binutils 2.36 change hasn't landed yet in Fedora rawhide and it is expected to be released upstream on the coming Sun, Jan 24th 2021 and @nickc said he could built it in rawhide on Sun and be ready for Monday. There is a prerelease of binutils 2.36 is available but he is only 90% confident that it will not affect the rebuilds (cant be for sure). Even if we go with the pre-release it wont be ready for today but it will be available for tomorrow.
Also, gcc "fixed version" just landed in Fedora rawhide and people are still working on building dwz and boost as I am writing this. RelEng can postpone the rebuild by 5 days without any request from FESCo, but I just wanna file the ticket and see what you all think.
+1 sounds reasonable
Oh, and there is the IMA signing thing, postponing will provide more time for it as well. Just a thought :smile:
I have no concerns with this. The mass rebuild has a robust cushion in the schedule, so starting on 25 Jan does not pose an undue risk to the rest of the schedule milestones.
+1
+1 here, there is also some slowness and python building issues with armv7 buiilders that I would like to try and debug...
binutils 2.36 change hasn't landed yet in Fedora rawhide
At this point, do we really need another last-last-last-minute change landing before the mass rebuild? :frowning:
binutils 2.36 change hasn't landed yet in Fedora rawhide At this point, do we really need another last-last-last-minute change landing before the mass rebuild? :frowning:
If we are going to run a mass rebuild today, gcc, dwz, boost all of them are just built now. So, its 3 vs 1 last minute builds. Pick your poison :stuck_out_tongue_winking_eye:
Pick your poison :stuck_out_tongue_winking_eye:
I'd rather not drink poison at all, but point taken. Will binutils 2.36 at least get a day or so to bake in rawhide until the mass rebuild starts, so koschei can pick up breakage before it affects lots of packages?
Other than that, fine, move mass rebuild however necessary. +1
Pick your poison :stuck_out_tongue_winking_eye: I'd rather not drink poison at all, but point taken. Will binutils 2.36 at least get a day or so to bake in rawhide until the mass rebuild starts, so koschei can pick up breakage before it affects lots of packages?
Hopefully, according to @nickc he will be building it on Sun itself, also it seems the compiler team is testing the pre-release as well on their test hardware, so they might catch some bugs as well.
Thanks.
+1. I think have a higher chance of being better off with the updated versions, and in general it seems better to delay the start of the mass rebuild a bit, too many things are in flux now.
The binutils proposal has been deferred to F35 by the owners.
Not this one ;)
I see 5 +1's here... so I think this is approved? Or does it need more?
+6, still one short. @dcantrell, @sgallagh, @ngompa opinion?
+1 (thought I already had, sorry)
That makes (+8, 0, 0) so this is approved under the fast track procedure.
Metadata Update from @zbyszek: - Issue tagged with: pending announcement
The mass rebuild is going now.
Announced: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org/message/TFLL7YSBN5YRDBGVCYRJKXP63YK332UK/
Metadata Update from @ngompa: - Issue close_status updated to: Accepted - Issue status updated to: Closed (was: Open)
Login to comment on this ticket.