16:00:34 <geppetto> #startmeeting fpc
16:00:34 <zodbot> Meeting started Thu Sep 24 16:00:34 2015 UTC.  The chair is geppetto. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:34 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:00:34 <geppetto> #meetingname fpc
16:00:34 <geppetto> #topic Roll Call
16:00:34 <zodbot> The meeting name has been set to 'fpc'
16:00:51 <orionp> Hello
16:01:12 <tomspur> Hi
16:01:19 <geppetto> #chair orionp
16:01:19 <zodbot> Current chairs: geppetto orionp
16:01:21 <geppetto> #chair tomspur
16:01:21 <zodbot> Current chairs: geppetto orionp tomspur
16:01:53 * RemiFedora here is question about the php tickets
16:01:58 <tibbs|w> Howdy.
16:02:02 <geppetto> #chair tibbs
16:02:02 <zodbot> Current chairs: geppetto orionp tibbs tomspur
16:02:38 <tibbs|w> Two weeks without quorum would not be good.
16:04:53 <geppetto> Matt said he couldn't make it
16:05:10 <geppetto> So that leaves Rathan and racor
16:05:34 <geppetto> limburgher racor Rathann SmootherFr0gZ: FPC ping
16:06:06 <geppetto> Rathann is on IRC
16:06:10 <Rathann> hi
16:06:12 <geppetto> #chair Rathann
16:06:12 <zodbot> Current chairs: Rathann geppetto orionp tibbs tomspur
16:06:13 <geppetto> hey
16:06:25 <geppetto> #topic Schedule
16:06:27 <geppetto> https://lists.fedoraproject.org/pipermail/packaging/2015-September/011020.html
16:06:43 <geppetto> #topic #572 	Mass review exception for php-zendframework
16:06:44 <geppetto> .fpc 572
16:06:44 <geppetto> https://fedorahosted.org/fpc/ticket/572
16:06:45 <zodbot> geppetto: #572 (Mass review exception for php-zendframework) – fpc - https://fedorahosted.org/fpc/ticket/572
16:06:54 <geppetto> Going to do this one first, as I think it's kind of trivial
16:07:42 <geppetto> Basically they had a single giant git repo. … and now have a bunch of smaller repos. … and don't want to rereview everything
16:08:08 <tibbs|w> +1 this is the whole point.
16:08:11 <geppetto> +1
16:08:19 <tibbs|w> (Whole point of the process.)
16:08:23 * geppetto nods
16:08:31 <geppetto> RemiFedora: You want/need to add anything?
16:08:39 <tibbs|w> And my comment about removing the cruft was merely a comment, not a condition.
16:09:00 <RemiFedora> just need to know what is next step, if approved, for SCM request
16:09:28 <tibbs|w> Well, three more +1's, first.
16:09:42 <RemiFedora> (as I said, if approved ;)
16:09:43 <tibbs|w> And then you'll get to be the first person to actually try this out.
16:09:55 <geppetto> oooh, lucky you :)
16:11:18 <RemiFedora> FYI, I think siwinski will own the stack with me
16:11:29 <tomspur> Isn't there a section in the guidelines about the new mass review exception? I cannot find it (and maybe I missed the meeting, where it was discussed)
16:12:42 <tibbs|w> tomspur: It's not in the guidelines, since it's not a guideline.
16:12:46 <tibbs|w> It's on the Committee page.
16:13:20 <tibbs|w> RemiFedora: What will end up happening is you'll open a bugzilla ticket (so that this is discoverable) and include a list of packages.
16:13:59 <Rathann> +1 from me
16:14:26 <tibbs|w> I'll create all of the repos from there and make sure that the necessary links to the policy and the meeting minutes are in there.  And then we'll be done.  This works as long as one FPC member has the necessary privileges.  Otherwise we'll have to point an admin at the ticket.
16:14:38 <orionp> I think the one thing I'd like to see is some kind of license and bundling audit for the whole collection - unless we're assuming 2.5 is very similar to 2.4
16:15:09 <geppetto> orionp: That was my assumption, and/or that RemiFedora would know if licensing had changed.
16:15:21 <tibbs|w> Yes, license audit has been my main concern about this.
