17:00:24 <geppetto> #startmeeting fpc
17:00:25 <zodbot> Meeting started Thu Mar  1 17:00:24 2018 UTC.  The chair is geppetto. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:00:25 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
17:00:25 <zodbot> The meeting name has been set to 'fpc'
17:00:25 <geppetto> #meetingname fpc
17:00:25 <geppetto> #topic Roll Call
17:00:25 <zodbot> The meeting name has been set to 'fpc'
17:00:28 <mbooth> hello
17:00:33 <geppetto> #chair mbooth
17:00:33 <zodbot> Current chairs: geppetto mbooth
17:00:37 <orionp> hello
17:00:44 <geppetto> #chair orionp
17:00:44 <zodbot> Current chairs: geppetto mbooth orionp
17:00:45 * limburgher HELO
17:00:49 <geppetto> #chair limburgher
17:00:49 <zodbot> Current chairs: geppetto limburgher mbooth orionp
17:01:00 <geppetto> limburgher: 200
17:01:42 <limburgher> geppetto: ^]
17:02:30 <mbooth> limburgher: 221 BYE
17:03:09 <nb> hello
17:03:17 <limburgher> mbooth: 500 FELICIA
17:03:40 <nb> 418 IMATEAPOT
17:06:39 <mbooth> ENOQUORUM
17:07:24 <geppetto> indeed
17:07:26 <geppetto> 404
17:07:38 <geppetto> #topic Open Floor
17:08:02 <geppetto> I sent out 3 applications earlier today … so hopefully everyone got those?
17:08:16 * mbooth checks email
17:08:17 <geppetto> Anything else anyone wants to talk about?
17:08:24 * nb is interested in applying
17:08:52 <limburgher> I did.
17:08:58 <mhroncok> I did as well.
17:09:01 <nb> I will send an application probably later today
17:09:45 <mhroncok> geppetto: how many do you get already?
17:09:47 <tibbs> Hey, folks.
17:09:55 <tibbs> Sorry, bad traffic this morning.
17:10:16 <geppetto> #chair tibbs
17:10:16 <zodbot> Current chairs: geppetto limburgher mbooth orionp tibbs
17:10:21 <limburgher> Freeway QoS is the literal worst.
17:10:25 <geppetto> go go FPC bots.
17:11:03 <geppetto> nb: cool, there's not a huge rush … it's open for two weeks
17:11:14 <geppetto> nb: And it's not first come first served ;)
17:11:19 <geppetto> mhroncok: 3 so far.
17:11:28 <geppetto> #topic Schedule
17:11:36 <geppetto> https://lists.fedoraproject.org/archives/list/packaging@lists.fedoraproject.org/message/V66SM3LTNII33D53WEJXAYX4IY2QCLZU/
17:11:46 <geppetto> #topic #752 metainfodir change breaks builds with appdata on F27
17:11:49 <geppetto> .fpc 752
17:11:51 <zodbot> geppetto: Issue #752: metainfodir change in F28 and breaks builds with appdata directories on F27 or lower - packaging-committee - Pagure - https://pagure.io/packaging-committee/issue/752
17:12:19 <tibbs> There's an interesting question underneath here, which is whether someone can ask FPC to add new macros.
17:13:02 <limburgher> I don't think so, since we don't control the macros. They can ask us to have the usage documented.
17:13:08 <tibbs> I guess I don't see why not, really, but traditionally those things have gone via bugs filed against redhat-rpm-config.
17:13:18 <geppetto> yeh
17:13:42 <geppetto> Add the macro somewhere … do the %{?foo} dance for compat. … and then update policy
17:13:53 <tibbs> But I have insinuated myself into some measure of redhat-rpm-config maintainership.
17:14:17 <tibbs> And when adding such things, I try to make sure they are available everywhere, including EPEL, so there's no question of compatibility.
17:14:29 <mbooth> Is there any other kind of maintainership than de-facto maintainership? :-o
17:14:49 <tibbs> I don't think it would be bad if people could ask us for those things and if we approved we'd just implement them.
17:14:52 <geppetto> Not in Fedora ;)
17:15:32 <tibbs> That package is weird in many ways, not just because of "redhat" in the package name.
17:15:44 <tibbs> I tried with fedora-rpm-macros but that didn't work out so well.
17:16:19 <geppetto> I'm somewhat reticent to just say we (FPC) own the macros package now, but obviously anything tibbs can do to make this smoother is great.
17:16:44 <geppetto> I'm also kind of confused about what this macro accomplishes … are they planning on changing the path?
17:16:50 <tibbs> Yeah, I don't think we want to pretend to own it.  Nominally I think it's Florian's.
