18:00:47 <sgallagh> #startmeeting FESCO (2015-10-07)
18:00:47 <zodbot> Meeting started Wed Oct  7 18:00:47 2015 UTC.  The chair is sgallagh. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:00:47 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
18:00:47 <sgallagh> #meetingname fesco
18:00:47 <zodbot> The meeting name has been set to 'fesco'
18:00:47 <sgallagh> #chair ajax dgilmore hguemar jwb nirik paragan rishi thozza sgallagh
18:00:47 <sgallagh> #topic init process
18:00:47 <zodbot> Current chairs: ajax dgilmore hguemar jwb nirik paragan rishi sgallagh thozza
18:00:56 <nirik> .hello kevin
18:00:58 <zodbot> nirik: kevin 'Kevin Fenzi' <kevin@scrye.com>
18:00:59 <jwb> present
18:01:01 <sgallagh> .hello sgallagh
18:01:03 <paragan> Hi
18:01:05 <zodbot> sgallagh: sgallagh 'Stephen Gallagher' <sgallagh@redhat.com>
18:01:08 <thozza> .hello thozza
18:01:09 <zodbot> thozza: thozza 'Tomas Hozza' <thozza@redhat.com>
18:02:35 <rishi> .hello rishi
18:02:36 <sgallagh> We have quorum, so shall we get started? I propose we cover the two "New Business" items first since they should be quick and then come back to the bundling question.
18:02:36 <zodbot> rishi: rishi 'Debarshi Ray' <debarshir@redhat.com>
18:02:38 <ajax> hi
18:02:53 <rishi> sgallagh: Sounds good.
18:03:11 <sgallagh> #topic #1484 unresponsive maintainer: chitlesh
18:03:11 <sgallagh> .fesco 1484
18:03:12 <zodbot> sgallagh: #1484 (unresponsive maintainer: chitlesh) – FESCo - https://fedorahosted.org/fesco/ticket/1484
18:03:53 <nirik> I mailed them (because it seems like people can't be bothered to do that anymore). But I didn't get a reply.
18:04:27 <sgallagh> Seems pretty straightforward, then.
18:04:38 <dgilmore> hiya
18:04:52 <jwb> unfortunate
18:05:04 <rishi> Yeah, I haven't seen Chitlesh around for a while.
18:05:06 <sgallagh> Unfortunate, yeah.
18:05:16 <sgallagh> Seems like he owned a lot of the electronics lab packages too
18:05:34 <rishi> He started the electronics spin a few years ago.
18:05:40 <rishi> So, understandable.
18:05:43 <dgilmore> sgallagh: we removed it a couple of releases ago, but I know someone wants to bring it back
18:05:43 <sgallagh> /me nods
18:05:50 <dgilmore> maybe she will pick up the packages
18:05:55 <sgallagh> dgilmore: Yeah, that's what I was wondering.
18:06:03 <nirik> if they can find a sponsor/mentor
18:06:15 <sgallagh> OK, so:
18:06:16 <sgallagh> Proposal: Orphan all packages that chitlesh is the PoC on
18:06:17 <sgallagh> +1
18:06:22 <paragan> +1
18:06:26 <thozza> +1
18:06:27 <dgilmore> but in thats case I am +1 to orphaning his packages
18:06:28 <rishi> +1
18:06:41 <jwb> +1
18:06:43 <dgilmore> nirik: I will sponsor/mentor her
18:06:57 <nirik> dgilmore: excellent.
18:06:58 <nirik> +1
18:07:38 <sgallagh> #agreed Orphan all packages that chitlesh is the PoC on (+7, 0, -0)
18:07:49 <sgallagh> nirik: Would you mind doing the necessary?
18:08:23 <nirik> sure, I can do it.
18:08:25 <sgallagh> Thanks
18:08:27 <sgallagh> #action nirik to perform the orphaning
18:08:32 <sgallagh> Next topic:
18:08:36 <sgallagh> #topic #1485 F24 System Wide Change: Python 3.5
18:08:37 <sgallagh> .fesco 1485
18:08:40 <zodbot> sgallagh: #1485 (F24 System Wide Change: Python 3.5) – FESCo - https://fedorahosted.org/fesco/ticket/1485
18:08:50 <sgallagh> So, I doubt this will be at all contentious.
