Catch2 fails to compile on Fedora Rawhide #2178

lzaoral · 2021-02-17T10:27:40Z

Describe the bug
Tests that use Catch2 fail to compile on Fedora Rawhide.

Expected behavior
Tests that use Catch2 compile and execute without any problems on Fedora Rawhide.

Reproduction steps
Try to compile the factorial example from the Catch2 tutorial:

$ c++ test.cpp
In file included from /usr/include/signal.h:315,
                 from catch.hpp:8034,
                 from test.cpp:2:
catch.hpp:10822:58: error: call to non-'constexpr' function 'long int sysconf(int)'
10822 |     static constexpr std::size_t sigStackSize = 32768 >= MINSIGSTKSZ ? 32768 : MINSIGSTKSZ;
      |                                                          ^~~~~~~~~~~
In file included from /usr/include/bits/sigstksz.h:24,
                 from /usr/include/signal.h:315,
                 from catch.hpp:8034,
                 from test.cpp:2:
/usr/include/unistd.h:630:17: note: 'long int sysconf(int)' declared here
  630 | extern long int sysconf (int __name) __THROW;
      |                 ^~~~~~~
In file included from test.cpp:2:
catch.hpp:10881:45: error: size of array 'altStackMem' is not an integral constant-expression
10881 |     char FatalConditionHandler::altStackMem[sigStackSize] = {};
      |                                             ^~~~~~~~~~~~

Platform information:

OS: latest Fedora Rawhide
Compiler+version: gcc (GCC) 11.0.0 20210210 (Red Hat 11.0.0-0)
Catch version: v2.13.4

Additional context
This behaviour seems to be related to this change in glibc: https://sourceware.org/git/?p=glibc.git;a=commit;h=6c57d320484988e87e446e2e60ce42816bf51d53

The text was updated successfully, but these errors were encountered:

horenmar · 2021-02-17T12:24:08Z

Humm, can't say I am happy that MINSIGSTKSZ became non-constant, but it should be easy enough to fix...

See catchorg/Catch2#2178.

tomhughes · 2021-02-19T10:25:33Z

It's also broken catch 1 which uses SIGSTKSZE which is also now non-constant.

tomhughes · 2021-02-19T11:16:57Z

I've pushed new builds of catch1 and catch for Fedora Rawhide now which hardwire the stack size as 32Kb until we have a better fix.

This patch is about fixing a compilation error that we are seeing on Fedora Rawhide with glibc 2.33.9000, which makes MINSIGSTKSZ not be a constant value anymore. Thus, the sigStackSize variable that is declared constexpr cannot use that MINSIGSTKSZ as initializer anymore. So as suggested in the issue catchorg/Catch2#2178 filed against 'catchorg' for this purpose, I am hardwiring the initialization value of sigStackSize for now. * tests/lib/catch.hpp: Update to v2.13.4 and initialize sigStackSize to 32768 for now, as suggested by catchorg/Catch2#2178. Signed-off-by: Dodji Seketeli <[email protected]>

Skip tests to workaround a compilation error same as catchorg/Catch2#2178

Because new glibc has changed `MINSIGSTKSZ` to be a syscall instead of being constant, the signal posix handling needed changes, as it used the value in constexpr context, for deciding size of an array. It would be simple to fix it by having the handler determine the signal handling stack size and allocate the memory every time the handler is being installed, but that would add another allocation and a syscall every time a test case is entered. Instead, I split apart the idea of preparing fatal error handlers, and engaging them, so that the memory can be allocated only once and still be guarded by RAII. Also turns out that Catch2's use of `MINSIGSTKSZ` was wrong, and we should've been using `SIGSTKSZ` the whole time, which we use now. Closes #2178

horenmar · 2021-04-04T16:27:46Z

Should be resolved in current commit on the v2.x branch. You will need to regenerate the single header to test it (use scripts/generateSingleHeader.py)

lzaoral · 2021-04-06T21:55:25Z

The fix works for me as well. Thanks!

horenmar · 2021-04-11T20:02:38Z

Closing as the fixed version has been released, and this won't autoclose since v2 is no longer main branch.

