[Bug]: 2.15.0: Failing tests on musl libc.

Kangie commented 3 months ago

What I did

meson test --print-errorlogs -C /var/tmp/portage/sys-libs/libmodulemd-2.15.0/work/libmodulemd-2.15.0-build --num-processes 32 --no-suite ci_valgrind

What I got

libmodulemd-build.log.txt

. . .
90/90 modulemd:smoketest+ci / test_import_headers                                                      OK               9.24s

Summary of Failures:

19/90 modulemd:ci+c+c_release / module_index_release                                           FAIL             0.19s   killed by signal 6 SIGABRT
20/90 modulemd:ci+c+c_release / obsoletes_release                                              FAIL             0.17s   killed by signal 6 SIGABRT
72/90 modulemd:ci+modulemd_validator / valid_modulemdv2_as_modulemdv2                          FAIL             0.09s   exit status 1
75/90 modulemd:ci+modulemd_validator / valid_obsoletesv1.yaml                                  FAIL             0.08s   exit status 1
77/90 modulemd:ci+modulemd_validator / valid_obsoletesv1_as_obsoletesv1.yaml                   FAIL             0.08s   exit status 1
89/90 modulemd:ci+c+c_release / module_index_merger_release                                    FAIL             0.94s   killed by signal 6 SIGABRT

Ok:                 50  
Expected Fail:      0   
Fail:               6   
Unexpected Pass:    0   
Skipped:            34  
Timeout:            0

What I Expected


Ok:                 56
Expected Fail:      0   
Fail:               0   
Unexpected Pass:    0   
Skipped:            34  
Timeout:            0

Environment

Kernel: Linux monolith 6.7.0-gentoo--bouncy-mouse #1 SMP PREEMPT_DYNAMIC Sat Jan 20 03:35:43 AEST 2024 x86_64 AMD Ryzen 9 5950X 16-Core Processor AuthenticAMD GNU/Linux OS Release: Gentoo Linux 2.14 (Rolling)

Thanks for the report. While I cannot exclude a bug in libmodulemd, I suspect the cause is in glib. Do glib tests pass for you on the same system? Could you give me more details about the system? E.g. what portage profile (17.1 or 23.0) is used. I'd like to reproduce it locally, but I need to know which stage to use. Is it http://ftp.linux.cz/pub/linux/gentoo/releases/amd64/autobuilds/current-stage3-amd64-musl-llvm/?

Kangie commented 3 months ago

That is a good question; I unpacked this chroot before the 23.0 profiles were stablised - a quick look says that it's a 17.0 profile:

(chroot) monolith / # eselect profile list
. . .
  [66]  default/linux/amd64/17.0/musl/clang (exp) *

You may be right about the root cause of these failures being glib:

Summary of Failures:

  1/343 glib:glib+core / collate                                               ERROR            0.19s   killed by signal 6 SIGABRT
  2/343 glib:glib+core / convert                                               ERROR            0.18s   killed by signal 6 SIGABRT
  3/343 glib:glib+core / date                                                  ERROR            0.18s   killed by signal 6 SIGABRT
  4/343 glib:glib+core+slow / gdatetime                                        ERROR            0.16s   killed by signal 6 SIGABRT
 27/343 glib:glib+core / option-context                                        ERROR            0.17s   killed by signal 6 SIGABRT
206/343 glib:gio / converter-stream                                            ERROR            0.59s   killed by signal 6 SIGABRT
309/343 glib:gio / resources                                                   ERROR            0.12s   killed by signal 6 SIGABRT

Ok:                 334 
Expected Fail:      0   
Fail:               7   
Unexpected Pass:    0   
Skipped:            2   
Timeout:            0

I'll see if I can find any info in our bug tracker about these failures, maybe they've already been reported upstream?

Edit: We have a Gentoo bug, which is linked to issues or merge requests for pretty much all of those issues. https://bugs.gentoo.org/864789

fedora-modularity / libmodulemd