18:09:03 <thozza> jkurik is sorry but he could not do it today to the meeting
18:09:06 <sgallagh> It's a shame it narrowly-missed F23, we should definitely focus on it for F24/
18:09:19 <number80> o/
18:09:20 <dgilmore> +1
18:09:24 <thozza> +1 from me
18:09:28 <rishi> +1
18:09:29 <sgallagh> +1
18:09:30 <nirik> +1
18:09:31 <paragan> +1
18:09:32 <ajax> +1
18:09:47 <jwb> +1
18:10:15 <sgallagh> number80: Was that a +1?
18:10:43 <number80> sgallagh: just joined, I don't even know the question
18:10:58 <thozza> number80: python 3.5 in F24
18:11:02 <sgallagh> number80: Ah, sorry. Approving Py 3.5?
18:11:05 <number80> +1
18:11:10 <sgallagh> #agreed F24 System Wide Change: Python 3.5 is approved (+9, 0, -0)
18:11:26 <sgallagh> OK, we have a full FESCo present, which is good because the next topic is:
18:11:40 <sgallagh> #topic #1483 Decision on bundling policy in the Fedora Package Collection
18:11:40 <sgallagh> .fesco 1483
18:11:42 <zodbot> sgallagh: #1483 (Decision on bundling policy in the Fedora Package Collection) – FESCo - https://fedorahosted.org/fesco/ticket/1483
18:12:19 <jwb> i'd like to be a bit more explicit in the "non-critpath pacakges do not need a bundling exception from FPC" part i guess, but otherwise i'm in favor
18:12:32 <sgallagh> The specific proposal is listed in https://fedorahosted.org/fesco/ticket/1483#comment:11
18:12:58 <nirik> were there some fpc folks that were going to send a counterproposal/another proposal?
18:13:06 <thozza> sgallagh: I would like to ask what was to initial reason you started this discussion in the first place?805333
18:13:19 <rishi> nirik: Yeah tibbs|w said he would.
18:13:27 <sgallagh> nirik: I pestered tibbs a couple times to send such a proposal but none materialized.
18:13:40 <nirik> if we use critpath for this, we should update the updates_policy and/or overwrite the old definition
18:14:00 <tibbs|w> Sorry, don't recall being pestered.  I've limited time, unfortunately.
18:14:05 <sgallagh> thozza: Less a specific incident and more a lot of straws on one camel
18:14:11 <tibbs|w> So I might have missed a ping or something.
18:14:16 <nirik> well, I guess https://fedoraproject.org/wiki/Critical_path_package
18:14:24 <number80> well, this is a long-running topic, I'm in favour of this proposal but I'd wait FPC counter-proposal if they have one
18:14:36 <tibbs|w> Anyway, the only thing that's important to me is that FPC not be cut out of the loop.
18:14:47 <tibbs|w> Though I'd welcome the shorter meetings which would result from that.
18:15:16 <sgallagh> tibbs|w: https://fedorahosted.org/fesco/ticket/1483#comment:11 retains the FPC on anything we consider critical path (term subject to change)
18:15:38 <dgilmore> I am not against the proposal. I do wonder what will acually be critical path.
18:15:48 <tibbs|w> Which implies that it would not be involved for anything not in the critical path.
18:15:53 <tibbs|w> Which I would be against.
18:15:55 <sgallagh> dgilmore: Part of the proposal requires that we (FESCo) assign a deadline to Base WG to define that.
18:16:34 <sgallagh> tibbs|w: Would a reasonable compromise be for an FPC sign-off on the critical path definition?
18:16:38 <dgilmore> sgallagh: well today it includes gnome, firefox and a whole bunch of things
18:16:40 <tibbs|w> Not to me, no.
18:17:11 <dgilmore> sgallagh: and In the proposal that is going to go away unless the WG for an edition requests its inclusion
18:17:14 <tibbs|w> Basically I'm concerned about just opening it up to a free for all.  With FPC oversight we'd at least have a chance to deal with issues of egregious bundling.
18:17:22 <dgilmore> I think KDE is also in critical path
18:17:42 <tibbs|w> And change FPC's approval criteria to basically "you tried hard enough to unbundle".
18:17:50 <thozza> BIND is in critical path
18:17:51 <sgallagh> dgilmore: Please see the part where I want it clearly understood that the current definition is unrelated and that the choice of terms wasn't ideal.
18:17:59 <dgilmore> sgallagh: so my concern is that critical path will significantly shrink
18:18:08 <nirik> actually all the desktops are...
18:18:15 <nirik> xfce/kde/etc...
18:18:25 <thozza> DHCP too IIRC
18:18:31 <tibbs|w> You could just leave it to FPC to evaluate the "importance" of a package, which would be fine with me.
