msys2 CI

[kassane]

cc: @andreasbuhr

Previous reference: #74

[kassane]

new error from Mingw:

D:/a/cppcoro/cppcoro/include/cppcoro/resume_on.hpp:120:81: error: insufficient contextual information to determine type
  120 |                 for (auto iter = co_await source.begin(); iter != source.end(); co_await ++iter)
      |                                                                                 ^~~~~~~~

[kassane]

New error on Mingw:

D:/a/cppcoro/cppcoro/test/file_tests.cpp: In lambda function:
D:/a/cppcoro/cppcoro/test/file_tests.cpp:98:35: warning: comparison of integer expressions of different signedness: 'int' and 'long long unsigned int' [-Wsign-compare]
   98 |                 for (int i = 0; i < sizeof(buffer); ++i, c = (c == 'z' ? 'a' : c + 1))
      |                                 ~~^~~~~~~~~~~~~~~~
D:/a/cppcoro/cppcoro/test/file_tests.cpp: In lambda function:
D:/a/cppcoro/cppcoro/test/file_tests.cpp:154:22: error: 'memset' is not a member of 'std'; did you mean 'wmemset'?
  154 |                 std::memset(buffer1, 0xAB, sizeof(buffer1));
      |                      ^~~~~~
      |                      wmemset
D:/a/cppcoro/cppcoro/test/file_tests.cpp:159:22: error: 'memset' is not a member of 'std'; did you mean 'wmemset'?
  159 |                 std::memset(buffer2, 0xCC, sizeof(buffer2));
      |                      ^~~~~~
      |                      wmemset
In file included from D:/a/cppcoro/cppcoro/test/doctest/cppcoro_doctest.h:9,
                 from D:/a/cppcoro/cppcoro/test/file_tests.cpp:24:
D:/a/cppcoro/cppcoro/test/file_tests.cpp:162:28: error: 'memcmp' is not a member of 'std'; did you mean 'wmemcmp'?
  162 |                 CHECK(std::memcmp(buffer1, buffer2, 50) == 0);
      |                            ^~~~~~
D:/a/cppcoro/cppcoro/test/doctest/doctest.h:2104:9: note: in definition of macro 'DOCTEST_WRAP_IN_TRY'
 2104 |         x;                                                                                         \
      |         ^
D:/a/cppcoro/cppcoro/test/doctest/doctest.h:2354:9: note: in expansion of macro 'DOCTEST_ASSERT_IMPLEMENT_2'
 2354 |         DOCTEST_ASSERT_IMPLEMENT_2(assert_type, __VA_ARGS__);                                      \
      |         ^~~~~~~~~~~~~~~~~~~~~~~~~~
D:/a/cppcoro/cppcoro/test/doctest/doctest.h:2398:28: note: in expansion of macro 'DOCTEST_ASSERT_IMPLEMENT_1'
 2398 | #define DOCTEST_CHECK(...) DOCTEST_ASSERT_IMPLEMENT_1(DT_CHECK, __VA_ARGS__)
      |                            ^~~~~~~~~~~~~~~~~~~~~~~~~~
D:/a/cppcoro/cppcoro/test/doctest/doctest.h:2927:20: note: in expansion of macro 'DOCTEST_CHECK'
 2927 | #define CHECK(...) DOCTEST_CHECK(__VA_ARGS__)
      |                    ^~~~~~~~~~~~~
D:/a/cppcoro/cppcoro/test/file_tests.cpp:162:17: note: in expansion of macro 'CHECK'
  162 |                 CHECK(std::memcmp(buffer1, buffer2, 50) == 0);
      |                 ^~~~~
D:/a/cppcoro/cppcoro/test/file_tests.cpp:165:28: error: 'memcmp' is not a member of 'std'; did you mean 'wmemcmp'?
  165 |                 CHECK(std::memcmp(buffer1 + 50, buffer2, 50) == 0);
      |                            ^~~~~~
D:/a/cppcoro/cppcoro/test/doctest/doctest.h:2104:9: note: in definition of macro 'DOCTEST_WRAP_IN_TRY'
 2104 |         x;                                                                                         \
      |         ^
D:/a/cppcoro/cppcoro/test/doctest/doctest.h:2354:9: note: in expansion of macro 'DOCTEST_ASSERT_IMPLEMENT_2'
 2354 |         DOCTEST_ASSERT_IMPLEMENT_2(assert_type, __VA_ARGS__);                                      \
      |         ^~~~~~~~~~~~~~~~~~~~~~~~~~
D:/a/cppcoro/cppcoro/test/doctest/doctest.h:2398:28: note: in expansion of macro 'DOCTEST_ASSERT_IMPLEMENT_1'
 2398 | #define DOCTEST_CHECK(...) DOCTEST_ASSERT_IMPLEMENT_1(DT_CHECK, __VA_ARGS__)
      |                            ^~~~~~~~~~~~~~~~~~~~~~~~~~
D:/a/cppcoro/cppcoro/test/doctest/doctest.h:2927:20: note: in expansion of macro 'DOCTEST_CHECK'
 2927 | #define CHECK(...) DOCTEST_CHECK(__VA_ARGS__)
      |                    ^~~~~~~~~~~~~
D:/a/cppcoro/cppcoro/test/file_tests.cpp:165:17: note: in expansion of macro 'CHECK'
  165 |                 CHECK(std::memcmp(buffer1 + 50, buffer2, 50) == 0);
      |                 ^~~~~
D:/a/cppcoro/cppcoro/test/doctest/doctest.h: In instantiation of 'doctest::detail::Expression_lhs<L>::operator doctest::detail::Result() [with L = const char (&)[27]]':
D:/a/cppcoro/cppcoro/test/file_tests.cpp:196:5:   required from here
D:/a/cppcoro/cppcoro/test/doctest/doctest.h:1469:42: warning: the compiler can assume that the address of 'doctest::detail::Expression_lhs<const char (&)[27]>::lhs' will never be NULL [-Waddress]
 1469 |             bool res = static_cast<bool>(lhs);
      |                                          ^~~
D:/a/cppcoro/cppcoro/test/doctest/doctest.h:1459:26: note: 'doctest::detail::Expression_lhs<const char (&)[27]>::lhs' declared here
 1459 |         L                lhs;
      |                          ^~~
D:/a/cppcoro/cppcoro/test/doctest/doctest.h: In instantiation of 'doctest::detail::Expression_lhs<L>::operator doctest::detail::Result() [with L = const char (&)[32]]':
D:/a/cppcoro/cppcoro/test/file_tests.cpp:207:4:   required from here
D:/a/cppcoro/cppcoro/test/doctest/doctest.h:1469:42: warning: the compiler can assume that the address of 'doctest::detail::Expression_lhs<const char (&)[32]>::lhs' will never be NULL [-Waddress]
 1469 |             bool res = static_cast<bool>(lhs);
      |                                          ^~~
D:/a/cppcoro/cppcoro/test/doctest/doctest.h:1459:26: note: 'doctest::detail::Expression_lhs<const char (&)[32]>::lhs' declared here
 1459 |         L                lhs;
      |                          ^~~

