16:00:20 <geppetto> #startmeeting fpc
16:00:20 <zodbot> Meeting started Thu Oct  8 16:00:20 2015 UTC.  The chair is geppetto. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:20 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:00:20 <geppetto> #meetingname fpc
16:00:20 <geppetto> #topic Roll Call
16:00:20 <zodbot> The meeting name has been set to 'fpc'
16:00:26 <orionp> hello
16:00:31 * tomspur waves
16:00:44 <geppetto> #chair orionp
16:00:44 <zodbot> Current chairs: geppetto orionp
16:00:47 <geppetto> #chair tomspur
16:00:47 <zodbot> Current chairs: geppetto orionp tomspur
16:00:51 <tibbs|w> Howdy.
16:00:54 <geppetto> #chair tibbs
16:00:54 <zodbot> Current chairs: geppetto orionp tibbs tomspur
16:00:57 <Rathann> hey
16:01:05 <geppetto> #chair Rathann
16:01:05 <zodbot> Current chairs: Rathann geppetto orionp tibbs tomspur
16:02:12 * racor is here, too
16:02:17 <geppetto> #chair racor
16:02:17 <zodbot> Current chairs: Rathann geppetto orionp racor tibbs tomspur
16:03:38 <geppetto> mbooth: FPC ping ?
16:03:46 <mbooth> Hey
16:03:52 <mbooth> Was in a world of my own
16:04:09 <geppetto> #chair mbooth
16:04:10 <zodbot> Current chairs: Rathann geppetto mbooth orionp racor tibbs tomspur
16:04:18 <geppetto> #topic Schedule
16:04:23 <geppetto> https://lists.fedoraproject.org/pipermail/packaging/2015-October/011043.html
16:04:39 <geppetto> #topic Rant about bundling
16:04:47 <geppetto> Ok, might as well get it out ;)
16:04:54 <tibbs|w> I ranted in the FESCo meeting, so I'm all ranted out.
16:05:00 * geppetto nods
16:05:01 <orionp> yeah, I'm good
16:05:05 <tibbs|w> I'm glad that at least one of us was there.
16:05:09 <nirik> note: I think fesco would be open to changes...
16:05:26 <orionp> I watched for awhile but had to leave...
16:06:20 <nirik> I was trying to gather info on what FPC role could be in a more relaxed bundling world, but then the vote just went on, so...
16:06:28 <tibbs|w> I think that if FESCo as a body was interested in FPC input, it would have waited to receive some, so...
16:06:49 <tibbs|w> Anyway, I did submit a draft to implement the policy as dictated.
16:07:16 <nirik> well, I think there was some of "I am tired of this, lets just do something" and also "lets do something and have people propose changes to it"
16:07:22 <geppetto> Ahh, https://fedorahosted.org/fpc/ticket/575 … you want to do that now?
16:07:53 <tomspur> It seems FPC has a very bad reputation marketing wise, which is now also solved as the FPC is not involved anymore
16:08:09 <geppetto> Yeh, people who say no often do
16:08:16 <tibbs|w> Indeed.
16:08:25 <geppetto> Esp. in Fedora where we are basically the only people who ever say no
16:08:37 * geppetto shrugs
16:08:40 <tibbs|w> We never even said no to the case which precipitated all of this.
16:08:48 <tibbs|w> Which makes it all the more hilarious.
16:08:54 <geppetto> Sure, but it was possible
16:09:18 <tibbs|w> I guess the possibility of us saying no to something in the next week or so was too much to bear.
16:09:22 <geppetto> Much better for everyone if they can just do whatever crazy thing they want and everyone else pays for it
16:09:31 <tibbs|w> Well, what's done is done.
16:09:38 <tomspur> Should we add to the new guidelines a kind of note, what "would be good to consider", when one bundles?
16:09:52 <racor> tomspur: FESCO having taken over the job and responsibility is the only advantage of their decision ;)
16:09:52 <tibbs|w> tomspur: Not our business any longer.
16:10:21 <tibbs|w> The policy as dictated says only that they have to ask upstream and record upstream's answer.
16:10:22 <racor> I am hoping and expecting it will back-fire at them.
16:10:30 <tomspur> It would still be nice to at least 'info' what we think about the policy.
