fpc
LOGS
16:59:29 <abadger1999> #startmeeting fpc
16:59:29 <zodbot> Meeting started Thu Jan  9 16:59:29 2014 UTC.  The chair is abadger1999. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:59:29 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:59:35 <abadger1999> #topic Roll Call
16:59:39 <abadger1999> who's here?
17:00:23 <abadger1999> #meetingname fpc
17:00:23 <zodbot> The meeting name has been set to 'fpc'
17:01:13 * RemiFedora here - Happy 2014 !
17:01:35 * Rathann here
17:02:31 <Rombobeorn> Hello!
17:02:33 <Rombobeorn> As usual ticket 374 isn't on the agenda, but that doesn't seem to stop you guys from discussing it.
17:05:07 * geppetto is here
17:05:47 <abadger1999> #chair RemiFedora Rathann geppetto
17:05:47 <zodbot> Current chairs: Rathann RemiFedora abadger1999 geppetto
17:06:12 <abadger1999> ping limburgher, tibbs, racor, SmootherFrOgZ  FPC meeting time
17:06:46 <limburgher> Here
17:07:37 <abadger1999> #chair limburgher
17:07:37 <zodbot> Current chairs: Rathann RemiFedora abadger1999 geppetto limburgher
17:07:44 <abadger1999> Well, that's enough for quorum.
17:07:57 <abadger1999> #topic #339     software collections in Fedora
17:08:03 <abadger1999> https://fedorahosted.org/fpc/ticket/339
17:08:13 <abadger1999> I tried to work on this over vacation, honest :-)
17:08:36 <abadger1999> I got a sample template done, a few minor renaming of terms.
17:09:01 <geppetto> Rombobeorn: You can ask me to add things to the agenda, and/or reply to the email to the lists.
17:09:22 <abadger1999> https://fedoraproject.org/w/index.php?title=User%3AToshio%2FSCL_Guidelines_%28draft%29&diff=366334&oldid=364381
17:09:39 <geppetto> Rombobeorn: For some reason it's not in the list of tickets from report 12 … which is weird.
17:10:07 <abadger1999> and https://fedoraproject.org/wiki/User:Toshio/SCL_Request_Template
17:10:43 <abadger1999> View the source for the SCL Request Template just like you would for the Fedora Change Template to see instructions for filling it out.
17:11:18 <abadger1999> What I tried to do over vacation but failed at was making an actual useful SCL.
17:11:21 <geppetto> Rombobeorn: Ahh, because it's status=writeup now.
17:11:39 <abadger1999> I tried to take the python2.4 from RHEL5 and turn it into an SCL for fedora.
17:11:39 <geppetto> abadger1999: Was it really hard, or did you just not get around to trying?
17:12:17 <tibbsphone> Hey folks; I am waiting for someone to be towed out of my parking spot.
17:12:26 <abadger1999> It turns out that SCLs are not a magic bullet for making old code run on newer libraries ;-/
17:12:32 <abadger1999> #chair tibbsphone
17:12:32 <zodbot> Current chairs: Rathann RemiFedora abadger1999 geppetto limburgher tibbsphone
17:12:44 <geppetto> abadger1999: ha :)
17:13:02 <Rombobeorn> geppetto, I've seen that some tickets have a status of "meeting", but I seem to be unable to set that status.
17:13:26 <limburgher> Yeah, building Python 2.5 RPMs on EL-6 is challenging.  Like, I've never been able to do it.
17:13:47 <abadger1999> So yeah -- That might be a little ambitious... I guess since this is just to test the packaging guidelines I'll try to make a python2.7 or python3 SCL instead.
17:14:01 <geppetto> Rombobeorn: That's a keyword, not the status … and yeh, that gets set sometimes by someone on the FPC.
17:14:29 <geppetto> abadger1999: *nods* … something that has a chance of building would probably be a better start :)
17:14:39 <abadger1999> limburgher: <nod>  And getting it to run on Fedora turns out to be worse... Gave up last night when I found that db4.8 has a different API than db4.7  (they made things that were mentioned as private actually be private).
