16:01:19 #startmeeting fpc 16:01:19 Meeting started Thu May 3 16:01:19 2018 UTC. 16:01:19 This meeting is logged and archived in a public location. 16:01:19 The chair is geppetto. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:01:19 Useful Commands: #action #agreed #halp #info #idea #link #topic. 16:01:19 The meeting name has been set to 'fpc' 16:01:19 #meetingname fpc 16:01:19 #topic Roll Call 16:01:19 The meeting name has been set to 'fpc' 16:01:29 Howdy. 16:01:30 hi there 16:01:33 #chair tibbs 16:01:33 Current chairs: geppetto tibbs 16:01:33 hi there 16:01:36 #chair mhroncok 16:01:36 Current chairs: geppetto mhroncok tibbs 16:01:41 #chair decathorpe 16:01:41 Current chairs: decathorpe geppetto mhroncok tibbs 16:02:33 redi said he might be a few minutes late 16:06:26 hello2 16:06:28 .hello2 16:06:29 ignatenkobrain: ignatenkobrain 'Igor Gnatenko' 16:06:35 hey! 16:06:38 how goes? 16:06:39 #chair ignatenkobrain 16:06:39 Current chairs: decathorpe geppetto ignatenkobrain mhroncok tibbs 16:06:51 #topic Schedule 16:07:11 https://lists.fedoraproject.org/archives/list/packaging@lists.fedoraproject.org/message/RESPZUIX76BHKC6DCATCSRHEQ7ZFPKHN/ 16:08:19 sorry I'm late 16:08:24 #chair redi 16:08:24 Current chairs: decathorpe geppetto ignatenkobrain mhroncok redi tibbs 16:08:30 No problem, close enough 16:08:46 No new tickets this week 16:08:48 #topic #691 noarch *sub*packages with arch-specific dependencies 16:08:51 .fpc 691 16:08:56 geppetto: Issue #691: guidelines change: noarch *sub*packages with arch-specific dependencies - packaging-committee - Pagure - https://pagure.io/packaging-committee/issue/691 16:09:03 So, just a minor wtf can we do update from tibbs here 16:09:11 And … I'm not sure. 16:09:39 * limburgher here 16:09:45 #chair limburgher 16:09:45 Current chairs: decathorpe geppetto ignatenkobrain limburgher mhroncok redi tibbs 16:10:06 This is mostly old because it needed someone to write drafts, probably after talking to rel-eng people and maybe doing tests 16:10:37 I guess nobody volunteered to do test build in koji? ;) 16:11:03 It's one of the things that people expect us to do, but which we aren't necessarily going to do. 16:11:29 Well, it's doing a bunch of different tests builds and seeing what the results are for each … esp. as some of the problems only happen sporadically 16:11:37 I don't know if anyone really knows what will happen, or even knows what is supposed to happen. 16:11:57 But you can't just do test builds if you want to test the whole process. 16:12:34 Imagine if we wrote something into the guidelines which, if followed, broke the rawhide compose. 16:13:08 so we'd have to create enough real "test" packages to fill the test matrix for exclusivearch noarch subpackages, and run each of the tests multiple times? 16:13:21 sounds like fun (not) 16:13:44 this still sounds more like a releng issue 16:13:48 That's pretty much it if we want to do the whole job. 16:14:20 And yes, it is fundamentally a releng issue. But people look to the packaging guidelines to document the packaging end of this kind of thing. 16:14:43 well, but the packaging stuff can't be documented if it's unclear what the tooling supports 16:15:08 11 months ago: ACTION: We need to ask someone in releng what policy should be and then write a draft (geppetto, 17:27:06) 16:15:17 indeed 16:15:18 did we actually ask anybody? 16:16:01 I mean, tibbs says "I haven't had much luck getting real answers", but is there a releng ticket or anyhting we can link and mark it as blocked? 16:16:03 I thought tibbs had spoken to someone, but they weren't 100% clear on the answer 16:16:18 I don't think we opened a rel eng ticket 16:16:32 tibbs: any better memories? 16:16:40 It's been a long time and some of this is bound up with the other changes related to just noarch packages (not subpackages). 16:17:44 Could we just open a ticket then and ask the questions that way? 16:17:56 We *can*... 16:18:01 Even if we get the answer … "Nobody knows" … at least that's something. 16:18:10 And we should probably at least start with that. 16:18:43 Any volunteers to open the ticket then? 16:19:18 I will file one with releng and see what happens. 16:19:33 But I'm not 100% sure what exactly needs to be asked. 16:19:52 Probably because I haven't gotten much sleep lately and have lost my reading comprehension. 16:20:07 That would be an impediment. 16:20:18 I don't want to dump everything on you then … mhroncok you think you could do it? 16:20:26 geppetto: sure thing 16:20:31 Ok, cool. 16:20:42 Basically I think it's "please tell us what we should tell packagers who have noarch packages and subpackages which have dependencies (either build or runtime) not available on all architectures" 16:21:05 #action mhroncok Will open a releng ticket to see if they know the answer to the koji questions. 