16:00:03 <geppetto> #startmeeting fpc
16:00:03 <zodbot> Meeting started Thu Aug 27 16:00:03 2020 UTC.
16:00:03 <zodbot> This meeting is logged and archived in a public location.
16:00:03 <zodbot> The chair is geppetto. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:03 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:00:03 <zodbot> The meeting name has been set to 'fpc'
16:00:03 <geppetto> #meetingname fpc
16:00:03 <geppetto> #topic Roll Call
16:00:03 <zodbot> The meeting name has been set to 'fpc'
16:01:07 <decathorpe> .hello2
16:01:08 <zodbot> decathorpe: decathorpe 'Fabio Valentini' <decathorpe@gmail.com>
16:01:11 <geppetto> #chair decathorpe
16:01:11 <zodbot> Current chairs: decathorpe geppetto
16:01:28 <decathorpe> hello geppetto :)
16:01:33 <geppetto> hey
16:01:45 <tibbs> Hey.
16:01:53 <geppetto> #chair tibbs
16:01:53 <zodbot> Current chairs: decathorpe geppetto tibbs
16:02:07 <tibbs> I still have power but didn't get a whole lot of sleep.
16:02:14 <carlwgeorge> .hello2
16:02:15 * geppetto nods
16:02:15 <zodbot> carlwgeorge: carlwgeorge 'None' <carl@redhat.com>
16:02:26 <geppetto> Hey, carl
16:02:39 <mhroncok> hey
16:02:44 <geppetto> #chair mhroncok
16:02:44 <zodbot> Current chairs: decathorpe geppetto mhroncok tibbs
16:03:04 * limburgher here
16:03:08 <geppetto> #chair limburgher
16:03:08 <zodbot> Current chairs: decathorpe geppetto limburgher mhroncok tibbs
16:03:40 <geppetto> tibbs: Is the uni opening in a normal way?
16:04:17 <tibbs> The university is basically closed.
16:04:18 <defolos> tibbs: besides not sleeping, you're ok?
16:05:08 <tibbs> But that was more than a precaution; the thing basically missed us completely and all we got was some wind.  I'm not close to the coast so the storm surge wasn't a concern (and it was only 70cm anyway).
16:05:39 * geppetto nods
16:06:34 <geppetto> #topic Schedule
16:06:37 <geppetto> #link https://lists.fedoraproject.org/archives/list/packaging@lists.fedoraproject.org/message/XNDM5XLLI2A337LEB32REEXLG352PHYE/
16:07:00 <limburgher> That's a relief.
16:07:22 <tibbs> All of western Louisiana is in very bad shape, though.  7 meters of storm surge in places and > 200kph winds.
16:07:32 <limburgher> Jesus.
16:07:39 <geppetto> So no new tickets/PRs … but we have some from last week … mhroncok asked people to vote on PR's 1011 and 1013, I think we are missing a couple of votes though
16:07:49 <geppetto> Anyone have any questions/concerns about those?
16:08:55 <geppetto> tibbs: Having that in kph is impressive :)
16:10:13 <geppetto> #topic #pr-1011 Python guidelines updates
16:10:13 <geppetto> https://pagure.io/packaging-committee/pull-request/1011
16:10:29 * decathorpe voted on both PRs
16:10:39 <geppetto> decathorpe++
16:11:23 <decathorpe> I didn't read all the details, but I've been looking at the "implementation" / "usage" side of things at they look good, and I trust Miro to write good Guidelines
16:11:46 <geppetto> yeh, that was basically my +1 too
16:12:31 <mhroncok> :)
16:12:58 <tibbs> The "Python Extras" bit is the meat of the change.
16:13:25 <mhroncok> the commits are spearted and can be reviewed one by one
16:13:36 <decathorpe> (this actually fixes bugs in two of my packages, so I like this)
16:13:39 <mhroncok> Python Extras indeed is the longest part and deserves the most attention
16:15:41 <tibbs> +1.  The Extras stuff is just documenting a change that was already acked through other means.  The rest is documenting existing things and cleaning up some stuff.
