[00:08:44] *** stalep has joined #jboss-as7
[00:28:09] *** fnasser has quit IRC
[00:31:44] *** laubai has joined #jboss-as7
[00:31:53] *** laubai has quit IRC
[00:32:29] *** rawblem has quit IRC
[00:33:17] <stuartdouglas> I am getting failures when I try and run the integrations tests: Provider org.apache.xerces.jaxp.datatype.DatatypeFactoryImpl not found
[00:33:30] <stuartdouglas> Nihility: could this be related to your changes the other day?
[00:36:27] <Nihility> Yes I can ex
[00:36:52] <Nihility> explain the details in couple hours
[00:37:05] <Nihility> (watching kid and cooking)
[00:37:08] <stuartdouglas> ok
[00:51:23] *** alesj has quit IRC
[00:54:20] *** frainone has joined #jboss-as7
[00:54:21] *** ChanServ sets mode: +v frainone
[01:00:10] <stuartdouglas> Nihility: never mind, I have figured it out, there was some code that was accidentally initialising jboss modules on the client side
[01:00:30] <stuartdouglas> and then when jboss modules tries to set up the redirect it all goes pear shaped
[01:00:49] *** pferraro has quit IRC
[01:00:54] *** rmaucher has quit IRC
[01:03:03] <smarlow> stuartdouglas:  Should org.jboss.as.ee.component.EEModuleDescription.componentsByName for a WEB archive be filled in yet?
[01:03:35] <stuartdouglas> yes
[01:03:55] <stuartdouglas> although web components don't really have names as such, so I have just been using the class name
[01:04:20] <stuartdouglas> you need the very latest upstream, it only got merged a few hours ago
[01:05:12] <smarlow> cool, I'll pull the latest, thanks for the timely fix for this ;)\
[01:06:15] *** jpearlin has joined #jboss-as7
[01:08:07] *** mbg|away is now known as mbg
[01:13:27] <smarlow> stuartdouglas:  Is the PARSE phase, too early for me to look at EEModuleDescription.componentsByName?
[01:13:40] <stuartdouglas> yea, unfortunatly
[01:13:46] <stuartdouglas> I had to move it all back to post module
[01:13:57] <stuartdouglas> as the web component annotation stuff loads the actual annotations
[01:14:35] <stuartdouglas> so If I load it in the parse phase I miss annotation based servlets
[01:18:59] <smarlow> do we have ordering within the phases?
[01:19:16] <stuartdouglas> yea, the Phase class defines the ordering
[01:20:09] <smarlow> So, I move my DUP to POST, with a higher number, you'll run first?
[01:20:16] <stuartdouglas> yes
[01:20:35] <stuartdouglas> I need to talk to the web guys about getting this changed
[01:20:41] <stuartdouglas> so we can move it back to post
[01:20:47] <stuartdouglas> s/post/parse/
[01:20:55] *** asaldhan has left #jboss-as7
[01:23:29] <jbossbot> git [jboss-as] push master 338cf4c.. Emanuel Muckenhuber don't rollback on startup
[01:23:29] <jbossbot> git [jboss-as] push master f1d0017.. bstansberry at jboss dot com Ignore ROLLOUT_PLAN param
[01:23:29] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/cbe11d5...f1d0017
[01:23:36] *** bstansberry is now known as bstans_afk
[01:38:43] <smarlow> stuartdouglas:  Its probably something in my branch but I'm getting testsuite deployment failures (no such file or directory) after pulling the latest changes.  Have you see anything like that today?  http://www.pastie.org/1645491
[01:39:05] <stuartdouglas> you don't have another as instance running in another terminal do you?
[01:40:01] <stuartdouglas> cause that is the error that you get if you leave an instance running and then rebuild
[01:40:32] <smarlow> jps says that I do, will kill that.  Good call!  :-)
[01:42:01] <stuartdouglas> I got into the habit of just doing killall java on my mac, which works great
[01:42:17] <stuartdouglas> but if you do that on linux it kills my ide as well :D
[01:45:29] <stuartdouglas> can someone merge my master? It has a few arquillian fixes
[01:48:01] <smarlow> stuartdouglas:  cool, I see 4 components by name/class now :)
[01:49:37] <stuartdouglas> cool
[01:56:27] <bstans_afk> stuartdouglas: I'll push those
[01:56:41] <stuartdouglas> thx
[01:56:44] *** bstans_afk is now known as bstansberry
[01:59:17] <jbossbot> git [jboss-as] push master ee0a63a.. Stuart Douglas minor arquillian pom fix
[01:59:17] <jbossbot> git [jboss-as] push master e34469e.. Stuart Douglas Remove unused static that forces jboss modules initalization
[01:59:17] <jbossbot> git [jboss-as] push master f9914f3.. Stuart Douglas Register a shutdown hook so the container is always stopped
[01:59:18] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/f1d0017...f9914f3
[02:09:38] <smarlow> Oops, got a "NCDFE: javax.persistence.EntityManager" during the install phase (http://www.pastie.org/1645588)
[02:12:07] <stuartdouglas> which CL is trying to load EntityManager?
[02:12:22] *** balunasj_busy has quit IRC
[02:12:25] <stuartdouglas> you probably need to add the java.persistence module to the deployment in a dependency processor
[02:14:53] <smarlow> stuartdouglas:  I thought I already was.  I'll try to catch it in a debugger.  I'm not sure of what happened here.  I'm guessing that it has something to do with the BindingDescription that I added to the ComponentDescription (which was for an annotation that referenced EntityManager
[02:16:39] <smarlow> hmm, I add JPA dependencies to the DUs MODULE_SPECIFICATION, maybe I need to add JPA dependencies to other attachments as well
[02:22:02] <smarlow> oh, I didn't inject the JPA dependencies because of a bug :)
[02:22:46] *** frainone has quit IRC
[02:34:12] <smarlow> that is pretty ugly, glad I tripped over it.  I put the persistence.xml in the wrong folder in my test jar, which means the JPA dependencies didn't get set.  That results in a later ugly NCDFE error
[02:36:47] *** jwulf has quit IRC
[02:38:35] *** jamezp has left #jboss-as7
[02:56:28] *** bgeorges has joined #jboss-as7
[02:56:28] *** ChanServ sets mode: +v bgeorges
[02:57:04] *** pferraro has joined #jboss-as7
[02:57:04] *** ChanServ sets mode: +v pferraro
[03:07:30] *** jpearlin has left #jboss-as7
[03:39:02] *** jwulf has joined #jboss-as7
[03:40:07] *** mbg is now known as mbg|away
[03:48:30] *** mbg|away is now known as mbg
[03:48:42] *** mbg has quit IRC
[03:49:33] *** pgier has quit IRC
[03:51:51] <Nihility> i hate eclipse
[05:00:27] *** jwulf has quit IRC
[05:37:28] *** irooskov has quit IRC
[05:41:39] <smarlow> I removed osgi from standalone.xml but still get "Failed to register driver for: org.h2.Driver"
[05:42:32] * stuartdouglas looks
[05:44:37] <smarlow> the rest of the exception chain is here http://www.pastie.org/1646062
[05:45:32] <stuartdouglas> do you have any kind of dep on the h2 deployment?
[05:46:08] <stuartdouglas> e.g. a class path entry?
[05:47:50] <smarlow> I have a JndiService dependency on the h2 deployment.  I manually copied jboss-7.0.0.Alpha2/modules/com/h2database/h2/main/h2-1.2.144.jar into standalone/deployment
[05:48:13] <smarlow> also copied jdbc-local.rar in as well
[05:48:58] <stuartdouglas> to actually load the class at the moment you will probably need a class-path entry referencing the h2 jar
[05:49:14] <stuartdouglas> what sort of JndiService dependency? Is this something you are setting up in JPA?
