17:00:22 <geppetto> #startmeeting fpc
17:00:22 <zodbot> Meeting started Thu Jan 28 17:00:22 2021 UTC.
17:00:22 <zodbot> This meeting is logged and archived in a public location.
17:00:22 <zodbot> The chair is geppetto. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:00:22 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
17:00:22 <zodbot> The meeting name has been set to 'fpc'
17:00:22 <geppetto> #meetingname fpc
17:00:22 <geppetto> #topic Roll Call
17:00:22 <zodbot> The meeting name has been set to 'fpc'
17:01:01 <ignatenkobrain> .ehllo2
17:01:06 <geppetto> #chair ignatenkobrain
17:01:06 <zodbot> Current chairs: geppetto ignatenkobrain
17:01:07 <geppetto> hey
17:01:08 <ignatenkobrain> .hello2
17:01:09 <zodbot> ignatenkobrain: ignatenkobrain 'Igor Raits' <igor.raits@gmail.com>
17:01:48 <tibbs> Hello.
17:02:08 <geppetto> #chair tibbs
17:02:08 <zodbot> Current chairs: geppetto ignatenkobrain tibbs
17:02:32 * carlwgeorge waves
17:02:38 <geppetto> #chair carlwgeorge
17:02:38 <zodbot> Current chairs: carlwgeorge geppetto ignatenkobrain tibbs
17:02:53 <decathorpe> .hello2
17:02:54 <zodbot> decathorpe: decathorpe 'Fabio Valentini' <decathorpe@gmail.com>
17:03:00 <geppetto> #chair decathorpe
17:03:00 <zodbot> Current chairs: carlwgeorge decathorpe geppetto ignatenkobrain tibbs
17:03:21 <mhroncok> hey
17:03:26 <mhroncok> I am here but I am unfortunatelly busy
17:05:02 <geppetto> #chair mhroncok
17:05:02 <zodbot> Current chairs: carlwgeorge decathorpe geppetto ignatenkobrain mhroncok tibbs
17:05:06 <geppetto> No big deal
17:05:30 <geppetto> #topic PR-#1045 WIP: Add discussion of macro names beginning with underscores.
17:05:43 <geppetto> #link https://pagure.io/packaging-committee/pull-request/1045
17:06:14 * King_InuYasha waves
17:06:18 <King_InuYasha> .hello ngompa
17:06:19 <zodbot> King_InuYasha: ngompa 'Neal Gompa' <ngompa13@gmail.com>
17:06:33 <King_InuYasha> sorry I'm late, work meeting ran over
17:06:51 * mhroncok is generally in favor of adding this, explicit suggestions added there
17:07:09 <geppetto> #chair King_InuYasha
17:07:09 <zodbot> Current chairs: King_InuYasha carlwgeorge decathorpe geppetto ignatenkobrain mhroncok tibbs
17:07:48 <geppetto> Yeh, I think I'm +1 … not sure if we want to be notified, or just have the problem fixed.
17:07:59 <geppetto> wrt to the last comment
17:08:10 <King_InuYasha> I'm okay with it as it stands +1
17:08:17 <mhroncok> geppetto: we want it fixed but we are in better posiion to work with the macro maintainers I'd say
17:08:22 <tibbs> Yeah, so the real question here is whether we want to say something that you can't actually follow.
17:08:24 <mhroncok> *position
17:08:47 <mhroncok> tibbs: we can add a note that explains this problem
17:09:06 <tibbs> Double-underscored macros appear in a number of places in the guidelines.
17:09:10 <geppetto> Sure I just don't want us to be in the way if someone could easily get it fixed without us
17:09:28 <mhroncok> NOTE: Sometimes, this is not possible to follow. We would like to elimiate sch cases. Let us a know and we'll try to make it possible. Your truly, FPC
17:09:59 <mhroncok> geppetto: good point
17:10:25 <tibbs> I tried to say "this is the general rule, but it's not always possible".
17:10:59 <tibbs> But there are so many cases where it's not possible that I have to wonder.
17:11:20 <tibbs> First, what about single-underscored macro names?  Right above we say you should use %_bindir and the like.
17:11:35 <mhroncok> :)
17:12:03 <mhroncok> I'd say start with dunderscored first and let the underscored be. they are omnipresent
17:12:15 <tibbs> And even the rust guidelines use %__install for some reason.
17:13:04 <tibbs> But, hey, the ocaml guidelines set BuildRoot.  So, uh, yeah, we just have cleanups to do.