17:15:01 <limburgher> abadger1999: Neeto.
17:15:02 <abadger1999> So that's the status there -- feel free to look over the template and give feedback.
17:15:07 <abadger1999> We can vote on that next week.
17:15:17 <RemiFedora> the template seems fine
17:15:28 <abadger1999> #topic #358     Please make some autotools guidelines.
17:15:32 <abadger1999> https://fedorahosted.org/fpc/ticket/358
17:16:01 <abadger1999> As mentioned i nthe agenda, defering this as no one's rounded up time to work on a draft.
17:16:08 <abadger1999> #topic #369     Guidance on dealing with the bundled libev in perl-EV
17:16:14 <abadger1999> https://fedorahosted.org/fpc/ticket/369
17:17:12 <abadger1999> So...I would not want to consider this a coopylib.
17:17:26 <abadger1999> -source is preferable to bundling.
17:17:44 <abadger1999> using system libraries is preferable to -source
17:18:07 <racor> abadger1999: ACK.
17:18:20 <abadger1999> so the question in my mind is whether we can get system libs working or have to go to the second best (-source in this case)
17:19:34 <RemiFedora> -source seems acceptable
17:19:51 <geppetto> From the comments it seems like upstream is pro. bundling, so using the system libs. would require a lot of work.
17:19:59 <abadger1999> So normally -- we ask that maintainers try to get changes merged to upstream or consider forking upstream.
17:20:02 <geppetto> So I guess we go -source route.
17:20:13 <abadger1999> For instance: https://bugzilla.redhat.com/show_bug.cgi?id=1049554#c2
17:20:32 <RemiFedora> I think there is also a php-ev (not in fedora), which also bundled this...
17:20:57 <abadger1999> but I think demanding that maintainers fork upstream is a grey area where we have decisions which are on either side of the line.
17:21:52 <Rathann> php-ev description says it provides an interface to the libev library... so on the surface, there's no reason to let it bundle
17:22:20 <abadger1999> -source is akin to static linking so it's the next compromise position if we can't get changes into (a new) upstrema.
17:22:27 <Rathann> bleh, it actually says libev is included with that extension
17:22:31 * geppetto nods
17:23:51 * SmootherFrOgZ is here now (sorry for the late)
17:27:13 <RemiFedora> problem with libev is the watchr structs which is design to be overrided at build time
17:27:23 <RemiFedora> https://bitbucket.org/osmanov/pecl-ev/src/30589504e7cbe14445053682e42c07f7d37f677d/embed.h?at=master
17:28:01 <abadger1999> Proposal: mention that copylibs would not be an option here.  say that we need to know why upstream doesn't want to merge changes to make this runtime changable
17:29:09 <geppetto> I don't mind waiting for that, but I suspect the answer is going to be "because he doesn't want to"
17:29:16 <abadger1999> The question would still be -- if upstream doesn't have a good reason to keep that from being runtime settable, what are we comfortable with.
17:29:23 <abadger1999> geppetto: yeah.
17:31:22 <abadger1999> RemiFedora: ouch, that would ned to be un-embedded and might be hard to disentangle.
17:31:42 <tibbs|w> Are we doing libev now or still on SCLs?
17:31:50 <geppetto> libev
17:31:50 <Rathann> tibbs: libev
17:31:50 <abadger1999> So we could vote on whether to allow -source now or vote on whether to allow -source now if the upstream refuses to make it runtime settable.
17:32:19 <geppetto> I guess I'm a reluctant +1 on allowing -source.
17:32:22 <tibbs|w> OK, so /topic is not syncing again or something.  Sorry, finally got into my parking space and to the office.
17:32:47 <geppetto> abadger1999: My guess is the leading space screws it up?
17:32:58 <abadger1999> oh, blah
17:33:04 <abadger1999> #topic #369     Guidance on dealing with the bundled libev in perl-EV
17:33:09 <abadger1999> geppetto: yep, you're right.