[05:51:00] <smarlow> Yes, in JPA, for the service that represents the particular PersistenceUnit definition/deployment.  I add a JndiService dependency
[05:51:25] <smarlow> something like:   JndiService.SERVICE_NAME_BASE.append(pu.getJtaDataSourceName() + JndiService.SERVICE_NAME_BASE.append(pu.getNonJtaDataSourceName())
[05:52:10] <smarlow> this test datasource is "java:/H2DS"
[05:53:00] <smarlow> I could switch back to the JCA Datasource service (or something like that)
[05:53:13] <stuartdouglas> I am not really sure what the JndiService is, it looks like it may actually be obsolete
[05:53:42] <stuartdouglas> try just adding a Class-Path: ../h2driver.jar to your deployment
[05:54:01] <smarlow> okay, I'll try that.
[06:01:43] *** bstansberry has quit IRC
[06:05:30] <smarlow> didn't seem to help, I tried Class-Path: ../h2-1.2.144.jar and Class-Path: h2-1.2.144.jar
[06:06:37] <stuartdouglas> Is this on github so I can take a look?
[06:07:35] *** miclorb_ has quit IRC
[06:08:31] *** ALR has quit IRC
[06:09:41] <smarlow> yep, everything is in https://github.com/scottmarlow/jboss-as and the source for this particular issue is here https://github.com/scottmarlow/jboss-as/blob/master/jpa/src/main/java/org/jboss/as/jpa/processor/PersistenceUnitDeploymentProcessor.java
[06:10:21] <smarlow> around line 171
[06:12:16] <stuartdouglas> what deployment are you trying to deploy?
[06:12:51] <smarlow> just a war I hacked together, should I email it to you?
[06:13:14] <stuartdouglas> that would be good, save me having to hack together my own one :-)
[06:16:43] <smarlow> should be on its way :)
[06:16:48] <stuartdouglas> just got it
[06:22:39] <Nihility> check your standalone.xml for h2
[06:23:01] <stuartdouglas> one problem is that the cp entry is in WEB_INF/classes/META_INF
[06:23:14] <stuartdouglas> rather than just META-INF
[06:23:22] <Nihility> ok do i made GlobalContextService
[06:23:25] <Nihility> so
[06:23:57] <Nihility> which allows for a different swapable context ot be used
[06:24:04] <Nihility> but with no selector
[06:24:07] <Nihility> (yet)
[06:24:08] <stuartdouglas> I wonder if we should just read a manifest in WEB-INF/classes-META-INF, it is probably only going to cause confusion ignoring it
[06:24:19] <stuartdouglas> do we need swapable contexts for global?
[06:24:52] <stuartdouglas> surely any swapping would make it no longer global, and therefore in violation of the spec?
[06:25:11] <Nihility> well its not swappable right now
[06:25:23] <Nihility> it could be in the case of app client perhaps
[06:25:28] * smarlow mving it in the test :)
[06:25:50] <Nihility> like talking to different servers or something
[06:26:18] <Nihility> but the other reason is that it allows the component install code
[06:26:24] <Nihility> to behave the same for all contexts
[06:30:19] <smarlow> h2 is in the standalone.xml but still not finding the driver.  I removed most of the standalond.xml lines that mentioned osgi as well
[06:31:32] <stuartdouglas> smarlow: I get that to, I will have a bit of a debug
[06:52:50] *** pferraro has quit IRC
[06:55:05] <smarlow> stuartdouglas:  at least the DS demo still works
[06:55:30] <stuartdouglas> there seems to be a problem with class-path, just fixing it now
[07:08:37] <stuartdouglas> smarlow: In my master I have fixed the Class-Path issue
[07:08:41] <stuartdouglas> now there is a different one
[07:11:32] <smarlow> stuartdouglas: thanks for looking!
[07:11:40] <stuartdouglas> now I get this: http://pastie.org/1646226
[07:13:03] <smarlow> interesting, I saw that when I ran the DS demo, while the other test was still 1/2 deployed
[07:13:29] <smarlow> I did a clean build and just ran the DS demo and didn't get that
[07:13:56] <smarlow> I wonder if we have multiple copies of H2 driver around
[07:14:52] <stuartdouglas> maybe, I am not really sure
[07:15:08] <stuartdouglas> I'd say it is more likely to be an as bug
[07:17:04] <stuartdouglas> I have to duck out for a bit, back in around 30 min
[07:17:12] <smarlow> I need to get some sleep anyway
[07:17:26] <Nihility> same
[07:17:28] <Nihility> :)
[07:17:41] <smarlow> good night :)
[07:18:03] * smarlow hoping my fedora suspends cleanly this time
[07:22:14] *** smarlow has quit IRC
[07:33:41] *** asoldano has joined #jboss-as7
[07:33:41] *** ChanServ sets mode: +v asoldano
[07:39:33] <Nihility> ok now really off to bed
[07:46:28] *** Nihility has quit IRC
[07:48:08] *** Nihility has joined #jboss-as7
[07:48:08] *** Nihility has joined #jboss-as7
[07:48:08] *** ChanServ sets mode: +v Nihility
[08:12:05] *** wolfc has joined #jboss-as7
[08:12:05] *** ChanServ sets mode: +v wolfc
[08:14:18] <wolfc> stuartdouglas, good evening
[08:17:05] *** jfclere has joined #jboss-as7
[08:17:05] *** ChanServ sets mode: +v jfclere
[08:17:36] *** Jaikiran has joined #jboss-as7
[08:17:36] *** ChanServ sets mode: +v Jaikiran
[08:18:45] <stuartdouglas> wolfc: hi
[08:19:35] <wolfc> stuartdouglas, did you find a spot to work upon?
[08:19:44] <wolfc> Jaikiran, good afternoon
[08:19:53] <Jaikiran> wolfc: good morning
[08:20:14] <Jaikiran> does dmlloyd
[08:20:26] <Jaikiran> does dmlloyd's ejb3-domain branch build for you?
[08:20:34] <Jaikiran> i just updated and ran into compile errors
[08:20:42] <wolfc> I can't rebase for some reason
[08:20:47] <Jaikiran> same here
[08:21:06] <Jaikiran> i just nuked my local branch (which didn't have any changes) and then created afresh from his branch
[08:21:18] <Jaikiran> but then i run into those compile errors
[08:21:20] <wolfc> yeah, did the sasme
[08:21:29] <wolfc> hmm, need more coffee
[08:21:41] <stuartdouglas> wolfc: Not really, I have been looking at a few other non-ejb related things
[08:22:22] <wolfc> stuartdouglas, you read through the ejb spec?
[08:22:48] <stuartdouglas> I have read bits of, I have never sat down and read it cover to cover though :-)
[08:24:21] <wolfc> that's okay. the goal for the next release is to get a SLSB smoke-test passing a.s.a.p. How much time do you have left to spent on it? Maybe it's better to start on next release stuff?
[08:24:45] <wolfc> spent on it == next release
[08:25:14] <wolfc> urgh, stupid English language, upcoming release
[08:25:24] <stuartdouglas> I only really have tonight and tomorrow this week
[08:25:28] <stuartdouglas> the I am at NHO
[08:25:37] <stuartdouglas> back on Friday night
[08:26:02] <stuartdouglas> and tomorrow I will be travelling for half of it
[08:26:12] <stuartdouglas> It is not the best timing :-)
[08:28:21] <wolfc> Yeah, like I yesterday only a spot of time to do some test cases.
[08:28:40] <wolfc> Ultimately I want to have something like: https://github.com/wolfc/jboss-as/commit/2a617925a2453f297cf87d21774da0555e7dccbc#L2R32
[08:28:41] <jbossbot> git [jboss-as] 2a61792.. Carlo de Wolf Using a custom pool from the deployment unit (exp)
[08:28:51] <wolfc> A MB being a building block for the EJB Component itself.
[08:29:09] <wolfc> (don't look at the branch itself, it's obsolete)
[08:29:29] <wolfc> Jaikiran, the branch does not compile for me
[08:29:46] <Jaikiran> wolfc: i fixed one set of compile errors but it runs into another compilation problem now
[08:29:58] <stuartdouglas> wolfc: nice
[08:31:26] <wolfc> another bit that I was contemplating was message inflow
[08:32:31] <wolfc> I don't think we have a hornetq rar yet, do you know if we have?
