Allow Building Tests as Submodule and GCC < 11 #777

jmartens-42 · 2023-10-25T14:11:49Z

moved BUILD_TESTS check outside of root cmake check to allow building tests when ETL is a submodule

added check for GNUC >= 11 for -Wstringop-overread ignore pragma since it isn't introduced until gcc/g++-11

… tests in submodules added check for GNUC >= 11 for -Wstringop-overread ignore pragma since it isn't introduced until gcc/g++-11

semanticdiff-com · 2023-10-25T14:11:54Z

Review changes with SemanticDiff

jwellbelove · 2023-10-26T09:00:36Z

There seems to be an issue with the Clang C++20 STL tests.
It seems to be unhappy with its own library.

In file included from /home/runner/work/etl/etl/test/test_bip_buffer_spsc_atomic.cpp:32:
/usr/bin/../lib/gcc/x86_64-linux-gnu/13/../../../../include/c++/13/chrono:2320:48: error: call to consteval function 'std::chrono::hh_mm_ss::_S_fractional_width' is not a constant expression
        static constexpr unsigned fractional_width = {_S_fractional_width()};
                                                      ^
/usr/bin/../lib/gcc/x86_64-linux-gnu/13/../../../../include/c++/13/chrono:2320:48: note: undefined function '_S_fractional_width' cannot be used in a constant expression
/usr/bin/../lib/gcc/x86_64-linux-gnu/13/../../../../include/c++/13/chrono:22[75](https://github.com/ETLCPP/etl/actions/runs/6641670039/job/18073735184?pr=777#step:3:76):2: note: declared here
        _S_fractional_width()

jmartens-42 · 2023-10-26T14:46:21Z

I haven't found a way to reproduce the issue. I can build with the following cmake command on Ubuntu 22.04.3 VM:

cmake -S . -B build -DBUILD_TESTS=ON -DNO_STL=OFF -DETL_USE_TYPE_TRAITS_BUILTINS=OFF -DETL_USER_DEFINED_TYPE_TRAITS=OFF -DETL_FORCE_TEST_CPP03=OFF -DETL_CXX_STANDARD=20 -DCMAKE_EXE_LINKER_FLAGS="-lstdc++ -lm" -DCMAKE_EXPORT_COMPILE_COMMANDS=ON

I'm attaching my compile commands file, but this is the command used for our troublemaker:

"/usr/bin/clang++-14 -DETL_DEBUG -I/magna/etl/test/../include -isystem /magna/etl/test/UnitTest++/.. -fno-omit-frame-pointer -fno-common -Wall -Wextra -Werror -std=gnu++2a -o CMakeFiles/etl_tests.dir/test_bip_buffer_spsc_atomic.cpp.o -c /magna/etl/test/test_bip_buffer_spsc_atomic.cpp"

Any thoughts on the best way to troubleshoot this?

compile_commands.json

…-GCC-less-than-11' of https://github.com/ETLCPP/etl into pull-request/#777-Allow-Building-Tests-as-Submodule-and-GCC-less-than-11

jwellbelove · 2023-10-26T18:54:28Z

It seems to be a Github Actions issue
actions/runner-images#8659

…-GCC-less-than-11' of https://github.com/ETLCPP/etl into pull-request/#777-Allow-Building-Tests-as-Submodule-and-GCC-less-than-11

…-GCC-less-than-11' of https://github.com/ETLCPP/etl into pull-request/#777-Allow-Building-Tests-as-Submodule-and-GCC-less-than-11 # Conflicts: # .github/workflows/clang.yml