16:15:41 <RemiFedora> yes, 2.5 is very similar to 2.4, not big change, and each compoenent have the same license that the old "big" one
16:15:47 <tomspur> +1 from me too. I trust RemiFedora to do the license and bundling audit just as in any other update.
16:16:29 <orionp> Okay, I'm +1 as this is an update
16:16:40 <geppetto> Ok, I believe that's +5
16:16:58 <geppetto> #action Mass review exception for php-zendframework (+1:5, 0:0, -1:0)
16:17:12 <geppetto> #topic #569 	texlive bundles lua
16:17:12 <geppetto> .fpc 569
16:17:12 <geppetto> https://fedorahosted.org/fpc/ticket/569
16:17:13 <zodbot> geppetto: #569 (texlive bundles lua) – fpc - https://fedorahosted.org/fpc/ticket/569
16:17:40 <geppetto> Spot not on IRC, oh well.
16:17:46 <orionp> I started looking into this last night.  I'd like to poke around a bit more
16:18:43 <orionp> There is also luajit in there too FWIW
16:18:46 <geppetto> I'm somewhat confused about the "texlive has probably been doing this since it was added to Fedora"
16:19:18 <geppetto> Was texlive added much more recently than I'd thought, or is lua much older than I'd thought?
16:19:48 <orionp> I can remember texlive replacing tetex
16:20:02 <orionp> but I've been around since before F1
16:20:31 <geppetto> I'm mostly willing to just trust spot, but this kind of feels like a bit of pushback upstream might stop the bundling :(
16:20:50 <orionp> http://www.luatex.org/
16:21:13 <orionp> I'd like to see some discussion with upstream
16:21:42 <orionp> but sure, it's in place, lets add the Provides: bundled, but keep poking
16:21:58 <geppetto> It looks like debian just ship it as a subpackage: https://packages.debian.org/sid/texlive-luatex
16:22:18 <orionp> so do we
16:22:33 <orionp> well, everything is a subpackage :)
16:22:53 <orionp> I'm curious as to who uses it
16:23:07 * tomspur too
16:23:37 <geppetto> You mean which TeX files use embeded lua?
16:23:53 * geppetto doesn't even know how you'd search for that
16:24:21 <tibbs|w> Well, we used to have tetex.
16:24:22 <tibbs|w> I'd definitely at least like to know exactly where it's bundled and have some indications of the modifications and such.
16:24:23 <tibbs|w> Knowing texlive there are probably several different bundled copies.  And probably more bundled stuff throughout the thing.
16:24:50 <tibbs|w> debian code search.
16:24:56 <orionp> It actually doesn't seem that bad, but there are a couple patches
16:25:06 <tibbs|w> Oh, I mean which actual executables are using bundled stuff.
16:25:28 <geppetto> from a cursory look, it seems more like there is a LuaLaTeX thing, which is TeX and lua … and that's it.
16:25:33 <geppetto> So fairly well contained
16:26:32 <orionp> but some random texlive packages seem to require it
16:27:14 <geppetto> Something not under texlive-*lua* ?
16:27:38 * orionp is looking
16:28:53 <orionp> ah maybe not - it's confusing that there are syles like "luabidi.sty", but they may not need luatex
16:29:33 <orionp> but luatex is brought in in a standard install
16:29:42 <orionp> at least it's on my machine
16:32:08 <orionp> So again, I'd like to do a bit more research, and I'm hoping I have time to help
16:32:15 <geppetto> ok
16:32:23 <geppetto> So just push it out one week?
16:32:49 <orionp> sure
16:33:07 <geppetto> Ok
16:33:09 <geppetto> #topic #570 	Allocating a soft static uid and gid for sssd
16:33:09 <geppetto> .fpc 570
16:33:10 <geppetto> https://fedorahosted.org/fpc/ticket/570
16:33:13 <zodbot> geppetto: #570 (Allocating a soft static uid and gid for sssd) – fpc - https://fedorahosted.org/fpc/ticket/570
16:35:42 <geppetto> This feels kind of weird
16:36:13 <geppetto> On the one hand sssd is maybe core enough that just giving it a uid might be fine.
16:36:16 <orionp> So, do we really have no way of preallocating UIDs in containers?
16:36:57 <geppetto> On the other, I'm not sure how Atomic is special here … and Atomic has a options to inject passwd entries into it's builds … not sure about container creation though.
