20:00:00 #startmeeting FESCO (2010-03-02) 20:00:00 #meetingname fesco 20:00:00 #chair dgilmore notting nirik skvidal Kevin_Kofler ajax pjones cwickert mjg59 20:00:00 #topic init process 20:00:01 Meeting started Tue Mar 2 20:00:00 2010 UTC. The chair is nirik. Information about MeetBot at http://wiki.debian.org/MeetBot. 20:00:03 Useful Commands: #action #agreed #halp #info #idea #link #topic. 20:00:05 * skvidal is here 20:00:06 The meeting name has been set to 'fesco' 20:00:07 Current chairs: Kevin_Kofler ajax cwickert dgilmore mjg59 nirik notting pjones skvidal 20:00:19 why hello there. 20:00:26 come on party people, throw your hands in the air. 20:00:32 * thomasj lurking 20:00:38 morning everyone. 20:00:40 * cwickert hides 20:00:41 come on party people, wave them like you don't care. 20:01:01 /o/ 20:01:03 \o\ 20:01:03 nirik, morning? 20:01:08 Present. 20:01:13 Hi 20:01:31 jwb: http://x0.no/4bl 20:01:32 jwb: it's a spherical planet. 20:01:46 jwb: he's hada long day 20:02:28 ok, shall we go ahead and start in. 20:02:51 #topic #343 cloture rule/procedure for fesco meetings 20:02:58 D: 20:03:03 ok, there was feedback in the ticket on this one... 20:03:40 how would folks feel about trying a '15minutes per topic' and require either a vote to continue or next topic at the end of that time? 20:03:44 * dgilmore is here 20:04:04 nirik: that would be fine w/me 20:04:05 I know we tend to discuss things at length sometimes... 20:04:05 nirik: that sounds a lot better than a cloture rule. 20:04:16 nirik: sometimes we need more than 15 minutes 20:04:27 but sure 20:04:27 i'm ok with a time limit. might need some tweaking per-issue 20:04:33 dgilmore: hence "continue" being an option 20:04:37 dgilmore: agreed, so we could vote then to continue discussion? simply majority of present members? 20:04:41 if all agree to continue, we can go on 20:04:54 simple majority is nice and... simple ;) 20:04:57 cwickert: all? or majority? 20:04:58 nirik: sounds fine 20:05:01 s/all/majority 20:05:07 majority sounds fine. 20:05:10 Sounds good to me 20:05:51 any objections? Kevin_Kofler? 20:05:59 +1 for 15 minutes with extension by simple majority vote. 20:06:13 +1 20:06:17 +1 20:06:17 15 minutes with possibility for extension makes sense to me. 20:06:19 +1 20:06:20 +1 20:06:24 0 20:06:29 +1 here. 20:06:30 +1 20:06:42 I.e. I don't think it's needed in the first place, but I can accept it. 20:06:43 i count +7, that's sufficient. 20:06:54 It's much better than shutting up folks with no discussion. 20:07:03 #agreed will use a '15 minutes per topic, majority vote of present members to extend discussion." 20:07:36 it's worth noting that upstream for the meetbot plugin we are using is working on setting up voting... so it gets reported better/the bot can tally, etc. 20:07:49 #topic #344 Policy proposal: contributing to Fedora should be FUN 20:07:55 .fesco 344 20:07:57 nirik: #344 (Policy proposal: contributing to Fedora should be FUN) - FESCo - Trac - https://fedorahosted.org/fesco/ticket/344 20:08:03 I added my thoughts to the ticket. 20:08:12 I think that contributing to fedora should be fun, and that hans should feel free to put that on the wiki someplace. 20:08:56 I think we should keep it in mind, but it's hard to measure and will be different for different people, so it's hard to quantify. 20:09:13 which makes it kind of a bad thing to encode as a rule or policy. 20:09:26 the proposal does not actually ask us to quantify it. 20:09:37 Sometimes things will occur that some group feels is necessary and which some other group feels reduces their fun 20:09:42 i agree with nirik that you can't really define 'fun' for everyone. for some, fun is starting flamewars or being a troll. 20:09:43 yeah, but this is the internet. people will argue it. 20:10:18 I think it's valuable to keep in mind that any policies produced by fesco impact a wide range of people, but codifying this as "fun" doesn't seem like a great way of ensuring that 20:10:19 It's kinda hard to define "fun" indeed, this proposal feels horribly vague to my mathematician soul. 20:10:20 notting: some? around here... ;) 20:10:43 and there will always be things in fedora that aren't considered 'fun', but are necessary 20:10:58 On the other hand, if keeping that principle in mind can reduce some of the useless bureaucracy, it might be a good thing. 20:11:18 I'm just a bit hesitant about voting something into policy which doesn't have a clear definition. 20:11:29 one could argue that the package review process isn't 'fun'. and yet i don't think we're going to remove it 20:11:36 It's not going to have any impact on "useless bureaucracy" 20:11:44 Freedom, Friends, Features, First, and Fun :P 20:11:50 If the people enacting it felt it was useless, they wouldn't be enacting it 20:12:06 or perhaps some people find useless bureaucracy fun? 20:12:12 Bureaucracy is not fun. 20:12:28 Except for sadomasochists, I guess. ^^ 20:12:32 you are all debating the thing you said can't be debated. move on 20:12:35 and bureaucrats. 