21:01:57 #startmeeting 21:02:04 Meeting started Wed Mar 16 21:01:57 2011 UTC. The chair is mojavelinux. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:02:04 Useful Commands: #action #agreed #help #info #idea #link #topic. 21:02:26 git [12forge] push 10master7 1564c17.. 6Lincoln Baxter, III began refactoring ShellImpl into smaller more maintainable pieces 21:02:30 git [12forge] push 10master7 e46c038.. 6Lincoln Baxter, III SEAMFORGE-70 21:02:35 jira [3SEAMFORGE-70] Ls and Rm do not function on JavaResource members [10Closed (Done) Bug,7 Major,6 Lincoln Baxter III] https://issues.jboss.org/browse/SEAMFORGE-70 21:02:39 git [12forge] push 10master URL: http://github.com/seam/forge/compare/ac99201...e46c038 21:02:43 #topic international concerns, vanity committing and release countdown 21:03:28 oops, sorry :) 21:03:33 we have stuart but shane is still snoozing...stuart, would one hour earlier be too painful? 21:03:53 we had to shift this week because ben franklin invented dst 21:04:13 Shane is having customer meetings in Sydney today 21:04:24 Sorry I'm late 21:04:47 no worries, we are still on time and date 21:05:00 One our earlier is 6:00am for Shane, 7:00am for me 21:05:07 yeah, kind of painful for shane 21:05:15 I am fine with it, but 6 is a bit early :-) 21:05:21 but we have guys on the east that are on the edge of the work day 21:05:29 6:00 is really painful 21:05:39 we are going to try this time, which is 7 for shane, 8 for you stuart 21:05:46 well, what about 6am eastern? 21:05:58 if I stay up one extra hour, I might make it :) 21:06:02 hahaha 21:06:18 lol 21:06:59 6 AM Eastern and you won't see Cody, or myself :) 21:07:04 #agreed try 5pm east coast to keep shane sane 21:07:12 won't see me at 6am est 21:07:20 yeah, hopefully that was a joke 21:07:29 no one else is an early bird? 21:07:37 not unless i'm still up from the night before 21:07:42 yep 21:07:46 i'm usually heading in to work at 6am... 21:07:51 okay, so we were discussing the SeamManaged annotation 21:08:11 which we moved into solder, but in thinking about solder as a general purpose library, it would make more sense to try to name the annotation according to it's fundtion 21:08:15 rather than overreach the brand 21:08:29 so, john and I were discussing ModuleManaged (john) or ExtensionManaged (dan) 21:08:44 that way, it's not restrictive to just seam modules (for others using solder) 21:08:54 it's the opposite of container-managed 21:10:25 so, any votes? 21:10:35 let's see them...we have money on it 21:10:40 ModuleManaged 21:10:50 ExtensionManaged 21:10:54 I like ModuleManaged better 21:11:03 +1 for ExtensionManaged 21:11:16 alright, come on voters 21:11:21 ExtensionManaged 21:11:25 ExtensionManaged 21:11:32 ExtensionManaged 21:11:40 sorry johnament 21:11:49 its cool 21:11:57 stuart, got a vote? 21:12:07 his vote doesn't count anymore 21:12:21 5/2 21:12:26 hey now, don't be dissin' the bread winner 21:12:30 just sayin' ;) 21:13:33 What's next? 21:13:42 okay, next up... 21:13:51 btw, we are running with a meeting bot thanks to jason 21:13:55 or should we record that result in the miutes somehow 21:13:59 for those that weren't at the party last night 21:14:04 Put an #agreed to that final vote please 21:14:20 I'm not a chair, so I can't. 21:14:45 #action SeamManaged annotation shall henceforth be referred to as ExtensionManaged 21:14:49 #agreed rename SeamManaged to ExtensionManaged to reflect it's function over branding, despite pleads by john to name it ModuleManaged 21:14:53 #info Vote for ModuleManged vs ExtetnionManged: 2/5 21:15:09 #action file a jira for previous agreed 21:15:27 sorry, I got a little meetbot keen there 21:15:36 bleathem: :) 21:15:41 sorry, I was having a shower, only just woke up :-) 21:16:14 ExtensionManaged is better than ModuleManaged though 21:16:20 #info let the record show that john and jose put in a good fight, and lincoln tried to insult our fearless bread winner 21:16:40 Ouch 21:16:49 So let it be written.... 21:16:59 #info vote now 2/10 in favor of ExtensionManaged 21:17:04 #rebuttal lincoln merely tried to state the fact that there wee too many votes to overcome 21:17:08 In bot blood 21:17:15 because lincoln's insult means that stuart's vote doubled the side 21:17:40 okay, upwards 21:17:45 * lincolnthree goes back to writing awesome forge code 21:17:50 #info onwards is now upwards 21:18:10 note to self, disregard info comments in meeting minutes :P 21:18:41 ken and I did much deliberating, and we think we worked out the division between i18n concerns in solder and international 21:19:02 solder is going to provide the annotations for typed loggers and message bundles 21:19:15 #topic division between i18n concerns in solder and international 21:19:20 and be based on the system locale 21:19:49 bleathem: Topic lasts for whole meeting, and only a chair can change it 21:19:57 and international handles all the configuration and client locale concerns, and extending the typed loggers and message bundles if we wanted to do stuff like load from a database or whatever 21:20:20 ignore bleathem feeble attempts to change topic :) 21:20:25 hehehe 21:20:29 Lol 21:20:33 lightguard_jp: oh, how do we track what we're talking about? 