Hi, I opened a dist-git PR against rawhide five weeks ago and it hasn't had any review, so I'm raising it here in case pagure PRs aren't on the maintainers' radar:
https://src.fedoraproject.org/rpms/rocm-smi/pull-request/4 Neither rocm-smi.spec nor the %cmake macro sets CMAKE_BUILD_TYPE, so NDEBUG is never defined and the shipped librocm_smi64 keeps its assert() calls. On unexpected sysfs data that aborts the calling process — btop dies with SIGABRT on affected machines — instead of returning RSMI_STATUS_UNEXPECTED_DATA, which is what the throw immediately after the assert is there for. Full analysis in rhbz#2417737 comment 14. The change is one line. NDEBUG goes in through CXXFLAGS rather than -DCMAKE_BUILD_TYPE=Release, because Release also enables upstream's post-build strip, which empties debugsourcefiles.list and fails the build. Verified in mock on rawhide (7.14.0): -DNDEBUG on every compile line, debuginfo and debugsource intact. All Packit checks pass. Happy to file the same for f44, or to rework it if there's a preferred way to solve this. Thanks, Igor Giamoniano -- _______________________________________________ devel mailing list -- [email protected] To unsubscribe send an email to [email protected] Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/[email protected] Do not reply to spam, report it: https://forge.fedoraproject.org/infra/tickets/issues/new