18:18:44 <tibbs|w> Instead of saying "ring 0" or "critpath" or whatever.
18:18:45 <dgilmore> sgallagh: I am concerned what you are proposing is a huge step backwards
18:19:09 <jwb> it sounds like people are uncomfortable approving this without getting the definition of what is and is not "critical".  maybe we should defer and focus on that definition
18:19:24 <number80> +1
18:19:25 <sgallagh> jwb: I think that's a catch-22
18:19:28 <nirik> tibbs|w: I guess the concern there is that many people wouldn't know that the FPC changed to 'you tried hard enough' and think they can't get an approval...
18:19:42 <sgallagh> No one is going to accurately be able to define "critical" without an understanding of the impact of that decision
18:19:46 <tibbs|w> Really FESCo could just ask FPC to come up with some guidelines which aren't so restrictive.  I'm in favor of relaxing things significantly.
18:19:47 <nirik> or don't want to 'try hard enough' I guess.
18:20:33 <number80> tibbs|w: do you think you could come up with relaxed guidelines in a reasonable delay?
18:20:39 <jwb> sgallagh, uh... what?
18:21:00 <tibbs|w> number80: Loose guidelines, certainly.  Something that will satisfy all nitpickers, probably not.
18:21:01 <jwb> sgallagh, specifically, what decision are you referring to with "that decision"
18:21:15 <sgallagh> jwb: Sorry, I mean that without a clear understanding of what being in the critical path means, no one can make an informed decision about what to put there.
18:21:25 <thozza> sgallagh: I agree
18:21:36 <number80> tibbs|w: well, I prefer consensus with FPC over any solution
18:21:45 <sgallagh> I'd like to mention that there's also more to this than the strict technical decision.
18:21:49 <tibbs|w> I can try to draft something tonight for the FPC meeting tomorrow, but that's awfully quick, and I'm on vacation for ten days starting Friday.
18:22:08 <tibbs|w> Though I'll be online enough to respond to email.
18:22:19 <sgallagh> Having a new high-level policy is something that we can announce and advertise and use as a mechanism for trying to increase participation in the project.
18:22:28 <jwb> sgallagh, sure but at the same time asking people to approve a proposal that hinges on that very undefined defintion is a bit silly.  i mean, this proposal is only one part of what "critpath" might mean
18:22:29 <tibbs|w> Unless rawhide nukes my laptop.
18:22:56 <sgallagh> This is significantly lessened by "Well, we're reducing a few of the hurdles" and makes it harder to advertise our more accommodating behavior.
18:23:05 <rishi> sgallagh: So, the new proposal redefines what a "critical path" package is, right? My questions is, how will this affect the rationale behind the current definition of critical path. eg., the "additional verification".
18:23:14 <tibbs|w> Instead of "critpath", how would you qualify packages where unbundling is important?
18:23:16 <jwb> rishi, it does not
18:23:26 <jwb> rishi, it depends on a defintion of critpath that is not set
18:23:45 <thozza> maybe we can start with the Base WG task? to define the ring 0 or critical components?
18:23:46 <number80> critpath here is a different thing
18:23:47 <sgallagh> tibbs|w: If you have a better term, please propose it. I couldn't come up with one and it's clearly causing the conversation to be confused.
18:24:02 <tibbs|w> sgallagh: I understand.  It's not at all easy.
18:24:02 <ajax> "pure set"
18:24:10 <ajax> "dogma-conformant set"
18:24:21 <number80> proposal: wait FPC feedback and review the question next week
18:24:23 <ajax> "screechingly-correct set"
18:24:24 <tibbs|w> I would instead say that there are several criteria and it's up to FPC to evaluate.
18:24:34 <rishi> jwb: Ok. Got it.
18:24:45 <dgilmore> number80: +1
18:24:50 <tibbs|w> But here's the question: what, broadly, does FESCo want from FPC with regards to this issue?
18:24:50 <sgallagh> number80: -1
18:24:55 <thozza> number80: +1
18:25:05 <sgallagh> Let's stop kicking this down the road.
18:25:05 <tibbs|w> "Please come up with guidelines which aren't so restrictive"?
18:25:09 <jwb> tibbs|w, to make bundling easier for cases where it's hard to unbundle
18:25:21 <sgallagh> Again, please don't think *only* of the technical problem.
18:25:38 <sgallagh> We also need to be able to fix *years* of negative sentiment about our policies.
