16:00:34 <geppetto> #startmeeting fpc
16:00:34 <zodbot> Meeting started Thu Oct  7 16:00:34 2021 UTC.
16:00:34 <zodbot> This meeting is logged and archived in a public location.
16:00:34 <zodbot> The chair is geppetto. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions.
16:00:34 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:00:34 <zodbot> The meeting name has been set to 'fpc'
16:00:35 <geppetto> #meetingname fpc
16:00:35 <zodbot> The meeting name has been set to 'fpc'
16:00:35 <geppetto> #topic Roll Call
16:00:48 <tibbs> Hey, folks.
16:00:49 * GwynCieslasheher here
16:00:53 <geppetto> #chair tibbs
16:00:53 <zodbot> Current chairs: geppetto tibbs
16:01:03 <decathorpe> hello o/
16:01:07 <geppetto> #chair decathorpe
16:01:07 <zodbot> Current chairs: decathorpe geppetto tibbs
16:01:14 <geppetto> #chair GwynCieslasheher
16:01:14 <zodbot> Current chairs: GwynCieslasheher decathorpe geppetto tibbs
16:03:15 <tibbs> For some reason they scheduled some kind of luncheon during the one hour a week I'm not flexible.
16:03:25 <tibbs> Not that I want to go to a luncheon with a bunch of people.
16:03:33 <geppetto> lol
16:03:42 <geppetto> also the centos dojo is on at the moment
16:03:58 <geppetto> I'd have skipped, but we have a few newish tickets that seem doable
16:05:30 <tibbs> I have a number of pending PRs.
16:05:40 * geppetto nods
16:05:45 <tibbs> Eventually I'll just merge them but I figured review doesn't hort.
16:06:06 <geppetto> sure
16:06:22 <tibbs> I don't think there's anything controversial except perhaps for removing the Eclipse plugin stuff.
16:06:41 <oturpe> hi! I am here because there were 2 of my old pull requests on the schedule
16:07:00 <GwynCieslasheher> Which...
16:07:11 <tibbs> One is 1074.
16:07:19 <oturpe> and pr-1046 Improve separate test suite sourcing instructions
16:07:23 <tibbs> The other is 1046
16:09:00 <oturpe> 1074 should be simple, either that is a rule and should be merged to the guidelines, or it is not, and the wiki should not say it is
16:09:44 <tibbs> It has three +1 votes currently.
16:10:25 <decathorpe> 1074 looks fine to me. do we need more than three +1 votes? ;)
16:10:57 <mhroncok> hey
16:11:03 <geppetto> I don't think so … but here's another +1 anyway :)
16:11:24 <GwynCieslasheher> +ed in the ticket
16:11:41 <tibbs> I'll get it merged.
16:11:53 <mhroncok> sorry, I overslept :)
16:11:57 <oturpe> great, thank you!
16:12:43 <geppetto> #topc Schedule
16:12:46 <geppetto> #link https://lists.fedoraproject.org/archives/list/packaging@lists.fedoraproject.org/message/ZYY52VKVGLJQJAPFULKRO63T5MDLIBLP/
16:12:57 <geppetto> #chair mhroncok
16:12:57 <zodbot> Current chairs: GwynCieslasheher decathorpe geppetto mhroncok tibbs
16:14:29 <geppetto> #topic #pr-#1094 Add documentation for build constraints macro
16:14:35 <geppetto> https://pagure.io/packaging-committee/pull-request/1094
16:15:26 <tibbs> So I think I can make something that is "better" but I haven't completed it yet.
16:15:37 <tibbs> Did anything happen on the FESCo side?
16:17:20 <mhroncok> decathorpe: prom experience: nobody responds to comments to closed tickets
16:17:23 <mhroncok> *from
16:17:27 <tibbs> Pour one out for zodbot....
16:17:37 <mhroncok> decathorpe: it has a tendency to be... lost
16:18:11 <decathorpe> true. but would opening a new ticket be better?
16:18:58 <mhroncok> decathorpe: my opinion is that fesco ticekts are not a good place for discussion. do you want discussion or fesco resolution about something?
16:20:06 <tibbs> Regardless of what happens, I'll keep playing with it.  I just ended up down a different rabbit hole this week.
16:20:20 <decathorpe> I think we discussed that here last week ... implementation is sufficiently different from what FESCo approved that we don't just want to document it as-is
16:20:35 <geppetto> Ok, so move on?
16:20:53 <tibbs> Yeah, until I come up with something.
16:21:00 <decathorpe> the question is: should we improve the implementation to be more like what was approved, or should we just document "the different thing"?
16:21:10 <geppetto> #topic #pr-#1097 Use caret in Obsoletes to simplify
16:21:12 <geppetto> https://pagure.io/packaging-committee/pull-request/1097
16:21:24 <tibbs> decathorpe: Technically both are possible.