16:21:26 #topic #694 Packaging guidelines for application independence 16:21:29 .fpc 694 16:21:31 geppetto: Issue #694: Packaging guidelines for application independence - packaging-committee - Pagure - https://pagure.io/packaging-committee/issue/694 16:21:33 New draft on this one. 16:21:45 https://fedoraproject.org/w/index.php?title=PackagingDrafts%2FIndependentApplications&type=revision&diff=517485&oldid=517484 16:22:44 I think the first paragraph puts things the wrong way. 16:22:52 generally I like idea 16:23:21 If a package contains both an application and a library, usually the "main" package contains the application and a "-libs" sub-package contains the libraries, not the other way round 16:23:35 decathorpe: +1 16:23:49 but yes, generally it's a good idea to clarify this 16:24:41 +1 16:25:02 decathorpe: I think they are talking more about Eg. client DB library vs. GUI application to manage it. 16:25:12 Maybe we shoudl rephrase that instead to cover both cases 16:25:17 decathorpe: Where a bunch of things would think of the client DB library as the main package. 16:25:26 sure: I was just saying that it's usually the other way around 16:25:35 geppetto: I think that's probably what they're trying to say, but that's not what it actually says. 16:25:42 Yeh, now's the time to tweak the wording :) 16:26:09 The problem is that "subpackage" can mean two things depending on context. 16:26:37 I think the "Library or Application" paragraph tries to differentiate the two cases 16:26:38 But from the perspective of someone writing a specfile, it means you have a %package statement and you put the application in that. 16:26:46 "Many packages, regardless of their primary purpose, include both applications and libraries. In this case, one or both SHOULD be packaged in a subpackage …" 16:27:14 yeah, sounds better 16:27:36 Yeah, though generally we do want the libraries to be the subpackage. I'm not sure if we say that. 16:28:10 do we? 16:28:23 libfoo providing libfoo-utils should be the exception. 16:28:27 I mean, that clearly is different for each case 16:28:30 Ok, I changed the draft 16:28:52 I don't think we do, though the maing guidelines do mention "-libs subpackages". I don't believe it's formalized anywhere. 16:29:02 In any case, that would be for a separate ticket. 16:29:38 yeh 16:31:12 I'm ok with the wording now 16:31:12 I like the current dfraft 16:31:16 should we vote? 16:31:28 I believe this is the diff now: https://fedoraproject.org/w/index.php?title=PackagingDrafts%2FIndependentApplications&type=revision&diff=517590&oldid=495753 16:31:32 We can vote on whether. . .nahhh. 16:31:35 Yeh 16:31:37 +1 16:31:40 +1 16:31:48 +1 16:31:53 +1 16:31:56 +1 16:32:13 ignatenkobrain: vote? 16:32:20 +1 16:32:34 redi: vote? 16:35:02 #action Packaging guidelines for application/lib independence (+1:6, 0:0, -1:0) 16:35:34 #topic #719 Simplify packaging of forge-hosted projects 16:35:38 .fpc 719 16:35:40 geppetto: Issue #719: Simplify packaging of forge-hosted projects - packaging-committee - Pagure - https://pagure.io/packaging-committee/issue/719 16:36:10 I really don't think this is cooked through yet 16:36:18 (see the comment I wrote today) 16:37:24 yeh, I agree with what you wrote 16:37:51 so we need them to tweak the draft to not use current date, I guess? 16:37:55 the "Release" tag can't just depend on when koji gets around to build the package ... 16:38:08 agreed 16:38:16 * ignatenkobrain reads 16:39:11 yeah, I definitely don't want it to use the current date 16:39:14 I think we can fix it by getting date of package by checking timestamp of SOURCE0 16:39:20 then it should be reproducible 16:39:58 that would probably be better, but it's still too opaque for me 16:39:58 That's assuming a lot about how SOURCE0 is put in place. 16:40:56 AFAICT, the only sure thing would be to have the packager set %commitdate (or %revdate or whatever) manually. 16:41:01 Bottom line is that this stuff is really complicated and needs a lot of review. 16:41:10 decathorpe: I think I agree with you. 16:41:13 but isn't SOURCE0 url generated by %forge macro? 16:41:33 Sure, but why does it matter what generated it? 16:41:36 its %forgeurl 16:42:03 but that doesn't change the fact that the date (and Release tag) is recalculated every time the src.rpm is rebuilt (in koji too) 16:44:24 move on? this being bad has been aid on the ticket 16:44:28 *said 16:44:39 Well no other tickets have updated 16:45:03 #topic Open Floor 16:45:11 Anything anyone else wants to talk about? 16:45:32 Well 723 did update, but I haven't had a chance to read it. 16:45:37 nope =( 16:45:50 still didn't have time to work on git-based guidelines 16:46:08 I did write up the ruby stuff but am not sure if I'm supposed to announce it or wait for some change to the related macro package. 16:46:21 Not here. 16:47:05 Ah, there was a reply on the ruby ticket. 16:47:22 I will take care of that, then. 16:47:24 tibbs: Ahh, that needs to move back into meeting 16:47:35 tibbs: 723, that is 16:47:47 I haven't marked 723 with meeting 16:47:54 becasue I supposed nobody had tiem to read that 16:47:56 I just did. 16:48:01 so i was wating for after this one 16:49:34 ok 16:50:16 We seem good then, I'll close 10 minutes early, and give everyone a few minutes of their day back. 16:50:20 I have been trying to touch a couple of old tickets every so often. 16:50:49 At least soon we will get under 40. There are probably a load of cruft that should just be closed at this point. 16:51:04 For sure. 16:51:36 But at least the ticket list fits on one page now. 16:51:50 yay 😂 16:52:03 \o/ 16:52:58 #endmeeting