16:16:21 <mhroncok> that was the idea
16:19:13 <geppetto> Ok, that's +5
16:19:20 <mhroncok> \o/
16:19:46 <geppetto> #action Merge pr-1011 Python guidelines updates (+1:5, 0:0, -1:0)
16:19:59 <geppetto> #topic #pr-1013 Document py_reproducible_pyc_path macro
16:19:59 <geppetto> https://pagure.io/packaging-committee/pull-request/1013
16:20:31 <geppetto> Again, everyone but tibbs has already voted
16:20:58 <geppetto> Feel free to have a look though carlwgeorge, and let us know if you see anything
16:21:57 <tibbs> This seems reasonable to me; it's obscure but in the appendix and if you are unlucky enough to hit the problem then it will be nice to have it documented somewhere.
16:22:00 <tibbs> +1
16:23:10 <carlwgeorge> makes sense to me
16:23:20 <carlwgeorge> +1
16:23:28 * mhroncok tried to fix the typo but cannot force push to the fork for some reason :/
16:23:48 <mhroncok> so I push it to master manually and close the PR
16:24:01 <geppetto> #action Merge #pr-1013 Document py_reproducible_pyc_path macro (+1:5, 0:0, -1:0)
16:24:01 <decathorpe> WFM :)
16:24:14 <geppetto> mhroncok: git being a git
16:24:19 <tibbs> Yeah, it does need a grammar pass but obviously minor.
16:24:43 <geppetto> #topic #pr-942 Recommend storing changelog entries in separate file.
16:24:44 <geppetto> https://pagure.io/packaging-committee/pull-request/942
16:24:55 <geppetto> Ok, so this is technically new … but it's been around a bit
16:24:56 <King_InuYasha> .hello ngompa
16:24:56 <zodbot> King_InuYasha: ngompa 'Neal Gompa' <ngompa13@gmail.com>
16:24:59 <King_InuYasha> sorry I'm late
16:25:01 <geppetto> Hey, neal
16:25:06 <King_InuYasha> work meeting ran over
16:25:29 <King_InuYasha> oh, we're talking about changelogs now, just in time I suppose
16:25:33 <geppetto> Also tried to ping sgallagh but failed.
16:25:52 <decathorpe> so ... I don't think PR 942 is a good idea on its own. I'd rather see this integrated with pingou's rpmautospec
16:25:59 <King_InuYasha> yes
16:26:00 <sgallagh> Hmm?
16:26:13 <King_InuYasha> in my opinion, this provides no benefits and adds complexity
16:26:41 <mhroncok> that
16:26:41 <geppetto> sgallagh: See https://pagure.io/packaging-committee/pull-request/942 … wanted to see your opinion on having changelogs not be in specfiles directly.
16:26:46 <tibbs> Don't we have like three competing implementations at this point?
16:26:46 <King_InuYasha> this is a situation that I deal with fairly regularly in openSUSE, and it's just not apprecibly different from just having the changelog in the main file
16:27:03 <King_InuYasha> it's arguably worse because it's easier to forget to update the changelog
16:27:14 <sgallagh> Will look. Eating lunch right now.
16:27:26 <geppetto> decathorpe: Yeh, there are a bunch of things trying to get here … lots of automated rebuilding.
16:27:27 <King_InuYasha> tibbs: yes
16:27:38 <decathorpe> tibbs: yeah, but at least the one that was deployed in koji-staging was opt-in and integrated with other workflows (other than rpmdev-bumpspec)
16:27:44 <King_InuYasha> yup
16:27:50 <King_InuYasha> and fixing it for rpmdev-bumpspec won't be hard
16:28:00 <King_InuYasha> just with no staging environment, no way to test
16:28:31 <geppetto> King_InuYasha: Does OBS do anything we could borrow for auto rebuilding etc.
16:28:54 <King_InuYasha> it has a spec parser and resolver that identifies the dep chains, sorts them, and rebuilds them correctly
16:29:12 * decathorpe shivers when hearing the words "spec parser"
16:29:14 <King_InuYasha> and the spec parser allows it to update the release field semi-intelligently on each rebuild
16:29:24 <King_InuYasha> well, OBS' parser predates the existence of rpmspec
16:29:31 <King_InuYasha> and an API in rpm for parsing spec files
16:30:39 <King_InuYasha> when Factory 2.0 was being worked on, I had hoped we'd get something like that in Fedora too
16:30:43 <geppetto> Yeh, I kind of feel like this PR and other efforts are trying to do something without having to have a spec parser
16:30:45 <King_InuYasha> but alas, my hopes were dashed
16:31:11 <King_InuYasha> there's no way around it, you *need* a spec parser to do this right
16:31:32 <geppetto> King_InuYasha: Just drop the last 4 words and we gud ;)
16:31:38 <King_InuYasha> :D
16:31:57 <tibbs> From a guideline standpoint, we don't care what happens under the hood; we only care about what packagers need to do.