17:17:04 * geppetto nods
17:17:35 <tibbs> I believe the problem is that we didn't realize that the guideline we implemented is actually incompatible with <= F27.
17:17:45 <tibbs> So if you try to follow it, you need ifdefs.
17:17:52 <geppetto> Ahh
17:18:20 <geppetto> I see … the path already changed and we are currently documenting F28+
17:18:28 * geppetto sighs
17:18:36 <tibbs> I really thought that the older releases looked in both places, but I guess the new directory doesn't even exist in F27.
17:19:15 <tibbs> If this can be cleaned up with the addition of one macro and a tweak to the guidelines, then I guess I don't see why we wouldn't just do it.  And get it into epel-rpm-macros as well (which I also mostly maintain).
17:19:43 <limburgher> Works for me.
17:19:51 <geppetto> sure +1
17:20:06 <tibbs> So the next time I'm in there I'll just fix that one up as well.
17:20:27 <mbooth> Yeah, I'm +1
17:21:54 <tibbs> Certainly I'm +1 if we're voting.
17:22:01 <tibbs> I wish the bot actually handled voting.
17:22:52 <tibbs> Have to run real quick and clear a paper jam.
17:23:00 <geppetto> Bot James does ;)
17:23:38 <mbooth> Mmmm, paper jam.
17:24:13 <geppetto> limburgher: Is that a +1?
17:24:17 <geppetto> orionp: vote?
17:24:21 <limburgher> +1 to being +1
17:24:38 * limburgher would present a challenge to a vote-counting bot.
17:24:41 <tibbs> Back.
17:24:44 <orionp> sorry, spaced that the meeting started...  reading..
17:25:09 <geppetto> orionp: Basically adding a metainfodir macro and updating the guidelines to use it.
17:25:57 <orionp> +1
17:26:25 <geppetto> limburgher: I'm going to assume your "works for me" is a +1 :p
17:26:30 <tibbs> And a sort of meta question as to whether we want to actually make decisions about adding macros like that.
17:26:33 <geppetto> #action Add metainfodir macro and change guidelines (+1:5, 0:0, -1:0)
17:26:41 <tibbs> Which sort of follows.
17:27:19 <geppetto> I'm fine for you to speed everything up, when you can … but I don't want to guarantee we'll do it for people and then not be able to
17:27:20 <tibbs> In retrospect that means we could have done a vote on nim's forge-hosted macro stuff as well.  (I just ended up merging that myself.)
17:27:46 <mbooth> I don't mind adding trivial macros as long as we all here agree, but for larger works, we still probably want contributions just due to resourcing
17:28:02 <tibbs> I guess the idea is that we shouldn't say that it's someone else's business if someone comes to us and presents a good case for something to be added.
17:28:29 <tibbs> But these macros can show up for all kinds of reasons.  Upstream rpm action, some uberpackager just adding them to redhat-rpm-config, etc.
17:28:51 <tibbs> No real reason why we couldn't just add things too if we think they would be good to have.
17:29:28 <geppetto> Sure. But see above. ¯\_(ツ)_/¯
17:31:16 <tibbs> That's about all I have to say about that one, at least.
17:33:16 <geppetto> ok, moving on
17:33:38 <geppetto> #topic #752 metainfodir change breaks builds with appdata on F27
17:33:46 <geppetto> #topic #753 python2-... should not be packaged when it's not needed
17:33:51 <geppetto> .fpc 753
17:33:54 <zodbot> geppetto: Issue #753: python2-... should not be packaged when it's not needed - packaging-committee - Pagure - https://pagure.io/packaging-committee/issue/753
17:34:27 <tibbs> So I think I've made my opinion pretty clear in the ticket.
17:34:34 <mhroncok> agreed
17:34:40 <mhroncok> what about the last comment?
17:34:51 <tibbs> Haven't read email this morning; let me re-read.
17:35:14 <tibbs> Thing is, I don't know if we know that the python2 stack will be going away.
17:35:29 <tibbs> I believe it's certain that your group will not be maintaining a python2 stack.
17:35:46 <tibbs> (at least in Fedora; enjoy RHEL6 for another several years)
17:35:46 <mhroncok> right. it might stay there for eternity
17:35:52 <geppetto> I'm fine with the last comment
17:36:10 <mhroncok> however we don't want to have more and more py2 stuff just because it's possible
17:36:34 <limburgher> I am too, and if someone still needs to run legacy proprietary py2 thing they can install it themselves.
17:37:03 <geppetto> Yeh, I'm fine with a change that makes it obvious packagers don't have to do py2 if they don't care about it … but if they want to for the next year or so that's fine too.
17:37:09 <racor> IMO, the pressure to keep py2 stuff in fedora originates from apps not from upstream py.
