21:08:30 #startmeeting 21:08:30 Meeting started Wed Jun 15 21:08:30 2011 UTC. The chair is lightguard_jp. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:08:30 Useful Commands: #action #agreed #help #info #idea #link #topic. 21:08:53 #topic Git Flow change target date 21:09:46 I think we're all in agreement we want to use git flow 21:09:53 +1 21:10:06 We need to set a date for module conversion 21:10:21 I'm thinking two or three weeks out. 21:10:27 What does everyone else think? 21:10:31 after JAXConf for sure 21:10:52 I am using git flow on Seam Reports already 21:10:52 sounds good 21:11:08 It rocks :) 21:11:47 Okay, three weeks it is 21:12:11 #agreed git flow conversion will be done for all modules by July 5, 2011 21:12:43 The core devs will fix the modules that aren't done by that week 21:12:45 update contributor page should be in there too 21:12:53 yeah 21:13:10 in addition, the contributor page should have some text about why and how you fork 21:13:14 that was just missing from before 21:13:29 #info contributor page needs to be updated by July 5, 2011 to reflect the change to git flow 21:13:32 wait, sorry 21:14:18 what is missing is how to pull in changes from upstream 21:14:24 it's a minor addition 21:14:33 I think we had talked about it on the mailinglist, but it never made it to the contribute page 21:14:43 the section should be "staying up to date" 21:15:03 #info pulling changes, and which branch to branch from need to be added to contributor page 21:15:07 though I may have heard this is automatic now...anyway, we should mention whatever reflects reality 21:15:46 Who did the current one, was that Cody? 21:15:48 Or Brian? 21:15:56 think it was bleathem 21:16:05 it was me 21:16:16 bleathem: Would you mind taking action item? 21:16:21 * bleathem hides 21:16:28 haha 21:16:37 Hiding because you know I'll follow up with you huh? 21:16:38 lol 21:17:04 #action bleathem will update the contributor page by July 5, 2011 21:17:11 So let it be written so let it be done 21:17:13 :) 21:17:41 Any further questions about git flow? 21:17:44 haha, the Ten commandments 21:17:54 Great show, great line 21:18:05 cool 21:18:14 Okay, moving on then. 21:18:17 No questions 21:18:22 #topic Design Work Group 21:18:28 kenfinnigan: You mind talking about this one? 21:18:34 And the ideas you sent to us? 21:18:34 sure 21:18:51 thanks 21:19:06 basically we're looking to put together a small working group of us to debate/decide on major API changes and future design ideas 21:19:24 my initial thoughts were for two co-chairs, one RH and one community elected 21:19:36 with a min of 4-5 inc the chairs at each mtg 21:19:53 with the pool of members for the meeting coming from the module leads and community integrator 21:20:19 Cool 21:20:20 with a couple of community member positions as well (that are not module leads already) 21:20:33 but after reading AoC today 21:20:41 I'm thinking we could make it a bit leaner 21:21:14 and actually make the DWG only 5-6 members, which consist of the two co-chairs, community integrator and a couple of positions from leads/community 21:21:26 maintaining an even balance between RH core dev and community 21:21:34 #agreed 21:21:42 got some of these ideas from Ubuntu's concept of a Technical Board 21:21:48 #agreed 21:22:00 with at least a monthly mtg, and at most two 21:22:16 if there are so many items that can't be dealt with in two mtgs a month 21:22:40 then we constitute action groups that are made up of people outside the DWG to put a proposal together based on the problem domain 21:22:49 and then the DWG will review and decide on that at a mtg 21:22:51 BTW, these will be short meetings, no more than an hour, most likely less. 21:22:55 all make sense? 21:23:02 lightguard_jp: good point 21:23:03 works for me 21:23:10 these are obviously initial musings 21:23:13 it seems a good idea 21:23:16 and are open to other ideas and thoughts 21:23:40 yep. and I think it needs to kick off with, perhaps as an input into it, the definition of seam project, which jason, rodney and I have been talking about pulling together 21:23:45 Sounds like everyone is in agreement 21:23:48 I think that helps us understand what the DWG is aiming for 21:23:54 more like the 20,000 ft view 21:24:23 definitely, from AoC, this group would need a well defined terms of operation, essentially, to prevent it getting stuck in mud 21:24:24 #agreed DWG will consist of 5-6 members which consist of the two-chairs, community integrator and a couple of positions from leads / community. 