16:00:57 <geppetto> #startmeeting fpc
16:00:57 <zodbot> Meeting started Thu Sep 17 16:00:57 2015 UTC.  The chair is geppetto. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:57 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:00:57 <geppetto> #meetingname fpc
16:00:57 <geppetto> #topic Roll Call
16:00:57 <zodbot> The meeting name has been set to 'fpc'
16:01:07 <geppetto> geppetto limburgher mbooth orionp racor Rathann SmootherFr0gZ tibbs|w tomspur: FPC ping
16:01:11 <tibbs|w> I'm around.
16:01:13 <orionp> morning
16:01:14 <tibbs|w> Sort of.
16:01:18 <geppetto> #chair tibbs
16:01:18 <zodbot> Current chairs: geppetto tibbs
16:01:20 <geppetto> #chair orionp
16:01:20 <zodbot> Current chairs: geppetto orionp tibbs
16:01:23 <mbooth> Evening
16:01:54 <geppetto> #chair mbooth
16:01:54 <zodbot> Current chairs: geppetto mbooth orionp tibbs
16:02:02 <mbooth> Looks like I missed an... eventful meeting last week :-)
16:02:24 <tibbs|w> Not the usual boring meeting indeed.
16:02:41 <geppetto> mbooth: Something like that :)
16:03:15 <geppetto> There was/is the giant thread on devel list though
16:03:45 <mbooth> I spent some time reading it today
16:03:54 <tibbs|w> I note that nobody has filed an FPC ticket.  I do have a proposal but haven't written it up.
16:04:13 <tibbs|w> Basically, make bundling "best effort" with FPC still the gatekeeper.
16:04:24 <geppetto> how is that different?
16:04:39 <geppetto> We imply we'll just let a lot more stuff through?
16:04:46 <tibbs|w> Essentially, yes.
16:04:51 <geppetto> hmmm
16:05:20 <tibbs|w> I know, I'm not sure I agree with my own proposal, but someone needs to actually throw up a strawman to compete with the original proposal.
16:05:41 <tibbs|w> Because I'll never go for getting rid of the restrictions altogether.
16:05:41 <geppetto> Well, I'm not sure we'll get anymore … as tom and rathan said they probably aren't going to make it
16:05:50 <tibbs|w> racor as well.
16:06:04 <tibbs|w> And we never get limburger these days.
16:06:19 <tibbs|w> Or Xavier.
16:07:05 <tibbs|w> Should probably think of trying to reschedule.  When is the next DST change, anyway?
16:07:26 <tibbs|w> Looks like Nov. 1
16:07:29 <geppetto> super late for the US
16:07:36 <geppetto> earlier for UK/etc.
16:07:41 <tibbs|w> Well, Nov 1 in idiot country.
16:07:55 <mbooth> Last weekend in Oct for UK, IIRC
16:08:18 <mbooth> So there will be some skew from my perspective
16:08:23 * geppetto nods … IIRC it's like 2 weeks before US
16:08:29 <geppetto> or 1 week, blah
16:08:39 * l0ft1369 is warming the butter for his popcorn
16:10:05 <geppetto> buttered popcorn is a really weird american thing
16:10:19 <l0ft1369> but it tastes so good though!!!
16:10:48 <tibbs|w> butter and parmesan cheese.
16:11:01 <l0ft1369> anyhow I thought I read somewhere about someone complaining about a developer or something and now with all the side chatter about last week's meeting...I'm excited for this one!
16:11:12 <l0ft1369> ew cheese on popcorn? really?
16:11:24 <tibbs|w> Well there's not going to be much of a meeting this week as we don't have quorum.
16:11:34 <geppetto> sugar is the only acceptable popcorn topping
16:11:43 * l0ft1369 frowns but eats his buttered popcorn anyhow
16:11:48 <tibbs|w> Cheese on popcorn is a New England thing.
16:11:53 <geppetto> Not really much tickets though, so it's not too bad
16:12:06 <tibbs|w> I still have a load of stuff to do.
16:12:06 <l0ft1369> no it's not...I spent 24 years on Cape Cod and have never heard of that
16:12:18 <tibbs|w> My family is from Maine.