16:32:04 <King_InuYasha> right
16:32:27 <King_InuYasha> from my perspective, we should reject pr-942
16:32:56 <decathorpe> it will hopefully be superseded by something "better" soon :)
16:33:07 <King_InuYasha> 1. we have no tooling for putting it back together for package builds and 2. this doesn't solve any problems compared to status quo
16:33:07 <mhroncok> +1 tot reject
16:33:09 <mhroncok> *to
16:33:12 <limburgher> I'm struggling to understand what the motivating benefit of this is. Maybe I missed something. What pain does this fix?
16:33:41 <decathorpe> limburgher: IIUC it doesn't, at least not without other changes that aren't there yet
16:33:43 <tibbs> I guess it makes specfiles shorter?  Maybe makes merges easier?
16:33:58 <geppetto> tibbs: I think merges are the main point
16:34:04 <limburgher> ...
16:34:05 <King_InuYasha> it doesn't fix merges because commits don't specifically care about files, they care about blobs
16:34:07 <tibbs> Also, does this %external_changelog macro even exist currently?
16:34:12 <King_InuYasha> it does not
16:34:19 <tibbs> Then, uh....
16:34:24 <geppetto> you can do a bunch of automation if there's no changelog and the release is handled by robots
16:34:29 <King_InuYasha> yup
16:34:29 <limburgher> Right, to git, it's not files, it's a pile of lines.
16:34:42 <King_InuYasha> geppetto: that's a big part of what OBS does
16:34:56 <geppetto> King_InuYasha: I've had lots of cherry pick merges fail due to changelog entries
16:35:02 <King_InuYasha> though since the OBS VCS sucks, they have a changelog file
16:35:16 <limburgher> What does this give us that rpmdev-bumpspec doesn't?
16:35:18 <King_InuYasha> and OBS auto-formats it into a %changelog section and appends it to spec
16:35:19 <carlwgeorge> second the cherry-pick thing
16:35:41 <King_InuYasha> carlwgeorge, geppetto: think about this: if it's a separate file but in the same commit, how will that help for merging?
16:36:04 <King_InuYasha> or cherrypicking
16:36:13 <carlwgeorge> i didn't claim putting it in a separate file helps, just that making cherry-picks work better is a good goal :D
16:36:13 <decathorpe> it changes 1 file with 2 merge conflicts into 2 files with 1 merge conflict each :)
16:36:19 <mhroncok> King_InuYasha: git restore CHNAGELOG --source=origin/master :D
16:36:20 <geppetto> King_InuYasha: was hoping the seperate file would be generated … or ignore merges on it … or something
16:36:29 <King_InuYasha> haha
16:36:36 <King_InuYasha> yeah, this PR does not propose any of that
16:36:54 <King_InuYasha> pingout's rpmautospec will actually provide a path to that, though
16:36:56 <limburgher> So the benefit of this PR would appear, with the current RPM to be nil?
16:36:59 <King_InuYasha> yes
16:37:12 <King_InuYasha> also, still no solution for releng tooling :)
16:37:20 <sgallagh> geppetto: I don't think I saw your ping, but I'm completely opposed to this approach.
16:37:22 <geppetto> I think it gets us to a place where the benefit could be non-nil … but eh
16:37:33 <geppetto> sgallagh: Fair enough :)
16:37:40 <limburgher> It's a lot of churn for *could*
16:37:54 <limburgher> I'm -1 unless I hear a compelling argument.
16:38:02 <carlwgeorge> -1 for me as well
16:38:03 <sgallagh> I'd much rather see an approach that just encodes the changelog text in the git commits.
