16:04:19 <geppetto> #startmeeting fpc
16:04:19 <zodbot> Meeting started Thu Oct 14 16:04:19 2021 UTC.
16:04:19 <zodbot> This meeting is logged and archived in a public location.
16:04:19 <zodbot> The chair is geppetto. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions.
16:04:19 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:04:19 <zodbot> The meeting name has been set to 'fpc'
16:04:19 <geppetto> #meetingname fpc
16:04:19 <zodbot> The meeting name has been set to 'fpc'
16:04:19 <geppetto> #topic Roll Call
16:04:30 <davdunc> apologies geppetto
16:04:32 <geppetto> #chair tibbs
16:04:32 <zodbot> Current chairs: geppetto tibbs
16:04:34 <geppetto> #chair Eighth_Doctor
16:04:34 <zodbot> Current chairs: Eighth_Doctor geppetto tibbs
16:04:43 <geppetto> davdunc: Again, no big deal
16:05:31 <carlwgeorge> .hi
16:05:32 <zodbot> carlwgeorge: carlwgeorge 'Carl George' <carl@redhat.com>
16:05:36 <geppetto> #chair carlwgeorge
16:05:36 <zodbot> Current chairs: Eighth_Doctor carlwgeorge geppetto tibbs
16:05:54 <geppetto> carlwgeorge: Than man, the myth, the epel ledgend ;)
16:05:55 <mcinglis> .hi
16:05:56 <zodbot> mcinglis: mcinglis 'Malcolm Inglis' <fedora@minglis.id.au>
16:06:03 <carlwgeorge> hehe
16:07:53 <Eighth_Doctor> :D
16:09:04 <geppetto> server split ftw
16:09:45 <tibbs> Timing.
16:10:40 <decathorpe> hello, sorry for being late
16:10:44 <geppetto> #chair decathorpe
16:10:44 <zodbot> Current chairs: Eighth_Doctor carlwgeorge decathorpe geppetto tibbs
16:10:49 <geppetto> Hey, no problem
16:10:56 <geppetto> #topic Schedule
16:11:08 <geppetto> #link https://lists.fedoraproject.org/archives/list/packaging@lists.fedoraproject.org/message/SFOJCUJTJ4QRABZA2A3F4I2R4IKA2FNF/
16:11:27 <geppetto> #topic #1099 Contradicting sections regarding duplicate files
16:11:27 <geppetto> .fpc 1099
16:11:27 <geppetto> https://pagure.io/packaging-committee/issue/1099
16:11:28 <zodbot> geppetto: Issue #1099: Contradicting sections regarding duplicate files - packaging-committee - Pagure.io - https://pagure.io/packaging-committee/issue/1099
16:11:55 <tibbs> That whole section is a mess.
16:12:15 <geppetto> we've talked about this a lil, in open floors … but we should really decide what to do
16:13:11 <dustymabe> .hi
16:13:12 <zodbot> dustymabe: dustymabe 'Dusty Mabe' <dusty@dustymabe.com>
16:13:23 <tibbs> I can't see a reason to keep the section at https://docs.fedoraproject.org/en-US/packaging-guidelines/#_duplicate_files
16:13:34 <carlwgeorge> agreed
16:14:11 * GwynCieslasheher here
16:14:11 <GwynCieslasheher> but distracted. :)
16:14:14 <tibbs> But really, all of https://docs.fedoraproject.org/en-US/packaging-guidelines/#_file_and_directory_ownership just needs a rewrite.
16:14:48 <geppetto> Seems fine to rm it … do we wan tot link to the other section somehow?
16:14:53 <decathorpe> hm, docs.fp.o is timing out for me right now, so I can't really comment :(
16:14:58 <geppetto> #chair GwynCieslasheher
16:14:58 <zodbot> Current chairs: Eighth_Doctor GwynCieslasheher carlwgeorge decathorpe geppetto tibbs
16:15:20 <tibbs> I mean, one is a subsection of the other so if you were at duplicate files you are just below the correct text.
16:16:05 <tibbs> Also, did you know that the main guidelines page is really long?  Because... it's really long.
16:16:28 * geppetto nods … fair enough … that page is big, so wasn't sure of the distance at first glance.
16:16:38 <geppetto> snap
16:17:15 <tibbs> At least the whole thing now has something resembling semantic breaks.
16:17:29 <decathorpe> 🎉
16:17:33 <geppetto> So we want to vote on just removing the duplicate_files section?
16:17:43 <tibbs> Unfortunately now the remaining PRs all have conflicts, but many were quite old and had conflicts anyway.