16:37:40 <orionp> Seems like that would be basic functionality, but have never played with any of this
16:37:56 <geppetto> Yeh, you'd think that was needed.
16:38:27 <geppetto> But from the bits I've played with docker … it kind of assumes something higher does that … and yum --installroot does nothing.
16:38:40 <geppetto> So I'd guess no.
16:39:22 <tibbs|w> My concern is only for the limited UID space.
16:39:52 <tibbs|w> I think since sssd will basically be on every system, there's no point in not allocating an ID for it since it's going to take one anyway.
16:40:25 <orionp> yeah, that seems a clincher
16:40:42 <orionp> So, I'm +1
16:40:44 * geppetto nods … I'm happy enough to +1 on those grounds
16:41:02 <tibbs|w> +1
16:41:03 <tomspur> +1 also
16:41:18 <geppetto> +1
16:41:31 <geppetto> Rathann: vote?
16:41:46 <Rathann> +1
16:43:09 <geppetto> #action Allow allocation of soft static UID/GID for sssd, due to always being installed (+1:5, 0:0, -1:0)
16:43:22 <geppetto> #topic #571 	sabre/dav bundle exception in Horde
16:43:23 <geppetto> .fpc 571
16:43:23 <geppetto> https://fedorahosted.org/fpc/ticket/571
16:43:25 <zodbot> geppetto: #571 (sabre/dav bundle exception in Horde) – fpc - https://fedorahosted.org/fpc/ticket/571
16:43:46 <geppetto> RemiFedora: This is you again, yeh?
16:43:52 <RemiFedora> yes
16:43:57 <tibbs|w> +1 for a temporary exception here, though I guess it would be nice to know how long "temporary" is.
16:44:07 <RemiFedora> (all those things related to owncould 8.1)
16:44:17 <tibbs|w> And yeah, I know that depends on Horde upstream, which is kind of tough to predict.
16:45:00 <geppetto> RemiFedora: You have any idea how long until 2.x can be used?
16:45:37 <RemiFedora> No really, upstream just have confirmed they plan to do the update for Horde 6 (as it raise dependency on PHP)
16:46:12 <tibbs|w> Ugh; I just managed to roll our Horde 5....
16:47:00 <Rathann> what's composer?
16:47:05 <geppetto> So tmp. upto and including Fedora-26?
16:47:11 <RemiFedora> notice, Horde_Dav with bundle sabre/dav library just install everything like the old php-sabredav* package
16:47:14 * geppetto picks number out of the air
16:47:14 <tibbs|w> Rathann: A replacement for pear, sort of.
16:47:21 <Rathann> ah
16:47:24 <Rathann> I see
16:47:28 <tibbs|w> PHP seems to like to invent these things, I guess.
16:47:38 <RemiFedora> so, this is mostly a rename/obsolete of the previous pear package
16:47:58 <orionp> I'm +1
16:48:01 <tibbs|w> That's how I'm seeing it, yes.
16:48:18 <RemiFedora> Rathann, composer is a "bundle everything factory" ;)
16:48:39 <Rathann> yuck
16:48:51 <Rathann> anyway, +1 from me
16:48:57 <orionp> Although perhaps by then owncloud 10 will be using saber 3.0
16:49:37 <tomspur> +1
16:49:53 <geppetto> Sure +1
16:51:26 <tibbs|w> I'm still +1.
16:51:35 <geppetto> Ok
16:51:37 <geppetto> #action Temporary bundling exception for sabre/dav in php-horde-Horde-Dav until version 2.x can be used. (+5, 0:0, -1:0)
16:51:52 <RemiFedora> thanks
16:51:54 <tibbs|w> RemiFedora: How much parallel installability is there for PHP libs/modules?
16:52:10 <tibbs|w> I've been thinking about parallel installs of things a lot lately.
16:52:19 <geppetto> #topic #567 	Packaging Python 3 applications and modules for EPEL 7+
16:52:19 <geppetto> .fpc 567
16:52:19 <geppetto> https://fedorahosted.org/fpc/ticket/567
16:52:20 <zodbot> geppetto: #567 (Packaging Python 3 applications and modules for EPEL 7+) – fpc - https://fedorahosted.org/fpc/ticket/567
16:52:39 <tibbs|w> I haven't had time to look at the macros much or play with this at all, which is sad.