16:38:07 <King_InuYasha> -1 from me as well
16:38:11 <decathorpe> -1 as well
16:38:30 <mhroncok> -1
16:38:43 <mhroncok> as a side note: https://github.com/encukou/rpm-spec-merge-driver
16:39:12 <geppetto> #action #pr-942 rejected … pretty much everyone doesn't like this approach to solving the problem. Speak to pingu, sgallagh, rpmbumpspec devs etc.
16:39:19 <tibbs> -1
16:39:33 <tibbs> It doesn't really solve any problem, unless the problem is that the specfiles are just too long.
16:40:21 <tibbs> I do still have the example where I was forbidden from stripping ancient changelog entries from a specfile.
16:40:58 <geppetto> Did you try to do that as a non-maintainer?
16:41:05 <geppetto> s/maintainer/owner/
16:41:37 <sgallagh> I would happily support a two-Fedora-release maximum on Changelog entries.
16:41:43 <tibbs> No, I co-maintain the package in question.
16:41:48 <geppetto> Eh, I probably don't care … too much of this was a good idea when we had 200 packages and now not so much.
16:42:15 <decathorpe> 200? good ol' times
16:42:17 <King_InuYasha> sgallagh: I wouldn't, but ehh
16:42:26 <geppetto> King_InuYasha: Shorter?
16:42:35 <limburgher> I personally like being able to read about what happened to a package in the olden days but I'm odd.
16:42:37 <sgallagh> geppetto++
16:42:37 <zodbot> sgallagh: Karma for james changed to 6 (for the current release cycle):  https://badges.fedoraproject.org/tags/cookie/any
16:42:38 <King_InuYasha> geppetto: I am generally not in favor of trimming changelogs
16:42:48 <tibbs> I had rewritten the specfile from scratch for a new major version of the software, but years and years ago Red Hat had incorporated someone else's specfile and promised that their changelog entries would remain in perpetuity.
16:42:51 <limburgher> King_InuYasha: same
16:43:02 <sgallagh> Trim the changelog into a file after two releases, maybe?
16:43:06 <sgallagh> That would be a reasonable compromise.
16:43:07 <King_InuYasha> I get screwed surprisingly often by long ago changes, and being able to figure that out is nice
16:43:15 <King_InuYasha> sgallagh: we already do that automatically with rpm
16:43:24 <sgallagh> We do?
16:43:26 <King_InuYasha> only the spec file has the complete changelog
16:43:33 <King_InuYasha> everything else is trimmed to two years
16:43:34 <tibbs> Yes, rpm just ignores old changelog entries.
16:43:40 <sgallagh> huh, I did not know that
16:43:49 <geppetto> King_InuYasha: At this point I feel like unless you can git blame it almost nobody will see it after 12 months.
16:43:54 <decathorpe> yup, rpmbuild cuts off all changelog entries that are older than a year I think?
16:44:05 <King_InuYasha> geppetto: I can't rely on git because of CentOS/RHEL
16:44:08 <mhroncok> which is funny when you build old SRPMs
16:44:09 <King_InuYasha> and openSUSE
16:44:23 <tibbs> But this is Fedora.
16:44:27 <geppetto> tibbs: I don't think that's true … createrepo trims them automatically, from being in the repo.
16:44:34 <mhroncok> this is Sparta!
16:44:45 <geppetto> But if you look at a package, or install it … all of it is still there.
16:44:46 <King_InuYasha> tibbs: and Fedora has EPEL, you're point?
16:44:51 <King_InuYasha> *your
16:44:59 <tibbs> redhat/macros:%_changelog_trimtime      %{lua:print(os.time() - 2 * 365 * 86400)}
16:45:07 <King_InuYasha> it applies to having to deal with CentOS/RHEL packages
16:45:14 <tibbs> This committee is not concerned with EPEL.
16:45:19 <King_InuYasha> and we don't exist in a vacuum
16:45:40 <geppetto> tibbs: huh … so that's an rpmbuild thing?
16:45:45 <King_InuYasha> geppetto: yep
16:45:47 <mhroncok> (this discussion is not getting us anywhere)
16:45:57 <King_InuYasha> it was added like five years ago, I think?
16:45:57 <geppetto> mhroncok: indeed
16:46:12 <King_InuYasha> and we turned it on three years ago?
16:46:15 <tibbs> Regardless, the PR under discussion was already rejected; this is just rambling.