16:21:58 <geppetto> So I'm confused about what is wrong with Obsoletes: <name> <= <oldest-released>
16:22:09 <tibbs> I don't know what to do with 1097.
16:22:13 <tibbs> I just don't know if it's better.
16:22:22 * decathorpe shrugs
16:22:22 <decathorpe> I think pmatilai should be involved.
16:23:07 <decathorpe> sure.
16:23:48 <geppetto> Yeh, if igor can post some reasons why it's better than the obvious above I'll think about it again … but for now I'm -1
16:24:13 <tibbs> Note that 1097 changed just recently.
16:24:36 <mhroncok> igor?
16:24:45 <mhroncok> recently as in today
16:25:23 <decathorpe> hm. the patch is also really hard to read - some insertions are sentence-per-line, some are paragraph-per-line
16:26:17 <tibbs> Yeah, sadly that part of the guidelines hasn't been reflowed and if I did it now then someone would have to merge manually.
16:26:55 <mhroncok> this is actually not that bad in the current form
16:28:48 <tibbs> I can say I completely agree with the commit made seven hours ago relating to fedora-obsolete-packages.  Not sure why that commit is bundled in here.
16:29:33 <mhroncok> it was there since the beginning
16:29:56 <tibbs> Yeah, the timestamps all changed when the PR was rebased.  I thought it was new.
16:31:24 <tibbs> Anyway, what I guess I'm not understanding is how this makes things better for packagers.
16:32:54 <mhroncok> I like the current wording about ^ much more. Not sure if we should offer two options for something that is extremely easy to get wrong, but I can live with this proposal if others think it is worth it.
16:33:02 <mhroncok> (All the other ^ unrelated changes seem good.)
16:33:57 <tibbs> It's so hard to read, but now it seems there are two schemes outlined that the packager needs to choose between and I don't see how someone is supposed to know what to do.
16:34:10 <mhroncok> exactly
16:34:14 <decathorpe> I'm not sure either. I agree that replacing one thing that easily go wrong with two different things that easily go wrong in different ways might not be optimal
16:34:35 <geppetto> lol
16:34:53 <geppetto> Again, I'm asking why using caret is better than just not?
16:35:17 <mhroncok> well, it is useful when you expect to keep the package bugfix-updated in previous releases
16:35:18 <decathorpe> maybe it would be better to make documentation for the current scheme better, and add some more examples?
16:35:39 <mhroncok> it often happened to me that we obsoleted 5.4.1-3
16:35:51 <mhroncok> and after 3 months, the previosu Fedora release was at 5.4.1-4
16:35:56 <geppetto> mhroncok: But you just update to the latest released, as you update the package
16:36:18 <decathorpe> well, you could always just obsolete 5.4.1-100 instead ;)
16:36:28 <mhroncok> geppetto: it requires always updating the obsoliting package in Fedora N when you update the obsoelted package in Fedora N-1/N-2
16:36:35 <tibbs> decathorpe:  Yes, that's the alternate "leave some space" method.
16:36:39 <mhroncok> decathorpe: yeah, that was my recommened hack
16:36:41 <geppetto> decathorpe: I'll take things that will go wrong in a different way for $500
16:36:56 <decathorpe> heh :)
16:37:06 <mhroncok> decathorpe: the ^ thing esentially acts like 5.4.1-infinity
16:37:12 <geppetto> yeh
16:37:13 <mhroncok> which is... dnagerous in so many ways
16:37:22 <tibbs> The problem is that we have to keep fedora-obsolete-packages in sync, and that gets missed so things stop being obsoleted.
16:37:36 <tibbs> And it would be nice to have a _default_ scheme that makes this easier.
16:37:36 <mhroncok> give or take
16:37:52 <tibbs> I'm just not sure how this change helps.
16:38:02 <geppetto> this might be easier, but I'm far from sure it's better
16:38:15 <mhroncok> it's easier, and more dangerous
16:38:29 <mhroncok> it's nicer if you are right
16:38:31 <decathorpe> right. because it makes introducing a package with the same version again impossible without epoch
16:38:37 <mhroncok> and it's bad if you realize you were wrong
16:38:59 <geppetto> decathorpe: AIUI you can use the caret on the package, so you don't need epoch
16:39:33 <geppetto> but like 1% of fedora will know about that, at best
16:39:33 <decathorpe> yay for 5.4.1^unretired-100.fc35 ?
16:39:41 <geppetto> yeh
16:40:02 <mhroncok> geppetto: that's almost a sbad as epoch, except it might eventually be removed with the next release (if it ever happens)
16:40:07 <tibbs> So what I'd like to do is lift the uncontroversial commit, then finish applying SemBr to index.adoc and hope that we can get this PR formatted in a way that it's actually readable.  There might be other uncontroversial changes but I just can't spot them because it's so hard to read.
