17:00:25 <geppetto> #startmeeting fpc
17:00:25 <zodbot> Meeting started Wed Jul 26 17:00:25 2017 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'
17:00:43 <tibbs> Too many channels.
17:00:48 <geppetto> Yeh
17:01:01 <geppetto> I hadn't realized I'd used meeting-2 in the calendar too
17:01:07 <geppetto> #chair tibbs
17:01:07 <zodbot> Current chairs: geppetto tibbs
17:01:20 <mbooth> Hi
17:01:31 <geppetto> #chair mbooth
17:01:31 <zodbot> Current chairs: geppetto mbooth tibbs
17:02:50 <mbooth> I'll just be in all the rooms all the time :-p
17:02:55 <geppetto> :)
17:04:09 <tibbs> I don't want my channel list to have to scroll.  Æsthetics are important.
17:04:39 <geppetto> ha
17:05:05 * mbooth has irc on his portrait monitor :-p
17:05:29 <tibbs> Me, too, along with two cameras, my music player and my browser.
17:05:50 <tibbs> But I have a 34" monitor in portrait mode.
17:05:56 <geppetto> two cameras?
17:06:00 <IgorGnatenko> .hello ignatenkobrain
17:06:01 <zodbot> IgorGnatenko: ignatenkobrain 'Igor Gnatenko' <ignatenko@redhat.com>
17:06:34 <Rathann> hi
17:06:35 <tibbs> I have security cameras all over the place here.
17:06:43 <geppetto> #chair Rathann
17:06:43 <zodbot> Current chairs: Rathann geppetto mbooth tibbs
17:09:13 <tibbs> limburgher said he would be here.
17:09:33 <geppetto> tibbs: so you switch between them on your camera outputs … like a ninja security guard ?
17:09:35 <geppetto> yeh
17:10:12 <geppetto> I've pinged him again
17:10:27 <tomspur> Ah, there it is
17:10:28 <tomspur> Hi
17:10:44 <tibbs> I just watch the space outside my office door and the nearest elevator lobby.
17:10:52 <tomspur> In the calender entry #fedora-meeting-2 is mentioned
17:10:59 <geppetto> #chair tomspur
17:10:59 <zodbot> Current chairs: Rathann geppetto mbooth tibbs tomspur
17:11:15 <Rathann> yay
17:11:28 <geppetto> yeh, I didn't realize I'd put different channels in the calendar and the emails
17:11:47 <Rathann> I was surprised we had -3 channel ;)
17:11:48 <geppetto> I guess we'll merge to -2 … for two weeks from now :)
17:12:29 <geppetto> #topic Schedule
17:12:31 <geppetto> https://lists.fedoraproject.org/archives/list/packaging@lists.fedoraproject.org/message/ITFHNHIPCXAEEMOF7OJRPQ3TXB3HFQ5I/
17:12:41 <geppetto> #topic #691 noarch *sub*packages with arch-specific dependencies
17:12:44 <geppetto> .fpc 691
17:12:46 <zodbot> geppetto: Issue #691: guidelines change: noarch *sub*packages with arch-specific dependencies - packaging-committee - Pagure - https://pagure.io/packaging-committee/issue/691
17:13:30 <geppetto> So, no changes since we talked about it last … anyone spoken to releng?
17:14:52 <Rathann> I don't think so
17:15:38 <geppetto> #topic #693 Wiki:Packaging:RPMMacros
17:15:43 <geppetto> .fpc 693
17:15:44 <zodbot> geppetto: Issue #693: Wiki:Packaging:RPMMacros - packaging-committee - Pagure - https://pagure.io/packaging-committee/issue/693
17:16:09 <geppetto> Ok, this one was on tibbs … did you manage anything with holidays etc?
17:16:18 <tibbs> I have basically done nothing at all, sadly.
17:16:27 * limburgher here, late, sorry, call ran over.
17:16:29 <tibbs> I did think about this some more.
17:16:43 <geppetto> #chair limburgher
17:16:43 <zodbot> Current chairs: Rathann geppetto limburgher mbooth tibbs tomspur
17:16:52 <tibbs> Thing is, we really do have to document some of these; I just don't know what needs to be in there.
17:16:56 <geppetto> limburgher: we waited for you, it's all good :)
17:17:06 <limburgher> w0000000h0000
17:17:42 <geppetto> tibbs: ok, want to leave it for next week or talk about some of it?
17:17:44 <tibbs> Seeing how the flatpack macros work by redefining many of these settings, I guess we do need to document the ones that people will be needing to use.
17:18:00 <Rathann> %{_prefix} for sure
17:18:01 <tibbs> The problem is, really, figuring out what to document and what to leave out.
