<@nirik:matrix.scrye.com>
16:00:05
!startmeeting riscv-sig
<@meetbot:fedora.im>
16:00:07
Meeting started at 2025-01-21 16:00:05 UTC
<@meetbot:fedora.im>
16:00:07
The Meeting name is 'riscv-sig'
<@nirik:matrix.scrye.com>
16:00:14
!topic init process / agenda gathering
<@nirik:matrix.scrye.com>
16:00:14
!meetingname riscv-sig
<@meetbot:fedora.im>
16:00:15
The Meeting Name is now riscv-sig
<@nirik:matrix.scrye.com>
16:00:20
morning everyone
<@davidlt:matrix.org>
16:00:29
👋hi
<@abologna:matrix.org>
16:00:57
!hi
<@zodbot:fedora.im>
16:00:59
None (abologna)
<@davidlt:matrix.org>
16:01:01
!hi
<@zodbot:fedora.im>
16:01:03
David Abdurachmanov (davidlt)
<@nhanlon:beeper.com>
16:01:10
!hi
<@zodbot:fedora.im>
16:01:11
Neil Hanlon (neil) - he / him / his
<@abologna:matrix.org>
16:01:18
how do I get zodbot to know my name?
<@abologna:matrix.org>
16:01:46
not that it matters that much
<@nirik:matrix.scrye.com>
16:01:50
abologna: likely you have the 'privacy' checkbox checked... so it can't get your real name.
<@nirik:matrix.scrye.com>
16:02:12
or you don't have your matrix id in your account info perhaps
<@davidlt:matrix.org>
16:02:33
Yes, it's most likely not having your Matrix information on FAS.
<@nirik:matrix.scrye.com>
16:03:22
Anyhow, I have a few status/info items... any other general topics we should add to the agenda?
<@davidlt:matrix.org>
16:04:43
GCC 12 + binutils 2.44 + glibc 2.41 (all snapshots, not released yet) are in Fedora/RISCV 42.
<@davidlt:matrix.org>
16:04:53
GCC 15 + binutils 2.44 + glibc 2.41 (all snapshots, not released yet) are in Fedora/RISCV 42.
<@nirik:matrix.scrye.com>
16:04:53
!topic status items
<@davidlt:matrix.org>
16:04:58
Sorry, GCC 15 😄
<@nirik:matrix.scrye.com>
16:05:59
There's a QE position they have opened up in my team (community linux engineering) with a riscv focus: https://redhat.wd5.myworkdayjobs.com/en-US/jobs/job/Software-Quality-Engineer---Fedora-Linux_R-041233-1
<@nirik:matrix.scrye.com>
16:06:21
interested parties are advised to apply. ;)
<@davidlt:matrix.org>
16:06:31
I posted this on the main channel a few days ago IIRC.
<@abologna:matrix.org>
16:06:55
no such thing as too much exposure :)
<@nirik:matrix.scrye.com>
16:07:05
Yeah, I should have last meeting too, but better late then never... yeah.
<@nirik:matrix.scrye.com>
16:07:49
Second info item I had was that I now have a work in progress PR for the secondary hub setup: https://pagure.io/fedora-infra/ansible/pull-request/2435 need to finish off a few things, but getting closer...
<@abologna:matrix.org>
16:08:12
amazing news
<@nirik:matrix.scrye.com>
16:09:02
whats the status on images?
<@abologna:matrix.org>
16:09:33
right. earlier today I shared my changes to kiwi recipes with davidlt
<@davidlt:matrix.org>
16:09:50
nirik: I think your PR needs fixing, see `roles/koji_hub/tasks/main.yml` (seems like not resolved merge conflict).
<@nirik:matrix.scrye.com>
16:10:29
yeah, I need to rebase it again, but also need to do some more stuff... like the proxy setup to reach it, check kerberos changes, etc.
<@abologna:matrix.org>
16:10:33
I figure it will take a few ways for him to look over them and we might have to iterate over some of the changes I've made if there's disagreement
<@davidlt:matrix.org>
16:10:46
>
<@davidlt:matrix.org>
16:10:46
> The server is temporarily unable to service your request due to maintenance downtime or capacity problems. Please try again later.
<@davidlt:matrix.org>
16:10:46
I didn't look at Kiwi change because Pagure is slow today, sometimes returning:
<@davidlt:matrix.org>
16:10:46
> Service Unavailable
<@nirik:matrix.scrye.com>
16:11:16
our constant battle with ai scrapers. ;(
<@abologna:matrix.org>
16:11:16
davidlt: hardly your fault considering that I only made you aware of them a couple of hours ago ;)
<@abologna:matrix.org>
16:11:30
FWIW pagure seems to work fine here
<@davidlt:matrix.org>
16:11:48
I didn't look much into F41 lately. Mostly focused on landing as many SOVERION bumps in F42 tag, and then GCC 15 and bin/sbin merge changes.
<@abologna:matrix.org>
16:11:50
I can push the branch somewhere else too if that can make things smoother
<@nirik:matrix.scrye.com>
16:12:01
seems ok now, but the scraping is sometimes really bad/creates high load.
<@davidlt:matrix.org>
16:12:20
Ah, you are also getting those? 😄 Fedora/RISCV suffers from them too.