17:13:34 <tibbs> I can add easy fixes to the PR, certainly.
17:13:51 <decathorpe> tibbs: they (Rust) shouldn't. is that fixed by my PR or not?
17:14:07 <tibbs> I don't know; it's in the rustgrep spec in the examples directory.
17:14:24 <tibbs> Sorry, ripgrep.
17:15:07 <tibbs> The CronFiles guidelines also uses %__install.  I will fix that one but avoid stepping on the rust rework.
17:16:59 <tibbs> Today I will learn how to reorder and selectively squash commits.
17:17:12 <ignatenkobrain> yeah, it has been fixed by Fabio Valentini (decathorpe) if I see it correctly
17:17:25 <mhroncok> sorry I need to go AFK
17:17:29 <geppetto> tibbs: Everytime I need to do that I have to google it again :)
17:17:47 <mhroncok> I use `git rebase -i origin/master`
17:17:49 <mhroncok> see you
17:18:30 <geppetto> Also always confuses me because you need to do it from the parent commit of the one I'd think
17:20:07 <geppetto> We want to officially vote, or just merge … or do any more changes before the merge?
17:20:37 <tibbs> It's still a WIP.  I certainly wouldn't merge it now.
17:20:43 * geppetto nods
17:20:53 <geppetto> Ok, I'll leav it to you then
17:20:56 <tibbs> Can vote on the sentiment if you like but otherwise I can just keep tinkering.
17:21:04 * geppetto nods
17:21:13 <geppetto> #topic Open Floor
17:21:19 <geppetto> Anything else we want to discuss?
17:21:24 <tibbs> I think we generally agree on the "only set __* macros" and it has the benefit of being Panu-approved.
17:21:45 <tibbs> I think there was one other thing; let me find it.
17:22:14 <tibbs> A couple of things, I guess.
17:22:30 <tibbs> There was a question about requiring AVX-512.
17:22:39 <tibbs> I don't think the distro as a whole has made a decision on this.
17:23:15 <geppetto> yeh, that doesn't feel like an us question
17:24:06 <King_InuYasha> the distro has generally said no to requiring AVX-512
17:24:11 <tibbs> I agree and will day that in the ticket.
17:24:36 <tibbs> It's a decision that needs to be repeatedly reconsidered as the processor landscape changes.
17:25:09 <tibbs> But not by us.  If Fedora does allow it then it would fall to us to figure out what packages would need to do when they do require it.
17:26:45 <tibbs> I don't think the question of a library (not application) which requires special CPU features has been answered in the past.
17:27:04 <tibbs> What I recall was about applications.
17:27:37 <geppetto> For libraries it's easy … you can do multiple builds and have the dynamic linker choose the right one etc.
17:27:41 <tibbs> The end result was always that things must transparently work for users.
17:27:57 <tibbs> Not for this library, though; it explicitly doesn't support old CPUs.
17:28:13 <geppetto> Yeh, I guess we just shrug at that point?
17:28:17 <tibbs> Applications which link it are supposed to do the right thing on those CPUs and call different functions.
17:28:30 <tibbs> I think it's at minimum a "kick to FESCo" thing.
17:29:21 <geppetto> To just not allow it in the distro? … seems harsh
17:29:45 <King_InuYasha> glibc supports alternate builds and selecting it properly
17:29:49 <King_InuYasha> but that's only for libraries
17:29:56 <King_InuYasha> for binaries, nothing exists like that
17:30:16 <geppetto> Yeh, doesn't help when it just doesn't work without it either
17:30:23 <tibbs> There is actually a good bit of supported magic here, but the question raised by this library seems to be different.
17:31:44 <tibbs> The other thing that was brought up recently is whether we want to allow the full time specification in the changelog.
17:32:32 <tibbs> I think maybe we should, but we should restrict the allowable set of timezones as the RPM devs have requested.
17:34:35 <tibbs> Was going to ask Miro about use of %__python3 in the python guidelines.
17:35:53 <tibbs> Anyway, that was all that I had.
17:36:28 <geppetto> Ok, I'll close in a couple unless anyone else has anything?
17:37:04 <geppetto> Anyone have any idea on non-FPC people we could volunteer to change the changelog quidelines?
17:37:09 <geppetto> ;)
17:39:25 <geppetto> Oh, one final thing … I might not be around next week
17:40:22 <tibbs> I doubt we'll have much to do.
17:40:32 <geppetto> #endmeeting