[08:33:30] <wolfc> stuartdouglas, maybe you can do a 'dummy' rar on some other even system so we can start on message inflow
[08:34:35] *** opalka has joined #jboss-as7
[08:34:35] *** ChanServ sets mode: +v opalka
[08:35:29] <Jaikiran> it's actually this commit which is causing the compile errors https://github.com/dmlloyd/jboss-as/commit/11e2fe8f9179404739047648b2659e5fb488a8bc
[08:35:30] <jbossbot> git [jboss-as] 11e2fe8.. David M. Lloyd Give every component a namespace selector, even if it is not gonna use it
[08:35:57] <Jaikiran> i don't know how it passed compilation for david
[08:36:06] <stuartdouglas> I am not sure if we have a hornetq rar
[08:36:08] <wolfc> Jaikiran, I think it's best to switch back to the older branch
[08:36:22] <Jaikiran> wolfc: which one?
[08:37:06] <wolfc> Jaikiran, https://github.com/wolfc/jboss-as/tree/ejb3-ee
[08:37:54] <wolfc> stuartdouglas, do you know if MSC has any events? we could do a service status inflow :-)
[08:40:11] *** miclorb_ has joined #jboss-as7
[08:40:48] <Jaikiran> wolfc: hmm, that branch builds for you?
[08:40:56] <Jaikiran> i see the same compile errors with that too
[08:41:19] <Jaikiran> which isn't surprising because that commit is part of that branch too
[08:41:21] <wolfc> oh crap, I had already reset it. Hold on.
[08:41:47] <stuartdouglas> sort of, you can set up service listeners that automatically get added to child services
[08:41:55] <stuartdouglas> so you probably could do something like that :-)
[08:42:12] <wolfc> as long as it generates nice events to play with
[08:42:38] <wolfc> just don't do quartz, we don't want a dependency on it in any way :-)
[08:42:49] <wolfc> Jaikiran, I've reset ejb3-ee back in time
[08:42:53] <Jaikiran> ok
[08:42:55] <Jaikiran> let me try now
[08:44:08] <wolfc> Jaikiran, that patch for the latest dmlloyd did you push it somewhere?
[08:44:31] <Jaikiran> wolfc: not yet. you want me to?
[08:45:44] <wolfc> Yeah just push it. I'll go briefly through the new branch and see if it can easily be brought up. It's not really worth the hassle...
[08:46:21] <Jaikiran> ok
[08:48:38] <wolfc> stuartdouglas, maybe it should not be MSC services, but AS Components
[08:49:05] <stuartdouglas> if it is just for test purposes, does it really matter?
[08:49:33] <wolfc> I can see this becoming one of toys that people will actually start to use :-)
[08:55:51] *** emuckenhuber has quit IRC
[08:56:03] <wolfc> stuartdouglas, okay I see a more pressing problem
[08:56:36] <wolfc> https://github.com/jbossas/jboss-as/commit/b528415d3e633c25d47eb535eadd29d430a7b2ef
[08:56:37] <jbossbot> git [jboss-as] b528415.. Stuart Douglas Add web components and CDI injection
[08:58:10] <wolfc> ah crock, interceptors and context are per instance. So the contract has changed again.
[08:58:19] <stuartdouglas> ?
[08:58:22] *** aloubyansky has joined #jboss-as7
[08:58:50] <wolfc> every time you change a constructor we get shit :-)
[09:00:28] <wolfc> https://github.com/dmlloyd/jboss-as/blob/ejb3-domain/ejb3/src/main/java/org/jboss/as/ejb3/component/stateless/StatelessSessionComponent.java#L84
[09:02:28] <wolfc> The contract of Component.createInstance & AbstractComponent.constructComponentInstance is a bit of a hassle
[09:04:17] <wolfc> Basically a ComponentInstance should only be made aware of two things, component and instance. The context and interceptors are details between Component and ComponentInstance I would say.
[09:05:40] <Jaikiran> wolfc: this fixes the compilation errors on dmlloyd's branch https://github.com/jaikiran/jboss-as/commits/ejb3
[09:05:43] <stuartdouglas> hmm
[09:06:02] <Jaikiran> i haven't really looked at any semantics of that, just fixed those compilation errors to get the build and smoke tests working
[09:06:52] <stuartdouglas> I made those changes to get around problems with lifecycle methods on interceptors being called on the wrong instance of the interceptor
[09:06:53] <wolfc> Jaikiran, I have the same here, only I don't pass the interceptors and context down yet.
[09:07:12] <stuartdouglas> I need to grab some dinner
[09:07:16] <stuartdouglas> will be back later
[09:07:39] <Jaikiran> yeah, those params keep changing so I decided to just pass them along for now and think about those later
[09:07:48] <wolfc> stuartdouglas, yes that is good. interceptors are tied to the instance.
[09:08:05] <stuartdouglas> there may be a better way however
[09:08:14] <wolfc> I just don't like the redundancy of constructInstance and constructComponentInstance
[09:08:19] <stuartdouglas> will have dinner and then will have a look
[09:08:46] <wolfc> Jaikiran, looks okay proceed on top of that patch
[09:08:53] <Jaikiran> ok
[09:13:40] *** Jaikiran has quit IRC
[09:25:15] *** emuckenhuber has joined #jboss-as7
[09:25:15] *** ChanServ sets mode: +v emuckenhuber
[09:26:29] *** mariolone has joined #jboss-as7
[09:26:30] *** mariolone has left #jboss-as7
[09:30:18] *** Jaikiran has joined #jboss-as7
[09:30:27] *** Jaikiran has quit IRC
[09:30:27] *** Jaikiran has joined #jboss-as7
[09:30:27] *** ChanServ sets mode: +v Jaikiran
[09:41:08] *** rmaucher has joined #jboss-as7
[09:41:52] *** miclorb_ has quit IRC
[09:45:59] *** Jaikiran is now known as Jaikiran|Lunch
[09:47:09] <wolfc> Jaikiran|Lunch, https://github.com/wolfc/jboss-as/commits/ejb3-ee
[09:48:00] <Jaikiran|Lunch> wolfc: I have the annotation processing currently under progress, so we should have something ready by today evening
[09:49:17] <wolfc> Is there some way to convince Intellij that @Resource.lookup exists and keep it that way? I moved the annotations module dependency above the JDK, but it doesn't stick.
[09:50:14] *** Jaikiran|Lunch has quit IRC
[09:50:25] *** Jaikiran has joined #jboss-as7
[09:50:28] *** Jaikiran has quit IRC
[09:50:28] *** Jaikiran has joined #jboss-as7
[09:50:28] *** ChanServ sets mode: +v Jaikiran
[10:04:03] *** maxandersen has joined #jboss-as7
[10:04:03] *** ChanServ sets mode: +v maxandersen
[10:14:39] *** maeste has joined #jboss-as7
[10:14:40] *** ChanServ sets mode: +v maeste
[10:17:41] *** ALR has joined #jboss-as7
[10:17:41] *** ChanServ sets mode: +v ALR
[10:23:16] *** jhalliday has joined #jboss-as7
[10:32:21] *** ALR1 has joined #jboss-as7
[10:34:41] *** ALR has quit IRC
[10:35:37] *** asoldano is now known as asoldano_away
[10:41:47] *** AndyTaylor has joined #jboss-as7
[10:41:47] *** ChanServ sets mode: +v AndyTaylor
[10:47:22] <stuartdouglas> wolfc: What do you think about moving all the component creation stuff out of Component and into ComponentInstance?
[10:47:34] <stuartdouglas> I think it would be a lot cleaner that what we have now
[10:47:51] <wolfc> stuartdouglas, sounds good
[10:48:23] <wolfc> can you type something up, then we can show it to dmlloyd?