<@davidlt:matrix.org>
16:12:54
I think it's Cloude and Amazon mainly generating the most traffic the last time I looked at it.
<@abologna:matrix.org>
16:12:54
my idea would be to get the images out soon (1-2 weeks?) then call F41 done and move on
<@nirik:matrix.scrye.com>
16:13:29
oh yeah, lots of them. ;( They are hard to block (random agent id, large cloud networks, ignores robots.txt)
<@davidlt:matrix.org>
16:14:46
One good outcome is that Fedora/RISCV should be well represented in some AI LLM models 😄
<@abologna:matrix.org>
16:14:54
slight correction: get the images out, *make some noise about them*, then move on ;)
<@nirik:matrix.scrye.com>
16:15:06
yeah. +1
<@nirik:matrix.scrye.com>
16:15:35
I didn't really have much else today...
<@davidlt:matrix.org>
16:16:11
Do we know when f42-rebuild is merged into f42?
<@davidlt:matrix.org>
16:16:39
I have notice a lot (annoying amount) of failures generic to GCC 15 change, mainly due to a new default C standard version.
<@nirik:matrix.scrye.com>
16:16:52
it should have been already, but I guess jnsamyak hasn't gotten to it yet. Will be today at any case.
<@nirik:matrix.scrye.com>
16:16:57
yep.
<@nirik:matrix.scrye.com>
16:17:15
https://kojipkgs.fedoraproject.org//mass-rebuild/f42-failures.html
<@nirik:matrix.scrye.com>
16:18:18
there was also some glang changes that broke a lot of golang packages.
<@abologna:matrix.org>
16:18:32
lovely
<@davidlt:matrix.org>
16:19:19
I will wait some time before our mass rebuild to let some of these get fixed.
<@abologna:matrix.org>
16:20:28
sounds like a good plan. not much point in chasing Rawhide failures
<@nirik:matrix.scrye.com>
16:20:29
I also gathered a list of packages that were not in failed, but were in needs rebuild... ie, in weird states, there is some discussion in https://pagure.io/releng/issue/12527 about it. Hopefully we can clean those up
<@nirik:matrix.scrye.com>
16:20:49
(that includes the golang packages you noted the other day)
<@davidlt:matrix.org>
16:21:41
nirik: it's seems mass rebuild pulled a new qt5 version too.
<@davidlt:matrix.org>
16:21:53
That's way all qt5 packages failed.
<@davidlt:matrix.org>
16:22:35
qt 5.15.16 is on a side tag, and mass rebuild also attempted to land that into f42-rebuild.
<@nirik:matrix.scrye.com>
16:22:39
Didn't see that one yet. Fun. ;(
<@davidlt:matrix.org>
16:22:44
https://koji.fedoraproject.org/koji/buildinfo?buildID=2634769
<@davidlt:matrix.org>
16:22:58
If you merge f42-rebuild as-is it will take quite some things.
<@nirik:matrix.scrye.com>
16:23:25
it sometimes does... we can sort them out as they happen.
<@davidlt:matrix.org>
16:23:36
At all other qt5 packages are still on another version, and all these RPMs have hard-coded dependencies.
<@nirik:matrix.scrye.com>
16:23:48
yeah, looks like that qt5 was just built before the mass rebuild started... so it wasn't considered by the script
<@nirik:matrix.scrye.com>
16:24:11
I'll ping the kde folks about it. we can just probibly untag the mass rebuild one for now
<@davidlt:matrix.org>
16:24:29
Yeah, I am not really "happy" how mass rebuild script built some things blindly as discussed before.
<@conan_kudo:matrix.org>
16:24:37
!hi
<@zodbot:fedora.im>
16:24:39
Neal Gompa (ngompa) - he / him / his
<@nirik:matrix.scrye.com>
16:24:46
it's not too smart... what blindly do you mean?
<@davidlt:matrix.org>
16:24:57
I understand that "main" branch must build, etc. but reality is like qt5, etc.
<@davidlt:matrix.org>
16:25:13
It doesn't consider last tagged build in f42 tag.
<@nirik:matrix.scrye.com>
16:25:22
it does.
<@davidlt:matrix.org>
16:25:28
You don't check NVR and content of package before bumping it.
<@nirik:matrix.scrye.com>
16:25:44
check nvr against what?
<@davidlt:matrix.org>
16:26:07
What's in Koji. What is "main" branch holds a new version (e.g. QT5).
<@nirik:matrix.scrye.com>
16:26:11
If there is a build after the time the mass rebuild info says, it won't do a bump or build... if there is not, it will
<@nirik:matrix.scrye.com>
16:26:56
the qt5 one was before what it considers.
<@nirik:matrix.scrye.com>
16:27:10
if it had been done just a bit later it would have been skipped
<@nirik:matrix.scrye.com>
16:28:45
anyhow, we can deal with the qt5 thing
<@nirik:matrix.scrye.com>
16:30:34
ok, anything else? or shall we get back to it ?
<@abologna:matrix.org>
16:31:32
nothing from my side
<@davidlt:matrix.org>
16:31:53
Nothing here.
<@nirik:matrix.scrye.com>
16:32:03
ok, thanks everyone!
<@nirik:matrix.scrye.com>
16:32:05
!endmeeting