20:12:45 If it's fun for someone, it's not useless? 20:12:49 jwb: but it's fun! ;) 20:12:49 pjones: Is that different? ^^ 20:12:56 Anyway 20:12:56 right, so whats the answer to this ticket? we will try and keep fun in mind, but can't codify it at this time? 20:13:02 nirik: That's my feeling 20:13:13 can we just close the ticket "worksforme" and go on? ;) 20:13:17 yawn, platitudes. -1 to having a Fun Policy. 20:13:18 i'd doubt we could codify it ever 20:13:22 I think we should put something somewhere that says that people should try to keep it in mind. 20:13:32 pjones: really? 20:13:34 no. 20:13:38 okay, good 20:13:49 but thanks for asking! 20:13:50 b/c it seems like something we can't really remind anyone of 20:14:04 "if you hate yourself everyday, clearly you need a new job/hobby" 20:14:14 hm 20:14:18 We're supposed to remind ourselves of it according to that proposal. 20:14:21 jwb: stop thinking 20:14:50 ok, any objections to closing this as 'we will try and keep fun in mind, but won't codify it' ? 20:14:56 I feel reminded already. 20:15:10 nirik: no objections here 20:15:12 if none, then I guess we don't need to vote? or should we vote for forms sake? 20:15:22 nirik: well my fun and yours are different 20:15:25 and might clash 20:15:27 nirik: I think that's the best we can really do. Especially as, you know, the engineering steering committee, not the Board. 20:15:30 dgilmore: indeed. 20:15:31 kinda hard to codify that 20:16:07 nirik: i approve of that closure text. 20:16:42 #agreed While fesco thinks fun should be kept in mind and contibuting to fedora be fun, we can't codify this. 20:16:55 #topic Fedora Engineering Services Tickets/Discussion 20:17:10 I'd like to take a few minutes to go over the FES tickets that are in progress/filed. 20:17:26 #topic FES: #2 SIGs roundup and pinging - jds2001 20:17:31 jds2001: you around? 20:17:47 * nirik suspects not. I meant to ping him before meeting, but failed. 20:18:03 i filed fes #7, and it seems to be moving along nicely 20:18:07 #topic FES: #8 Document Fedora as android devel platform - stickster 20:18:18 stickster is in another meeting, but working on it. 20:18:30 #topic FES: #5 Fix broken dependencies - mmcgrath 20:18:34 mmcgrath: ? 20:18:47 I'm going through this now, focusing on a release at a time starting with rawhide. 20:19:01 cool. 20:19:05 I've gotten a couple fixed already, one major offender is ecore 20:19:13 It's being worked on and should be fixed soon. 20:19:20 are you finding that most maintainers are aware of the problem? or not? 20:19:35 One question I had is if it's rude to request a rebuild on stuff that just needs to be built against newer libs or not. 20:19:48 nirik: I think I've surprised some of them yeah. 20:19:49 doesn't sound rude to me. 20:19:59 * nirik doesn't think it's rude at all. 20:20:12 mmcgrath: Just bump, commit and submit the build. 20:20:13 mmcgrath: interesting. Did they file/filter the broken deps mail? 20:20:17 I've done that a lot. 20:20:23 and I think people have generally have responded well to someone actually contacting them personally instead of just random nagmail. 20:20:34 If the dependency is broken and the maintainer doesn't care, what else should we do? 20:20:42 nirik: not sure. and I'm not even sure how often that goes out. 20:20:43 the impersonal nag mails are... easy to ignore. 20:20:56 mmcgrath: nightly with rawhide. 20:21:05 * thomasj knows it goes out daily 20:21:10 and that sends the report mail and a personal email right? 20:21:14 mmcgrath: yes 20:21:20 for the stable ones, whenever mshwent runs it. 20:21:25 I've gotten the individual emails 20:21:29 nirik: Sorry to be late here -- but I actually talked with a friend today who gave me some helpful links and pointers, so even though I haven't written anything, I've started thinking about it :-) 20:21:30 when I broke yum :) 20:21:38 * nirik just totally failed that name. 20:21:46 stickster: excellent. 20:21:57 * stickster back to other meeting, ping if needed 20:22:02 mmcgrath: anything more on this ? the next ticket is yours too... 20:22:13 #topic FES: #7 spec cleanup task: fix the need for perl (etc) in scriptlets - mmcgrath 20:22:22 nirik: that's all I've got. If I get yelled at for rebuilding something at least I know some FESCo people have my back :) 20:22:34 That was an interesting one to work on because it was tricky to actually look for that stuff. 20:22:47 mmcgrath: im my experience 99% of people dont mind 20:22:57 fortunately the only real bad apples were both java, I've sent an email to panu to figure out how rpm lua macros work. 20:23:19 cool. 20:23:23 And I've got to say, I really like this format. I've said I'm going to work 4 hours a week, I work my 4 hours and I'm done. 20:23:27 mmcgrath: the biggest issue with the rpm lua macros is going to get people to start using them beyond that 20:23:31 I like that there's interesting things to do, but I know exactly what my commitment is. 20:23:49 skvidal: yeah. still, every bit helps. 