16:01:05 <geppetto> #startmeeting fpc
16:01:05 <zodbot> Meeting started Thu Oct 12 16:01:05 2023 UTC.
16:01:05 <zodbot> This meeting is logged and archived in a public location.
16:01:05 <zodbot> The chair is geppetto. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions.
16:01:05 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:01:05 <zodbot> The meeting name has been set to 'fpc'
16:01:05 <geppetto> #meetingname fpc
16:01:05 <zodbot> The meeting name has been set to 'fpc'
16:01:05 <geppetto> #topic Roll Call
16:01:25 <michel-slm> .hello salimma
16:01:26 <zodbot> michel-slm: salimma 'Michel Lind' <michel@michel-slm.name>
16:02:06 <carlwgeorge> .hi
16:02:07 <zodbot> carlwgeorge: carlwgeorge 'Carl George' <carl@redhat.com>
16:02:15 <geppetto> #chair carlwgeorge
16:02:15 <zodbot> Current chairs: carlwgeorge geppetto
16:08:03 <michel-slm> this is turning out to be a quiet meeting :)
16:08:31 <geppetto> Yeh, probably will soon be a non-existant meeting
16:08:51 <geppetto> #topic Open Floor
16:09:15 <geppetto> carlwgeorge: You have anything I need to hear or look at?
16:11:37 <carlwgeorge> i had an idea for a guidelines pr but nothing of substance to look at yet.  basically i want to document the usage of `~^` for pre-release snapshots.
16:11:59 <geppetto> Ok, I thought tibbs was working on a version of that.
16:12:09 <carlwgeorge> it's used fairly regularly, and it matches my interpretation of the guidelines, but it's not explicitly laid out
16:12:18 <geppetto> I actually thought we'd published it too
16:12:20 <carlwgeorge> oh interesting, i should sync up and make sure i'm not duplicating effort
16:12:40 * geppetto nods
16:12:41 <carlwgeorge> i've seen enough people ask me about why some of my updates have that that it's worth documenting
16:12:45 <tibbs> Hey, folks
16:12:50 <geppetto> #chair tibbs
16:12:50 <zodbot> Current chairs: carlwgeorge geppetto tibbs
16:13:00 <geppetto> Hey, just in time ;)
16:13:04 <carlwgeorge> hey tibbs, see above regarding pre-release snapshots
16:13:23 <carlwgeorge> i was thinking of writing up a clarification about it, but geppetto says you might already be working on that
16:13:35 <tibbs> If I was working on that, I have completely forgotten about it.
16:14:00 <tibbs> I do remember there were questions and it came down to "yes it's allowed but we just don't make that completely obvious".
16:14:05 <carlwgeorge> i can drive it then, i have a pretty clear idea about how to describe it
16:14:48 <carlwgeorge> pretty much that, the guidelines cover post-release snapshots and tagged pre-releases, but not untagged pre-releases
16:15:01 <tibbs> Good, because at this point I definitely do not have a clear idea.
16:15:14 <tibbs> Besides just saying "yes you can combine them; it looks like this".
16:15:59 <carlwgeorge> i'll work up a pr for it and we can refine the exact language if needed
16:16:25 <tibbs> Hmm, is this a typo in the guidelines: "All snapshots must contain a snapshot information field (<snapinfo>:) in the Release: tag."
16:16:31 <tibbs> The colon after <snapinfo>
16:17:01 <carlwgeorge> on a related note, currently the guidelines say that packagers MAY use the old deprecated style.  any thoughts on making that SHOULD NOT to nudge folks along to updating to the new style?
16:17:08 <tibbs> I think that must be extraneous.  That's what I get for looking.
16:17:41 <carlwgeorge> i think it's really only necessary for epel7 at this point
16:18:36 <carlwgeorge> actually i need to double check, i have a vague recollection of tilde and caret not being available in the same version of rpm, so epel8 may need to use the old style as well
16:18:44 <tibbs> Did we just lose geppetto?
16:19:20 <geppetto> Well, desktop crash ftw
16:19:43 <tibbs> That hasn't happened to me in forever.
16:20:00 <michel-slm> carlwgeorge: if you're clarifying that, please also provide an example for how to use the ^~ combination
16:20:20 <michel-slm> it did not occur to me when reading the guidelines that that's how it was supposed to work until I saw your example, but it makes sense sorting-wise
16:20:24 <carlwgeorge> i don't know what a valid use of ^~ in that order would be
16:20:31 <michel-slm> I guess I mean ~^
16:20:44 <tibbs> I do think we were trying to be maximally kind to anyone who had packages using the old scheme, and also isn't there a heavily macroized language ecosystem which is hardcoded to use the old method?