Fixes compilation with new glibc on Fedora Rawhide. See: catchorg/Catch2#2178

Because new glibc has changed `MINSIGSTKSZ` to be a syscall instead of being constant, the signal posix handling needed changes, as it used the value in constexpr context, for deciding size of an array. It would be simple to fix it by having the handler determine the signal handling stack size and allocate the memory every time the handler is being installed, but that would add another allocation and a syscall every time a test case is entered. Instead, I split apart the idea of preparing fatal error handlers, and engaging them, so that the memory can be allocated only once and still be guarded by RAII. Also turns out that Catch2's use of `MINSIGSTKSZ` was wrong, and we should've been using `SIGSTKSZ` the whole time, which we use now. Closes #2178

see catchorg/Catch2#2178 Signed-off-by: Dennis Gilmore <[email protected]>

Based on this patch: woodard/libabigail@8ae8dcb Please refer to the catch issue #2178 for details catchorg/Catch2#2178 This might not work on all platforms Signed-off-by: Christian Ege <[email protected]>

Fixes catchorg/Catch2#2178 on Ubunto 22.04

The currently used version still suffers from the issue that apparently MINSIGSTKSZ is no longer a constant expression in newer glibc versions. Upstream bug report: catchorg/Catch2#2178

catch.hpp:10822:58: error: call to non-'constexpr' function 'long int sysconf(int)' followed by: catch.hpp:10881:45: error: size of array 'altStackMem' is not an integral constant-expression

-TODO:Move dwarf versions testing to os-level/Dockerfile. This avoids having to rely on expiremental features of DWARF support for some gcc versions/DWARF version combinations. -Update Catch2 for Ubuntu22. catchorg/Catch2#2178

* -Write macro records correctly. Fixes #7 * -Cleanup * -Update unit testing for draco patterns. WIP. * -Update eclipse project * -Minimally functional target_symbol * -Do not store row id for symbol target updates * -Update github CI * -Docker config for CI * -Point submodules to github * -Cleanup * -Add symbols and encoding mappings. WIP. * -Add symbols and encoding mappings. WIP. * Writer DWARF5 ecodings to encodings table * -Add minimally functional encodings implementation. -TODO:Cleanup. * -Update Makefile * -Update gitignore * -Cleanup * -Remove Ubuntu18 from github actions since it is not available anymore. -https://github.blog/changelog/2022-08-09-github-actions-the-ubuntu-18-04-actions-runner-image-is-being-deprecated-and-will-be-removed-by-12-1-22/ * -Remove dead code * -Cleanup * -Cleanup * -Cleanup * -Include unit test code in format checks. -Format unit test code. * -Update unit tests * -Configure coveralls.io for CI * -Configure coveralls.io for CI * -Configure coveralls.io for CI * -Configure coveralls.io for CI * -Configure coveralls.io for CI * -Configure coveralls.io for CI * -Configure coveralls.io for CI * -Configure coveralls.io for CI * -Update docs * -Update unit tests for new schema. WIP. -TODO:Add function to follow target symbols to test typdef'd types. -TODO:Add unit tests for new macro and elf image features. * -Add followTargetSymbol. Useful for testing typedef'd types. -Update unit tests. * -Use CFE_ES_HousekeepingTlm_Payload for unit testing. * -Update docs * -Update docs * -Extract ELF image data from 64-bit targets properly. -Add ElfClass to ElfFile * -Update docs * -Update unit-test test files * -Remove libdwarf from unit test code test file * -Remove dead code * -Remove dead code * -Add 32-bit make recipes. -Add 32-bit unit tests. * -Update 32-bit unit testing. * -Update docs * -Remove dead code * -Update Logger unit tests -Remove dead code * -Update unit tests * -Update docs * -Cleanup * -Cleanup * -Cleanup * -Add macro unit test. -Update Dockerfile * -Create test files that comply with DWARF4 and DWARF5. Allows us to verify support of DWARF4 as more features are added in the future. * -Unit test DWARF4 and DWARF5. WIP. * -Add DWARF4 testing to Dockerfile * -Cleanup * -Update docs * -Update docs * -Update Catch 2 due to issue on Ubuntu22:catchorg/Catch2#2178 * -pathIndex handling for DWARF5. TODO:This changes for DWARF4, so add alternative for DWARF4. -Remove DisplayDie function for now as it blows up the stack with big c++ names inside DWARF. -TODO:Rewrite DisplayDie function -Add include to src/Artifact.h, needed for Ubuntu22. * -Handle DWARF dbg source files for DWARF4 and 5. -TODO:Figure out what to do about Ubuntu 20/22 behavior producing different DWARF src file line number info for the same DWARF version. * -Add group number as CLI argument. WIP. -Issues of interest:#36, #35 * -Group number argument. Useful for getting DWARF data from multiple COMBAT sections, like macros inside of unlinked ELF files. * -Add unit tests for macros across multiple COMDAT groups. * -Cleanup * -Update docs * -Update unit tests. * -Update Dockerfile for Ubuntu22 * -Update CI * -Update Dockerfiles for Ubuntu20 and Ubuntu22 * -Update Makefile * -Update CI * -Update CI * -Test DWARF version in unit tests * -Update docs * -Update docs * -Update docs * -Update docs. WIP. * -Update docs * -Cleanup * -Cleanup * -Update docs. -Remove DECL code from pointer types. Pointer types do not have DECL coords. * -Add unit tests for artifacts * -Cleanup * -Update test_file1 * -Remove dead code * -Update tests in Docker files * -Update docs * -Update docs * -Update docs * -Update unit tests * -Fix Field constructors * -Minimal testing for bitfields. WIP. * -Update Docker files -Update Makefile * -Add Ubuntu18 to CI. -Update docs * -Update docs * -Update docs * -Update docs * -Update docs * -Update clang_format_all * -Add format checks to Dockerfile.ubuntu20 * -Update clang_format_all * -Format code * -Cleanup * -Add error-checking for groupNumber * -Push coveralls report * -Update docs * -Update CI --------- Co-authored-by: Mathew Benson <[email protected]>