16:17:51 <geppetto> Or just have tibbs do it and pretend we never saw anything ;)
16:18:01 <decathorpe> geppetto: dropping that section is fine with me
16:18:45 <tibbs> I'll just file a PR and merge it unless there are objections.  It seems the reasonable thing to do until one of us can find the time to try and make that section a bit clearer.
16:18:57 <Eighth_Doctor> I'm good with it :)
16:18:57 <geppetto> Seems good
16:19:14 <geppetto> #action tibbs to remove the dup files section.
16:20:15 <tibbs> Thanks to Miro I now have a much better comprehension of git so I'm far more able to get things done.
16:20:26 <geppetto> #topic #pr-1077 Introduce %sysusers_create
16:20:26 <geppetto> https://pagure.io/packaging-committee/pull-request/1077
16:21:17 <geppetto> s/macroses/macros/
16:21:27 <geppetto> But apart from that minor thing, I think this is fine.
16:21:39 <geppetto> Although I've not looked at the macro to see if it does anything unholy
16:22:10 <tibbs> So obviously at least a little work is needed to fix up the wording, but since I've not used this stuff I really can't say much.
16:23:34 <tibbs> Two questions: do we want the a file dependency and not a package dependency in Requires(pre)?
16:24:00 <tibbs> And why is %sysusers_create marked as deprecated inside of the actual macro file?
16:24:01 <geppetto> it's a /usr/bin one … so eh
16:24:06 <tibbs> # Deprecated. Use %sysusers_create_package instead
16:24:59 <geppetto> Is that on F33?
16:25:02 <tibbs> Seems... unwise to put something in the guidelines that isn't supposed to be used.
16:25:10 <tibbs> This is an up to date F34 machine.
16:25:24 <decathorpe> Neal might know more?
16:25:34 <tibbs> systemd-rpm-macros-248.7-1.fc34.noarch
16:26:00 <decathorpe> we could also just drop "starting with Fedora 33" - Fedora 33 is already the oldest supported Release.
16:26:15 <tibbs> Obviously if systemd-sysusers is the way forward then that's what we should be recommending but we should use what is actually recommended.
16:26:22 <decathorpe> tibbs: and that too (maybe even put that stuff into a macro, like the one used for %systemd_requires
16:26:56 <decathorpe> that's ... odd?
16:27:21 <tibbs> I don't know how much macroization is possible, really, since you need things in five separate places in the spec.
16:28:46 <geppetto> Yeh, this is from 2018
16:28:47 <geppetto> https://lists.freedesktop.org/archives/systemd-devel/2018-March/040469.html
16:29:56 <decathorpe> So, should we ask the PR author why they documented the deprecated one instead of the recommended new one?
16:30:02 <geppetto> yeh
16:30:29 <geppetto> sysusers_create_package seems like the correct solution
16:30:36 <tibbs> Is there any example of a package using this properly?
16:30:48 <Eighth_Doctor> yes
16:31:17 <tibbs> I see only three examples in my old spec checkout.
16:31:50 <tibbs> I'm happy to just fix the PR if I can get a good example.
16:32:37 <tibbs> I note that %tmpfiles_create is also deprecated.
16:32:58 <tibbs> But we don't document that either.
16:33:04 <tibbs> So probably two things to do there.
16:34:21 <Eighth_Doctor> the most recent example I know of is the one I did for frr
16:34:21 <Eighth_Doctor> https://src.fedoraproject.org/rpms/frr
16:34:31 <geppetto> Eighth_Doctor++
16:34:56 <Eighth_Doctor> tmpfiles are done via filetrigger now
16:35:06 <tibbs> Ah, that's useful.
16:35:45 <tibbs> frr uses susysers_create_compat, though.
16:36:32 <decathorpe> tibbs: tmpfiles is documented: https://docs.fedoraproject.org/en-US/packaging-guidelines/Tmpfiles.d/
16:36:59 <decathorpe> and it looks like that page is not out of date.
16:37:05 <tibbs> Indeed it is.
16:37:24 <tibbs> I was grepping for the macros, which obviously aren't needed.
16:37:48 <tibbs> I do wish %_tmpfilesdir was just %tmpfilesdir, though.
16:38:07 <Eighth_Doctor> I was told to use that to avoid the systemd-sysusers dep
16:38:09 <Eighth_Doctor> 🤷‍♂️
16:38:23 <tibbs> Hun.
16:38:45 <geppetto> Well at least everything is wrong in this PR then
16:38:48 <tibbs> Anyway, I don't think there's anything we can do with this PR as is.
16:39:04 <tibbs> We can ask them to resubmit or we can just figure out what the correct thing is and document it.
16:39:50 * geppetto nods
16:39:51 <tibbs> Any chance someone who knows this stuff could do that?
16:40:15 <tibbs> Oh man I finally read the RPM 4.17 release notes....