[10:48:37] <stuartdouglas> sure
[10:49:40] *** mariolone has joined #jboss-as7
[10:49:41] *** mariolone has left #jboss-as7
[10:52:53] <stuartdouglas> hmm, this may not actually be better
[10:58:31] *** alesj has joined #jboss-as7
[10:59:57] *** kkhan has joined #jboss-as7
[10:59:57] *** ChanServ sets mode: +v kkhan
[11:02:01] *** epbernard has joined #jboss-as7
[11:02:01] *** ChanServ sets mode: +v epbernard
[11:02:27] *** epbernard is now known as emmanuel
[11:04:12] <stuartdouglas> wolfc: I think pushing the creation logic into the subclasses makes things a bit cleaner https://github.com/stuartwdouglas/jboss-as/commit/3069600f419392f1f5ae7212159d525e35af3630
[11:04:13] <jbossbot> git [jboss-as] 3069600.. Stuart Douglas Push creation logic into sub classes of AbstractComponent
[11:04:28] <stuartdouglas> as the creation logic will probably be quite different for ejb's
[11:05:22] <wolfc> stuartdouglas, the logic should be the same for all EE component types.
[11:05:40] <wolfc> delegate to CDI for instantiation, inject, postConstruct
[11:08:03] <stuartdouglas> well in that case I don't know
[11:08:19] *** asoldano_away is now known as asoldano
[11:09:57] *** miclorb_ has joined #jboss-as7
[11:10:58] *** ALR has joined #jboss-as7
[11:10:58] *** ChanServ sets mode: +v ALR
[11:11:14] <wolfc> stuartdouglas, I was more or less expecting that the logic would be in AbstractComponent and be called from the ComponentInstance constructor.
[11:12:02] <wolfc> You could even make the ComponentInstance responsible for the lifecycle of the real instance.
[11:12:09] <stuartdouglas> having it in the constructor just seems kinda messy
[11:12:30] *** jcosta has joined #jboss-as7
[11:12:30] *** ChanServ sets mode: +v jcosta
[11:12:33] <stuartdouglas> as if you put it in AbstractComponentInstance's constructor you loose all flexibility to override
[11:13:15] <stuartdouglas> and if you move it out to the individual instances it is pretty similar to what I had above
[11:13:20] <stuartdouglas> If we were only dealing with proxied components we could stuff the extra data we need like pre destroy interceptors into the proxy
[11:13:36] <stuartdouglas> and it would make this a lot simpler
[11:14:20] <stuartdouglas> but as it is the ComponentInstance has to hold on to stuff like pre destroy methods for non-proxied components
[11:15:58] <wolfc> we don't have any proxies components
[11:16:04] <wolfc> hmm, weirdness
[11:16:46] <wolfc> we have two definitions of proxies floating
[11:17:00] <wolfc> an EJB proxy can't hold onto interceptors or context or anything in any way
[11:17:16] <stuartdouglas> for managed beans it can
[11:17:18] <wolfc> it only has a session id (if stateful)
[11:18:03] <wolfc> in that case the proxy == component instance, not?
[11:18:11] <stuartdouglas> that session id should be enough state though
[11:19:27] <stuartdouglas> basically I have been trying to avoid an identity map of instance -> interceptors
[11:19:34] <stuartdouglas> for the web components
[11:19:42] <stuartdouglas> but it may be inevitable
[11:19:45] *** Jaikiran has quit IRC
[11:19:54] *** Jaikiran has joined #jboss-as7
[11:19:54] *** ChanServ sets mode: +v Jaikiran
[11:20:18] <wolfc> why would you want to proxy web components?
[11:20:58] <stuartdouglas> they can be managed beans or CDI beans, but actually that does need a redesign
[11:21:40] <stuartdouglas> cause in the simple non-proxied case it will be enough to just call the pre-destroy methods on the class
[11:21:49] <stuartdouglas> the problem then becomes CDI injections
[11:22:30] <stuartdouglas> as injected @Dependent scoped objects need to be destroyed when the web component is destroyed
[11:23:24] <wolfc> those are just lifecycle interceptors on the instance, it still doesn't warrant proxying
[11:25:05] <wolfc> I think the misconception comes for jboss-classrewriter or jboss-invocation proxies.
[11:25:28] <wolfc> Those are low level beasties in my book.
[11:27:16] <wolfc> stuartdouglas, do you put the ComponentInstance in that low level beasty?
[11:27:32] <stuartdouglas> I guess what I am trying to say is that this is the sort of info that can be attached to instances that are only accessed though a proxy quite easily
[11:28:21] <stuartdouglas> for managed beans the I think the component instance is accessible through the proxy handler
[11:29:42] <wolfc> I would for MB and Web, the ComponentInstance is the 'session identifier', for EJB it is the session id.
[11:29:53] <wolfc> Then the proxy just calls upon the Component with that session identifier.
[11:29:57] <stuartdouglas> session identifier?
[11:30:21] <stuartdouglas> MB's don't even have a well defined lifecycle all the time, only when they are injected into something that does
[11:30:48] <wolfc> So... treat everything the same.
[11:31:10] <wolfc> Don't think too much about that non-defined lifecycle.
[11:33:54] <stuartdouglas> so in the case of a SLSB, will there be a one to one relationship between ComponentInstance and SLSB's in the pool?
[11:34:51] <wolfc> The pool contains ComponentInstances
[11:38:23] <wolfc> Ehr, the answer is yes.
[11:41:42] * stuartdouglas has a G & T and thinks about it
[11:42:14] <wolfc> G&T?
[11:42:34] <stuartdouglas> Gin & Tonic
[11:42:58] <wolfc> Ah better :-)
[11:44:22] <stuartdouglas> basically at the moment all the state that ComponentInstance stores at the moment is stored/setup in the InterceptorFactoryContext and hidden behind Interceptor instances
[11:44:36] <stuartdouglas> I am not sure if that is to low level or not
[11:44:44] <alesj> wolfc: I see stuartdouglas is going down your alco path … G&T …
[11:45:04] <wolfc> yes I know. In the proxy one of the interceptors is an instance association interceptor.
[11:45:56] <wolfc> But those are component level interceptors. Some interceptor further down the line shoves in the instance interceptors, not?
[11:47:22] <wolfc> Hmm, we're missing some bits here. The EJB proxy should not have a direct connection with Component (and thus component level interceptors)
[11:47:34] <stuartdouglas> yea
[11:48:39] <wolfc> alesj, don't distract me, now I've made a bugger of my git branch :-)
[11:49:10] <wolfc> Jaikiran, my commit on ejb3 demo was bad I had to redo it. I've pushed a new ejb3-ee with TransactionManagement processing.
[11:49:40] <Jaikiran> wolfc: ok, let me pull that in
[11:53:25] <alesj> wolfc: geez, and I'm asking you for git advice …
[11:53:42] <alesj> wolfc: if you need any git help let me know :-)
[11:54:20] <stuartdouglas> wolfc: it looks like at the moment the component interceptors are tied very closely to the view
[11:56:25] *** bgeorges has quit IRC
[12:22:49] *** miclorb_ has quit IRC
[12:37:08] *** miclorb has joined #jboss-as7
[12:37:10] *** darranl has joined #jboss-as7
[12:37:10] *** darranl has joined #jboss-as7
[12:37:10] *** ChanServ sets mode: +v darranl
[12:42:00] *** miclorb has quit IRC
[12:56:19] <Jaikiran> wolfc: i just pushed a commit to my branch to include your tx processing changes https://github.com/jaikiran/jboss-as/commits/ejb3
[12:56:31] <Jaikiran> it's WIP but I pushed it so that we can work off the same code
[12:56:45] <wolfc> cool
[12:57:12] *** AndyTaylor has quit IRC
[12:57:16] <wolfc> what are all the in between commits on your branch/
[12:57:40] <Jaikiran> the one from stuart?