* -Write macro records correctly. Fixes #7 * -Cleanup * -Update unit testing for draco patterns. WIP. * -Update eclipse project * -Minimally functional target_symbol * -Do not store row id for symbol target updates * -Update github CI * -Docker config for CI * -Point submodules to github * -Cleanup * -Add symbols and encoding mappings. WIP. * -Add symbols and encoding mappings. WIP. * Writer DWARF5 ecodings to encodings table * -Add minimally functional encodings implementation. -TODO:Cleanup. * -Update Makefile * -Update gitignore * -Cleanup * -Remove Ubuntu18 from github actions since it is not available anymore. -https://github.blog/changelog/2022-08-09-github-actions-the-ubuntu-18-04-actions-runner-image-is-being-deprecated-and-will-be-removed-by-12-1-22/ * -Remove dead code * -Cleanup * -Cleanup * -Cleanup * -Include unit test code in format checks. -Format unit test code. * -Update unit tests * -Configure coveralls.io for CI * -Configure coveralls.io for CI * -Configure coveralls.io for CI * -Configure coveralls.io for CI * -Configure coveralls.io for CI * -Configure coveralls.io for CI * -Configure coveralls.io for CI * -Configure coveralls.io for CI * -Update docs * -Update unit tests for new schema. WIP. -TODO:Add function to follow target symbols to test typdef'd types. -TODO:Add unit tests for new macro and elf image features. * -Add followTargetSymbol. Useful for testing typedef'd types. -Update unit tests. * -Use CFE_ES_HousekeepingTlm_Payload for unit testing. * -Update docs * -Update docs * -Extract ELF image data from 64-bit targets properly. -Add ElfClass to ElfFile * -Update docs * -Update unit-test test files * -Remove libdwarf from unit test code test file * -Remove dead code * -Remove dead code * -Add 32-bit make recipes. -Add 32-bit unit tests. * -Update 32-bit unit testing. * -Update docs * -Remove dead code * -Update Logger unit tests -Remove dead code * -Update unit tests * -Update docs * -Cleanup * -Cleanup * -Cleanup * -Add macro unit test. -Update Dockerfile * -Create test files that comply with DWARF4 and DWARF5. Allows us to verify support of DWARF4 as more features are added in the future. * -Unit test DWARF4 and DWARF5. WIP. * -Add DWARF4 testing to Dockerfile * -Cleanup * -Update docs * -Update docs * -Update Catch 2 due to issue on Ubuntu22:catchorg/Catch2#2178 * -pathIndex handling for DWARF5. TODO:This changes for DWARF4, so add alternative for DWARF4. -Remove DisplayDie function for now as it blows up the stack with big c++ names inside DWARF. -TODO:Rewrite DisplayDie function -Add include to src/Artifact.h, needed for Ubuntu22. * -Handle DWARF dbg source files for DWARF4 and 5. -TODO:Figure out what to do about Ubuntu 20/22 behavior producing different DWARF src file line number info for the same DWARF version. * -Add group number as CLI argument. WIP. -Issues of interest:#36, #35 * -Group number argument. Useful for getting DWARF data from multiple COMBAT sections, like macros inside of unlinked ELF files. * -Add unit tests for macros across multiple COMDAT groups. * -Cleanup * -Update docs * -Update unit tests. * -Update Dockerfile for Ubuntu22 * -Update CI * -Update Dockerfiles for Ubuntu20 and Ubuntu22 * -Update Makefile * -Update CI * -Update CI * -Test DWARF version in unit tests * -Update docs * -Update docs * -Update docs * -Update docs. WIP. * -Update docs * -Cleanup * -Cleanup * -Update docs. -Remove DECL code from pointer types. Pointer types do not have DECL coords. * -Add unit tests for artifacts * -Cleanup * -Update test_file1 * -Remove dead code * -Update tests in Docker files * -Update docs * -Update docs * -Update docs * -Update unit tests * -Add support for unions. WIP. * -Add support for unions. WIP. * -Update CI to build Ubuntu20 first. * -Update CI to build Ubuntu20 first. * -Format files * -Update union object used for unit tests * -Add user to docker dev configs * -Add support for unions. WIP. -Update unit tests for unions. WIP. * -Minimally functional union support. -TODO:Upgrade to C++17 and std::optional for union fields, which will be NULL (or empty). * -Remove dead code * -Upgrade to C++17. -Use optional for byte offset of fields. -When byte offset value is empty, write NULL to db. This allows to make a distinction between unions and structs. -TODO:Finish union unit testing. -TODO:Add union support documentation. * -Update docs * -Remove duplicate recipes from Makefile * -Handle default case in getdbgSourceFile * -Update unit testing for union support * -Update unit testing for union support * -Update docs * -Update docs * -Use std::optional for symbol encoding * -Add vscode config * -Add tests for padding. Fixes #21 * -Give root access to dev docker user * -Use std::optional for getdbgSourceFile * -Do not attempt to add padding for unions. -Add checks to padding function. * -Add warning message when field byte offsets do not exist * -Add warning message when field byte offsets do not exist * -Use git tag and latest commit as version value * -Log warning when there is no DW_AT_name attribute --------- Co-authored-by: Mathew Benson <[email protected]>