21:24:31 or trying to do what it wasn't designed for 21:24:42 #info one chair will be a RH employee, the other will be a community elected position. 21:25:14 I vote for jose_freitas as the community member :) 21:25:53 #info send lightguard_jp your nominations for community member elections 21:25:57 similar to the community integrator, the elected positions would have a set term 21:26:06 yup 21:26:07 but you could re-nominate for the same position 21:26:14 #agreed 21:26:33 #info we'll do a vote on Monday-Wednedsay next week. Submissions must be in before Monday 21:27:25 Will all community spots be elected or just the co-chair? 21:27:38 good question 21:28:02 obviously the RH chair will be decided internally 21:28:16 but as for the other positions, probably makes sense to nominate for them as well 21:28:30 as we wouldn't want to just assign a lead the position and they don't want it 21:29:13 but as an initial start, the co-chairs could seek out people they see as being appropriate for the roles and approach directly 21:29:22 to save time and not need another round of voting 21:29:31 that's what Ubuntu did for their first councils 21:29:50 #agreed with kenfinnigan 21:30:31 Okay, I'm good with that idea 21:31:07 Everyone else like the idea? 21:31:28 I do 21:31:54 I think that's most people in that case 21:32:16 #agreed the two co-chairs will will directly approach others to be on the initial DWG 21:32:21 sbryzak: Are you here? 21:33:06 Guess not. 21:33:34 Are there other questions about the DWG idea? 21:34:52 Maybe our Community Integrator can talk about the next item 21:35:05 What is the next item ? 21:35:09 #topic Combined jar 21:35:19 oh 21:35:30 IIRC it was Dan, Shane, George and myself talking about this on IRC a few days ago 21:36:07 I believe Shane should have more to tell than myself :) 21:36:08 The idea has been presented to drop the combined jar 21:36:18 I don't think Shane is present though 21:36:36 And to rename the impl jars to just the name of the module. 21:36:36 I like the the Profile idea 21:36:53 gastaldi: Care to explain so everyone is on board? 21:36:59 Or on the same page rather 21:37:03 Ok, I´ll try it 21:37:56 After all the discussion, mojavelinux came with the idea to have a "Seam Web Profile" JAR that bundles several Seam modules altogether 21:38:25 That could make easier to start running a Seam app without needing to list them separately 21:38:26 think I saw some of that on the mailing list too 21:38:43 But I think the discussion stopped there :) 21:38:45 Might have been on the list too, I don't remember 21:39:15 I like the idea of targeted combined jars 21:39:20 as opposed to one for each module 21:39:29 can we seperate this into two issues 21:39:36 1) dropping the combined jar 21:39:39 There are some complaints with that, that could be discussed also. I think johnament was the first to disagree with that 21:39:40 2) having profile jars 21:39:44 true 21:39:50 bleathem: Agreed 21:39:50 they're two different things 21:39:52 The current combined jars for each module is cumbersome, adds complexity and the questions remains with what we do with multiple impl modules like Reports and Cron 21:39:57 Or Mail 21:40:00 hey guys, sorry i'm late 21:40:08 welcome back sbryzak ! 21:40:39 Oh cool. 21:40:43 Just in time. 21:40:46 I personally like the merged jar idea 21:40:48 couldn't we have a combined jar for modules with just one impl, and combined jars for each implementations? 