…-Tests-as-Submodule-and-GCC-less-than-11' into development

* Resolve all C++ -O2 warnings Resolve all C++ -O2 warnings Resolve all C++ -O2 warnings * Updated release notes and version * Made unit test suite names unique * Added etl::bit_v for C++17 * add stringify macro (#741) * Add include/linux/stringify.h from Linux kernel 2.6.12-rc2 (cherry picked from commit torvalds/linux@1da177e) * Make __stringify support variable argument macros too For example: __stringify(__entry->irq, __entry->ret) will now convert it to: "REC->irq, REC->ret" It also still supports single arguments as the old macro did. Signed-off-by: Zhao Lei <zhaolei@cn.fujitsu.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Steven Rostedt <rostedt@goodmis.org> LKML-Reference: <49DC6751.30308@cn.fujitsu.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> (cherry picked from commit torvalds/linux@8f7c2c3) * Add documentation. * Adjust names in order to satisfy naming convention. * Use __VA_ARGS__ instead a GNU extension as this works with more compilers. Works with the newest versions of ICCAVR, GCC, CLANG and MSCV. C++11 may be required. * Adjust to ETL folder structure. * Change include guard to the one usual in ETL. * Add definition guard for STRINGIFY. --------- Co-authored-by: Linus Torvalds <torvalds@ppc970.osdl.org> Co-authored-by: Zhaolei <zhaolei@cn.fujitsu.com> * Updated stringify macro for variadic parameters * Edited comments * Added multi char type strcmp, strncmp, strcpy & strncpy * Added support for self transitions * Remove some unreachable code in bitset_new.h (#743) Warning[Pe111]: statement is unreachable * Re-enabled tests for etl::string_ext * Changes require for non-virtual messages * MutexMessageRouter example * Allow C++03 delegate in etl::callback_timer * #745 Assertion triggered in hash.h when compiling for MSP430 * Removed redundant #includes * Removed unused variable Minor format edit * Added time_to_next() status function * Added 'sentence case' versions of constants * Added C++03 compatible etl::move or etl::move_backward * Added c++03 zero parameter emplace member functions * Fix issue of release of an object on empty an pool * Remove used variable * Remove used variable * Change to require semicolons after debug count macros. Added error assert for vector reserve greater than capacity. * Added destructor to CMSIS_RTOS2 mutex * Updated version and release notes * Removed semicolon warnings * Updated vesrion and release notes * #753 [porting] IAR's overload resolution struggles with etl::span::span(etl::array&) #754 ambiguous etl::bit_stream_reader::bit_stream_reader() for some etl::span * #755 warning about constexpr in array.h * Re-enabled tests * const_reference at(size_t i) const is only constexpr from C++14 * Added conversion operators to std::span * MacOS clang span issue * MacOS clang span issue * MacOS clang span issue * MacOS clang span issue * MacOS clang span issue * Updated version and release notes * Possible fix for #756 * Fixed: move assignment not working for non-empty containers (#759) * Updated release notes * Make "private" includes relativ (#761) Avoid "include not found" when indirectly including etl-header Co-authored-by: Andreas Pelczer <pelczer@stoerk-tronic.com> * Make "private" includes relativ (#761) Avoid "include not found" when indirectly including etl-header Co-authored-by: Andreas Pelczer <pelczer@stoerk-tronic.com> * Remove BOM from random.h (#762) random.h is the only file with UTF8-BOM and IAR can not handle it. For compatibility reasons it should be removed. Co-authored-by: Andreas Pelczer <pelczer@stoerk-tronic.com> * Enabled etl::overload for C++11 & C++14 * Added conversion operators to std::span MacOS clang span issue MacOS clang span issue MacOS clang span issue MacOS clang span issue MacOS clang span issue * Updated version and release notes * Possible fix for #756 * Fixed: move assignment not working for non-empty containers (#759) * Updated release notes * Make "private" includes relativ (#761) Avoid "include not found" when indirectly including etl-header Co-authored-by: Andreas Pelczer <pelczer@stoerk-tronic.com> * Remove BOM from random.h (#762) random.h is the only file with UTF8-BOM and IAR can not handle it. For compatibility reasons it should be removed. Co-authored-by: Andreas Pelczer <pelczer@stoerk-tronic.com> * Enabled etl::overload for C++11 & C++14 * Commets added * #766 etl::unique_ptr::reset() can call deleter with nullptr argument * Refactored storage for etl::optional * Updated release notes * Updated release notes * Test name updates * Fixed alignment functionality * Updated versions * Updated versions * Initial code * Final changes * Fix maybe-uninitialized warning * Fix maybe-uninitialized warning * Added etl::string::fill() * Added etl::string interface to etl::base64 * static_cast for string_view & array_view size() member functions * Return raw pointer for `etl::multi_span::operator->()`. (#773) - using `operator*()` in case getting the current element is changed one day - using `&` to get the address Simply returning `p_value` may break in case the internal type of `p_value` would change. This way it is more robust to changes. * Added -> operator test * #773 Return raw pointer for etl::multi_span::operator->() * static_cast for string_view & array_view size() member functions * Initial code * Final changes * Fix maybe-uninitialized warning * Added etl::string::fill() * Added etl::string interface to etl::base64 * Add missing semicolons (#776) * #776 Add missing semicolons * moved BUILD_TESTS check outside of root cmake check to allow building tests in submodules (#777) added check for GNUC >= 11 for -Wstringop-overread ignore pragma since it isn't introduced until gcc/g++-11 Co-authored-by: Josh <joshua.martens@magna.com> * Try to fix library error for Github Actions for Clang C++20 * Try to fix library error for Github Actions for Clang C++20 * Added pul-request branches to CI * Removed temporary fix * Refactored swap * Added add_lvalue_reference for rvalue reference * Revert "Added add_lvalue_reference for rvalue reference" This reverts commit a09efff. * Updated version and release notes * Updated version and release notes * Try to fix library error for Github Actions for Clang C++20 (cherry picked from commit f208eb4) The line got lost in 95c7f44. * Add helper macro which enables function-like macro overload with 1 or 2 arguments. Inspired by https://stackoverflow.com/a/11763277/ Warning: This requires support for variadic macros (`__VA_ARGS__`). These were officially added in C99 and C++11. * Add unit test for static_assert.h (cherry picked from commit 91e7f46) * Add static assert overload for only 1 argument. (cherry picked from commit ae8c19e) * Add test case for static_assert with only 1 argument. (cherry picked from commit 4d6e7a6) * Spelling. --------- Co-authored-by: John Wellbelove <jwellbelove@users.noreply.github.com> Co-authored-by: David Hebbeker <dhebbeker@users.noreply.github.com> Co-authored-by: Linus Torvalds <torvalds@ppc970.osdl.org> Co-authored-by: Zhaolei <zhaolei@cn.fujitsu.com> Co-authored-by: John Wellbelove <john.wellbelove@etlcpp.com> Co-authored-by: Soma Veszelovszki <soma.veszelovszki@gmail.com> Co-authored-by: Andreas <and1p@web.de> Co-authored-by: Andreas Pelczer <pelczer@stoerk-tronic.com> Co-authored-by: H1MSK <ksda47832338@Outlook.com> Co-authored-by: jmartens-42 <jmartens@jayhawk.us> Co-authored-by: Josh <joshua.martens@magna.com> Co-authored-by: John Wellbelove <john.wellbelove@etlcpp.co.uk>