16:52:46 <RemiFedora> tibbs|w, if we rely on default include_path : 2 (in /usr/share/pear, and /usr/share/php), but else as many as we want
16:53:58 <geppetto> ok
16:54:04 <tibbs|w> I've finished one big project and am almost done with another here at the office.
16:54:12 * geppetto nods
16:54:14 <geppetto> #topic #566 	RPM file triggers
16:54:14 <geppetto> .fpc 566
16:54:14 <geppetto> https://fedorahosted.org/fpc/ticket/566
16:54:15 <zodbot> geppetto: #566 (RPM file triggers) – fpc - https://fedorahosted.org/fpc/ticket/566
16:54:51 <geppetto> There appears to have been a bunch of work here … there's now around 10 packages in rawhide that contain %filetrigger scripts
16:54:57 <geppetto> Mostly desktop related.
16:55:51 <tibbs|w> That's great.  Do we have a list?
16:56:14 <tibbs|w> glibc would be the big one, I think.
16:56:15 <geppetto> I generated one … alas. I didn't save it and it takes a couple of minutes to generate
16:56:21 <tibbs|w> Ouch.
16:58:49 <tibbs|w> Anyway, progress.  I'd just paste the list into the ticket once you have it and I can work on noting this in the guidelines.
16:59:26 <geppetto> almost done
16:59:57 <tibbs|w> For fun, I estimate that about 15% of the texlive spec would disappear if it grew the triggers.
17:00:05 <tibbs|w> Might even be 20%.
17:00:45 <geppetto> http://paste.fedoraproject.org/271129/11404114/
17:00:56 <geppetto> cool
17:00:57 <tibbs|w> spot didn't want to wade in that far, though.  I guess it just needs some testing.
17:01:03 * Rathann notes we're on one hour mark
17:01:07 * geppetto nods
17:01:14 <geppetto> #topic Open Floor
17:01:32 <geppetto> Anything anyone wants to talk about?
17:01:38 <tibbs|w> Haven't filed a ticket on it, but does anyone have any opinions on banning /usr/bin/env in python packages?
17:01:45 <Rathann> all these triggers look reasonable
17:01:47 <tibbs|w> I kind of thought that was already the case.
17:02:10 <geppetto> tibbs: I'm 99% sure that's banned due to security
17:02:27 <tibbs|w> Well, it doesn't appear to actually be banned.
17:02:38 <geppetto> Maybe just banned in RHEL?
17:02:49 * geppetto is sure he's had bugs about it in the past.
17:03:00 <tibbs|w> At least not explicitly.  One sentence near the beginning of the python guidelines and maybe a rpmlint check would be enough.
17:03:40 <tibbs|w> Also, for fun just cram a script or link to /bin/false in ~/bin/python and see what breaks.
17:03:51 * geppetto nods … I'm happy to +1 that
17:04:22 <tibbs|w> I'm +1 as well; there are so many reasons not to use env.
17:04:31 <Rathann> https://fedoraproject.org/wiki/Features/SystemPythonExecutablesUseSystemPython
17:04:37 <Rathann> this was a feature in F16
17:04:51 <tibbs|w> proposed feature.
17:04:58 <tibbs|w> Percentage of completion: 0%
17:05:05 <tibbs|w> Never accepted by FESCo in any case.
17:05:06 <Rathann> yes
17:06:14 <geppetto> Hmm
17:06:56 <geppetto> Well, if you propose something for next week we can +1 it and do it that way :)
17:07:19 <tibbs|w> "Executable python scripts must directly call the appropriate python binary instead of using /usr/bin/env to locate one."
17:07:28 <tomspur> :)
17:07:39 <geppetto> Is that the proposal, where would it go?
17:07:51 <tibbs|w> In its own section at the very beginning, I think.
17:08:22 <Rathann> s/call/specify/
17:08:22 <tibbs|w> What to name the section, though?
17:08:40 <tibbs|w> Rathann: Indeed, that makes more sense.
17:09:08 <tibbs|w> I can open a ticket if we'd like.  Or I could just write this up as soon as I can find to write everything else up.
17:09:22 <Rathann> and if they work equally well with python3 and python2, they should use python3, right?
17:09:35 <tibbs|w> Yes, though that should already be in the guidelines.