18:25:41 <tibbs|w> jwb: And of course "hard" is hard to determine with precision.  Of course we can evaluate.
18:25:59 <dgilmore> ring 0 and ring 1 have to be unbundled? outer rings can bundle with appropriate documentation?
18:26:19 <number80> sgallagh: I agree with you but if we can FPC approval, it's worth waiting
18:26:24 <tibbs|w> Can FPC leave it to someone with more marketing focus to deal with negative sentiment?
18:26:25 <number80> *get
18:26:26 <sgallagh> dgilmore: That's pretty much my proposal as it stands, only I didn't use the ring 0 and ring 1 terminology because that's a rathole
18:26:49 <sgallagh> number80: I don't think punting before we decide what we as FESCo wants is useful
18:26:51 <tibbs|w> Any such terminology is going to be a rathold.
18:27:03 <dgilmore> sgallagh: that is not what you proposed, at least how I read it.
18:27:03 <nirik> I guess it seems to me like people want a easier way to allow bundling in cases where upstream bundles as part of their development process and has no desire to change
18:27:13 <tibbs|w> But I'll be happy to know what FESCo wants and try to whip something up.
18:27:14 <dgilmore> sgallagh: what you have proposed is a rathole also
18:27:15 <jwb> nirik, right
18:27:49 <sgallagh> dgilmore: I attempted to leave the exact implementation of ring0/1 out of the proposal. If that happens to mirror what I was calling "critical path", great.
18:28:15 <jwb> nirik, the base proposal then would be "allow bundling in cases where upstream has no desire to unbundle without FPC approval".  sgallagh's proposal attempts to restrict it further and document things
18:28:17 <sgallagh> dgilmore: I was *very* clear in the proposal that what I am calling "critical path" in the proposal *IS NOT* the previous definition and needs to be considered as a separate concept
18:28:26 <rishi> sgallagh: Since GNOME is full of libgd-like cases that people consider bundling, I wonder how pure ring 1 can really be.
18:28:41 <jwb> sgallagh, yes.  you were.  that does not mean it is at all clear what *IS* the definition
18:28:57 <jwb> and that is what people are stuck on
18:29:28 <sgallagh> jwb: The definition was intentionally left up to the Base WG to try to avoid ratholing on where the line is in this conversation.
18:29:43 <sgallagh> I'd be fine if we wanted to s/Base WG/Some other entity or entities/
18:29:59 <nirik> jwb: right, I like the documentation, I am trying to decide if the restriction is perhaps something we could drop since it's causing lots of confusion. ;)
18:30:12 <jwb> nirik, right.  personally, i think the base proposal would be fine
18:30:15 <sgallagh> nirik: Which restriction(s)?
18:30:22 <jwb> sgallagh, the critpath restriction
18:30:31 <nirik> sgallagh: critpath cannot bundle without fpc approval
18:30:40 <sgallagh> Ah, ok
18:31:04 <sgallagh> I'm fine with that personally, but that was my attempt to compromise to the more conservative community members.
18:31:11 <nirik> So the fear there is that we have some amazing thing that bundles it gets into the collection and starts getting used for things and has problems that bundling brings?
18:31:40 * nirik suspects this may be self correcting
18:32:25 <sgallagh> Unbundling usually solves more problems than it causes, but we also need to be aware that it *does* cause problems (both politically and technically)
18:32:34 <nirik> I completely agree.
18:32:38 <jwb> it is arguable that Workstation would consider darktable "critpath" for their edition.  if the intention was to make it easier for that package to get into the collection without FPC approval, then this proposal fails to lower the hurdle.
18:32:49 <number80> enforcing FPC current guidelines to declare bundled libs could help secteam and other groups to track these issues
18:32:49 <jwb> as an example package fwiw
18:33:19 <jwb> number80, the proposal states nothing about enforcement
18:33:19 <dgilmore> jwb: indeed
18:33:35 <sgallagh> jwb: That situation catalyzed this discussion but isn't representative of the complete problem
18:33:59 <jwb> sgallagh, i agree.  but i'm giving you a literal example of how people are balking at approving without knowing what critpath means
18:34:06 <sgallagh> Fair
18:34:18 <rishi> jwb: I thought that is addressed by the part which says "each WG can add their own bits to the sacred set".
18:34:40 <jwb> rishi, yes, but if added to said set, then you have to get an FPC approval to bundle
18:34:46 <sgallagh> I thought so too, but I guess my legalese is a bit accented
18:34:46 <jwb> because it's in the sacred set
18:34:53 <jwb> defeating the purpose
18:35:05 <sgallagh> jwb: Well, my assumption there is that it would be used sparingly
18:35:21 <sgallagh> Such as to address cases that have historically been problematic in this regard.