moved BUILD_TESTS check outside of root cmake check to allow building…

1b907b3

… tests in submodules added check for GNUC >= 11 for -Wstringop-overread ignore pragma since it isn't introduced until gcc/g++-11

jwellbelove added the enhancement label Oct 26, 2023

jwellbelove changed the base branch from master to pull-request/#777-Allow-Building-Tests-as-Submodule-and-GCC-less-than-11 October 26, 2023 18:07

jwellbelove merged commit 13759c1 into ETLCPP:pull-request/#777-Allow-Building-Tests-as-Submodule-and-GCC-less-than-11 Oct 26, 2023
54 of 58 checks passed

jwellbelove pushed a commit that referenced this pull request Oct 28, 2023

Merge remote-tracking branch 'origin/pull-request/#777-Allow-Building…

39c3ca8

…-Tests-as-Submodule-and-GCC-less-than-11' into development

dhebbeker mentioned this pull request Nov 15, 2023

Fix library error for Github Actions for Clang C++20 #784

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Allow Building Tests as Submodule and GCC < 11 #777

Allow Building Tests as Submodule and GCC < 11 #777

jmartens-42 commented Oct 25, 2023

semanticdiff-com bot commented Oct 25, 2023

jwellbelove commented Oct 26, 2023

jmartens-42 commented Oct 26, 2023

jwellbelove commented Oct 26, 2023 •

edited

Loading

Allow Building Tests as Submodule and GCC < 11 #777

Allow Building Tests as Submodule and GCC < 11 #777

Conversation

jmartens-42 commented Oct 25, 2023

semanticdiff-com bot commented Oct 25, 2023

jwellbelove commented Oct 26, 2023

jmartens-42 commented Oct 26, 2023

jwellbelove commented Oct 26, 2023 • edited Loading

jwellbelove commented Oct 26, 2023 •

edited

Loading