16:10:45 <tomspur> racor: This would be one 'info' :)
16:11:09 <racor> tomspur: I don't you nor FESCO want to know what I think about this decision!
16:11:10 <tibbs|w> tomspur: I mean, if someone wants to draft an "official FPC response" or something, great.
16:11:13 <orionp> At this point I think we should just try to keep the guidelines simple
16:11:24 <racor> ... don't think ...
16:11:31 <tibbs|w> But I go on vacation tomorrow and am looking forward to not worrying about this.
16:12:14 <geppetto> orionp: +1 … just write what FESCo voted on
16:12:29 <geppetto> Anything else like seems pointless.
16:12:56 <tomspur> It would be enough for me to say, that it is too easy to just bundle and like FESCo to reconsider. Then implement the decision from yesterday...
16:13:41 <tibbs|w> Anyway, that's what the draft I submitted does.
16:14:29 <tibbs|w> I wish I had rewritten all of the bundling stuff earlier; if it had been clearer and perhaps more concise, we might not have been in this situation.
16:14:36 <geppetto> doubt it
16:15:18 <mbooth> tibbs|w: I'm not convinced about that :-)
16:15:49 <tibbs|w> Well, there were multiple different sets of complaints and at least some of them were more about our pages than the actual policy.
16:16:21 <tibbs|w> But now we get to delete them all.  So how about it?  Anyone take issue with my draft in 575?
16:16:49 <orionp> +1
16:17:01 <geppetto> No, it looks fine to me
16:17:02 <geppetto> +1
16:17:06 <tibbs|w> https://fedoraproject.org/wiki/User:Tibbs/BundlingDraft2 as well as "deleting" the two standalone pages on bundling.
16:17:09 <tibbs|w> +1 obviously.
16:17:11 <tomspur> tibbs|w: The draft doesn't say why a packager needs to "make every effort to avoid bundling".
16:18:05 <geppetto> tomspur: I don't think that's a thing anymore
16:18:06 <racor> +1 ... under protest!
16:18:08 <tibbs|w> tomspur: That sentence was in the original policy.  I can delete it as well, but usually we don't include justification.
16:18:27 <tibbs|w> I guess my leaving it in was a bit of passive aggression.
16:18:44 <tomspur> But I can understand, that you want to keep it simple and wait for the "release early, release often" attitude of FESCo.
16:19:28 <tibbs|w> +1 for the record.
16:20:20 <tomspur> Well, +1...
16:20:35 <geppetto> mbooth: Rathann: vote?
16:20:51 <Rathann> eh
16:21:23 <Rathann> what if it's not possible to determine the version?
16:21:36 <tibbs|w> Anyway, I expect we'll need to refine things further and I don't think FESCo would complain as long as we don't put ourselves back in the loop.
16:21:49 <tibbs|w> Rathann: No clue; that's the exact tag FESCo told us to use.
16:22:10 <tibbs|w> At this point I didn't want to deviate materially from the language they instructed us to implement.
16:22:17 <mbooth> Yeah +1
16:23:07 <Rathann> *shrug* +1
16:23:49 <geppetto> #action Bundling Guidelines Overhaul (+1:7, 0:0, -1:0)
16:24:38 <geppetto> #info For what it's worth nobody in FPC seems to be of the opinion that this was a good change.
16:24:44 * geppetto shrugs
16:24:59 <tomspur> geppetto: thanks
16:25:02 <geppetto> #topic #567 	Packaging Python 3 applications and modules for EPEL 7+
16:25:03 <geppetto> .fpc 567
16:25:03 <geppetto> https://fedorahosted.org/fpc/ticket/567
16:25:04 <zodbot> geppetto: #567 (Packaging Python 3 applications and modules for EPEL 7+) – fpc - https://fedorahosted.org/fpc/ticket/567
16:25:13 <geppetto> And now back to python policy :)
16:26:06 <tibbs|w> So, tomspur and I have been playing with this.
16:26:36 <tibbs|w> It's not done; in particular it doesn't yet handle %files or additional dependencies for the subpackages.
16:26:54 <tibbs|w> But it's getting there. and personally I think it's kind of nice.
16:27:10 <tomspur> spectool-py looks nice.
16:28:27 <tibbs|w> tomspur: Yeah.  Right now it doesn't download, but it at least gets the right sources and patches for me where often spectool didn't.