16:46:21 <mhroncok> King_InuYasha: the changalog for when this was added is long gone :D
16:46:22 <geppetto> King_InuYasha: So became usable a few months ago in rhel8?
16:46:29 * geppetto ᕕ( ᐛ )ᕗ
16:46:37 <King_InuYasha> geppetto: yeah, it's in RHEL 8 already :D
16:46:49 <geppetto> #topic #977 Get new members?
16:46:49 <geppetto> .fpc 977
16:46:49 <geppetto> https://pagure.io/packaging-committee/issue/977
16:46:50 <zodbot> geppetto: Issue #977: Get new members? - packaging-committee - Pagure.io - https://pagure.io/packaging-committee/issue/977
16:47:03 <geppetto> That kind of leads us to this …
16:47:06 * King_InuYasha keeps thinking of "Got milk?" campaign with that statement
16:47:11 <geppetto> carlwgeorge and King_InuYasha wave to everyone
16:47:13 <mhroncok> so I assume we got carlwgeorge and King_InuYasha?
16:47:17 * carlwgeorge waves
16:47:20 * King_InuYasha waves to the flat crowd
16:48:02 <geppetto> mhroncok: technically we haven't all officially voted … but I've not seen anyone say anything negative either, so I'm guessing we'll be adding both
16:48:05 <decathorpe> should we make it official?
16:48:17 <geppetto> decathorpe: That's kind of what we do :)
16:48:33 <carlwgeorge> put a ring on it :D
16:48:37 * decathorpe goes to fetch the big rubber stamp
16:48:39 <mhroncok> I assume that we would replce mbooth, right?
16:49:20 <decathorpe> yeah, I think he already stepped down a few months ago?
16:49:33 <geppetto> Then yeh
16:49:41 <mhroncok> he said "if you got somebody else, feel free to swap me"
16:49:45 <geppetto> we have 8 atm … aiming for 9
16:49:57 <geppetto> 8 includes mbooth
16:50:42 <King_InuYasha> so we still need another?
16:50:52 <mhroncok> ok, proposal: replace mbooth with carlwgeorge and King_InuYasha
16:50:58 <geppetto> King_InuYasha: 7
16:50:59 <mhroncok> King_InuYasha: not really
16:51:04 <geppetto> King_InuYasha: 7+2 == 9 :)
16:51:08 <King_InuYasha> :D
16:51:19 <King_InuYasha> math is hard, doncha know? :P
16:51:20 <decathorpe> mhroncok: +1
16:51:40 <tibbs> No complaints from me.
16:51:45 <limburgher> Sounds good.
16:52:05 <carlwgeorge> \o/
16:52:07 <mhroncok> geppetto: ?
16:52:18 <geppetto> Ok, I'm going to just stamp it then … collect your rings from the front desk.
16:52:30 <mhroncok> that is basically +5
16:52:35 <sgallagh> That would be the SufferRing, yes?
16:52:38 <geppetto> #action Add carlwgeorge and King_InuYasha as new members
16:52:49 <carlwgeorge> sgallagh: hahaha
16:53:07 <decathorpe> you won't be laughing for long ;)
16:53:14 * King_InuYasha laughs and sighs
16:54:05 <geppetto> sgallagh: We generally prefer to wait for at least the second time they are here officially before letting them know what a terrible idea it was
16:54:34 <limburgher> Come play with us carlwgeorge and King_InuYasha .... come play with us forrrreeeeeevvveeeerrrrrrrr
16:54:34 <mhroncok> carlwgeorge: you are carlgeorge in FAS, rght?