16:12:24 <tibbs|w> Well, half of it.
16:12:27 <l0ft1369> ah...that's why haha
16:12:36 <geppetto> I guess it's a maine thing then?
16:12:37 <tibbs|w> Yes, they're weird up there.
16:12:42 <l0ft1369> I mean
16:12:54 <l0ft1369> I feel like those are your words...and I'm not gonna disagree with them ;-)
16:13:27 <mbooth> So, same bat-time, same bat-place next week, then?
16:13:47 <l0ft1369> that's it?!
16:13:55 * l0ft1369 exclaims and popcorn goes everywhere
16:14:09 <tibbs|w> Well, occasionally we talk about things.
16:14:16 <l0ft1369> haha
16:14:17 <mbooth> l0ft1369: I don't think we have quorum :-(
16:14:22 <geppetto> We don
16:14:25 <geppetto> 't
16:14:25 <tibbs|w> Can probably brainstorm about the python macro thing.
16:14:42 <orionp> Feedback would be welcome
16:14:50 <tibbs|w> I think it's possible for the almost all of it to be generated by the macros.
16:14:52 <l0ft1369> mbooth: let me google translate that q word to English real quick
16:15:01 <geppetto> But, yeh, we can talk about either the python thing or the file triggers thing … or the bundling thing … or anything else
16:15:01 <tibbs|w> I just have to have some time to sit down and play with it.
16:15:11 <tibbs|w> file triggers, well, there's not much to do.
16:15:46 <tibbs|w> If someone wants to make a list of which packages would be responsible for owning/including the triggers for each scriptlet on our pages, that would be great.
16:15:51 <geppetto> For file triggers I've got some work time to help with the conversion
16:16:00 <geppetto> So that should help it move along
16:16:08 <tibbs|w> Then we can track those to see when triggers get added and note that in the guidelines.
16:16:23 <tibbs|w> But if geppetto is getting paid to do it, then I think he just volunteered.
16:16:30 <geppetto> Yeh, pretty much
16:16:44 <tibbs|w> Eventually I think we'll want to file bugs unless you're just going to go ahead and make the necessary commits.
16:16:50 <tibbs|w> Which I hope you will.
16:17:36 <geppetto> both, my current plan is to help with commits for the "core" packages and just file BZs for the non-core ones
16:17:41 <l0ft1369> ha! googled it! :-( I don't count?
16:18:17 <geppetto> l0ft1369: You aren't on FPC committee … so, no :)
16:18:43 <tibbs|w> I'm happy for anyone to work on stuff.
16:18:43 <l0ft1369> geppetto: awww. Well if I had the time and the skills...haha
16:19:43 <tibbs|w> geppetto: So is that a plan for file triggers?  I guess you can attack that however you like.  But it would be super nice to have this in place for F23 assuming that RPM version stays in there.
16:20:54 <geppetto> tibbs: Yeh, not sure we can get it converted in time for F23
16:21:01 <geppetto> tibbs: But, yeh, that's my plan
16:21:13 <tibbs|w> Just getting the triggers added would be enough.
16:21:49 <tibbs|w> Though my understanding is that if the packages still have their own scriptlets, then everything will run twice.
16:21:57 <geppetto> Hmmm, end of Oct. for release
16:22:08 <tibbs|w> Well, maybe not twice if the trigger runs only at the end of the transaction.
16:22:21 <geppetto> Yeh, I know what you mean
16:22:22 <tibbs|w> And it can technically be done after release.
16:24:34 <geppetto> Yeh, we can def. do it in stages
16:24:52 <tibbs|w> And the python macro thing, I have an idea of what I'd like it to look like but I need to paste the example spec up.
16:25:01 <geppetto> Ok
16:27:21 <orionp> What would you like it to look like?
16:27:43 <tibbs|w> Pasting it up....
16:27:46 <tibbs|w> http://paste.fedoraproject.org/268523/50723114/
16:28:08 <tibbs|w> I mean, have the %python2_package and %python3_package macros expand to everything in the header.