17:33:34 <SmootherFrOgZ> +1 on allow -source if upstream refuses
17:35:06 <abadger1999> Proposal: Allow libev build -source package and packages which bundle libev presently need to switch to using the code from the -source package instead.  Follow the static library guidelines for making sure that the packages which use -source stay up to date with libev changes.
17:35:21 <abadger1999> Reluctant +1
17:35:33 <RemiFedora> +1
17:35:49 <Rathann> +1 if upstream refuses
17:35:53 <abadger1999> I'll explain about the hierachy of preferences so that people know why copylibs doesn't cover this as well.
17:36:38 <tibbs|w> +1 I guess.  I still don't get why this library needs to be so problematic, though.
17:36:59 <geppetto> reluctant +1
17:37:27 <racor> +1
17:37:32 <abadger1999> Rathann: k.  To accomodate your vote, I'll ask the package maintainer to post why upstream doesn't want to merge changes to make things runtime settable instead of buildtime settable.
17:38:07 <abadger1999> Rathann: (Which could be... "I just don't want to")  Is that acceptable to you?
17:38:22 <Rathann> yes
17:38:38 <racor> abadger1999: I am asking myself, why the Fedora maintainer is so reluctant on change the code himself.
17:40:31 <abadger1999> racor: It would be a fork.  I don't think all the places where libev is consumed have been evaluated to see what would need to be made runtime settable are.  I think the package maintainer is only interested in the apps using libev; libev itself is code that he isn't familiar with.
17:41:51 <racor> abadger1999: That's the normal job as package maintainer :)
17:41:56 <abadger1999> #info Proposal to allow libev-source (with conditions -- see summary in ticket) approved. (+1:6, 0:0, -1:0)
17:42:08 <abadger1999> #topic #375     Update to the Addon Packages naming guidelines for Python 3 Modules
17:42:13 <abadger1999> https://fedorahosted.org/fpc/ticket/375
17:42:47 <SmootherFrOgZ> abadger1999: actually, it's +1:7
17:42:54 <geppetto> abadger1999: spaces :-o
17:43:02 <abadger1999> #undo
17:43:02 <zodbot> Removing item from minutes: <MeetBot.items.Link object at 0x2e81da10>
17:43:09 <geppetto> abadger1999: You want me to format the agenda differently?
17:43:13 <abadger1999> #info Proposal to allow libev-source (with conditions -- see summary in ticket) approved. (+1:7, 0:0, -1:0)
17:43:22 <abadger1999> #topic #375     Update to the Addon Packages naming guidelines for Python 3 Modules
17:43:30 <abadger1999> whitespace will be the death of me
17:43:41 <geppetto> :)
17:43:51 <abadger1999> geppetto: nah, it's just the way I'm cutting and pasting
17:44:09 <abadger1999> https://fedorahosted.org/fpc/ticket/375
17:45:44 <tibbs|w> I'm having trouble seeing this as anything other than rules lawyering.
17:45:52 <abadger1999> Let's see... I think  for this one we wanted to make clear that new packages need to follow python-/python3-NAME
17:46:30 <Rathann> not python2/3-NAME?
17:46:53 <abadger1999> but also make clear that packages which already exist and are NAME-python need to be NAME-python3
17:47:08 <Rathann> and python-NAME with python2/3-NAME subpackages if it supports both
17:47:14 <abadger1999> Rathann: well... that's coming up in bkabrda's proposal.
17:48:47 <abadger1999> Proposal: Treat this as a clarification to mean new packages need to follow python-/python3-NAME and old packages that are named NAME-python need to have NAME-python3 counterparts
17:49:08 <abadger1999> As a clarificatoin, any of us can write that up once we get the 'round tuits
17:49:29 <abadger1999> (and it might be moot after we discuss and vote on bkabrda's proposal)
17:51:54 <abadger1999> Does the proposal work for everyone?
17:51:55 <tibbs|w> I guess that's a good idea in any case, but I'm not sure it will satisfy the original request.