[kassane]

New issue on build: Error: could not load cache - no detect msys2 shell

[kassane]

There are still some warnings that can be corrected later.

Wow! It was fun refactoring the CI that reported some errors and learning a bit more about the API.

What do you think @andreasbuhr?

[andreasbuhr]

Wow, that's cool, thanks for all your effort!

Why did you disable the configure, build, and test steps? If I understand the diff correctly, you are doing all (including configure, build and test) in the 'Install Mingw' step. Was this necessary?

Before merging, I'd appreciate if you cleaned up the patch. Currently this pull request changes a lot of lines which are technically unchanged. This messes up the history and "git blame" does not give sensible results for these lines afterwards. It would be nice if this pull request touched only lines in needs to touch. I usually do such cleanups by creating a cleanup commit, undoing the non-necessary changes, then use use the 'git rebase -i' command to fixup this cleanup commit into the commit doing the unnecessary changes and then 'git push --force' the result to GitHub.

[kassane]

Wow, that's cool, thanks for all your effort!

Why did you disable the configure, build, and test steps? If I understand the diff correctly, you are doing all (including configure, build and test) in the 'Install Mingw' step. Was this necessary?

By default setup-msys2 requires passing own shell, the build and test steps didn't identify the files even passing msys2-env in: D:/a/_temp/setup-msys2/msys2.cmd

Before merging, I'd appreciate if you cleaned up the patch. Currently this pull request changes a lot of lines which are technically unchanged. This messes up the history and "git blame" does not give sensible results for these lines afterwards. It would be nice if this pull request touched only lines in needs to touch. I usually do such cleanups by creating a cleanup commit, undoing the non-necessary changes, then use use the 'git rebase -i' command to fixup this cleanup commit into the commit doing the unnecessary changes and then 'git push --force' the result to GitHub.

Ok.

[kassane]

Commands:

git remote add andreasbuhr git@github.com:andreasbuhr/cppcoro.git
git fetch andreasbuhr
git rebase andreasbuhr/main
git rebase -i HEAD~2
git push -f origin msys2-CI

[andreasbuhr]

Great! It's bedtime in my timezone now, I'll have a closer look tomorrow.

[andreasbuhr]

Thank you very much, this is good to have.