[12:57:44] <Jaikiran> i don't know actually
[12:57:56] * Jaikiran checks
[12:58:29] <Jaikiran> hmm looks like those logs are just reordered or something like that
[12:58:41] <wolfc> somehow upstream/master has been intermixed
[12:59:24] <Jaikiran> i don't remember playing with upstream
[12:59:27] <Jaikiran> let me check though
[13:00:23] <wolfc> I don't like the EjbDeploymentMarker
[13:00:51] <wolfc> I think its better to just look at the type of ComponentDescription
[13:01:20] <Jaikiran> yeah, that was how i initially implemented it locally i.e. instanceof check
[13:01:22] <Jaikiran> but didn't like it
[13:01:23] <wolfc> Jaikiran, did you actually see a CCE?
[13:01:32] <Jaikiran> wolfc: yeah, on the managed bean smoke tests
[13:01:40] <wolfc> Weird, smoke tests passed here
[13:02:38] *** ALR has quit IRC
[13:03:40] <Jaikiran> the CCE http://pastebin.com/jL6YgUqh
[13:03:57] <Jaikiran> wolfc: btw, what's wrong with the marker?
[13:04:02] <Jaikiran> i mean conceptually
[13:04:13] <Jaikiran> i see there's other techs too using similar construct
[13:04:55] <Jaikiran> wolfc: i think your smoke tests passed because you did not have the DUP which adds the EJB component description
[13:05:02] <Jaikiran> i just introduced that one
[13:05:13] <Jaikiran> ah nah! ignore that
[13:05:45] * Jaikiran now goes back to figure out why those upstream changes came in
[13:07:53] <wolfc> Jaikiran, a deployment may contain multiple component types
[13:08:04] <wolfc> So we need to have that instanceof
[13:08:34] <wolfc> wicked, smoke-tests are passing locally for me. It must be one of those upstream commits.
[13:09:34] <Jaikiran> hmm the only thing that remotely was related to upstream master was my local branch from which i did a git checkout -b ejb3-ee
[13:10:06] <Jaikiran> and then a rebase -p wolfc/ejb3-ee
[13:10:16] <Jaikiran> other than that i don't remember doing anything with upstream
[13:10:29] <Jaikiran> let me figure out and get rid of those upstream commits in that branch
[13:11:07] <wolfc> I got it on my ejb3-ee branch
[13:12:52] <Jaikiran> ok cool, i'll just use that afresh then
[13:17:12] *** aloubyansky has quit IRC
[13:18:14] <wolfc> Jaikiran, the EjbAnnotationProcessor is working the wrong way round
[13:19:06] <wolfc> Once you know the annotation (@Stateless) it should proceed on that knowledge, not build up and query again
[13:19:40] <Jaikiran> query the index?
[13:19:57] <Jaikiran> it queries it just once
[13:20:24] *** asoldano is now known as asoldano_lunch
[13:20:51] <wolfc> The for loop over sessionBeanAnnotations is silly
[13:23:00] <Jaikiran> we do need to loop over those annotations
[13:23:31] <wolfc> No, you have the index to immediately get the annotations
[13:23:59] <Jaikiran> right and we do get the annotations
[13:24:16] <Jaikiran> via the getSessionBeans call
[13:24:34] <wolfc> where you make a copy and loop over the copy. That's bad practice :-)
[13:26:02] <Jaikiran> we can get rid of the copy, but i still don't see how (or why) we can get rid of the loop :)
[13:26:15] *** davidbos has joined #jboss-as7
[13:28:16] <wolfc> Jaikiran, yes loop over annotations directly, not over a copy. I pushed a new ejb3-ee it does an EJB deployment. It goes NYI at EJBComponentDescription.createComponentConfiguration
[13:28:32] <Jaikiran> ok, let me get in those changes
[13:32:28] <wolfc> I think now is the time to get rid of SessionComponentFactory
[13:32:32] <Jaikiran> wolfc: btw, the git checkout -b ejb3-ee was the culprit which caused the upstream commits to come in
[13:32:46] <Jaikiran> i was on my local master when i did that
[13:32:51] <wolfc> hehe
[13:32:52] <Jaikiran> and it brought in the upstream changes
[13:33:11] <wolfc> I still find it weird that the smoke-tests passed for me locally
[13:33:20] <wolfc> That should not have happened
[13:35:21] *** davidbos has quit IRC
[13:35:21] *** Jaikiran has quit IRC
[13:35:21] *** nickarls has quit IRC
[13:40:50] *** davidbos has joined #jboss-as7
[13:40:50] *** Jaikiran has joined #jboss-as7
[13:40:50] *** nickarls has joined #jboss-as7
[13:40:50] *** asimov.freenode.net sets mode: +v Jaikiran
[13:48:07] <darranl> Do we have a preferred class in AS7 for Base64 handling?
[13:48:32] *** maxandersen has quit IRC
[13:53:24] *** maxandersen has joined #jboss-as7
[13:53:24] *** ChanServ sets mode: +v maxandersen
[13:54:22] *** maxandersen has quit IRC
[13:54:50] *** maxandersen has joined #jboss-as7
[13:54:50] *** ChanServ sets mode: +v maxandersen
[14:06:49] *** jpederse has joined #jboss-as7
[14:07:35] *** jpederse has quit IRC
[14:07:35] *** jpederse has joined #jboss-as7
[14:09:08] *** smarlow has joined #jboss-as7
[14:09:08] *** ChanServ sets mode: +v smarlow
[14:18:18] <smarlow> stuartdouglas:  Hi, any luck with the JDBC driver classpath issue?
[14:22:25] <Jaikiran> wolfc: regarding this: (06:01:01  IST) wolfc: I think now is the time to get rid of SessionComponentFactory
[14:22:34] <Jaikiran> I don't see that class
[14:22:53] <Jaikiran> or did you mean the StatelessSessionComponentFactory?
[14:23:22] <wolfc> Yes, both of them
[14:25:16] *** asoldano_lunch is now known as asoldano
[14:29:05] *** frainone has joined #jboss-as7
[14:29:05] *** ChanServ sets mode: +v frainone
[14:31:09] <smarlow> stuartdouglas, Nihility :  I was just talking with jpederse + maeste about JBAS-8925, which might be the same issue we were hitting last night.
[14:31:12] <jbossbot> jira [JBAS-8925] Failure in the rar demo smoke test [Open (Unresolved) Bug, Blocker, Jason Greene] https://issues.jboss.org/browse/JBAS-8925
[14:34:19] *** AndyTaylor has joined #jboss-as7
[14:34:19] *** ChanServ sets mode: +v AndyTaylor
[14:34:23] <maeste> smarlow, stuartdouglas , Nihility : yep it's related, even if not exactly the same. As said when someone will have time to look at these cl issues I can help discussion and analysis with all the info I've collected. If you are going to do that during evening/night of europe TZ, just let me know a bit of time before, and I'll arrange to be online.
[14:34:42] <smarlow> stuartdouglas, Nihility: that might be easier to recreate than the JPA testcase
[14:39:08] <Jaikiran> wolfc: i have got it past the component creation issues. now i'll get the bindings and view processing working
[14:40:24] <wolfc> maeste, speaking of rars, is there a hornetq rar for AS 7? Hmm, it should actually be the standard hornetq rar, right?
[14:42:13] <maeste> wolfc: yup I think so. jpederse do you agree?
[14:42:56] <wolfc> maeste, there was one issue where the hornetq rar would fire up the endpoint even when the underlying resource is not available yet, is that resolved?
[14:43:06] <jpederse> wolfc: the HQ rar will be available in the next official drop from the team
[14:43:21] <jpederse> wolfc: I assume they will then integrate it in AS 7
[14:43:36] <jpederse> wolfc: we are using a snapshot tag atm for testing
[14:44:08] <jpederse> AndyTaylor: right ? ^
[14:45:21] <AndyTaylor> jpederse: ? if ones needed then i guess we will provide one, not sure what you actually mean tho?