21:20:37 even though his actions were noble 21:21:08 so, here is how the annotations will fall, we are looking for objections only 21:21:12 lol 21:21:16 bleathem: I'd have to go over the commands again 21:21:25 #commands 21:21:29 Available commands: #accept #accepted #action #agree #agreed #chair #commands #endmeeting #halp #help #idea #info #link #lurk #meetingname #meetingtopic #nick #rejected #restrictlogs #save #startmeeting #topic #unchair #undo #unlurk 21:21:33 org.jboss.seam.solder.i18n.Message 21:21:37 org.jboss.seam.solder.i18n.MessageBundle 21:21:42 org.jboss.seam.solder.i18n.Locale 21:21:46 org.jboss.seam.solder.i18n.Formatter 21:21:50 org.jboss.seam.solder.logging.Log 21:21:54 org.jboss.seam.solder.logging.MessageLogger 21:21:58 org.jboss.seam.solder.logging.Category 21:22:03 org.jboss.seam.solder.logging.Cause 21:22:07 oh, there is #meetingtopic and #topic 21:22:12 so I guess #topic is short term? well, anyway, we'll see 21:22:16 oops 21:22:20 sorry, that was a mistake 21:22:25 paste again 21:22:29 Thought Message was in messaging not i18n? 21:22:43 org.jboss.seam.solder.messages.Message 21:22:47 org.jboss.seam.solder.messages.MessageBundle 21:22:52 org.jboss.seam.solder.messages.Locale 21:22:56 org.jboss.seam.solder.messages.Formatter 21:23:01 org.jboss.seam.solder.logging.Log 21:23:05 org.jboss.seam.solder.logging.MessageLogger 21:23:10 org.jboss.seam.solder.logging.Category 21:23:14 org.jboss.seam.solder.logging.Cause 21:23:18 there we go 21:23:23 and now we have this 21:23:27 Ok, thought I was going crazy for a sec 21:23:41 @MessageLogger 21:23:45 public interface MyLogMessages { 21:23:50 @Log(level = INFO) @Message("Seam rocks") 21:23:54 void seamRocks(); 21:23:58 } 21:24:02 so some optimizations and cleanups of packaging 21:24:07 and the status messages will be like this 21:24:59 @StatusReporter public interface BookingStatusReporter { 21:25:04 @Status @Message("Hotel booked!") 21:25:08 void hotelBooked(); 21:25:12 } 21:25:16 ken, we didn't work out what @StatusReporter would be yet, something parallel to @MessageLogger 21:25:27 i'll let you ponder that unless you've worked it out already 21:25:31 Not @MessageBundle? 21:25:47 That is the current usage 21:25:56 oh, yeah, I guess we can double it up 21:26:00 k 21:26:15 cool, so that's coming along, just some heads up about that 21:26:36 so we've been beat-up a bunch about the glassfish issues 21:27:13 I did a whole bunch of digging on that and if I can ever finish it, I have a blog entry that will restate our committment to portability and explain why glassfish is giving us trouble 21:27:25 I have written some additional tests in the solder "compat" package 21:27:34 many of which fail on glassfish 3.1 due to the same issue 21:27:47 #topic glassfish 3.1 issues 21:27:52 i'll have details, but in short here, it's the alpha bravo issue 21:28:20 extensions and beans in an alpha.jar can see beans in the bravo.jar 21:28:24 but not the other way around 21:28:29 has to do with alphabetical order of jars 21:28:33 alpha bravo? 21:28:37 a b 21:28:41 alphabetical, really!?! 21:28:47 yep 21:28:52 yep, that's why we had problems with @Inject Messages 21:28:56 wow 21:29:00 that is retarded 21:29:05 makes sense 21:29:09 holy sh** 21:29:13 because international comes after faces 21:29:17 and faces was like huh? 21:29:22 lol 21:29:26 where is Messages? 21:29:30 motion to rename faces to zFaces 21:29:35 hahaha 21:29:39 Genius 21:29:43 rename solder to alpha 21:29:47 yep, and that's why the Loggers are not satisified 21:29:51 because solder can't process classes in servlet 21:29:56 The fix for this will be in Weld 1.1.1? 21:30:01 correct, according to ales 21:30:09 any timeframe? 21:30:31 not sure, just "soon" 21:30:35 bleathem: It's not a problem in weld 21:30:42 (ie. before Seam 3.0.0.Final) 21:30:47 It's the integration code in glassfish 21:30:51 actually, it is a problem in weld, in a sense 21:30:55 Really? 