On Thu, Mar 05, 2020 at 10:26:56AM +0100, Miro Hrončok wrote: > On 05. 03. 20 9:12, jkonecny@xxxxxxxxxx wrote: > > Hello everyone, > > > > I'm co-maintaining package of kakoune and this package have failing > > build on F32 and Rawhide because of tests. > > > > https://bugzilla.redhat.com/show_bug.cgi?id=1799560 > > > > As you can see in the last comment there is an upstream bug on gcc-10 > > which is blocking the package tests. > > > > What I want to know is what should be my next steps right now. > > > > - Should I create bugzilla on Fedora side on gcc and block my bug on > > this bug? > > You can. You don't need to. > > > - Should I disable those tests? > > You can. You don't need to, unless you need to rebuild the package for some reason. I would *not* disable the tests, as IIUC, the tests are demonstrating that code is mis-compiled & thus not going to be reliable at runtime. The broken tests correctly detected the compilation problem, preventing a bad build getting into rawhide repos. > > - Should I wait if gcc-10 upstream will solve this issue in time for > > F32? > > You can. Or they can solve it later. > > > - Should I propose Fedora 32 blocker because of the bug in gcc? > > Why would you do that? How is Fedora 32 broken if we release it with kakoune > not rebuilt yet? Is this FTBFS blocking other libraries from being updated > as planned for Fedora 32? The linked GCC bug indicates this is a case that GCC 10 is mis-compiling the code. We were lucky in this case that kakoune has a test suite that detected the mis-compiled code by seeing incorrect test behaviour, and thus failed the build. So we still have the previous correctly compiled kakoune RPM in the repos. The risk here is that there might be other packages in Fedora now built with GCC 10 that suffer the same mis-compilation, but which have not got a test suite to catch the issue. These incorrectly build apps may then suffer runtime crashes for users. I don't know enough about the GCC 10 to say whether this particular issue is likely to be a serious widespread issue. In the most serious cases these kind of bugs can require a mass rebuild of affected packages to solve it. I'd suggest we file a BZ against GCC in Fedora, linking to the GCC 10 upstream bug, and then let GCC maintainers give their opinion on whether it is bad enough to warrant blocker status. Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :| _______________________________________________ devel mailing list -- devel@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe send an email to devel-leave@xxxxxxxxxxxxxxxxxxxxxxx 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/devel@xxxxxxxxxxxxxxxxxxxxxxx