[14:45:34] <AndyTaylor> jpederse: your not talking about the RA are you?
[14:45:56] <jpederse> AndyTaylor: yeah, pulling the .rar from the Nexus repo into the AS 7 dist
[14:46:27] <AndyTaylor> jpederse: jpederse the hornetq ra rar?
[14:46:33] <jpederse> AndyTaylor: ypu
[14:46:36] <jpederse> * yup
[14:46:53] <AndyTaylor> jpederse: i think the last QA build we did had that, so yeah, we'll provide it at some point
[14:47:28] <jpederse> AndyTaylor: yeah, precisely - its just not in AS 7 yet, but will be part of your next pull request
[14:47:46] <AndyTaylor> jpederse: yip
[14:49:59] *** mbg has joined #jboss-as7
[14:49:59] *** ChanServ sets mode: +v mbg
[15:04:26] *** pgier has joined #jboss-as7
[15:04:27] *** ChanServ sets mode: +v pgier
[15:11:59] *** fnasser has joined #jboss-as7
[15:19:53] *** smarlow has quit IRC
[15:22:03] *** maxandersen has quit IRC
[15:24:14] *** bstansberry has joined #jboss-as7
[15:24:14] *** ChanServ sets mode: +v bstansberry
[15:25:20] *** maxandersen has joined #jboss-as7
[15:25:27] *** ChanServ sets mode: +v maxandersen
[15:27:13] *** pferraro has joined #jboss-as7
[15:27:13] *** ChanServ sets mode: +v pferraro
[15:31:15] *** smarlow has joined #jboss-as7
[15:31:15] *** ChanServ sets mode: +v smarlow
[15:31:38] *** maxandersen has quit IRC
[15:34:22] *** maxandersen has joined #jboss-as7
[15:34:22] *** ChanServ sets mode: +v maxandersen
[15:36:56] <dmlloyd> good morning
[15:37:18] <kkhan> morning
[15:37:58] <opalka> afternoon
[15:43:19] <darranl> I know we have a copy of the HTTP server for domain management in git - is that the version we are currently using?
[15:56:17] *** aslak has joined #jboss-as7
[15:56:17] *** aslak has joined #jboss-as7
[15:56:17] *** ChanServ sets mode: +v aslak
[15:58:12] *** frainone has quit IRC
[16:04:11] *** asaldhan has joined #jboss-as7
[16:04:11] *** ChanServ sets mode: +v asaldhan
[16:04:30] *** aslak has quit IRC
[16:04:41] <wolfc> Jaikiran, how are you going?
[16:04:57] <wolfc> dmlloyd, ping
[16:05:03] *** aslak has joined #jboss-as7
[16:05:03] *** aslak has joined #jboss-as7
[16:05:03] *** ChanServ sets mode: +v aslak
[16:05:04] <dmlloyd> what's up
[16:05:39] <wolfc> dmlloyd, I want to be more strict in defining the responsibility of Component and ComponentInstance. Right now we keep having to change little bits and pieces.
[16:05:52] <dmlloyd> stuartdouglas: ping me tomorrow morning when your up and remind me to explain how proxy association works
[16:06:10] <dmlloyd> wolfc: example?
[16:06:23] <wolfc> From what I recall it should be an interceptor in the proxy itself.
[16:07:04] <dmlloyd> what should be?
[16:07:15] <dmlloyd> oh association?  yeah I was just reading what stuartdouglas wrote last night
[16:07:17] <wolfc> instance association
[16:07:27] <wolfc> https://github.com/wolfc/jboss-as/commit/49fc484f599e9e6de3b7ff37c8e085760be5ad8a
[16:07:28] <jbossbot> git [jboss-as] 49fc484.. jaikiran Fix compilation errors
[16:07:30] <dmlloyd> and I realized I never took him through that
[16:07:54] <dmlloyd> what branch is this wolfc?
[16:07:56] <wolfc> we got too many construct instance methods
[16:08:04] <wolfc> an extension of yours
[16:08:24] *** aslak has quit IRC
[16:08:28] <alesj> dmlloyd: how much should JSF work in AS7?
[16:08:52] <dmlloyd> alesj: it should work pretty well at this point, stan had sent out a status a while back
[16:09:23] <alesj> java.lang.IllegalStateException: Application was not properly initialized at startup, could not find Factory: javax.faces.context.FacesContextFactory
[16:09:23] <alesj> 	at javax.faces.FactoryFinder$FactoryManager.getFactory(FactoryFinder.java:804)
[16:09:23] <alesj> 	at javax.faces.FactoryFinder.getFactory(FactoryFinder.java:306)
[16:09:30] <dmlloyd> wolfc: I don't see anything unusual here
[16:09:34] <alesj> java.lang.IllegalStateException: Application was not properly initialized at startup, could not find Factory: javax.faces.context.FacesContextFactory
[16:09:34] <alesj> 	at javax.faces.FactoryFinder$FactoryManager.getFactory(FactoryFinder.java:804)
[16:09:36] <wolfc> stuartdouglas modified it to https://github.com/stuartwdouglas/jboss-as/commit/3069600f419392f1f5ae7212159d525e35af3630
[16:09:37] <jbossbot> git [jboss-as] 3069600.. Stuart Douglas Push creation logic into sub classes of AbstractComponent
[16:09:48] <wolfc> as an experiment
[16:10:10] <dmlloyd> for what purpose
[16:10:13] <wolfc> the changing constructor keeps bugging me
[16:10:23] <wolfc> and maybe that is just it
[16:10:50] <dmlloyd> oh so his goal here is to have a simpler constructor signature?
[16:11:17] <dmlloyd> at the cost of a more complex constructComponentInstance() method
[16:11:24] <dmlloyd> dunno this change doesn't seem that important to me
[16:11:26] <wolfc> or even give the ComponentInstance class some responsibility
[16:11:31] <dmlloyd> seems more like scratching an itch
[16:11:38] <wolfc> yeah
[16:11:47] <wolfc> a nagging itch :-)
[16:11:55] *** bgeorges has joined #jboss-as7
[16:12:19] <dmlloyd> the ComponentInstance class doesn't need responsibility
[16:12:33] <dmlloyd> I don't think any of this should change until after all of our component types are implemented
[16:12:47] <dmlloyd> then we can use the superior design power of hindsight :)
[16:12:51] *** aloubyansky has joined #jboss-as7
[16:13:23] <wolfc> Yeah, because 3069600f41 starts duplicating the logic. Not good either.
[16:14:45] <wolfc> dmlloyd, just give a heads up beforehand when a branch doesn't compile anymore. It brings back AS 4 nightmares.
[16:15:41] <wolfc> luckily we're not on svn anymore
[16:17:27] <dmlloyd> okay well I really want to get this merged into upstream asap
[16:18:10] <dmlloyd> really as soon as we have the build working (with smoke tests, even if we have to disable EJB's for now), and the server starting with at least no regressions wrt upstream, then it should go in
[16:18:19] <dmlloyd> I feel we've missed 2-3 opportunities to do this :)
[16:18:37] <dmlloyd> then we can work with minimal patches against upstream instead of a big branch
[16:20:19] <wolfc> yup, that would alleviate a lot of trouble
[16:20:41] <wolfc> having diverging branches is just as bad as too many components ;-)
[16:21:03] <dmlloyd> yeah I never intended to be a maintainer for this thing
[16:21:10] <emuckenhuber> bstansberry: a while ago we discussed not exposing some of our internal operations in the description - so can you review: https://github.com/emuckenhuber/jboss-as/commit/4ab2f38122a1acbb52bac89b8bdf477e2dd161fe
[16:21:11] <jbossbot> git [jboss-as] 4ab2f38.. Emanuel Muckenhuber don't expose internal operations
[16:21:12] <dmlloyd> just want to get it in asap
[16:25:33] <bstansberry> emuckenhuber: would this affect anything being merged in? (e.g. EJB?)  A quick look says no; you've just added a method
[16:26:17] <emuckenhuber> bstansberry: yes, it's an additional method - so it should not affect existing integration code
[16:29:31] *** opalka has quit IRC
[16:33:41] <Jaikiran> wolfc: sorry, missed that note
[16:33:55] <wolfc> How far are you?
