16:00:15 #startmeeting 16:00:15 Meeting started Thu Apr 2 16:00:15 2015 UTC. The chair is tibbs|w. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:15 Useful Commands: #action #agreed #halp #info #idea #link #topic. 16:00:22 #meetingname FPC 16:00:22 The meeting name has been set to 'fpc' 16:00:24 #meetingtopic Packaging Committee meeting 16:00:31 Anyone around? 16:00:38 I am here 16:01:52 SmootherFrOgZ and spot are at least in the channel. 16:02:12 I'll wait a few. 16:03:00 Not particularly optimistic about making quorum, but at least I have an agenda ready. 16:03:27 * racor is here, too 16:03:47 * SmootherFrOgZ here 16:03:52 That's four. 16:05:16 Why did I ping spot? geppetto pinged him last week but he's not actually on the committee. 16:05:36 ... but it's quite likely I have to leave early. 16:05:36 Should check before cut-n-paste, I guess. 16:06:12 Assuming we don't hit quorum, anyone want to at least look at the one piece of new business this week? 16:06:18 #topic #520 [Guidelines Draft] Per-Product Configuration Defaults v2 16:06:20 .fpc 520 16:06:21 https://fedorahosted.org/fpc/ticket/520 16:07:10 Argh, someone at the door. 16:07:22 Anyway, this was a mess before and it's less of a mess now. 16:07:44 /me is around if there are questions 16:08:20 Diff is https://fedoraproject.org/w/index.php?title=User%3ASgallagh%2FPer-Product_Configuration_Packaging_Draft&diff=408394&oldid=408358 16:08:54 With mbooth we have 5. 16:09:13 mbooth: looking over https://fedorahosted.org/fpc/ticket/520 16:09:22 The diff is https://fedoraproject.org/w/index.php?title=User%3ASgallagh%2FPer-Product_Configuration_Packaging_Draft&diff=408394&oldid=408358 16:10:06 Personally I'm fine with this but I haven't linted it. 16:10:13 * mbooth looks 16:13:40 sgallagh: Do we expect the list of product deliverables to change often enough? 16:13:52 Well, often enough that it would be annoying to update the list in the guidelines? 16:14:07 We could just refer to an external list somewhere if one exists and it changes often. 16:14:30 Adding new Products requires an Act of Council, so should be extremely rare 16:14:37 Cool. 16:14:52 Hmm, I have no real complaints about this 16:15:21 So basically the big change is getting rid of all of the conflicts stuff and instead just moving the right thing into place in %posttrans. 16:15:31 yes 16:15:34 Which I heartily support as conflict are really annoying. 16:15:40 Agreed 16:15:49 However, can that be scripted? 16:15:51 And changed in subtle ways with DNF 16:16:05 Not realistically no 16:16:07 As in, call out to something in %posttrans that does this, because that's a lot of code. 16:16:24 And, really, way too much code. 16:16:43 Since you may be working with a subset of the Products or have multiple files that need to be munged (like Firewalld does) 16:17:13 So I can't think of a way to script this that wouldn't take just about the same amount of code in arguments 16:17:56 I'm surprised there's no rpm macro to get the product variant 16:17:58 I guess I could write a script for the trivial case... 16:18:07 I guess there aren't too many packages that would need this. Is there a list? 16:18:16 mbooth: I could write one of those, I suppose. 16:18:28 It would just be doing exactly what that sed command does 16:18:29 And, yes, at least getting the variant should really be a macro. 16:18:43 You mean awk command, I guess. 16:18:51 tibbs|w: As of today: firewalld 16:18:58 Heh, so just the one. 16:19:07 yes 16:19:08 Basically, anything we can make simpler, we should make simpler. 16:19:24 openssh-server may do this in the future as well, but that's ambiguous right now 16:19:36 I know /etc/os-release isn't going to change format, but even hardcoding that in the spec is something we should avoid. 16:19:49 /me nods 16:20:27 So, I'm generally +1 but anything that can make this any cleaner would be nice. 16:20:36 me too 16:20:53 Also trying to remember why you need to backwhack the ! in the first line of posttrans. 16:21:12 I copy-pasted that from elsewhere and just trusted it 16:21:26 I forget exactly where 16:21:26 how about %ghost'ing the symlinks? 16:21:38 I don't think bash does history expansion in shell scripts. 16:21:49 racor: Did I forget to write that? 16:21:55 Yeah, that's necessary. Oops 16:22:00 There's no %files section at least. 16:22:15 systemd and alternative apply similar approach, and ... the result is havoc 16:22:20 I did make a couple of minor formatting tweaks. 16:22:33 racor: Sorry, similar to what? 16:22:35 s/alternative/alternative/ 16:23:01 They apply symlinks as configuration items and do not %ghost them 16:23:15 Ugh. 16:23:44 I didn't realize that systemd did that. With alternatives I'm not sure there's a better way, but that's off-topic. 