.8 (7.3.7)
19______pypy-3.7_____macos-latest_____x64.txt:XPASS ../../tests/test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
19______pypy-3.7_____macos-latest_____x64.txt:XPASS ../../tests/test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
19______pypy-3.7_____macos-latest_____x64.txt:XPASS ../../tests/test_exceptions.py::test_python_alreadyset_in_destructor Failure on PyPy 3.8 (7.3.7)
20______pypy-3.8_____macos-latest_____x64.txt:XPASS test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
20______pypy-3.8_____macos-latest_____x64.txt:XPASS ../../tests/test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
20______pypy-3.8_____macos-latest_____x64.txt:XPASS ../../tests/test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
21______pypy-3.9_____macos-latest_____x64.txt:XPASS test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
21______pypy-3.9_____macos-latest_____x64.txt:XPASS ../../tests/test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy
21______pypy-3.9_____macos-latest_____x64.txt:XPASS ../../tests/test_call_policies.py::test_keep_alive_argument sometimes comes out 1 off on PyPy

@charlesbeattie

* Add test_namespace_visibility To probe environment/toolchain/platform-specific behavior under the exact same conditions as normal tests. (An earlier version of this code was used to inform PR #4043.) * Disable flake8 in ubench/holder_comparison_*.py, to suppress new & useless diagnostics. * Disable namespace_visibility_1s.cpp (tosee if that resolves the MSVC and CUDA `test_cross_module_exception_translator` failures). * Turn off flake8 completely for ubench (the Strip unnecessary `# noqa`s action un-helpfully removed the added noqa). * Disable test_namespace_visibility completely. Just keep the two .cpp files, only setting the module docstring and doing nothing else. * Rename test_namespace_visibility.py to test_exc_namespace_visibility.py, so that it is imported by pytest before test_exceptions.py * Add `set_property(SOURCE namespace_visibility_1s.cpp PROPERTY LANGUAGE CUDA)` * Add reference to PR #4054 * Complete the documentation (comments in test_exc_namespace_visibility.py). * Rename namespace_visibility.h to namespace_visibility.inl, as suggested by @charlesbeattie

PYBIND11_NAMESPACE consistency fixes.

760be32

rwgk mentioned this pull request Jul 6, 2022

Add test_namespace_visibility (experiment, not for merging) #4044

Closed

rwgk marked this pull request as ready for review July 6, 2022 20:51

rwgk requested review from henryiii and Skylion007 July 6, 2022 20:51

Skylion007 approved these changes Jul 6, 2022

View reviewed changes

rwgk merged commit cd08869 into pybind:master Jul 6, 2022

rwgk deleted the PYBIND11_NAMESPACE_consistency_fixes branch July 6, 2022 21:29

github-actions bot added the needs changelog Possibly needs a changelog entry label Jul 6, 2022

rwgk mentioned this pull request Jul 8, 2022

[smart_holder] Add test_namespace_visibility #4050

Merged

henryiii removed the needs changelog Possibly needs a changelog entry label Oct 20, 2022

rwgk mentioned this pull request Oct 22, 2022

[BUG]: ABI incompatibility between 2.9.2 and 2.10 #4105

Closed

3 tasks

rwgk mentioned this pull request Feb 10, 2023

FWD pybind11 google/pybind11clif#4043

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

`PYBIND11_NAMESPACE` consistency fixes. #4043

`PYBIND11_NAMESPACE` consistency fixes. #4043

Uh oh!

rwgk commented Jul 5, 2022 •

edited by henryiii

Loading

Uh oh!

rwgk commented Jul 6, 2022

Uh oh!

rwgk commented Jul 6, 2022

Uh oh!

rwgk commented Jul 8, 2022

Uh oh!

Uh oh!

PYBIND11_NAMESPACE consistency fixes. #4043

PYBIND11_NAMESPACE consistency fixes. #4043

Uh oh!

Conversation

rwgk commented Jul 5, 2022 • edited by henryiii Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Suggested changelog entry:

Uh oh!

rwgk commented Jul 6, 2022

Uh oh!

rwgk commented Jul 6, 2022

Uh oh!

rwgk commented Jul 8, 2022

Uh oh!

Uh oh!

`PYBIND11_NAMESPACE` consistency fixes. #4043

`PYBIND11_NAMESPACE` consistency fixes. #4043

rwgk commented Jul 5, 2022 •

edited by henryiii

Loading