16:28:38 <orionp> That's be nice.  No idea how to do it.
16:28:41 <tibbs|w> For EPEL, %py3_build/install can expand to the commands needed to build for both versions.
16:29:04 <tibbs|w> So at the end you have to deal with %files and %check specially.
16:29:08 <orionp> Ah, that's a good idea
16:29:17 <tibbs|w> I think it can be done, honestly.
16:29:41 <tibbs|w> The macros can take arguments and options if something needs to be overridden.
16:29:55 <tibbs|w> But I have to learn the freaking macro language, or lua, first.
16:30:06 <orionp> I don't see how you get the header info
16:30:10 <tibbs|w> And find the time.  I'm almost done with one big project.
16:30:18 <tibbs|w> Well, what info is in the header?
16:30:33 <orionp> All BR/Rs
16:30:46 <orionp> any special provides/obsoletes
16:30:46 <tibbs|w> No, those go up top.
16:31:15 <tibbs|w> Special provides/obsoletes can be passed or pre-defined.
16:31:28 <orionp> so how do the subpackages pick up the Rs?
16:31:39 <orionp> Requires
16:31:44 <tibbs|w> We'd have to pass individual requirements in the macro.
16:31:51 <tibbs|w> Or, to the macro, rather.
16:32:16 <tibbs|w> Either that or define them and the macro can add python2- or python3- as appropriate.
16:32:29 <tibbs|w> Which would work for most cases.  The corner cases will have to be looked at a bit.
16:32:41 <tibbs|w> And since python2- providing is not yet universal, that would be slightly fun.
16:33:12 <tibbs|w> But I think it can be worked out.  It's just so damn difficult to get python macro changes pushed out.
16:33:41 <orionp> I think we need to put the macros in a separate packge
16:33:51 <tibbs|w> I've been saying that from day 1.
16:33:59 <tibbs|w> Anyway, how do you actually hack on this stuff?
16:34:12 <tibbs|w> RPM won't show you the expanded spec file.
16:34:27 <geppetto> rpmspec command
16:34:29 <tibbs|w> You have to rebuild the whole damn python package and stick it in a local repo to test things in mock.
16:34:29 <orionp> rpmspec --parse
16:34:44 <tibbs|w> Oh, cool.  When did that start existing?
16:34:58 <geppetto> Ages ago
16:35:00 <tibbs|w> And can one macro file override another?
16:35:04 <orionp> Locally I just edit the macros.python file
16:35:23 <orionp> You could build a local python-macros package that overrode the python one perhaps
16:35:37 <tibbs|w> Yes, but it would be good to do proper mock testing so we (I) don't destroy rawhide.
16:36:07 <orionp> hmm, requires specific python-macros version unfortunately
16:36:14 <tibbs|w> Yeah, ugh.
16:36:23 <tibbs|w> If we wanted it split, what would we need to do?
16:36:58 <tibbs|w> I guess the necessary dependencies are already in place.
16:36:59 <orionp> Just drop the version info in the python2-devel requires
16:37:19 <orionp> make a separate python-macros srpm
16:38:48 <orionp> gotta go for a bit...
16:39:27 <tibbs|w> I still don't know if people would puke at the sight.
16:39:44 <tibbs|w> But rpm itself has worse macros already, like the one which generates the debuginfo package.
16:40:18 * geppetto nods … that is invisible though
16:40:30 <tibbs|w> Well, it is until it gives you problems....
16:41:00 <tibbs|w> But indeed.
16:41:16 <tibbs|w> I don't think anything we use  parses spec files for %package declarations.
16:41:42 <tibbs|w> And maybe the macros should be %package_python2/3 so they look more like regular %package.
16:43:03 <tibbs|w> Anyway, I'm seeing the light at the end of the tunnel for my big projects but at this point I'm not going to make any promises about this.
16:43:30 <geppetto> ok
16:44:14 <geppetto> Anyway … it's been 45 minutes, and we still don't have 5 … so we want to go do other stuff until next week?
16:49:42 <geppetto> Ok, going to close in a minute
16:49:49 <geppetto> See you all next week
16:50:25 <geppetto> #endmeeting