17:11:53 <RemiFedora> notice: /usr/bin/env make sense in some other language
17:11:57 <tibbs|w> I'll just open a ticket and make sure that the "py3 by default" think is in there as well.
17:12:02 <tibbs|w> RemiFedora: Maybe, but this is python.
17:12:08 <Rathann> maybe call the section "Python interpreter in scripts" or something like that
17:12:33 <tibbs|w> And, also, if it makes sense in some other language then how on earth does it avoid brokenness with stuff in my ~/bin?
17:12:52 <RemiFedora> tibbs|w, but just a side note to avoid a strict rule to refuse its usage "globally"
17:12:56 <tibbs|w> I just can't see how that would work, and why that would ever be allowed.  But that's a different discussion, since this is python specific.
17:12:58 <tomspur> How about "specifying the python binary it was built with", so we can rely on py3 by default from the other section and don't say it twice?
17:13:21 <tibbs|w> tomspur: I would probably just reorganize that bit to avoid stating it twice.
17:13:28 <tibbs|w> I'll open a ticket with a draft.
17:13:51 <tibbs|w> BTW, just edited the python guideline page to fix a typo.
17:14:51 <Rathann> tomspur: not sure what you mean
17:15:00 <geppetto> Ok, you want to vote on it next week or wait a few mins. and do it this?
17:15:15 <tibbs|w> Next week is good.
17:15:41 <geppetto> ok
17:15:52 <geppetto> If there's nothing else then I'll close in a couple of minutes.
17:15:56 <tomspur> Rathann: When you build the package with "%{__python} setup.py ..." the script should use %{__python}
17:16:08 <tibbs|w> Yes, that happens automatically for setuptools stuff.
17:16:21 <tibbs|w> But some stuff doesn't use setuptools, so....
17:16:29 * tomspur nods
17:16:49 <tibbs|w> Anyway, did folks see my bad draft about the applications/libraries distinction and when to split packages?
17:17:17 <tibbs|w> It's too long but I'm lost as to how to say all of that without the verbiage.
17:17:47 <tibbs|w> The "draft" is https://fedoraproject.org/wiki/User:Tibbs/AppsVsLibs
17:18:06 <tibbs|w> It's in 558 which by the title seems completely unrelated.
17:18:08 <geppetto> 558?
17:18:12 * geppetto nods
17:18:19 <tibbs|w> .fpc 558
17:18:22 <Rathann> yes, it's not too bad
17:18:22 <zodbot> tibbs|w: #558 (Switch order of install macros) – fpc - https://fedorahosted.org/fpc/ticket/558
17:18:59 <orionp> Also, FWIW I think we're running into trouble not having the latest python macros in earlier versions of Fedora
17:19:06 <tibbs|w> I think I'm saying the right things there but need to make a couple more passes.
17:19:24 <tibbs|w> orionp: I thought they were in there, but the situation has gotten far too complex.
17:20:55 <orionp> indeed
17:22:09 <tibbs|w> tomspur would probably more knowledge about why F21 python doesn't have those.
17:22:27 <tibbs|w> But we really, really need to split them out.  I'm just not in the loop on that.
17:23:57 * tomspur thought there would be an update
17:24:21 <tomspur> searching in new bodhi for python is not a good idea
17:24:36 <orionp> :)
17:25:07 <orionp> I know python3 is being blocked due to a test failure, I think in F21
17:25:47 <tibbs|w> Aargh.
17:26:45 <tomspur> python2 is stable (I hope that are the latest python macros with all fixes): https://bodhi.fedoraproject.org/updates/FEDORA-2015-11384
17:27:08 <tomspur> Lately, I had way to much other stuff to do and might have missed some stuff... :/
17:27:45 <orionp> Doesn't look up to date - missing the obsoletes, epoch, etc. in python_provide
17:28:04 <tomspur> :(
17:28:06 <orionp> same in F22
17:28:31 <orionp> I pushed stuff to F23, but didn't feel comfortable doing F21/22
17:28:43 <tibbs|w> OK, then at least the three of us need to get on the python sig list and work this damn thing out.
17:29:02 * tomspur will have a look at the python2 stuff later on then...
17:30:57 <tibbs|w> Anyway, I'm done.
17:31:03 * geppetto nods
17:31:08 <geppetto> Anyone else have anything?
17:33:11 <geppetto> #endmeeting