21:40:52 We're talking about the combined jar issue(s) 21:41:07 jose_freitas: Yes, that´s the point 21:41:21 jose_freitas: I think that would get too complicated 21:41:32 and then pose problems if a module went from single to multiple impls 21:41:38 like for seam faces we could have one combined jar called seam-faces, and for seam-reporting, we could have seam-reporting-jasper, and seam-reporting-birt 21:41:49 the combined jar discussion had gotten deferred prior to the release of seam 3. my concern at that time, and still today, is that because of transitive dependencies, users will accidently include a combined jar and an individual jar (api or impl) when they start adding seam modules 21:42:07 The current build system would be amazingly complex for that, it could be done, but I don't think that's the way we want to go 21:42:08 it's just too easy to screw it up, because maven is not giving you any warning when you have a mix 21:42:13 jose_freitas: a combined jar wouldn't work in that scenario 21:42:26 just include any of the impls, and maven deps will bring in the api 21:42:34 the combined "stack" jars get away from this problem since you are choosing just that one thing, and nothing else (presumably) 21:42:37 is it more complicated than that? 21:42:40 bleathem: Yeah, I could live with that 21:42:46 exactly, that's what we should be leveraging 21:42:50 bleathem: that's the idea 21:43:06 then let's do it, the profile jars can be a seperate decision 21:43:19 yeah, and the combined jars are really just stack poms 21:43:20 I believe that the number of jars is not a valid argument for this 21:43:24 I agree with that. that way we manage just one dependency 21:43:27 so they depend on the impl and api of various modules 21:43:31 again, no risk of conflict 21:43:42 we shouldn't add additional artifacts into the system (jar artifacts that is) 21:43:45 going from one har to 2 is not a concern, this is not exponentially scaling 21:44:00 ^jar, not har 21:44:02 also, there could be beans.xml on both jars, so to merge them into one only would be complicated 21:44:07 besides, it doesn't really matter how many jars you have in your deployment, because in the end, it's the same number of KB 21:44:16 mojavelinux: agreed 21:44:30 right, esp with AS 7's modular classloader, we will get different behavior if we mash stuff into one jar 21:44:41 so everyone agrees to drop the combined jar? 21:44:43 God save AS7 21:44:43 the way bleathem exposed, works pefectly for both sides. 21:44:51 #agreed 21:44:59 will we do this for the 3.1 release, or earlier? 21:45:18 When is 3.1 scheduled to ? 21:45:34 we'll do it for 3.1, but if any modules have a prior release we could probably do it then 21:45:39 (for that module) 21:45:53 well, seam reports is using that already :) 21:46:11 ok, so one module at a time, with all done by the 3,.1 release 21:46:24 sounds good to me :) 21:46:27 if there's dependencies on other modules then we'll have to do another release for 3.1 21:46:41 cool 21:46:47 actually, no we won't.. it should "just work" 21:47:03 yeah 21:47:09 the bom will take care of it 21:48:09 the bom takes care of all 21:48:24 all hail the bom 21:48:36 #agreed the combined jar will go away 21:48:45 who's going to do that? 21:48:48 Are we agreed to renaming the impl jar to just the module name? 21:48:50 sounds like a lot of maven messy work 21:48:57 #agreed 21:49:10 +1 to rename 21:49:14 e.g. seam-mail, seam-servlet, seam-catch etc 21:49:27 +1 to rename 21:49:40 #agreed renaming impl jars to just the name of the module 21:49:48 The rename should be such that people's pom's don't break if they used the combined jar previously 21:50:04 That's part of the reason for doing it this way, yes. 21:50:06 * bleathem is participating more than he intended to 21:50:12 Actually, probably the biggest reason. 21:50:14 then +1 to rename 21:50:33 i'll handle it when we do the next major seam release 21:50:52 bleathem: but don't forget that it will break for those not using combined jar 21:51:08 #info the rename and abolishing of the combined jar will happen for all modules by Seam 3.1 but if there are module releases before that they will adopt the new standard. 21:51:09 jose_freitas: we can't when them all :P 21:51:11 but it's for good 21:51:30 #action sbryzak will handle the renames / combined jar pom configs as they come up 21:52:20 jose_freitas: I think in reality that number will be very low 21:52:34 mostly because maven is lazy with it's dependencies. 