16:23:50 racor: Yeah, that was an oversight in the proposal. The %ghost is there in the real firewalld package 16:24:25 In cases of mistakes/bugs/breakdowns you typically end up with dangling/unowned symlinks, not knowing if something is still in use or what else. 16:24:43 So, add %files section, perhaps macroize the variant bit, and revisit next week? 16:25:20 With a macro, being able to write "case %{product_variant} in" would eliminate 33% of your steps in "Applying Configuration" ;-) 16:25:20 With that done I'm happy to vote +1 in the ticket. Not sure if this needs to get in urgently since it's just one package currently. Which I'm sure has already been updated. 16:25:26 cf /etc/systemd/system/* and /etc/alternatives/* 16:25:52 And I guess maybe check whether the ! really needs the backwhack. 16:26:40 Right, firewalld has already made these changes and it's working nicely. 16:28:10 So, anyone have any other comments? sgallagh: are those things OK? 16:28:27 (I really don't want to use the phrase "action items". Except I just did. I feel dirty.) 16:28:27 tibbs|w: Sounds reasonable to me. 16:28:47 #action sgallagh to add %files section and create a macro for product variant 16:28:59 Ninja'd. 16:29:01 Hmm, I'm probably not chaired 16:29:15 Oh, hell, I don't even know how to do that. 16:29:17 #action sgallagh to add %files section and create a macro for product variant 16:29:39 tibbs|w: "#chair " 16:29:56 #chair mbooth 16:29:56 Current chairs: mbooth tibbs|w 16:29:59 #chair racor 16:29:59 Current chairs: mbooth racor tibbs|w 16:29:59 .fas corey84 16:30:00 Corey84: corey84 'Corey84' 16:30:18 #chair SmootherFrOgZ 16:30:18 Current chairs: SmootherFrOgZ mbooth racor tibbs|w 16:30:19 tibbs|w: "#chair [*]" actually 16:30:37 Hopefully geppetto will come back soon. 16:30:45 Basically sets the people who are allowed to issue commands during the meeting 16:31:02 It doesn't acknowledge on action items anyway, I don't think. 16:31:39 So, we have a few tickets from past meetings we could look at. 16:31:52 tibbs|w: No, but it also wouldn't have put it in the minutes. 16:31:58 /me shuts up 16:32:23 Yeah, but I actioned it so it should be there. 16:32:47 Anyway, anyone want to look at any of the old ones? 16:32:56 mbooth, you did something with 513, I think. 16:33:03 #topic #513 Use python -Es in shbang meeting defect 16:33:04 .fpc 513 16:33:06 https://fedorahosted.org/fpc/ticket/513 16:33:25 Hmm, did I? 16:33:33 Man, I thought so. 16:33:54 It says "ACTION: tibbs Will work on macros to vote on next week." :-) 16:34:13 sorry, I spaced 16:34:20 #chair orionp 16:34:21 Current chairs: SmootherFrOgZ mbooth orionp racor tibbs|w 16:34:31 mbooth: Sorry, I'm dumb. It was 281. 16:34:40 #topic #281 New Python Macros for Easier Packaging 16:34:41 .fpc 281 16:34:43 https://fedorahosted.org/fpc/ticket/281 16:34:59 And it was orion. I know someone did something. 16:36:18 * orionp the macro conditionals make my eyes bleed... 16:36:24 orionp: Your macros look reasonable to me, but I haven't tested. 16:36:33 me neither :) 16:36:44 I guess that would be the next step. 16:37:27 And yeah, the conditionals are unpleasant, but doing it once saves a load of cruft in a big load of packages. 16:37:54 And lets us mess with the default options without having to change every package. 16:38:20 Though I suspect the first three macros should just be two macros. 16:39:14 Because I don't suspect both pythons will have the same macros for long, so I wouldn't want people getting used to setting just one of them. 16:40:30 Not really sure about the naming, but I don't really care. And I guess I'd use %optflags instead of the yelling equivalent, but that's just me. 16:41:12 Anyway, I guess a couple of us test this and make sure we're good? 16:41:34 Sounds like a plan to me 16:41:38 I have a couple of my python packages I can run this through. 16:41:42 sounds like a plan. 16:41:54 #action Test this a bit and revisit next week. 16:42:17 Anyone want to look at the other tickets? 16:42:37 I don't know what to do about scintilla; I guess someone needs to contact all of the involved maintainers and try to sort something out. 16:43:11 I don't understand bcond and the like well enough to deal with 509. 16:43:46 scintilla is https://fedorahosted.org/fpc/ticket/126, https://fedorahosted.org/fpc/ticket/509 is the bootstrapping thing. 16:44:15 I don't think any other tickets need immediate attention -- just all waiting on someone to do something 16:44:38 Yeah, but some of those might be waiting on us to do something. I'm not really sure. 16:44:58 I may look into the bcond/bootstrapping one since I have some packages that have bootstrapping 16:45:12 Anyway, if nobody has any further input on any of the followups, I guess we're done. 16:45:21 #topic Open Floor 16:45:38 Anyone have anything else they'd like to discuss? Anything I missed? 16:46:19 Sorry if I'm railroading, but I like short meetings. 16:46:56 Me too :) 16:46:57 And I have a pile of work to get through today. 16:47:17 Me too, feel free to end and enjoy the long easter weekend 16:47:52 Yeah, if you celebrate that or just enjoy a national holiday, enjoy it. 16:48:03 I'll close in a couple of nobody has anything. 16:48:13 We're all good with the time now that the world is back in sync, right? 16:48:30 Should be 16:49:31 OK, folks, thanks for coming. See everyone next week. 16:49:37 #endmeeting