17:18:20 <tibbs> Rathann: It's way more than just %_prefix.
17:18:47 <tibbs> I mean, just that one thing may be redefined, but that changes the definitions of other macros which people will be using.
17:20:00 <Rathann> indeed
17:20:14 <tibbs> But the rest of the work there is still the same: rip out the stuff that really doesn't need to be in there.  Rewrite the stuff that does.
17:21:08 <tibbs> Also, I really wish that we could have far more useful names for most of these.
17:21:37 <tibbs> They're so inconsistent.  Some have "dir" appended.  Some have "path" appended.
17:22:02 <geppetto> lol, good luck fixing any of that
17:22:10 <tibbs> Just saying use %_var %_usr %_lib %_bin is at least readable.
17:22:34 <tibbs> Some of those at least exist now; they're just de-emphasized.
17:22:42 * geppetto nods
17:22:47 <Rathann> right
17:22:49 <tibbs> It's part of the reason I just don't use them (except for %_lib).
17:23:17 <tibbs> Seeing the actual paths everywhere makes things so much more understandable.
17:23:36 <geppetto> yeh, it tends to be pretty ugly with the macros
17:23:52 <tibbs> Right, and poor first time packagers get crazy confused.
17:24:03 <geppetto> And so much of it is useless anyway, just one fix in rpm and 95%+ of it can go
17:24:13 <tibbs> Are you supposed to use them?  If several of them mean the same thing, which should I use?
17:24:24 <geppetto> whatever the reviewer wants ;)
17:24:43 <tibbs> I understand that some of it comes from autoconf but... there is plenty of stuff that doesn't use autoconf at all.
17:25:50 <tibbs> I did at least make a start at https://fedoraproject.org/wiki/User:Tibbs/RPMMacros
17:25:53 <tibbs> But it's just a start.
17:25:59 <geppetto> cool
17:26:54 <tibbs> So my plan is to stop emphasizing autoconf-relatedness and instead try to list the macros which are closest to the actual directory name.
17:27:06 <geppetto> Sounds fine to me
17:27:14 <Rathann> good plan
17:27:23 <tibbs> And if I grep the specfiles and can find only a few packages using one of the macros then I just won't list it.
17:27:36 <tibbs> Heh, %_initddir.....
17:27:50 <tibbs> Because we still use init.
17:28:10 <Rathann> move to EPEL:Packaging ;)
17:28:36 <tibbs> Can we just pretend %_localstatedir doesn't exist?
17:28:42 <Rathann> hehe
17:28:48 <Rathann> sure
17:28:50 <limburgher> But now I can't install proprietary-crap-from-2005-14.3-1.i686.rpm!!!!!
17:28:52 <tibbs> I mean, I type fast but "/var"....
17:29:22 <tibbs> Anyway, I have a few minutes today while I do another kernel build so I'll put a little more work into this.
17:29:51 <limburgher> On old systems I call initscripts with /etc/rc.d/init.d/FOO because god knows why.
17:29:56 <geppetto> ok, cool. Can come back to it next week
17:30:14 <geppetto> #topic #696 Many packages are not following the Guidelines for bundled libraries
17:30:19 <geppetto> .fpc 696
17:30:20 <zodbot> geppetto: Issue #696: Many packages are not following the Guidelines for bundled libraries - packaging-committee - Pagure - https://pagure.io/packaging-committee/issue/696
17:30:52 <tibbs> limburgher: Because habit.
17:31:20 <tibbs> So I did reply in the ticket.
17:31:36 <geppetto> Yeh, I agree with everything you said
17:31:58 <tibbs> I am really trying to avoid being passive-aggressive with this, though.
17:32:25 <Rathann> can we point to community-maintained wiki page for the list of bundled names that are not packaged in Fedora (and are probably never going to be)?
17:32:39 <tibbs> Does such a page exist?
17:33:18 <tibbs> Basically, I think we should add a note to attempt to follow the versioning and naming guidelines for determine <libname> and <version> in that guideline.
17:33:21 <tibbs> https://fedoraproject.org/wiki/Packaging:Guidelines#Bundling_and_Duplication_of_system_libraries
17:33:29 <Rathann> yes, I kept it when we deleted everything about bundling from Packaging:
17:33:31 <Rathann> https://fedoraproject.org/wiki/Bundled_Libraries_Virtual_Provides
17:34:05 <tibbs> But I'm not sure that "Provides: bundled(foo) = 0" is any better than "Provides: bundled(foo)".
17:34:46 <Rathann> "Version can be omitted if it cannot be determined with reasonable effort."