17:37:21 <geppetto> racor: Hey racor!
17:37:22 <limburgher> Right.
17:37:24 <geppetto> #chair racor
17:37:25 <zodbot> Current chairs: geppetto limburgher mbooth orionp racor tibbs
17:37:29 <tibbs> We do already say that there's no requirement that python2 stuff be packaged, but... we could certainly make that more clear.
17:37:50 <mhroncok> racor: and keeping packages needed for apps is absolutely fine
17:37:54 <geppetto> tibbs: You happy with the changes in the last comment?
17:37:54 <racor> geppetto: Yeah, I've had technical probs.
17:38:04 <mhroncok> what I'm afraid of is that packagers of python modules juts keep adding py2
17:38:19 <tibbs> And I guess it wouldn't technically hurt to say that the future of python2 support past the upstream EOL date is not clear, though really, that's not a guidelines thing.
17:38:35 <tibbs> I probably wouldn't state it the same way as in that least comment, but I don't disagree with the sentiment.
17:41:22 * geppetto nods … did you want to try rewording it?
17:42:21 <tibbs> "It is possible that the python2 stack will be removed from Fedora when support for the python2 interpreter officially ends in 2020."
17:42:36 <tibbs> Instead of "Hosever the python2 stack will be removed in the forseeable future".
17:43:19 <geppetto> Sure. +1
17:43:24 <limburgher> +!
17:43:32 <limburgher> =1
17:43:33 <mhroncok> (works for me as well)
17:43:35 <tibbs> And change "mistake" to something like "oversight", though even then that's not correct.
17:43:35 <racor> +1
17:43:37 <limburgher> +1
17:43:40 <limburgher> Finally.
17:43:53 <geppetto> limburgher: 👍
17:43:56 <mbooth> +1
17:43:57 <mhroncok> tibbs: what about "problem"
17:44:12 <tibbs> That also sticks a date on it. Needless to say we will have to have loads of public discussion about what happens to the python2 stack anyway.
17:44:50 <tibbs> mhroncok: Sure, "problem" or even "issue".  It's bikeshedding at this point, but "mistake" is actively negative.
17:45:05 <mhroncok> ok
17:45:19 <limburgher> "situation"?
17:45:21 <tibbs> I do think it would be good to at least start talking about the future of python2 in public.
17:45:36 <limburgher> "jwoww"?
17:45:43 <geppetto> #action Change Packaging:Python to make it clear py2 is nearling upstream EOL and isn't required (+1:6, 0:0, -1:0)
17:45:43 <tibbs> Instead of internal to the python maintainer group or in FPC tickets.
17:45:53 <mhroncok> tibbs: do you think we are not public enough about it?
17:46:10 <limburgher> Right. Like I assume ansible will speak py3 and still work on py2-only hosts by then, but I don't know for sure.
17:46:19 <tibbs> I didn't intend to imply that anything underhanded was happening.
17:46:30 <tibbs> But it would be better if nobody was surprised.
17:47:01 <geppetto> Yeh, I don't think public is right … more like when to start directly pinging packagers of py2 packages?
17:47:02 <mhroncok> BTW we intend to send a reminder to devel@ maybe announce, that right time for removing python2 crap from Fedora is now
17:47:10 <tibbs> And if we want to ramp up to an active discouragement of python2 packages, or a ban on them unless needed as dependencies for applications, then we can move that.
17:47:24 <geppetto> mhroncok: seems reasonable
17:47:24 <mhroncok> we want to identify leave packages in portingdb and politely ask maintainers to consider removing those
17:47:30 <tibbs> Thing is, the python maintainer group needs to telegraph its intentions really loudly.
17:47:52 <tibbs> And make room for others in the community to step in and do that if they want to do so.
17:47:58 <limburgher> I'm thinking Morse on a trashbin with a hammer.
17:48:01 <racor> mhroncok: Are you sure, we won't see a preserve "phy2" initiative somewhere?
17:48:20 <geppetto> limburgher: is that heavy metal?
17:48:31 <tibbs> I fully expect to see some community initiative to keep python2 alive.
17:48:34 <mhroncok> racor: you mean within Fedora?
17:48:35 <tibbs> Or to package tauthon.
17:48:41 <limburgher> geppetto: Needn't be, could be aluminium.
17:48:52 <tibbs> Or, I don't know, pypy still supports py2 and probably will forever.
17:49:10 <mbooth> Also jython 3 will never arrive at this rate
17:49:15 <geppetto> I'm pretty sure yum will ... ᕕ( ᐛ )ᕗ
17:49:27 <racor> mhroncok: Not necessarily. May happen in Fedora, may happen by some $$$-company ....
17:49:33 <limburgher> 4Suite should be done. . .