16:40:32 <geppetto> tibbs: I'm fine with you doing that
16:40:42 <mhroncok> there are many good changes in here and the oyl controversial part is the ^ thing
16:41:12 <decathorpe> sounds like a good plan; +1
16:41:31 <tibbs> mhroncok: Do you want to lift those good parts out?  I am really bad at git so I would end up just retyping it.
16:42:06 <mhroncok> Sure
16:42:12 <tibbs> Alternately if someone tells me how to separate out the good stuff in git while preserving the attribution then I'll do it.  Always willing to learn.
16:42:33 <mhroncok> tibbs: I can do it together with you after the meeting if you want to
16:42:49 <tibbs> Sure.  I would love to know how to do it.
16:43:35 <geppetto> tibbs: I would pull the commit locally and use --amend
16:43:44 <geppetto> There might well be a better way
16:43:52 <mhroncok> that's what I would do
16:43:54 <decathorpe> I'd do "git cherry-pick" :)
16:44:13 <mhroncok> decathorpe: and then?
16:44:17 <tibbs> I have only used amend to fix a typo in an unpushed commit message.
16:44:26 <decathorpe> git push?
16:44:35 <mhroncok> decathorpe: and the changes?
16:44:43 <tibbs> Let's not get too deep into git here, though.
16:44:47 <mhroncok> ack
16:45:07 <geppetto> #topic #1107 Soft-static allocation of gids for "tcb passwd shadowing"
16:45:17 <decathorpe> ah yeah, you want to do changes on top of both commits. then cherry-pick won't help.
16:45:24 <geppetto> .fpc 1107
16:45:25 <zodbot> geppetto: Issue #1107: Soft-static allocation of gids for "tcb password shadowing scheme" - packaging-committee - Pagure.io - https://pagure.io/packaging-committee/issue/1107
16:45:33 <geppetto> https://pagure.io/packaging-committee/issue/1107
16:45:45 <mhroncok> I haven't read that just yet
16:45:56 <tibbs> +1
16:46:46 <besser82[m]> If there are questions about the ticket, I'm here.
16:47:50 <geppetto> Seems simple, I'm kind of surprised there weren't allocations from pam_unix
16:47:59 <decathorpe> I have no idea what pam, pam_unix, or pam_tcb are or what they do, but if a soft static allocation for those groups is required to make the change work, then I'm all for it ...
16:47:59 <geppetto> but, yeh, +1
16:48:28 <decathorpe> though maybe we should wait with actually pushing the allocation until the FESCo change is approved? so we don't allocate things that won't end up shipping (and free numbers below 500 already being sparse)
16:48:50 <mhroncok> we can allocate conditionally
16:49:41 <tibbs> We don't actually make the changes to the setup package anyway, so I assume that nothing else will happen until the change goes through.
16:50:03 <decathorpe> +1 then
16:50:33 <GwynCieslasheher> +1 for me as well.
16:51:09 <geppetto> #action #1107 Soft-static allocation of gids for "tcb password shadowing scheme" (+1:5, 0:0, -1:0)
16:51:48 <geppetto> #topic Open Floor
16:52:02 <geppetto> Ok, with less than 10 mins. I'm going to just do open floor
16:52:02 <mhroncok> +1
16:52:20 <geppetto> We didn't get to: #1099 Contradicting sections regarding duplicate files
16:52:38 <geppetto> But we can get to that another week
16:52:58 <geppetto> Anything else anyone wants to talk about?
16:53:18 <GwynCieslasheher> Not I, said the duck.
16:53:21 <tibbs> I did a bunch of cleanup.  I doubt anyone cares if I just merge them but I will leave the eclipse thing alone until we can talk about.
16:53:29 <tibbs> it
16:53:59 <geppetto> If you are pretty sure, and it's just cleanup … I'd just merge it
16:54:09 <geppetto> how bad could it be
16:54:17 <geppetto> ᕕ( ᐛ )ᕗ
16:54:36 <tibbs> I changed links to https, fixed or remove bad links, added a makefile target for checking links.
16:54:47 <mhroncok> tibbs: merge it
16:55:07 <tibbs> But there's a bad link in the eclipse plugin docs, and since we don't have eclipse at all it seems kind of dumb to fix the document instead of just removing it.
16:55:24 <decathorpe> I briefly looked at some diffs and they all looked good, so merge away 💯
16:56:11 <tibbs> OK, can always revert if eclipse comes back I guess.
16:56:52 <tibbs> There is still a whole lot of work to do but things are getting a bit cleaner.
16:57:20 <tibbs> And the issue list is at least slightly shorter, although there are still far too many pending issues.
16:57:39 * geppetto nods … sounds good
16:59:27 <geppetto> Ok, on that note happy end of meeting
16:59:30 <geppetto> #endmeeting