[16:33:56] <Jaikiran> wolfc: yeah, the @Local binding is almost done
[16:33:59] <wolfc> Cool
[16:34:05] <Jaikiran> i'll push the changes in the next hour or so
[16:34:13] <wolfc> The demo assumes that I have a default local interface
[16:34:23] <Jaikiran> yeah, the corner cases
[16:34:28] <wolfc> :-)
[16:36:00] <bstansberry> emuckenhuber, kkhan: WDYT about having operation registration look more like attribute registration? provide an indication whether the op is RO or RW?
[16:36:06] <bstansberry> post M2
[16:36:42] <bstansberry> when the master DC handles an op against a domain-level resource, it's important to know if it is RO
[16:37:03] <kkhan> Maybe add isReadOnly() to the OperationHandler interface
[16:37:21] <dmlloyd> adding constant queries to interfaces is a red flag imo
[16:37:27] <dmlloyd> maybe a @ReadOnly annotation though ;)
[16:37:35] <bstansberry> there you go!
[16:37:38] <bstansberry> and @Roles
[16:37:44] <dmlloyd> sure
[16:37:52] <dmlloyd> you could really go to town with it
[16:38:12] *** AndyTaylor has left #jboss-as7
[16:38:12] <dmlloyd> could even use jandex to auto-register them.... no... someone stop me
[16:38:25] <bstansberry> I keep forgetting we have these marker interfaces already
[16:38:41] <bstansberry> ModelUpdateOperationHandler etc
[16:38:42] <dmlloyd> yeah marker interfaces are a great evil.  I am ashamed
[16:38:48] <kkhan> Weren't you going to do something last week with those?
[16:39:05] <bstansberry> like I said, I keep forgetting!
[16:39:28] <kkhan> Repeat 100000000000000 times after me ;-)
[16:39:52] <bstansberry> those interfaces are a great evil; we should come up with something better after M2
[16:40:30] <bstansberry> but, the pre-M2 pile is big enough, so back into my cave I go
[16:43:17] <bstansberry> emuckenhuber: please rebase
[16:43:23] <bstansberry> that branch is way behind
[16:44:09] <emuckenhuber> hmm, ok i thought i did that
[16:47:29] *** mbg is now known as mbg|away
[16:50:12] <alesj> dmlloyd, Nihility: can we somehow confirm that JSF works?
[16:50:38] *** mbg|away is now known as mbg
[16:50:40] <dmlloyd> I don't have any JSF apps handy
[16:50:55] <dmlloyd> might want to ping stan, I'm sure he has some test apps
[16:51:58] *** maxandersen has quit IRC
[16:52:12] <alesj> dmlloyd: afaik, stan is on vacation?
[16:52:29] <dmlloyd> ah
[16:52:34] <dmlloyd> don't know
[16:52:34] <kkhan> stan's on holiday
[16:52:52] <kkhan> JSF is meant to work
[16:53:16] <alesj> kkhan: java.lang.IllegalStateException: Application was not properly initialized at startup, could not find Factory: javax.faces.context.FacesContextFactory
[16:53:17] <kkhan> http://community.jboss.org/docs/DOC-16546
[16:53:18] <alesj> :-(
[16:53:58] <kkhan> I am trying to get http://community.jboss.org/wiki/JSFUnit200GettingStartedExample ported to a smoke test
[16:54:04] <kkhan> But easier said than done
[16:54:29] <kkhan> Still sorting out the arquillian parts
[16:54:36] <kkhan> but the app seems to deploy
[16:54:46] <kkhan> Haven't tried calling it yet
[16:55:06] <kkhan> Apart from that I know f**k-all about jsf :-)
[16:55:41] <alesj> kkhan: :-)
[16:55:44] *** jpederse has quit IRC
[16:55:50] <alesj> did Stan already merged his work?
[16:55:52] *** davidbos has quit IRC
[16:55:58] *** jpederse has joined #jboss-as7
[16:56:03] <alesj> or is that DOC related to some branch of his?
[16:56:39] <kkhan> AFAICT all the commits from https://github.com/ssilvert/jboss-as/tree/JBAS-8876_JSF_with_Bean_Validation are in master
[16:56:40] <jbossbot> jira [JBAS-8876] Add JSF to AS7 [Resolved (Done) Task, Major, Stan Silvert] https://issues.jboss.org/browse/JBAS-8876
[16:56:47] *** jpederse has quit IRC
[16:56:47] *** jpederse has joined #jboss-as7
[16:58:14] *** slaboure has joined #jboss-as7
[16:59:57] *** maxandersen has joined #jboss-as7
[16:59:57] *** ChanServ sets mode: +v maxandersen
[17:04:03] *** Jaikiran is now known as Jaikiran|Dinner
[17:06:27] *** smarlow has quit IRC
[17:06:50] <Jaikiran|Dinner> wolfc: fyi - i'll be back after dinner and will push my changes. i am debugging an issue where the bindings aren't making into jndi
[17:06:54] <Jaikiran|Dinner> will get past that soon
[17:06:58] *** smarlow has joined #jboss-as7
[17:06:58] *** ChanServ sets mode: +v smarlow
[17:07:29] <aloubyansky> bstansberry: are there any restrictions for operation names? and for property names?
[17:08:32] *** bgeorges has quit IRC
[17:09:29] <bstansberry> not enforced, but they should follow the same rules as child types
[17:10:21] <bstansberry> for ops, everything is alpha-numeric + '-'
[17:11:15] <bstansberry> we need to be able to translate them into mbean attributes and operations via a simple algorithm
[17:12:50] <aloubyansky> so, for parameter names i can re-use the node type pattern?
[17:13:43] <bstansberry> yes
[17:14:13] <aloubyansky> ok, thanks
[17:14:23] *** balunasj has joined #jboss-as7
[17:14:24] *** balunasj has joined #jboss-as7
[17:14:24] *** ChanServ sets mode: +v balunasj
[17:16:59] <kkhan> dmlloyd: Re my earlier mail, I guess that dependencies from the manifest are imported without META-INF/services?
[17:18:10] <dmlloyd> right
[17:18:21] <dmlloyd> but services bundled should be seen
[17:18:42] <kkhan> I'm trying another approach from earlier
[17:18:44] <emuckenhuber> bstansberry: ok, rebased against master: https://github.com/emuckenhuber/jboss-as/commit/ed1eaae5959bf2881c681fe10c8daec3aba81166
[17:18:45] <jbossbot> git [jboss-as] ed1eaae.. Emanuel Muckenhuber don't expose internal operations
[17:19:42] <kkhan> The JSF unit stuff was bundling all the dependencies into the jar, I'm trying to create jsfunit modules so that it works more similarly to how the other arquillian tests work
[17:19:50] *** ALR has joined #jboss-as7
[17:19:50] *** ChanServ sets mode: +v ALR
[17:20:10] <bstansberry> emuckenhuber: thanks
[17:20:11] <kkhan> But for that to work I think I would need to be able to see services from other modules
[17:20:23] <kkhan> Say I have
[17:20:35] <kkhan> Dependencies: some.module1,some.module2
[17:20:44] <kkhan> and I want the services for some.module2
[17:21:07] <kkhan> Would you object to me adding something like the following to manifest.mf
[17:21:16] <kkhan> Services: some.module2
[17:23:08] <Nihility> alesj: is this with jbossweb
[17:23:08] <Nihility> ?
[17:24:00] <kkhan> Nexus is down
[17:24:20] <Nihility> pgier mentioned an upgrade was happening soon
[17:24:43] <kkhan> pgier: Any ETA on when nexus is up again?