21:30:59 Thought it was gf integration code 21:31:04 I think it is actually a bit of both 21:31:15 yep, probably best said that way, it's sort of both needed to change to get past the problem 21:31:22 because of the way bdas are handled in glassfish 21:31:29 jboss as 6 is flat, so it didn't matter 21:31:35 as 7 would have had the same issue 21:31:44 given that it's more structured, right stuart? 21:32:12 yea, it has the weld related problems, but the visibility code is all correct 21:32:27 AS6 did not see the problems cause it did not implement the spec correctly 21:32:41 #info glassfish 3.1 out of the box processes bdas in such a way that alphabetical order of jars affects visibility of extensions and beans 21:32:54 GF also has a problem where WEB-INF/lib beans cannot see WEB-INF/classes beans 21:33:02 and this is all them 21:33:07 yep, that's another issue, I call that something 21:33:56 i don't know what it was, the "don't look over your shoulder" or something hehehe 21:34:04 glassfish 3.0.1 is definitely out for seam 3 21:34:08 in a major way 21:34:32 it has a serious bug in weld...where weld fails the first time it sees a reference to any class that doesn't exist 21:34:45 which means about 50% of the seam modules will just fail outright 21:34:56 that bug is still there in 1.1 21:35:00 such as persistence, which has references to Hibernate classes 21:35:05 It just does not show up as often 21:35:09 ah 21:35:14 well, I can at least use persistence on glassfish 3.1 21:35:19 so maybe they swallow it 21:35:23 I did a fix for it yesterday 21:35:28 good 21:35:35 that's called the overeager scanner problem 21:35:40 Hope it makes it in to 1.1.1 21:35:44 my hero 21:35:49 hyperactive scanner 21:35:54 no, now it only happens if a superclass of a bean references a non-existant class 21:35:58 I have names for all this :) 21:36:10 we had a dangling web fragment reference in seam servlet 21:36:14 which I committed a fix for 21:36:18 lame 21:36:39 in relative ordering, if you reference a web fragment like PrettyFaces 21:36:43 and you aren't using PrettyFaces 21:36:47 deployment fails 21:36:56 so it sort of defeats the whole point of relative ordering 21:37:02 that's pretty lame 21:37:06 which happens to be kind of an important thing 21:37:13 yep, so they committed a fix to glassfish 21:37:18 jboss as fixed it in AS 6 21:37:22 final 21:37:27 and jetty fixed it in 8.0.0.M1 21:37:32 but Tomcat still fails 21:37:36 who would be using pretty faces?? 21:37:47 johnament you don't want to know :P 21:37:53 actually, the real purpose was to order in relation to Weld Servlet 21:38:00 because that matters or else Seam Servlet fails 21:38:09 i heard the guy who used to work on it is now busy with CLIs. 21:38:14 however, Weld Servlet doesn't even use a web fragment yet, so that was a premature optimization anyway 21:39:05 our problem with not having concrete typed logger and typed message bundles classes is now behind us, thanks to james and ken 21:39:09 they kicked ass 21:39:20 the annotation scanner is portable to any annotations you want to use 21:39:30 so solder now has it's own annotations for defining these things 21:39:40 My pleasure. Though most of it was James 21:39:57 and the annotation processor can be used for generating any sort of typed stuff 21:40:02 as we go down the road 21:40:16 good reuse of code...for now just focused on the original goal and also for the typed stuff in seam international 21:40:20 so that can move forward 21:40:50 more importantly, no more of this having to tell people to use the jboss.i18n.generate-proxies system property just to use seam modules 21:40:54 Indeed. Lots of ideas to put together 21:41:03 so that should solve a lot of complaining 21:41:22 now, to get seam to run on glassfish 3.1 out of the box without any visibility problems and no upgrade to weld 21:41:27 which you know people are going to try 21:41:32 you know it 21:41:38 one solution is to shade all the seam modules together 21:41:50 the other is to deal with each problem in turn 21:41:56 which I've done all except for one problem in seam servlet 21:42:01 which is just a pita 21:42:36 Does anyone know where we are with regard to a weld 1.1 release ? 