18:35:33 <sgallagh> But yes, I probably should have spelled that out more clearly.
18:35:35 * rishi notes that "sacred" can be typoed as "scared" :)
18:36:05 <jwb> my assumption is that it would be used to bring in pacakges that people have avoided attempting to package because of the bundling rules.
18:36:09 <jwb> see how assumptions cut both ways?
18:36:36 <dgilmore> jwb: such as chromium? as an example
18:36:42 <nirik> tibbs|w: what role do you see FPC playing here? checking that someone really tried to unbundle? or ?
18:36:46 <sgallagh> jwb: I'm not sure I parsed that correctly.
18:36:53 <jwb> dgilmore, possibly
18:37:08 <sgallagh> I think you defined the intention of this proposal: to make it easier for packages that are useful but hard to unbundle to get into Fedora.
18:37:22 <tibbs|w> nirik: Essentially.  Enforcing restrictions on packages which are "too important" from the standpoint of security or such.
18:37:34 <nirik> if I package foo and it bundles and during review my reviewer notes that and I contact upstream and they say "this is the way we work" and I record that in a file and add "Provides: bundled(bar" what would FPC add here?
18:38:11 <tibbs|w> In other words, FPC makes sure that unbundling happens if possible, and is documented if not, and basically decides on what is in the "critical set" or whatever.
18:38:20 <dgilmore> nirik: especially when bar is a fork with patches upstream bar will not take
18:38:36 <rishi> jwb: Re-reading your darktable-critpath bit, I guess you confused critpath-that-need-more-testing with critpath-that-cant-bundle.
18:38:43 <tibbs|w> It is not entirely uncommon for unbundling to be found to be easy.
18:38:47 <nirik> well, if we drop the last one, does it really make sense to spend your time on the first 2?
18:38:53 <sgallagh> tibbs|w: So if I changed the proposal to require that it be FPC that did the upstream contact and negotiation, would that work?
18:39:04 <sgallagh> (And FPC commits to doing that work...)
18:39:09 <rishi> jwb: While darktable might be something that Workstation might want to test, they may not want to unbundle everything that it has.
18:39:14 <tibbs|w> Erm, not really; packagers should still be the primary contact with upstream.
18:39:22 <tibbs|w> FPC has helped in the past, of course.
18:39:25 <jwb> rishi, i did not.  because there's no definition of critpath.  i was suggesting critpath-as-in-the-product-showcases-this
18:39:31 <sgallagh> tibbs|w: I didn't think so, but I'm trying to figure out what you're angling for.
18:40:10 <number80> can someone suggest a different wording for the critpath concept?
18:40:14 <tibbs|w> I thought it was pretty clear, honestly.  Would folks like me to try and communicate a different way?
18:40:22 <rishi> jwb: Well, there is https://fedoraproject.org/wiki/Critical_path_package and the new thing in sgallagh 's proposal. (sorry for going off in this tangent)
18:40:29 <jwb> number80, i believe nirik and i are angling for dropping it all together.
18:40:31 <nirik> Is "makes sure that unbundling happens if possible, and is documented if not" not something a reviewer could do? or you just think it's more complex than most reviewers would bother with?
18:40:42 * nirik nods to jwb
18:41:02 <number80> jwb: ack
18:42:17 <nirik> I guess there's somewhat of a middle area here where it's easy to unbundle and carry that in fedora package but upstream has no interest in taking that (not sure how often that happens)
18:42:24 <sgallagh> jwb, nirik: Something like: "Packagers must unbundle if it's currently possible. If not, they must try to contact upstream and advocate for it. If that fails, Provides: bundled(foo)" ?
18:42:33 <tibbs|w> nirik: It happens pretty often in my experience.
18:42:44 <dgilmore> number80: everything in the deliverables for the basewg and the editions
18:42:57 <jwb> sgallagh, right.  basically your proposal, but no restrictions on package "set"
18:43:02 <ajax> upstream is never going to have an interest in unbundling
18:43:03 <rishi> sgallagh: I think they should contact upstream first, beforetrying to unbundle themselves.
18:43:16 <number80> dgilmore: ack, should be clearer for the sake of the discussion
18:43:25 <ajax> if they thought that was a valuable thing they'd have not bothered to bundle in the first place
18:43:31 <rishi> sgallagh: Because certain upstreams might be offended because unbundling will be seen as breaking their code.