17:53:34 * RemiFedora can't find where confusion is. Current guidelines seems clear about "a now removed exception"
17:54:02 <abadger1999> yeah -- I think it's just the examples that he wants clarified.
17:55:01 <tibbs|w> I don't know; there's that wall of text but nothing actually saying what he'd like changed.
17:55:50 <tibbs|w> So, yeah, clarify the examples as much as possible but don't actually change any guidelines unless someone tosses in a proposal for what they actually want changed.
17:58:20 <abadger1999> Okay.
17:58:29 <abadger1999> I guess that's where we'll leave that then.
17:58:54 <tibbs|w> I'm sure we'll get complaints (and still no draft) if that wasn't what was being requested.
17:59:06 <abadger1999> #topic 374 Ada guidelines changes for Comfignat and runpaths
17:59:11 <abadger1999> https://fedorahosted.org/fpc/ticket/374
17:59:30 <Rombobeorn> Running check-rpaths from %check seems to work well, so I agree to make it mandatory for packages that use GNAT_add_rpath.
17:59:36 <Rombobeorn> I also propose to recommend it for other Ada packages, since the risk of accidental runpaths is fairly high in Ada packages for several concurrent reasons.
18:00:30 <tibbs|w> I have no problems with this.  I've kind of wondered why we don't just run check-rpaths on every build.
18:00:47 <abadger1999> +1 to the latest proposal
18:01:08 <limburgher> tibbs|w:  Indeed.
18:01:15 <abadger1999> tibbs|w: I'm not sure if check-rpaths would trigger on private libraries.
18:01:59 <tibbs|w> Off topic, yes, but I'd say that you can disable it if it gives false positives.  But that's a separate discussion.
18:02:03 <tibbs|w> +1 to this proposal.
18:02:13 <limburgher> +1 as well.
18:03:03 <SmootherFrOgZ> +1 from me.
18:03:39 <racor> what are we voting on? https://fedoraproject.org/w/index.php?title=User%3ARombobeorn%2FAda_Guidelines_Changes_2&diff=365584&oldid=365581 ?
18:03:48 <racor> In that case, my vote -1
18:04:26 <abadger1999> racor: The changes highlight what's changed from his original proposal.
18:05:34 <Rombobeorn> No, the changes are relative to the current guideline page, after the Comfignat part was added.
18:05:36 <racor> abadger1999: the later 2 change block to me read as card-blanche to allow rpaths in Ada.
18:05:51 <abadger1999> Rombobeorn: ah -- now I see i nthe history where you synced to the current guidelines.
18:07:22 <geppetto> I'm +1
18:07:37 <abadger1999> That's +5
18:07:39 <geppetto> racor: You arne't fine with it even with the "must use check-rpath"?
18:08:00 <Rathann> +1 from me
18:08:09 <abadger1999> #info Latest GNAT_add_rpath changes approved (+1:6, 0:0, -1:1)
18:08:09 <Rombobeorn> Then you need to read more carefully, racor. The proposal is to require check-rpaths precisely to AVOID runpaths in installed binaries.
18:08:38 * RemiFedora cas confused by racor -1
18:09:09 <abadger1999> #topic #377     tmpfiles.d guidelines update
18:09:11 <Rombobeorn> Thanks guys! Please also fix the list structure as shown at https://fedoraproject.org/w/index.php?title=User%3ARombobeorn%2FAda_Guidelines_Changes_2&diff=365584&oldid=365581 as the current page has a broken XML structure.
18:09:40 <abadger1999> Rombobeorn: Thanks.  I'll do my best to remember to do that.
18:09:45 <abadger1999> :-)
18:09:56 <abadger1999> https://fedorahosted.org/fpc/ticket/377
18:09:56 <racor> My vote: -1, the changes are not-selfexplantory and will be subject to misunderstandings.
18:10:20 * abadger1999 runs over to see if his dryer is about to catch fire.  brb
18:11:43 <limburgher> I'd rather have file permissions specified in RPMs, no more error prone than having to run another tool.