horenmar added the Bug label Feb 17, 2021

agiersch added a commit to simgrid/simgrid that referenced this issue Feb 18, 2021

Fix build error with non-const MINSIGSTKSZ.

d98472e

See catchorg/Catch2#2178.

vmihalko mentioned this issue Feb 18, 2021

CBMC fails to build on Fedora Rawhide due to Catch2 diffblue/cbmc#5840

Closed

james-d-mitchell mentioned this issue Feb 28, 2021

MINSIGSTKSZ may not be constant libsemigroups/libsemigroups#271

Closed

musicinmybrain mentioned this issue Mar 9, 2021

Tests fail to compile with pre-release glibc and leatherman-vendored Catch puppetlabs/cpp-hocon#132

Open

sixg0000d added a commit to sixg0000d-copr/shadowsocksr-uvw that referenced this issue Mar 17, 2021

Update shadowsocksr-uvw.spec

cb9dd8f

Skip tests to workaround a compilation error same as catchorg/Catch2#2178

horenmar mentioned this issue Mar 24, 2021

2.13.4: build fails #2192

Closed

xvitaly mentioned this issue Apr 2, 2021

1.8.5 build failure on GCC 11 gabime/spdlog#1899

Closed

horenmar added the Resolved - pending review Issue waiting for feedback from the original author label Apr 4, 2021

sixg0000d mentioned this issue Apr 11, 2021

在最新版 gcc 中构建单元测试失败，建议更新 test/catch.hpp Qv2ray/Qv2ray#1398

Closed

1 task

horenmar closed this as completed Apr 11, 2021

horenmar removed the Resolved - pending review Issue waiting for feedback from the original author label Apr 11, 2021

lzaoral added a commit to lzaoral/dg that referenced this issue Apr 12, 2021

tests: Update Catch2 to 2.13.5

0b8e558

Fixes compilation with new glibc on Fedora Rawhide. See: catchorg/Catch2#2178

lzaoral mentioned this issue Apr 12, 2021

Update Catch2 to 2.13.5 mchalupa/dg#388

Merged

mchalupa pushed a commit to mchalupa/dg that referenced this issue Apr 13, 2021

tests: Update Catch2 to 2.13.5

425f634

Fixes compilation with new glibc on Fedora Rawhide. See: catchorg/Catch2#2178

liheinz mentioned this issue May 8, 2021

Build of testthat fails r-lib/testthat#1373

Closed

ausil added a commit to ausil/PrusaSlicer that referenced this issue May 16, 2021

Fix build error with non-const MINSIGSTKSZ.