18:43:49 <nirik> http://paste.fedoraproject.org/276064/44243383/ is sgallaghs proposal without the critpath distinction
18:43:54 <tibbs|w> ajax: That's not entirely true; we have had instances in the past where upstream has understood the issue.
18:44:04 <jwb> dgilmore, fwiw, i think that's far too large of a set
18:44:16 <jwb> nirik, yes.  exactly that.
18:44:16 <nirik> there's a lot of responses I'm sure...
18:44:32 <dgilmore> jwb: perhaps
18:44:34 <tibbs|w> jwb: Too large of a set for FPC to handle, or too large of a set to ban bundling completely?
18:44:38 <ajax> tibbs|w: not entirely, but certainly mostly.
18:44:41 <jwb> tibbs|w, the latter
18:44:46 <sgallagh> tibbs|w: True. Node.js for example was amenable to unbundling for Fedora, provided they didn't have to do the work :)
18:45:23 <sgallagh> nirik: I'd be perfectly happy approving that version of the proposal.
18:45:35 <rishi> nirik: sgallagh: That one looks good to me.
18:45:48 <sgallagh> Like I said, I added the critpath mumbo-jumbo mostly to appease the hardcore conservative faction of Fedora, but it looks like that may have been misguided.
18:46:00 <nirik> sgallagh: so from your side adding a FPC ack step would be too much process? or ?
18:46:08 * rishi smiles at "mumbo-jumbo"
18:46:27 <jwb> nirik, i don't see how adding an FPC ack step is any different that today's process
18:46:30 <dgilmore> I do not think we should ban bundling. just that we should be very careful about it in the things we showcase
18:46:50 <dgilmore> and in the other bits we can be a bit more lenient
18:46:57 <tibbs|w> jwb: I don't think we should change the process.  I think things should be relaxed on the FPC end.
18:46:58 <nirik> jwb: well, if we changed the critera to whats there it would be very different
18:47:06 <rishi> sgallagh: nirik: Although, I am worried about the "upstream bundles, but is dead" case.
18:47:07 <sgallagh> Nothing in this process *mandates* bundling, to be clear.
18:47:21 <sgallagh> Any place that the maintainers feel is better served by unbundling is always welcome
18:47:25 <rishi> In that case, it is probably better to stick to the old rules. ie. must unbundle.
18:47:27 <number80> rishi: dead upstream stuff should be dropped
18:47:31 <thozza> tibbs|w: that sound more reasonable than complicating processes with another process
18:47:36 <jwb> nirik, so FPC would be acking what?  the packager followed those steps?
18:47:39 <rishi> Because nobody would be around to react to security bugs.
18:48:29 <dgilmore> I need to run soon to pick up my daughter from school
18:48:45 <nirik> jwb: yeah, I guess so...
18:48:46 <tibbs|w> thozza: That was always my stance.  Somehow this became a "have to do this now!" issue and the proposal came here instead of FPC.
18:48:56 <sgallagh> Can we just run a straw poll on nirik's current proposal and see if it would pass?
18:49:09 <jwb> why would we do that?  why wouldn't we just vote?
18:49:11 <rishi> number80: Well, we do have a lot of stuff with dying upstreams, don't we?
18:49:18 <nirik> tibbs|w: how would relaxing look like onn the fpc end?
18:49:19 <sgallagh> jwb: Fine, let's vote.
18:49:44 <number80> rishi: well, not that I disagree with enforcing strict unbundling but we can't for every case
18:49:47 <nirik> just always approve if upstream refuses to unbundle and the submitter did the process correctly?
18:49:50 <sgallagh> Proposal: Adjust the packaging policy as described in http://paste.fedoraproject.org/276064/44243383/
18:49:53 <sgallagh> +1 from me
18:50:19 <tibbs|w> nirik: Well, in broad terms, very broad acceptance for things which don't appear to be security sensitive, aren't bundling something with known security issues, and where bundling won't really cause problems for anything.
18:50:43 <rishi> number80: I am not talking about "every case".
18:50:43 <tibbs|w> But even the most leaf node package shouldn't bundle glibc.
18:51:00 <rishi> I am talking about an unresponsive upstream who might not react to security issues.
18:51:10 <ajax> tibbs|w: i'm comfortable handling that kind of abuse by auditing the bundled() provides every so often
18:51:12 <thozza> sgallagh: how is this different from your original proposal other than not including the critical path part?