21:52:43 well, I defend combined jar, but I mostly use it api and impl 21:52:43 hehehe 21:52:58 but yes, I believe it too 21:54:02 Do we want the profile jar idea as it's own topic? 21:54:14 * bleathem now has the *best* bash prompt including git details 21:54:33 * lightguard_jp thinks bleathem had better screenshot it for all to see 21:54:44 something for the DWG? 21:54:55 lightguard_jp: yes, it's own topic 21:54:59 lightguard_jp: I belive so 21:55:03 or yes, good for the DWG 21:55:13 all hail the DWG 21:55:15 It sounds like a good discussion for the DWG 21:55:31 That'll give people some time to think about it as well. 21:56:06 minor segway back to DWG briefly 21:56:25 my thoughts are that the agenda is public for each meeting and topics can be submitted by anyone 21:56:45 agreed 21:56:55 and there is nothing saying that if a non DWG member has strong opinions about an agenda item they can submit their thoughts to a DWG member for inclusion 21:57:24 we're not creating a club of exclusivity here, just a means to make global seam design decisions 21:57:28 #info agendas will be public, topics can be submitted by anyone. 21:57:31 nice kenfinnigan 21:57:43 is there some kind of tool to "officialize" that? 21:57:49 Do we want a separate mailing list for this or just use the seam-dev list? 21:57:52 or it will be by email? 21:58:09 i was thinking about that 21:58:25 I think it probably needs to be separate, with decisions and actions posted to the main list 21:58:33 The seam-dev list would be fine I guess 21:58:40 agreed with gastaldi 21:58:54 It´s related to it 21:59:00 they're seam-dev discussions 21:59:14 fine with using seam-dev, but would need some kind of subject key maybe to distinguish it from other traffic 21:59:30 Yep 21:59:37 [DWG] ? 21:59:39 I think one bonus of the DWG is to streamline the decision making process 21:59:44 that works 21:59:53 and having all dicsussion on the sem-dev mailing list will load on extra baggae 22:00:00 seam-dev gets so little traffic now that i don't think it will get lost 22:00:05 =1 for seperate mailing list for DWG 22:00:09 ^ +1 22:00:12 if we use [DWG], than it's better to keep a differnt list 22:00:54 +1 for [DWG] 22:00:54 the thing is to keep the feed to seam-dev 22:01:41 i have a question.. what is the design work group going to do that's different to what we do on seam-dev already? 22:01:44 We're pretty divided on the mailing list :) 22:01:51 Let´s keep in seam-dev then, if it gets messy, split into another one 22:02:01 Everyone agrees with that ? 22:02:09 sbryzak: actually come to a conclusion on issues 22:02:15 It's a streamlined group for large architecture discussion and design. 22:02:22 sbryzak: it seems things atll a lot on seam-dev 22:02:30 ^stall a lot 22:02:39 bleathem: any particular examples? 22:02:50 the combined jars most recently 22:02:57 profile jars 22:03:05 i don't think it stalled 22:03:12 we planned to discuss it at today's irc meeting 22:03:22 and we've now reached a conclusion 22:03:28 would that make a difference being on another mailing list ? 22:03:29 it didn't take that long ;) 22:03:44 ok, you convinced me 22:03:45 we're not creating another mailing list 22:03:50 :) 22:03:54 we already have one for this exact purpose 22:04:19 if you think something is stalling, then bring it up directly with the people you think it should involve 22:04:34 or discuss it with them on irc 22:04:41 a lot of decisions get made during irc discussions 22:05:07 It´s amazing IRC is still useful nowadays :) 22:05:16 i don't think that creating a new subset group of people is going to make any difference 22:05:34 and besides, our team size just isn't big enough to warrant it 22:05:35 ok, I think everyone agrees to use the seam-dev mail list then 22:05:54 next topic / 22:05:55 ? 22:05:59 It may not now, but if we had every module lead in attendance to these meetings it certainly would. 22:06:05 (since I was probably the only one who didn't agree initially :P ) 22:06:05 Don't have one 22:06:09 That's it 22:06:14 Cool 22:06:17 We´re done in time 22:06:25 yeah! 22:06:33 Good meeting all! 22:06:34 i think the weekly irc meeting is the time to bring up stuff like that 22:06:36 getting better every weel 22:06:38 week 22:06:43 #endmeeting