18:13:21 <Rathann> sorry, I have to leave now
18:14:25 <RemiFedora> in all case, the file must be list in rpm to be correctly own
18:14:52 <RemiFedora> so if rpm don't have the file it need to %ghost it... just awfull.
18:15:06 <RemiFedora> so I agree with rdieter comment
18:22:50 <geppetto> Yeh, so this is confusing
18:22:54 <abadger1999> Hmm...
18:22:55 <geppetto> But I'm pretty sure I'm -1
18:23:08 <geppetto> I think this bit:
18:23:10 <geppetto> 'Yes, that's the way it's done currently. It just feels error-prone to have to duplicate this list of files in two places. '
18:23:16 <abadger1999> Okay I've just read through those bugs and don't see how the proposed guideline update addresses them.
18:23:36 <geppetto> Basically means "yeh, it's annoying to have to put the data in rpm … much more fun to not have it in rpm so that rpm -V etc. doesn't work"
18:23:52 <limburgher> geppetto: <nods quickly>
18:24:15 <tibbs|w> I'm not a fan.
18:24:44 <limburgher> Neither am I.
18:24:52 <tibbs|w> Plus requiring a scriptlet to do something that doesn't now require one seems a step backwards.
18:25:53 <RemiFedora> and the will be error-prone for dir/file ownership, so -1
18:26:28 <abadger1999> tibbs|w: +1
18:27:37 <tibbs|w> -1 for the record.
18:27:51 <limburgher> -1 while we're at it.
18:27:58 <abadger1999> -1
18:28:31 <SmootherFrOgZ> is also -1
18:28:31 <abadger1999> with two more -1's, this would be definitively rejected (instead of failing for now due to lack of 5 +1s)
18:28:32 * RemiFedora will even think to add a "files create by tmpfile must be own by the package"
18:28:47 <racor> -1
18:29:15 <racor> my time is up, i've got to quit, bye.
18:29:18 <abadger1999> geppetto: You still -1?
18:29:22 <abadger1999> racor: Thanks for coming
18:30:53 <abadger1999> #info Proposal to use systemd-tmpfiles --create and not have packages specify the directories under /var/lock in rpm's %files rejected.  Vote to approve the proposal was (+1:0, 0:0, -1:5)
18:31:04 <abadger1999> #topic Open Floor
18:31:20 <abadger1999> That's everything from the agenda.  Anyone have other tickets/issues/problems to bring up?
18:31:51 <tibbs|w> I may actually have more time to get involved again soon.
18:31:53 <geppetto> abadger1999: yeh
18:32:01 <limburgher> I have a lot of meetings at $DAYJOB today, can someone assist?
18:32:05 <abadger1999> tibbs|w: Excellent!
18:32:17 <abadger1999> limburgher: Sure.  I'll trade you mine for yours ;-)
18:32:23 <geppetto> abadger1999: There were two more
18:32:29 <geppetto> abadger1999: 379 and 380
18:32:31 <tibbs|w> The kid is off to the dorms on Saturday and then my life can start to settle down.
18:32:36 <geppetto> abadger1999: But we can leave them for next week, I think :)
18:33:01 <abadger1999> tibbs|w: congratulations to you and to the kid on reaching that milestone!
18:33:08 <limburgher> abadger1999:  You don't know where I work or what they're about.  Be very, very careful about making that offer.
18:33:21 <limburgher> tibbs|w:  Wow.  Cool!
18:33:31 <abadger1999> limburgher: Be very very careful about letting someone ignorant of everything attend meetings in your place ;-)
18:33:46 <limburgher> abadger1999:  :)
18:33:53 <abadger1999> "Yeah, limburgher would be *happy* to do that by tomorrow"
18:33:57 <abadger1999> ;-)
18:34:09 * limburgher gets out large mallet
18:34:36 <tibbs|w> Of course, it's the dorm at the university where I work, so I'm not completely free or anything, but at least some normalcy can return to my home life.
18:35:27 <abadger1999> Okay, if nothing else, I'll close in 60s
18:36:17 <tibbs|w> Nothing from me.
18:36:47 <abadger1999> #endmeeting