20:23:51 skvidal: are lua macros extended in rpm itself? or can I write my own lua macro in the spec file? 20:23:58 mmcgrath: both 20:23:59 mmcgrath: the latter. 20:24:00 ajax: not disputing that - just an uphill battle 20:24:01 * mmcgrath has 0 experience in that at the moment. 20:24:13 ah 20:24:16 mmcgrath: having some good examples wouldn't make anyone cry a river. 20:24:27 well I have some further education to do then 20:24:27 fixing things like this also means others can look at examples / see how to do it. 20:24:29 skvidal: do you know of any? 20:24:41 mmcgrath: istr some lua examples in the rpm.org docs 20:24:45 mmcgrath: rpm --showrc | grep lua, finds three semi-trivial examples. 20:24:52 ajax: excellent, thanks. 20:24:56 skvidal: my concern has always been "if you have to resort to lua, please doublecheck what you're doing first, because ideally you shouldn't" 20:25:04 so I'll keep working on that. I expect to have it done next week. 20:25:12 notting: well - lua in rpm is better than pulling in perl :) 20:25:14 mmcgrath: excellent. 20:25:20 notting: in the java example it's an absolute path to relative path replacement thing. 20:25:45 #topic Unassigned FES tickets 20:26:02 we also have 2 unassigned tickets for folks thinking of siging up to work on them: 20:26:06 #4 tool idea: script to evaluate buildroot poisoning 20:26:06 #6 Fix packages that fail to build from source 20:26:21 see mmcgrath to join FES. :) 20:26:34 mmcgrath: have you looked at advertising more now? or want to keep it growing slowly? 20:26:44 anyone who writes #4 wins their beverage of choice from me 20:26:45 mmcgrath: I've been thinking about #4 some 20:26:51 ajax: I'm big on 'water' 20:26:56 mmcgrath: https://bugzilla.redhat.com/show_bug.cgi?id=490740 also has some (sortof) examples 20:27:02 skvidal: it'll be the best dang water you ever had! 20:27:06 nirik: actually I think I have enough to do more advertising now. If we end up with people looking for work and not able to do it, I'll throw some random scrips their way to work on (locking, logging, error handling, etc) 20:27:10 pjones: thanks 20:27:18 ajax: the biggest trick is knowing the evr it was built against 20:27:30 ajax: vs just finding any pkg built with any of these pkgnames 20:27:37 mmcgrath: I can file some more too. I'm full of ideas. ;) 20:27:40 nirik: I have some primitive script to look up buildroot poisoning. 20:27:56 Kevin_Kofler: cool. :) Add link/script to that ticket? 20:27:57 skvidal: koji can tell you the list of BRs an NEVR was ever in, i'm told. 20:28:02 But the problem is, it can do millions of Koji lookups if the period of poisoning was more than a couple days. 20:28:06 skvidal: I haven't even looked at that one yet but I'll make sure to ping when the time comes. 20:28:08 So it doesn't terminate in any decent time. 20:28:14 https://fedorahosted.org/fedora-engineering-services/ticket/4 20:28:31 i'm okay with this kind of thing running slowly. after all, if we get to the point where we need it, we're in a bad place already. 20:29:43 ok, so if anyone has good ideas for tickets file them... 20:29:50 anything else on FES today? 20:30:24 ajax: Not if "slowly" is days or weeks. 20:30:42 Doing Koji queries all the time so poor Koji gets hammered with requests. 20:31:01 Kevin_Kofler: sounds like you need to look into multicall 20:31:14 Kevin_Kofler: perhaps your script could be optimized. 20:31:22 weeks would be unpleasant, but again: if we really need this to work, we probably don't want koji accepting tasks at all. 20:31:51 Kevin_Kofler: then post your script - and we can start from there 20:31:52 The cases I used it for was finding stuff being built against a bad Qt or kdelibs, which would break the package, but not cause security issues. 20:31:59 I'll post what I have. 20:32:33 ok, we have 3 new tickets that were filed after the agenda went out... do we want to discuss any of those today? or leave them for next week? 20:32:49 (and I ran the checker after a fixed build went in, so blocking Koji wouldn't have helped at all) 20:33:15 two of them were just filed, so should probibly wait... we could discuss "#345 Dependency chain painpoints" perhaps though if folks would like. 20:33:24 i'd prefer next week, i'm just a touch busy today. 20:33:28 likewise. 20:33:31 nirik: i'm ok with discussing them, but can wait too 20:33:33 sounds fine to me. 20:33:40 * nirik is ok either way as well. 20:33:47 #topic Open Floor 20:33:56 Anyone have anything for open floor? 20:34:03 i don't mind if people discuss, but if they do, i'll be reading it later. 20:34:51 I don't think any of those items are urgent... 20:35:36 ok, if nothing comes up in the next minute or less I will close out the meeting... 20:36:57 ok, thanks for coming everyone@ 20:37:04 #endmeeting