On Thu, 2020-08-06 at 17:54 +0200, Fabio Valentini wrote: > On Thu, Aug 6, 2020 at 4:51 PM Jeff Law <l...@redhat.com> wrote: > > On Thu, 2020-08-06 at 06:17 -0500, Richard Shaw wrote: > > > On Thu, Aug 6, 2020 at 5:08 AM Fabio Valentini <decatho...@gmail.com> > > > wrote: > > > > I'd say it's pretty safe to submit things to rawhide again. I haven't > > > > seen any lingering buildroot issues for the past 2-3 days, except for > > > > some longer wait/build times in koji due to batched ELN build > > > > submissions (but those are finished now too). > > > > > > Ok, good deal. I've been trying to knock out 2-3 packages a day so I'll > > > get to the bottom of the pile in a week or so :) > > I'm maybe 40% of the way through the failures looking for things that are > > obviously LTO related and reassigning those to myself. > > > > Based on what I've seen the cmake stuff is still the most common failure. > > I've > > still seen a few s390x infrastructure issues. Lots of noarch packages that > > have > > failed (haven't looked at those at all since they can't be LTO failures). > > When resubmitting failures caused by intermittent and infra issues, I > came across a few of those "lto-wrapper failures" on some arches as > well. > I can compile a list, if it helps. Sure that helps.
> > I also think LTO can definitely cause issues in noarch builds, if > mis-compiled archful dependency crashes during the build. Those kinds of secondary failures do happen. I've even had tertiary failures (gcc mis-compiles gas which in turn mis-compiled some X library and your xclock would render as a mirror image of what it should have -- yes, that's really happened to me). Clearly someone is going to need to look at the .noarch things. But I'm really focused on first order LTO issues right now. I'm sure that if there's a second order issue where LTO is mis-compiling something which in turn causes a .noarch issue, they'll let the world know... But the most likely places where an LTO issue is going to show up is in packages that build code with GCC or LLVM, so that's where I need to focus my time right now. jeff > _______________________________________________ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org 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@lists.fedoraproject.org