17:34:49 <tibbs> We could also add a note that it is better to use the version from which the bundled version was forked than to use nothing.
17:34:50 <Rathann> something like that?
17:35:15 <tibbs> Thing is, I don't know how these bundled(x) provides are actually being used.
17:35:38 <geppetto> I don't think they are being used by non-humans, are they?
17:35:40 <tibbs> What or who wants to use them?  What would actually help them?
17:35:54 <geppetto> I thought they were all just data for repoquery and nothing else
17:36:10 <tibbs> Sure, but this issue was raised for some reason.
17:36:31 <tibbs> I hope it was just more than someone doing a query, finding these things without versions and deciding to ask us to do work.
17:37:08 <geppetto> I think that's exactly what happened
17:37:09 <tibbs> I'd rather not make something up which turns out not to be useful for whatever wants to consume these things.
17:38:23 <tomspur> Can we agree on omitting it, when "it cannot be determined with reasonable effort"?
17:38:48 <geppetto> sure, at least I'm fine with that
17:39:02 <tomspur> If FESCO wants to have a 0 in that case, that might work too... up to FESCO
17:39:11 <geppetto> "= 0" might be better if some tools were looking at it … but AFAIK there are no tools
17:39:44 <ignatenkobrain> I would prefer to omit version at all
17:40:10 <ignatenkobrain> I don't see any reason why "= 0" is better than without it
17:40:31 <tomspur> +1
17:40:47 <tibbs> +1 to omitting when it can't be reasonably determined.
17:41:23 <tibbs> But I do believe that _some_ information is better than none.
17:41:42 <Rathann> I think we all agree on that
17:42:17 <sgallagh> tibbs: Perhaps with some leeway that if a *specific* version cannot be determined, an approximate version with a trailing .0 makes the most sense?
17:42:19 <tibbs> The version of the upstream source which was last merged.  Or if there has never been a re-merge of upstream, the version that upstream when the original fork happened.
17:42:37 <tibbs> sgallagh: Sadly 1.1.0 is strictly newer than just "1.1".
17:42:47 <sgallagh> e.g. if a package has forked some other package at 1.2.3, but it includes some backported packages from 1.2.6 and 1.2.7, we just call it bundled at 1.2 ?
17:42:57 <sgallagh> tibbs: Sorry, I phrased that poorly.
17:43:04 <tibbs> So adding a trailing zero would mean that looking for all foo <= 1.1 wouldn't find 1.1.0.
17:43:09 <sgallagh> I meant to go for "lowest common denominator" where reasonable
17:43:32 <limburgher> sgallagh: If it won't mess up sorting, yes, but won't it?
17:44:04 <sgallagh> Not sure. I was just trying to figure out how we could maximize information on a search while minimizing packager effort.
17:44:11 <sgallagh> If that's not a valid answer, so be it
17:44:35 <tibbs> sgallagh: If you last merged or forked version 1.1, then just saying that you bundle 1.1 should be quite good for all reasonable uses of this bundling information.
17:44:53 <sgallagh> Right, that's kind of what I was trying to say.
17:45:03 <sgallagh> Though clearly not phrased well :)
17:45:08 <tibbs> If someone finds a security bug in 1.1, that tells them that they need to at least look at your package to see if it has the bug.
17:45:21 <sgallagh> tibbs: Yeah, that was exactly where I was going with that
17:45:46 <tibbs> My understanding is that is what these bundled(x) tags are really for.
17:45:52 <sgallagh> yes
17:46:13 <tibbs> So just saying "the version that was last merged or forked" should be sufficient, yes?
17:46:48 <sgallagh> tibbs: Yes, I agree. That was exactly what I was trying (and failing) to communicate when I said "lowest common denominator")
17:47:06 <Rathann> tibbs: +1 to that
17:47:11 <geppetto> yeh
17:47:20 <tibbs> OK, that's what I'll use.
17:48:11 <geppetto> #info We discussed some wording based on tibbs's comment.
17:48:34 <geppetto> #action tibbs Write up a page for recommendations on provides.
17:48:50 <geppetto> Good?
17:48:50 <tibbs> Fortunately the section is short so this should be easy.
17:54:59 <geppetto> Ok
17:55:04 <geppetto> #topic Open Floor
17:55:18 <geppetto> Anything else anyone wants to talk about in the last 5 minutes?
17:55:36 <tibbs> Not from me.  Looks like I have stuff to do.
17:57:34 <limburgher> Not particularly.
17:58:31 <tibbs> Was that really it?
17:59:32 <geppetto> yeh
17:59:36 <geppetto> #endmeeting