The condition was wrong because `github.event_name` will never be
`labeled` since this field always refers to the actual type of the
event, which in this case will always be `pull_request`. We need to
determine which type of activity has triggered the workflow via the
`github.event.action` context.
MSVC needs two matchers: One general matcher for regular
single-line warnings, and one multi-line matcher for template
warnings that originate from a template in MSVC's library.
In the latter case, MSVC prints the path to the actual file
that should be annotated several lines down and all lines
in between the first line and that line need to be matched
by one of the regexps.
Since recently, rockylinux:9 failed with linker errors against bz2, lzma
and zstd. Installing the relevant devel packages fixed the builds.
For reasons, cmake has started to add the -lbz2, -llzma, and -lzstd
linker flags. Since Icinga 2 usually does not need those, the relevant
devel packages were not installed. This may be due to some other
transitively linked dependency.
> 2025-06-05T13:12:59.5866282Z : && /usr/lib64/ccache/c++ -O2 -flto=auto -ffat-lto-objects -fexceptions -g -grecord-gcc-switches -pipe -Wall -Werror=format-security -Wp,-D_FORTIFY_SOURCE=2 -Wp,-D_GLIBCXX_ASSERTIONS -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -fstack-protector-strong -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -m64 -march=x86-64-v2 -mtune=generic -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection -Wsuggest-override -Wrange-loop-construct -g -pthread -Winvalid-pch -O2 -g -DNDEBUG -Wl,-z,relro -Wl,--as-needed -Wl,-z,now -specs=/usr/lib/rpm/redhat/redhat-hardened-ld -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -Wl,--export-dynamic -rdynamic third-party/mmatch/CMakeFiles/mmatch.dir/mmatch.c.o third-party/socketpair/CMakeFiles/socketpair.dir/socketpair.c.o lib/base/CMakeFiles/base.dir/application-version.cpp.o lib/base/CMakeFiles/base.dir/journaldlogger.cpp.o lib/base/CMakeFiles/base.dir/base_unity.cpp.o third-party/execvpe/CMakeFiles/execvpe.dir/execvpe.c.o lib/config/CMakeFiles/config.dir/config_lexer.cc.o lib/config/CMakeFiles/config.dir/config_parser.cc.o lib/config/CMakeFiles/config.dir/config_unity.cpp.o lib/remote/CMakeFiles/remote.dir/remote_unity.cpp.o plugins/CMakeFiles/check_nscp_api.dir/check_nscp_api.cpp.o -o Bin/RelWithDebInfo/check_nscp_api -Wl,-rpath,:::::::::::::::::::::::: -ldl /usr/lib64/libboost_coroutine.so.1.75.0 /usr/lib64/libboost_context.so.1.75.0 /usr/lib64/libboost_date_time.so.1.75.0 /usr/lib64/libboost_filesystem.so.1.75.0 /usr/lib64/libboost_iostreams.so.1.75.0 /usr/lib64/libboost_thread.so.1.75.0 /usr/lib64/libboost_system.so.1.75.0 /usr/lib64/libboost_program_options.so.1.75.0 /usr/lib64/libboost_regex.so.1.75.0 /usr/lib64/libssl.so /usr/lib64/libcrypto.so -lsystemd /lib64/libedit.so -ltermcap /usr/lib64/libboost_chrono.so.1.75.0 -lbz2 -llzma -lz -lzstd -licudata -licui18n -licuuc && :
> 2025-06-05T13:12:59.5872347Z /usr/bin/ld: cannot find -lbz2
> 2025-06-05T13:12:59.5872577Z /usr/bin/ld: cannot find -llzma
> 2025-06-05T13:12:59.5872802Z /usr/bin/ld: cannot find -lzstd
> 2025-06-05T13:12:59.5873039Z collect2: error: ld returned 1 exit status
Hopefully, this issue may resolve itself in the near future, but for the
time being this satisfies our CI run on Rocky Linux 9.
If CPack fails, it may write the actual errors to a dedicated log file:
EXEC : CPack error : Problem running WiX. Please check 'D:/a/icinga2/icinga2/Build/_CPack_Packages/win64/WIX/wix.log' for errors. [D:\a\icinga2\icinga2\Build\PACKAGE.vcxproj]
Show all `*.log` files as part of the job output so that it doesn't get lost.
Since recently, the amazonlinux:2023 job in the Linux action fails due
to conflichting 'mariadb1*-devel' packages.
> package mariadb1011-devel-3:10.11.11-1.amzn2023.0.1.x86_64 from amazonlinux conflicts with mariadb105-devel provided by mariadb105-devel-3:10.5.16-1.amzn2023.0.7.x86_64 from amazonlinux
It seems like Amazon Linux added mariadb1011 packages next to mariadb105
packages, resulting in a conflict due to the wildcard. On prior runs,
the mariadb105 packages was installed.
This change installs mariadb-connector-c-devel instead of a specific
mariadb1*-devel package, as suggested by the package description.
The test was disabled due to a difference in behavior of mktime() between glibc
and musl. This inconsistency was fixed with the introduction of
Utility::NormalizeTm() and Alpine no longer needs this special case.