17:50:14 <tibbs> But my whole point was that the current python stack maintainers don't completely control the future of python2 in Fedora.
17:50:26 <mhroncok> there are plenty of scenarious that might happen, having tauthon or relying on pypy2, why not. nevertheless the current py2 stack in fedora is cpython based, so something needs to be done anyway
17:50:35 <racor> mhroncok: just have a look at the situation perl is in. perl6 is widely ignored by many people ;)
17:50:43 <tibbs> All we can really say is that it's going to be a huge mess.
17:50:56 <mhroncok> for sure
17:51:01 <geppetto> racor: we seem to have mostly got past that point with py2 => py3
17:51:22 <mhroncok> also, most of the large upstream projects move quite fast to py3
17:51:30 <racor> geppetto: Are you sure? I doubt it.
17:51:32 <mhroncok> see Django now
17:51:41 <limburgher> Yeah, that's in a lot of changelogs for recent major releases.
17:51:48 <mhroncok> I don't see taht much Perl 6 only projects out there
17:52:10 <mhroncok> but I don't follow Perl that closely
17:52:52 <geppetto> racor: I'm not saying everyone has moved, but it's way better than perl6 … and I'd guess over 50% of python has either moved or has near term plans to do so
17:52:53 <racor> mhroncok: But there also are packages with "almost dead" upstreams in Fedora.
17:53:20 <mhroncok> racor: there are also plenty of "almost dead" packagers
17:53:43 <mhroncok> often "mainatining" those mostly dead packages
17:53:56 <geppetto> Anyway … we already voted on the wording change here, and I think we are good?
17:54:04 <racor> mhroncok: And you want to be the one who pulls the trigger to kill them?
17:54:06 <tibbs> Yeah.
17:54:11 <geppetto> #topic Open Floor
17:54:20 <mhroncok> racor: well i tought about it
17:54:22 <geppetto> Ok, so we have 5 minutes … anyone have anything?
17:54:32 <limburgher> Not I.
17:54:33 <mhroncok> I want to ask for an early feedback on something
17:54:39 <mhroncok> https://fedoraproject.org/wiki/User:Churchyard/Packaging:PythonMustMayMacros
17:54:41 <racor> mhroncok; Users will hate you and Fedora for this.
17:54:43 <limburgher> Get closer to the amp.
17:54:49 <mhroncok> packagers asked us for %has_python3 etc. macros
17:54:58 <tibbs> Sorr6y about last week, by the way.  I guess we still had a meeting but I've been too busy to look.
17:55:08 <tibbs> I will clear the pending writeups hopefully today.
17:55:31 <limburgher> mhroncok, I'd love to see spec examples where these would be used.
17:55:35 <tibbs> mhroncok: Anything that makes it easier to do things in epel is great.
17:55:43 <limburgher> tibbs, THIS
17:55:45 <mhroncok> limburgher: sure, I can add it
17:56:09 <geppetto> mhroncok: looks fine to me :)
17:56:17 <mhroncok> also, this is important for any hyphotetical downstreams of Fedora that might only have python3 but have some mainstainers who would want to share the specfile
17:56:22 <tibbs> But yeah, some idea of what that would look like, and what macro spaghetti it replaces would be great.
17:56:44 <tibbs> But I think I can already see most of it.
17:56:57 <mhroncok> one thing I didn't know
17:57:02 <limburgher> That sounds like giant pasta, like X1000.
17:57:03 <mhroncok> when one packages python-foo to epel
17:57:20 <mhroncok> is tehre a guideline that says it has to have python2 subpackage? or packaging py3 only is fine?
17:57:35 <mhroncok> (I mean when upstream supports both)
17:57:36 <tibbs> The latter.
17:57:57 <tibbs> EPEL has no specific guideline about python stuff, and the Fedora document says you MAY package for python2.
17:58:04 <mhroncok> ok
17:58:14 <tibbs> EPEL could certainly use a page about python stuff, though.
17:58:36 <tibbs> Also note that I have been adding a lot of stub python2-* packages to EPEL.
17:58:41 <mhroncok> well the python34 and python36 situation is certainly... messy
17:58:48 <mhroncok> tibbs: noted that
17:58:53 <tibbs> All they do is depend on python-* (or whatever the base package is called).
17:59:24 <tibbs> Current status (and the place to add your requests) is at https://fedoraproject.org/wiki/User:Tibbs/EPELPythonStubPackages
18:01:16 <tibbs> I think if you include what's currently done in buildroot overrides, the biggest problem packages are already taken care of.
18:01:54 <tibbs> Anyway, I think we're out of time.
18:01:59 * geppetto nods
18:02:04 <geppetto> #endmeeting