16:20:46 <carlwgeorge> oh ok yes, that is what my original topic was
16:20:55 <michel-slm> (case in point, this needs to be written down)
16:22:26 <carlwgeorge> i think the forge macros are designed around the old style, but it's fairly easy to ignore that part
16:23:14 <carlwgeorge> found an answer to my own question, tilde was added in rpm 4.10, caret wasn't until 4.15
16:23:40 <carlwgeorge> gonna check for backports but based solely on the versions that means only epel9 has access to both
16:24:00 <carlwgeorge> interestingly it means that tilde works all the way back to epel7
16:24:22 <geppetto> If I search for "tibbs fedora version cleanup" I get a wiki page.
16:24:36 <tibbs> That was a long time ago.
16:24:41 * geppetto nods
16:24:53 <geppetto> That is probably what my old brain is remembering
16:25:14 <tibbs> That was the initial work on introducing tilde back before the guidelines migrated out fo the wiki.
16:28:40 <carlwgeorge> it does appear that caret support was backport to rhel8's rpm, based on some rpmdev-vercmp prodding
16:29:42 <carlwgeorge> done in 4.14.2-34, so back in 8.2
16:30:08 <tibbs> I do recall that Panu backported it.
16:30:46 <carlwgeorge> maybe the rhel7 eol in june next year is the right time to make using the old style SHOULD NOT
16:31:06 <carlwgeorge> then we can say the new style works in all branches
16:31:19 <carlwgeorge> *current branches
16:31:27 <tibbs> Yeah, I think when EL7 is gone, we should consider suggesting it a bit harder.
16:31:44 <carlwgeorge> i'm fine with that, mainly was gauging peoples thoughts on it
16:32:10 <geppetto> sure, seems good
16:32:18 <carlwgeorge> i'll have to look at the golang macros these days and see how they are written regarding using the forge macro snapshotting or the new style
16:33:00 <tibbs> The forge macros are at least being developed now so there is a chance.
16:33:28 <carlwgeorge> i heard they were split out so they could get developed, but i hadn't heard about any progress yet
16:33:33 <tibbs> Also, I am not up to date on the golang issue.
16:33:48 <tibbs> With their novel multi-package-versioning scheme.
16:34:33 <carlwgeorge> .fpc 1307
16:34:34 <zodbot> carlwgeorge: Issue #1307: Clarification needed regarding the naming of Golang packages - packaging-committee - Pagure.io - https://pagure.io/packaging-committee/issue/1307
16:35:00 <tibbs> I'm sorry about missing last week; I didn't even snap to the fact that today was Thursday.
16:35:34 <carlwgeorge> i think i made my case there that simply not using %goname (which is already the case for some packages) is an easy fix allows the package to be named any way we choose
16:36:39 <carlwgeorge> yes some stuff would have to be renamed, but that's the result of not following the guidelines from the start
16:36:52 <tibbs> Seems so, but there was some commentary that the method used in the guidelines is somehow outdated or based on incorrect assumptions or something like that.
16:37:20 <tibbs> And that we should just change to allowing or even recommending using a separator, and I don't know if we should be talking about that.
16:38:34 <carlwgeorge> i just have a hard time getting on board with "you should change to rules to match how i've been breaking them"
16:39:01 <tibbs> I agree, but there's no point in being obstinate of there are good arguments for changing.
16:39:19 <tibbs> The question is whether there are good arguments besides "well we did it that way"
16:39:45 <carlwgeorge> i hadn't seen any, unless i missed them
16:40:58 <geppetto> Eh, I assume the intention was good and to not break the rules … maybe meant to have the -vX but then thought it looked better without the v part, but didn't realize that meant dropping the hyphen too
16:41:04 <tibbs> There was some mention that the currently mandated scheme was based on the outdated concept that multi-versioning is a rare thing.
16:41:56 <tibbs> But I don't see why one scheme is really any better than the other, only that we should be consistent.  And we already picked a method, so being consistent means just sticking with that method.
16:42:19 <geppetto> Yeh
16:42:20 <carlwgeorge> do you recall where that was, or if it was actually substantiated
16:42:46 <geppetto> I thought it went all the way back to glib/glib2
16:43:19 <geppetto> When everyone still used C and tried to use linker scripts etc. to not change API versions.
16:43:19 <tibbs> I was specifically referring to https://pagure.io/packaging-committee/issue/1307#comment-876450
16:45:13 <geppetto> Anyway … you two want to talk outside the meeting, and I can close?
16:45:18 <carlwgeorge> seems to be more speculation about past intent than anything
16:45:54 <carlwgeorge> we can table it till the next meeting
16:46:01 * geppetto nods
16:46:07 <tibbs> Fine with me.  I just wanted to be on the record as being happy to listen to good arguments for changing things.
16:46:57 * geppetto nods
16:47:05 <geppetto> Ok, see you next week
16:47:07 <geppetto> #endmeeting