581fef4

see catchorg/Catch2#2178 Signed-off-by: Dennis Gilmore <[email protected]>

ausil mentioned this issue May 16, 2021

Fix build error with non-const MINSIGSTKSZ. prusa3d/PrusaSlicer#6518

Closed

bigfooted mentioned this issue Jul 26, 2022

SU2 does not compile with the provided catch2 on newer ubuntu linux distributions su2code/SU2#1724

Closed

h3poteto mentioned this issue Jul 31, 2022

Update deps catch2 version to 2.13.9 ibc/libsdptransform#22

Merged

This was referenced Aug 15, 2022

updating catch to cache2 latest release. Edgio/is2#28

Merged

Fixes for 22 04 Edgio/hurl#49

Merged

myrqs mentioned this issue Aug 23, 2022

changed catch2 version to fix build for newer glibc version BenjaminHinchliff/dungeonator#1

Merged

dvd0101 mentioned this issue Oct 3, 2022

[BUG] bootstrap fails on a recent linux distribution vector-of-bool/bpt#147

Open

Psypher9 mentioned this issue Oct 26, 2022

Catch 2.11.0 fails to build tests rezalas/riftshadow#177

Closed

gotnone mentioned this issue Nov 7, 2022

Failure to compile on glibc 2.34 and later TartanLlama/expected#114

Closed

mmore500 added a commit to devosoft/Empirical that referenced this issue Nov 21, 2022

Bump Catch pin

a74e69d

Fixes catchorg/Catch2#2178 on Ubunto 22.04

4kangjc mentioned this issue Dec 9, 2022

compile fail Tencent/flare#64

Closed

choll added a commit to choll/xtr that referenced this issue Dec 10, 2022

Upgrade catch2 for catchorg/Catch2#2178

ebb0dc4

choll added a commit to choll/xtr that referenced this issue Dec 10, 2022

Upgrade catch2 for catchorg/Catch2#2178

de5e860

This was referenced Dec 23, 2022

chore(test): update catch2 to 2.13.10 whoahq/squall#3

Merged

chore(test): update catch2 to 2.13.10 whoahq/typhoon#3

Merged

fnc12 mentioned this issue Jan 17, 2023

build error: call to non-‘constexpr’ function ‘long int sysconf(int)’ fnc12/sqlite_orm#1122

Closed

andrewfowlie mentioned this issue Mar 13, 2023

Compile error from catch 2 with ubuntu 21.04 PhaseTracer/PhaseTracer#18

Closed

hfiguiere mentioned this issue Mar 23, 2023

Build failure with recent glibc junrrein/pdfslicer#234

Open

Krzmbrzl mentioned this issue Mar 27, 2023

Update Catch2 to latest v2 release standardese/standardese#240

Merged

arintanen mentioned this issue May 22, 2023

Catch2 version outdated Aalto-CFD/DLBFoam#35

Closed

Stikus mentioned this issue Dec 8, 2023

Issues building from source with cmake Illumina/REViewer#57

Open

jackforman1024 mentioned this issue Aug 21, 2024

Update catch version to resolve build issue on linux mmorse1217/lean-vtk#9

Merged

lorenzo-gomez-windhover added a commit to WindhoverLabs/juicer that referenced this issue Sep 11, 2024

-Update Catch 2 due to issue on Ubuntu22:catchorg/Catch2#2178

8bdbfa4

Marcvi19 mentioned this issue Nov 29, 2024

fix: v8.0.0 release version on latest machine[Ubuntu 24.04] verilog-to-routing/vtr-verilog-to-routing#2826

Open

7 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Catch2 fails to compile on Fedora Rawhide #2178

Catch2 fails to compile on Fedora Rawhide #2178

lzaoral commented Feb 17, 2021

horenmar commented Feb 17, 2021

tomhughes commented Feb 19, 2021

tomhughes commented Feb 19, 2021

horenmar commented Apr 4, 2021

lzaoral commented Apr 6, 2021

horenmar commented Apr 11, 2021

Catch2 fails to compile on Fedora Rawhide #2178

Catch2 fails to compile on Fedora Rawhide #2178

Comments

lzaoral commented Feb 17, 2021

horenmar commented Feb 17, 2021

tomhughes commented Feb 19, 2021

tomhughes commented Feb 19, 2021

horenmar commented Apr 4, 2021

lzaoral commented Apr 6, 2021

horenmar commented Apr 11, 2021