18:51:26 <sgallagh> thozza: It's not. That's the point.
18:51:36 <nirik> I think I am +0 for the modified proposal right now, as I would like some more discussion and don't see this issue as time critical.
18:52:02 <thozza> sgallagh: I don't understand why are you pushing so hard to approve this ASAP?
18:52:03 <jwb> thozza, it isn't.  it is exactly the original except the critpath part
18:52:12 <nirik> tibbs|w: do you think that has support in FPC? or hard to say?
18:52:23 <dgilmore> i am +0 as I would rather get this right
18:52:30 <jwb> i'm +1 fwiw
18:52:38 <sgallagh> thozza: Because it has been a serious problem for years and FESCo keeps kicking it down the road to continue to get worse
18:52:40 <nirik> thozza: I think he just wants it over with so we can move on to other things we have to do. ;)
18:52:44 <tibbs|w> nirik: I think it's a matter of degrees.  I can see an immediate -1 from a certain person.  Probably enough to pass.
18:52:50 <ajax> +1, we can always amend later and this is an improvement.
18:52:52 <number80> +1 though I would prefer trying to get FPC approval
18:53:06 <paragan> +1 to nirik proposal
18:53:12 <jwb> that 5
18:53:12 <sgallagh> Yes, I'd like to get this over and done with. If it's not perfect we can revise as we go along. Release early, release often
18:53:19 <nirik> thats a pass. ;)
18:53:21 <rishi> I am +0 too. I would rather have the FPC sign off on this.
18:53:26 <thozza> -1
18:53:38 <thozza> I would also like FPC involved
18:54:01 <tibbs|w> I really don't know why FESCo didn't start with "FPC, we'd like you to come up with some way to relax the restrictions."
18:54:10 <tibbs|w> I guess because this became time critical.
18:54:10 <rishi> I will also note that we were reluctant to overrule the FPC over the darktable issue, but now suddenly seem to be keen to push this through.
18:54:39 <ajax> i'm more keen to stop talking in circles about it than anything
18:54:53 <ajax> i don't think there's any material disagreement, here.
18:54:55 <number80> I think we're trying to have a general agreement and then fix all the details
18:54:56 <sgallagh> I count +5, 2, -1
18:54:59 <sgallagh> Who's missing?
18:55:14 <rishi> paragn, maybe?
18:55:15 <tibbs|w> I guess that means I don't need to draft anything.
18:55:23 <paragan> voted +1
18:55:37 <rishi> Oh, sorry paragan
18:55:40 <sgallagh> Oh, sorry. I missed rishi
18:55:45 <sgallagh> So that's +5, 3, -1
18:56:18 <sgallagh> I'll wait 60s in case anyone wants to change their vote.
18:56:44 * nirik takes this chance to get more coffee.
18:57:44 <sgallagh> #agreed Adjust the packaging policy as described in http://paste.fedoraproject.org/276064/44243383/ (+5, 3, -1)
18:58:13 <thozza> so happy bundling :)
18:58:13 <tibbs|w> I find this terribly disappointing.  FESCo really shouldn't have just steamrolled over FPC like that.
18:58:33 <sgallagh> OK, as I brought this to the table:
18:58:33 <sgallagh> #action sgallagh to submit wiki page update drafts
18:58:39 <thozza> tibbs|w: I agree with you
18:58:40 <tibbs|w> I'll let FPC know and start working on removing the anti-bundling stuff from the guidleines.
18:58:59 <sgallagh> #undo
18:58:59 <zodbot> Removing item from minutes: ACTION by sgallagh at 18:58:33 : sgallagh to submit wiki page update drafts
18:59:17 <sgallagh> #action tibbs|w to inform FPC and work on removing the anti-bundling stuff from the guidelines
18:59:20 <tibbs|w> This kind of thing would never have happened with the FESCo of old; they used to work with their committees instead of pretending they don't exist when there's a contentious issue.
19:00:02 <sgallagh> #topic Next week's chair
19:00:03 <ajax> tibbs|w: i wasn't aware this was being phrased as overruling fpc
19:00:06 <jwb> that's kind of a strange statement to make, considering the people that are in fesco have been for literally years with a few minor exceptions.
19:00:17 <tibbs|w> And nobody seemed to consider how that makes is feel.  "Someone got pissed off at a decision you actually didn't make, and so now you just aren't involved in that any longer."
19:00:22 <jwb> minor in number, not in people or contribution
19:00:24 <tibbs|w> But it's your committee.