21:42:41 1.1.1 21:42:45 +1 for fully shaded jar that can be used by those that want everything 21:42:49 so seam booking works if you compile the branch temporary-logger-fix of seam servlet 21:43:14 Might also make it easier to incorporate in EAP going forward 21:43:18 I think the seam 3.0 final release should be shaded together as an option for people that just want the "one jar to rule them all" approach 21:43:36 plus, it will shut the codi guys up :) not that they are annoying me over this issue 21:44:08 we are not sure where we are with a 1.1.1 release, we haven't gotten word from ales about a date 21:44:12 he just said "soon" 21:44:18 Maybe have a "web profile" shaded jar, with "essential" modules in it 21:44:25 exactly what I was thinking 21:44:30 you read my mind 21:44:34 :D 21:44:39 seam-web-profile.jar 21:44:53 I'd use that 21:44:57 Sounds good 21:45:01 or we can just call it 21:45:15 compare-this-to-codi-gehard.jar 21:45:21 #idea seam 3.0 final release should be shaded together as an option for people that just want the "one jar to rule them all" approach 21:45:29 #idea have a "web profile" shaded jar, with "essential" modules in it 21:45:36 lol 21:45:41 I like that name better 21:45:58 I just love the chart where he says 21:46:09 codi seam-faces 21:46:17 persistence check no 21:46:29 right, because you get persistence from seam-faces 21:46:56 The uber-seam-jar would have an inverse looking chart! 21:47:03 when compared to codi 21:47:11 yeah, we can just shade all 28 modules 21:47:15 :) 21:47:36 boom! 21:47:41 may there be an problem with dependencies? 21:47:52 Validation e.g. depends on HV 4.2Beta2 21:47:57 haha, we are just kidding :) 21:48:34 but I think having a stack jar of seam would be an attractive idea, and it's come to my mind a couple of times in the past, for the people who don't like the a la carte option 21:48:43 microwave dinner 21:48:59 arquillian alpha5 release is emminent 21:49:11 I think the idea is generally good, too 21:49:15 and jason got the glassfish 3.1 container adapter working for that 21:49:19 Might be easier for those moving from Seam 2 where it was all in one 21:49:34 yep, just an option, it wouldn't shade dependencies, just the seam bits 21:49:44 the dependencies would be dependencies of the shaded thing 21:50:56 ah, almost forgot 21:51:04 we've moved transactions into org.jboss.seam.transaction package 21:51:13 but I think we should just go ahead and make the transaction module now 21:51:18 because jcr is going to need it anyway 21:51:55 #action create transaction module for seam 3.0 final release 21:52:14 seam booking needs security in it, for two reasons 21:52:23 1. to prove that it isn't causing deployment problems (always a good thing) 21:52:27 2. to demonstrate it 21:53:30 oh 21:53:40 I'll be looking at SEAMFACES-33 next 21:53:44 #action lincoln needs to do a proper release of render so cody can integrate it into mail 21:53:49 jira [3SEAMFACES-33] Create a solution for consolidated page-flow, transactional control, security constraints and URL-rewriting configuration [10Open (Unresolved) Feature Request,7 Blocker,6 Brian Leathem] https://issues.jboss.org/browse/SEAMFACES-33 21:53:59 if that relates to the security in booking 21:54:03 yep, and once that seam-faces issue is resolved, we'll integrate that into booking as well 21:54:07 there are two parts 21:54:12 authentication part can be done now 21:54:20 authorization part can be done after 33 21:54:47 I would push url-rewriting 21:54:54 because that's a bigger issue imho 21:55:11 and right now we are using pretty faces in booking to solve that problem...we just need to look into how to make that configuration integrated 21:55:24 Do we have an issue for it? 21:55:28 because we don't want to start over with url rewriting :) 21:55:33 not sure, but I think it needs to be 101 21:55:37 if not 21:56:20 he stuart, is as 7 ready to host a cdi + jpa app? 21:56:31 as of yesterday, yes 21:56:44 cause we should test the forge scaffold and the archetypes on it 21:56:48 see whether they fly 21:56:53 As of beta1? 21:57:00 likely two example apps people will try 21:57:04 also, we should try booking 21:57:20 btw, awesome 21:57:45 however given that JPA was added about 24 hours before the beta went out, and the weld integration was added about 4 hours before the beta went out, I am not giving any guarentee's 21:58:20 hehe, got it 22:00:54 okay, time is up...if you have extra cycles, it would be good to focus on getting bugs resolved, it would be super appreciated 22:01:54 and before we wrap up, we have stellar news 22:02:09 lightguard_jp is joining the Seam Team at Red Hat 22:02:14 :) 22:02:21 congratulations 22:02:25 * lightguard_jp cheers 22:02:30 Congrats! 22:02:34 congrats :-) 22:02:51 Thank you all 22:03:00 Looking forward to hepling out more and moving things along 22:03:08 he's been more patient than you can possibly imagine while we worked through HR 22:03:21 Ciao guys. Dinner time. Should be back later 22:03:26 lightguard_jp congrats! 22:03:37 kenfinnigan: See ya 22:03:42 yep, now we are really going to get stuff done! 22:04:12 #endmeeting