16:40:22 <geppetto> I guess is the developers wanted to they'd have sumbitted PRs themself.
16:40:50 <geppetto> So you and Eighth_Doctor might be it … maybe the submitter of this PR wants to chase down wtf to do and update the PR?
16:41:26 <geppetto> Anyway, I'm going to move on …
16:41:46 <geppetto> #info PR seems to be based on old data, needs a big update.
16:41:53 <geppetto> #topic Open Floor
16:42:01 <geppetto> Anything anyone wants to talk about?
16:42:26 <tibbs> So I did come up with a version of the build constraints macro thing which is at least on the right track according to Panu.
16:43:34 <tibbs> So I will refine that and at least get it into rawhide.
16:43:38 <decathorpe> why? %_datadir, %_bindir, all have single underscore prefixes too
16:43:46 <tibbs> Yeah, I know.
16:44:11 <tibbs> Am I lagging terribly?
16:44:26 <geppetto> I don't think so?
16:44:37 <tibbs> It feels like I'm getting responses to things I typed minutes ago.
16:45:23 <geppetto> I'm not sure what line decathorpe was responding to
16:45:30 <geppetto> So not sure if I missed it
16:45:49 <tibbs> Something about %_tmpfilesdir from earlier in the discussion, I assume.
16:46:04 <geppetto> Although I will confirm that the leading _'s confuse me
16:46:27 <Eighth_Doctor> I'm pretty sure the underscore is there to make it less confusing when working with shell variables in the same script
16:46:30 <geppetto> Ahh "<tibbs> I do wish %_tmpfilesdir was just %tmpfilesdir, though.
16:47:21 <Eighth_Doctor> yeah
16:47:23 <tibbs> Yeah, I just wish that new additions wouldn't use underscores to keep the separation between internal implementation details and things we expect packagers to use.
16:47:25 <Eighth_Doctor> it probably warrants a completely new one
16:47:46 <tibbs> But that's a separate PR relating to underscores that I need to get back to one day.
16:47:58 <tibbs> Anyway, the build constraints macro thing.
16:48:08 <geppetto> good luck, if you want to fix all of that
16:48:39 <geppetto> yeh, you think you'll have a pr in the next couple of weeks?
16:49:02 <tibbs> We have a way forward, and Panu likes it in concept so I just need to split it into an rpm 4.17 version and a backwards compatible version and then get it into redhat-rpm-config.
16:49:24 <tibbs> I expect to have that all done soon if my current rate of free time continues for a few more days.
16:50:36 <geppetto> cool
16:50:46 <Eighth_Doctor> it's probably better that tibbs and I figure this out
16:50:46 <tibbs> Hopefully I can just get it upstreamed, but we can obviously carry it ourselves until that happens.
16:50:57 <Eighth_Doctor> clearly everything is all over the place now
16:51:03 <geppetto> on that note, I won't be here next week
16:52:07 * decathorpe is solent
16:52:14 <decathorpe> *silent and can't type
16:53:03 <decathorpe> tibbs: I briefly looked at it, though the "you're on the right track" comment wasn't really helpful? I mean, on the right track to ... where?
16:53:43 <tibbs> Well we know the end goal, but the earlier implementation didn't really get there.
16:54:03 <tibbs> Sadly the discussion is in multiple separate places.
16:54:12 <decathorpe> I don't think so?
16:54:33 <decathorpe> 🤯
16:55:24 <tibbs> Sorry, it was quite clear to me but maybe that's because I've been following all of the discussion.  I can try to document what this does and dig up various links to the previous attempt and why it was problematic.
16:55:37 <decathorpe> ah, it looks like the libera IRC -> matrix direction is delaying messages, but the other direction seems to work fine.
16:56:36 * geppetto nods … currently I'm on both, but use irc directly for the meeting
16:56:54 <decathorpe> if you tell me what response you are referring to, I can tell you what it was a response to :)
16:57:03 <geppetto> kind of thinking about a day in the near future where I'm just on matrix though
16:57:18 <tibbs> If only the clients were good.
16:57:39 <geppetto> The one client, you mean
16:57:57 <tibbs> Everything seems to want to look like discord with big icons and such when I just want a basic channel list and text.
16:58:27 <decathorpe> ugh. uhm, looking at the discord bridge, my history is missing multiple messages from the IRC side of things, so sorry if I'm out of sync :/
16:58:52 <tibbs> I guess the bot will have a log but I can also supply one if that would help.
16:59:50 <tibbs> Anyway, I know what I need to do with respect to this macro, so I'll be trying to work on that today and tomorrow.
16:59:55 <geppetto> cool
16:59:57 <geppetto> thanks tibbs
17:00:09 <geppetto> And on that note I'll end the meeting
17:00:16 <geppetto> #endmeeting