19:00:42 <tibbs|w> ajax: Doesn't really matter how it's phrased, does it?  That's the end result.
19:00:45 <jwb> i don't think people are pissed off.  i think people are tired.
19:00:46 <sgallagh> Who wants to chair next week?
19:01:10 <bconoboy> tibbs|w: It's still perfectly reasonable to propose further refinements for vote, we're just seeing some incremental progress.
19:01:49 * nirik suspects some folks on the list may be mad too as we approved something that had no discussion in it's current form, but oh well.
19:02:06 <tibbs|w> It's not FPC's problem any longer, though.  Hopefully zlib won't happen again.
19:02:06 <rishi> tibbs|w: I will note that as one of the newer members of FESCo I didn't vote +1 for this. Although I did  +1 the proposal to overrule the initial darktable removal.
19:02:18 <sgallagh> nirik: It's not far from my original email, honestly. The only difference is that an attempt to contact upstream has to be made.
19:02:24 <tibbs|w> rishi: FPC never voted against darktable.
19:02:38 <nirik> sgallagh: well, no critpath stuff...
19:02:43 <tibbs|w> They just didn't initially vote for it.
19:02:54 <tibbs|w> Maybe one day someone can tell us what critpath means.
19:03:03 <jwb> it doesn't matter any longer
19:03:05 <sgallagh> nirik: I'm talking about my original email dated 2015-09-10
19:03:05 <nirik> it has a definition right now. ;)
19:03:08 <jwb> at least for these purposes
19:03:33 <nirik> https://fedoraproject.org/wiki/Critical_path_package
19:03:35 <tibbs|w> Anyway, I've registered my disappointment with the process.
19:03:46 <tibbs|w> And now I have work to do.
19:04:30 <sgallagh> OK, so next week's chair?
19:04:58 <paragan> I can do it
19:05:15 <sgallagh> #info paragan to chair next week's meeting
19:05:19 <sgallagh> #topic Open Floor
19:05:53 <thozza> when should the elections for next FESCo start?
19:06:16 <thozza> since we are approaching 23 GA
19:06:36 <paragan> guess after F23 is released and jkurik can answer that
19:06:55 <sgallagh> Unfortunately jkurik isn't here this week.
19:07:07 <sgallagh> thozza: Could you open a FESCo ticket and make sure jkurik is CCed?
19:07:38 <thozza> sure
19:07:48 <rishi> I have a quick question.
19:07:57 <rishi> Now that io.js has merged  back with node.js, is there something to be done in Fedora? I recall a io.js ticket from a while ago. Does anybody remember the details off the top of their heads?
19:08:14 <nirik> I think fedora kept the node.js names...
19:08:18 <thozza> #action thozza to open FESCo ticket about elections after F23 GA and CC jkurik
19:08:24 <jwb> that sounds like something the package maintainers should be asked :)
19:08:26 <nirik> so nothing to do FWIW.
19:08:28 <paragan> right we have not started with io.js
19:08:49 <rishi> ok
19:09:09 <sgallagh> rishi: The package maintainer actually had to go AWOL for a bit, so that missed F23 in any cas
19:09:25 <paragan> also the Change owner has not updated Change page
19:09:47 <paragan> ah ok
19:10:05 <sgallagh> I've spoken to him. His dayjob attacked hard for a few months
19:10:07 <thozza> based on versions we never had io.js in fedora
19:10:27 <thozza> we still have nodejs of an old version
19:10:27 <rishi> thozza: sgallagh: paragan: I see, ok.
19:10:34 <rishi> Thanks.
19:10:56 <thozza> rawhide contains version released on 2015-01-30
19:11:46 <sgallagh> thozza: Like I said; primary maintainer got swamped and no one backed him up.
19:12:08 <thozza> I'm just stating facts
19:12:25 <thozza> not saying anything about the maintainer
19:12:52 <sgallagh> I don't see FESCo having anything to do here unless someone is proposing a non-responsive maintainer decision
19:12:57 <sgallagh> (Which I don't think anyone is)
19:13:07 <thozza> no
19:13:40 <rishi> No, no.
19:14:07 <rishi> I just read the io.js news recently and remembered the FESCo ticket. Nothing more than that.
19:14:27 <sgallagh> ok
19:14:33 <sgallagh> Anything else for Open Floor?
19:14:47 <rishi> None from me.
19:15:10 <number80> let's wrap it
19:15:34 <sgallagh> ok, thanks folks. I know this was a very high-stress meeting.
19:16:03 <sgallagh> #endmeeting