16:54:36 <mhroncok> *right
16:54:52 <carlwgeorge> mhroncok: nope, carlwgeorge
16:54:59 <decathorpe> limburgher: wow that was dark
16:55:01 <mhroncok> carlwgeorge: oh, sorry
16:55:14 <limburgher> decathorpe: I'm that girl. :shrugs:
16:55:17 <geppetto> #topic Open Floor
16:55:18 <carlwgeorge> carlgeorge was my old fas, but there is no way to change a fas so i started over at some point
16:55:48 <bcotton> the "w" stands for "with a w now". it's very meta
16:55:48 <King_InuYasha> new AAA system might let us actually do renames :D
16:55:54 <geppetto> Ok, going to try to end before the hour this week so bcotton doesn't have to #action me ;)
16:56:09 <King_InuYasha> Council is in #fedora-meeting now
16:56:11 <bcotton> geppetto: no council meeting this week, so you're okay :-) and we moved to #fedora-meetnig in the future
16:56:12 <King_InuYasha> so he's not going to
16:56:25 <bcotton> King_InuYasha: oh i still might ;-)
16:56:29 <King_InuYasha> hah
16:56:37 <geppetto> Ohh, back to dealines making a nice whooshing sound then
16:56:45 <mhroncok> https://fedoraproject.org/wiki/Packaging_Committee#Members
16:57:19 <decathorpe> mhroncok++
16:57:51 <mhroncok> also https://pagure.io/packaging-committee
16:57:56 <mhroncok> there is no FAS group, right?
16:58:44 <limburgher> IIRC we only needed that for Trac.
16:58:49 * limburgher has dated herself
16:59:01 <mhroncok> carlwgeorge: https://fedoraproject.org/w/index.php?title=User:Carlgeorge&redirect=no
16:59:55 <geppetto> limburgher: bah, trac wasn't that long ago … was it?
16:59:58 <carlwgeorge> heh, some out of date stuff there
17:00:21 <limburgher> geppetto: after this year, who can remember?
17:00:39 <geppetto> limburgher: indeed, everything is now the before times.
17:01:10 <limburgher> Remember kids, commit to CVS before submitting a build to plague!
17:01:20 <King_InuYasha> oh god
17:01:22 <King_InuYasha> I remember that
17:01:29 <limburgher> LOL
17:01:38 <geppetto> wow, limburgher with the black and white TV references
17:01:44 <decathorpe> King_InuYasha: I'm somehow glad I don't :)
17:01:57 <limburgher> Sort of astounding that we still got stuff done...
17:01:57 <King_InuYasha> plague was in the process of being replaced with Koji when I started contributing to Fedora
17:02:02 * mhroncok only remembers CVS in rpmfusion
17:02:20 <King_InuYasha> the CVS to Git change happened a few years later
17:02:25 <limburgher> Yeah.
17:02:31 <tibbs> The plague days were so much simpler.
17:02:44 <smooge> "Do you remember when we had to cvs commit for plague builds? Well RPM committee remembers"
17:02:53 * limburgher dying
17:03:10 * King_InuYasha cracks up
17:03:19 <geppetto> smooge: Oh, for sure you remember too
17:03:34 <limburgher> Barring head trauma...
17:03:43 <smooge> thinks cvs was a move too modern
17:03:58 <bcotton> RIP limburgher
17:04:12 * limburgher haunts bcotton
17:04:17 <smooge> RCS over rsh gave you everything you ever needed
17:04:58 <smooge> then again still uses emacs ~N~ as his version control system on his desktop
17:05:15 <limburgher> Unless someone has something else, I'll go eat my sepia-toned lunch in the drawing room...
17:05:24 <smooge> ls /etc does take a couple of hours.. but I can see all the changes
17:05:33 * geppetto checks for smooge in ~/.rhosts
17:05:36 <decathorpe> :O
17:05:45 <King_InuYasha> X_X
17:05:54 <geppetto> Sounds good
17:05:59 <smooge> have a good lunch
17:06:10 <limburgher> Thanks all, stay safe!
17:06:10 <mhroncok> enojoy your lucnh
17:06:16 <mhroncok> it's wine time for me already :)
17:06:22 <smooge> maybe we can relive the version control wars of 20 years ago next time
17:06:32 <geppetto> smooge: SCCS ftw
17:06:35 <tibbs> Bitkeeper forever.
17:06:42 <smooge> I don't disagree geppetto
17:06:50 <mhroncok> :)
17:06:51 <geppetto> haha
17:06:53 <King_InuYasha> heh
17:07:10 <mhroncok> Now available as Open Source, didn't know that
17:07:34 <smooge> I actually learned that one first, then RCS because we lost our licence to use the SCCS binaries
17:07:40 <geppetto> mhroncok: yeh, I downloaded it when it went that way … a few years too late
17:08:04 <smooge> #endmeeting
17:08:20 <mhroncok> #endmeeting