16:28:33 <tomspur> Hmm, does it make sense to have denyhosts built for several implementations?
16:28:50 <tibbs|w> tomspur: No, that's just a spec I was using to test the spectool thing.
16:29:03 <tibbs|w> Not the macros thing.
16:29:05 <tomspur> ok
16:29:21 <tibbs|w> Anyway, I've set out to make sure these macros don't conflict with any existing usage.
16:29:57 <tibbs|w> I need to do a mass rebuild of all python packages locally to make certain that they don't confuse anything.
16:30:46 <orionp> How do you envision specifying requires looking like?
16:30:48 <tibbs|w> Anyway, more work to do on that, but if anyone has any suggestions, please let me know.
16:31:13 <tibbs|w> orionp: Probably either arguments to the subpackage section or some other macro to set them.
16:31:31 <tibbs|w> Auto-substituted for different python versions, of course
16:31:57 <orionp> can you set a variable/macro in lua that can then be referenced later in another macro?
16:32:18 <tibbs|w> Yes.
16:32:25 <geppetto> really?
16:32:26 <tibbs|w> Look at %py_init.  It sets global functions.
16:32:31 <geppetto> So it's the same lua instance for each package?
16:32:36 <tibbs|w> Indeed it is.
16:32:51 <tibbs|w> There's actually a whole lot you can do that nobody seems to know about.
16:33:24 <tibbs|w> in lua, everything's global anyway unless you stick local in front of it, and that global table persists.
16:36:00 <tibbs|w> Anyway, I tried to comment everything so the magic isn't quite so magical.
16:36:10 * geppetto nods
16:36:15 <orionp> It's pretty nifty
16:36:24 <geppetto> Is there anything we want to vote on this week?
16:38:15 <tibbs|w> Unless folks want to say "please don't do this", I guess.
16:38:43 <tibbs|w> I'm basically going to throw in every convenience macro I possibly can and then if people think that hides too much then I can take that out of the package.
16:38:46 <Rathann> nope, I think it's looking good, thanks for working on it
16:38:57 <tomspur> Would it make sense to have a general fpc-macros package and not just python-macros?
16:39:14 <geppetto> Yeh, keep going we'll see how far the rabbit hole goes :)
16:39:33 <tibbs|w> I'll also draft the necessary guidelines changes to go with this.
16:39:59 <tibbs|w> I am actually finding this kind of fun.  I'll have a set of requests to the RPM devs once I'm done, that's for sure.
16:40:08 <geppetto> cool
16:41:31 <geppetto> #topic #558 	Application/Library distinction and package splitting
16:41:31 <geppetto> .fpc 558
16:41:32 <geppetto> https://fedorahosted.org/fpc/ticket/558
16:41:33 <zodbot> geppetto: #558 (Application/Library distinction and package splitting) – fpc - https://fedorahosted.org/fpc/ticket/558
16:41:55 <geppetto> So … more python stuff :)
16:42:11 <tibbs|w> I haven't messed with this.
16:42:21 <tibbs|w> Besides changing the title to make more sense.
16:43:00 <Rathann> not just python, this is generic
16:45:29 <tibbs|w> Anyway, I bodged up that draft but I'd really like it if someone took a pass over it.
16:45:58 * geppetto nods … although I think python is somewhat special here as it's somewhat common for a library to also contain the program bits if called as main etc.
16:46:06 * geppetto nods
16:46:33 <geppetto> #topic Open Floor
16:47:03 <geppetto> Anything anyone wants to bring up?
16:48:31 <tibbs|w> geppetto: Pretty common to most interpreted languages, really.
16:48:44 <tibbs|w> I closed all of the existing bundling tickets, so the count went down.
16:48:50 <geppetto> kind of
16:49:13 <geppetto> I feel like perl had a clearer split
16:49:39 <tomspur> orionp: Could you maybe test the python_provide macro etc on epel7?
16:49:40 <geppetto> between modules and programs … where a lot of python the "program" is just a couple of lines to run the module in program mode.
16:49:58 <geppetto> Maybe just me and my memory though
16:50:07 <tomspur> I'm about to push the latest changes of Fedora's python_provide macro there and all current macros should be out everywhere.
16:50:21 <geppetto> Anyway … thanks for coming!
16:50:31 <orionp> tomspur - I'll try
16:50:41 <geppetto> #endmeeting