[17:25:37] <jbossbot> git [jboss-as] push master ed1eaae.. Emanuel Muckenhuber don't expose internal operations
[17:25:37] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/f9914f3...ed1eaae
[17:28:18] <pgier> kkhan: no ETA, eng-ops is working on it
[17:28:39] <kkhan> ok thanks
[17:28:41] <pgier> it seems to be the whole server that is down
[17:28:59] <pgier> I don't know if Nexus did that or not
[17:29:28] *** frainone has joined #jboss-as7
[17:29:29] *** ChanServ sets mode: +v frainone
[17:29:36] *** jamezp has joined #jboss-as7
[17:31:23] <kkhan> I rolled back the latest commits that wanted new stuff from nexus, so I'm ok for now
[17:35:49] <Nihility> i just downloaded some stuff
[17:35:53] <Nihility> from nexus
[17:37:08] *** mbg is now known as mbg|away
[17:37:32] <kkhan> Yeah, it is up again
[17:42:16] <aloubyansky> bstansberry: about escaping special chars... what if in that case we require quotes for the whole node name? otherwise the quotes would be optional
[17:44:02] <bstansberry> that sounds simple and clear to explain, and easy to parse, so +1
[17:45:04] <aloubyansky> i'm gonna do that
[17:47:47] *** mbg|away is now known as mbg
[17:57:05] <pgier> kkhan: did you release a new version of the forked surefire plugin?
[17:57:42] <pgier> Maven 3.0.3 has been released and the as7 build currently fails with the new version of Maven
[17:58:17] *** AndyTaylor has joined #jboss-as7
[17:58:17] *** ChanServ sets mode: +v AndyTaylor
[17:58:21] <kkhan> pgier: I can release it now
[17:59:19] <Nihility> great
[17:59:27] <Nihility> pgier: any idea why?
[17:59:41] <kkhan> pgier gave me a patch some days ago
[17:59:46] *** balunasj has quit IRC
[18:00:26] <pgier> Nihility: there was a change in the way Maven handles plugin params (http://jira.codehaus.org/browse/MNG-5027)
[18:00:27] <jbossbot> jira [MNG-5027] Problem executing custom surefire implementation in Maven 3.0.3-RC1 [Closed (Not A Bug) Bug, Major, Benjamin Bentmann] http://jira.codehaus.org/browse/MNG-5027
[18:01:15] * dmlloyd pets jbossbot
[18:01:41] <Nihility> pgier: hmm cant we patch that then in our forked plugin
[18:04:05] *** AndyTaylor has quit IRC
[18:04:26] <asoldano> Nihility, hi
[18:04:43] <pgier> Nihility: yes, I gave kkhan a patch for that last week or week before ^^
[18:06:38] *** jcosta has quit IRC
[18:07:09] <Nihility> wow eclipse just royaly pissed me off
[18:07:18] <Nihility> it would not refresh its class index
[18:07:25] <Nihility> even though i clean/build
[18:07:32] <Nihility> told it to update project config from maven
[18:07:36] <Nihility> told it to update deps
[18:07:41] <Nihility> cleaned on the command
[18:07:43] <Nihility> line
[18:07:44] <Nihility> etc etc
[18:07:51] <darranl> Nihility, Eclipse regularly pisses me off - especially as it's formatting doesn't match chackstyle !!
[18:07:53] <Nihility> i actually had to "close" all projects
[18:07:56] <Nihility> and reopen it
[18:07:58] <kkhan> I've downloaded intellij but haven't started using it yet
[18:08:06] <Nihility> its total shit
[18:08:12] <Nihility> the combination of eclipse + m2e
[18:08:19] <darranl> Nihility, and what is the point of the press F5 to refresh this file message in Eclipse all about - does it server any purpose?
[18:08:30] <Nihility> lol
[18:08:35] <Nihility> darranl: to annoy you
[18:08:46] <Nihility> yeah im done with eclipse
[18:08:50] <darranl> It can tell it needs to be refreshed so why doesn't it just go ahead and refresh it !!
[18:08:50] <Nihility> switching to intellij
[18:08:57] *** aslak has joined #jboss-as7
[18:08:57] *** aslak has joined #jboss-as7
[18:08:57] *** ChanServ sets mode: +v aslak
[18:09:11] <Nihility> random hangs
[18:09:13] <Nihility> crashes
[18:09:18] <Nihility> performance problems
[18:09:23] <Nihility> broken behavior
[18:10:01] <darranl> And every time I create a new Eclipse workspace I get a series of things that want to share what I am doing in an anonymous way !
[18:10:50] <Nihility> maxandersen: i honestly dont know how our users get by using eclipse
[18:11:03] <Nihility> maxandersen: its become an awful platform to work on
[18:13:59] <asoldano> Nihility, when you have few minutes, please take a look at my email on as7 dev mailing list... still on your jaxp-related changes in modules
[18:18:42] *** aloubyansky has quit IRC
[18:18:58] <kkhan> Nihility: https://github.com/kabir/jboss-as/commit/e18acb5a42d2d28b06cf7a4a482d83d831889b59
[18:18:59] <jbossbot> git [jboss-as] e18acb5.. kabir Upgrade to 1.0.0.Alpha3 of modular surefire plugin to avoid issues on maven 3.0.3
[18:19:02] <kkhan> pgier: ^^
[18:19:34] <pgier> kkhan: thanks
[18:20:00] *** emuckenhuber has quit IRC
[18:25:46] <dmlloyd> hm, we leak module loaders on reload
[18:25:49] * dmlloyd makes a note for later
[18:32:30] *** jfclere has quit IRC
[18:34:52] <darranl> Nihility, BTW that message telling you to refresh the file by pressing F5 - F5 never works for me so I always have to do File->Refresh just to make it that little more interactive ;-)
[18:39:46] <alesj> Nihility: that JSF is with jetty :-)
[18:39:51] <asoldano> darranl, that's because the focus is not on the editor area... ?
[18:39:54] <Nihility> darranl: it works you just have to click the file tab on the left
[18:40:26] <alesj> Nihility: i'll try and debug, but any hint is appreciated ...
[18:40:33] <darranl> Nihility, asoldano - so to refresh my file, even with a given keyboard shortcut I still need the mouse !!!
[18:40:41] <asoldano> darranl, right ;-)
[18:40:56] <Nihility> alesj: i am going to write a detailed clasloading wiki today
[18:40:57] <alesj> Nihility:  see email for more details ;-)
[18:41:11] <alesj> Nihility: nah, i'm fine with Modules
[18:41:21] <Nihility> yeah i mean with factories and TCCL
[18:41:22] <alesj> Nihility: been hacking on them for C...
[18:41:40] <Nihility> is the issue that it cant find JSF-API
[18:41:43] <Nihility> or the JSF impl
[18:41:46] <Nihility> let me see the email
[18:42:01] <alesj> Nihility: gotta go, i'll be back in the evening
[18:42:04] <Nihility> the module should have JSF-api imported
[18:42:11] <alesj> push me an email if you find something out
[18:42:14] <Nihility> ok
[18:42:17] <alesj> thanks
[18:42:19] <alesj> ttyl
[18:42:28] *** alesj has quit IRC
[18:59:48] *** alesj has joined #jboss-as7
[19:00:18] *** smarlow has quit IRC
[19:08:06] *** smarlow has joined #jboss-as7
[19:08:06] *** ChanServ sets mode: +v smarlow
[19:11:30] <alesj> Nihility: WarClassloadingDependencyProcessor adds JSF impl dependency
[19:11:41] <alesj> which probably already depends on JSF api
[19:11:54] <Nihility> ah thats wrong
[19:12:11] <alesj> dunno for the api dependency pull in
[19:12:13] <alesj> just guessing
[19:12:20] <Nihility> yeah the way everything should work is
[19:12:26] <alesj>         ModuleDependency jsf = new ModuleDependency(moduleLoader, JSF_IMPL, false, false, false);
[19:12:27] <alesj>         jsf.addImportFilter(PathFilters.getMetaInfFilter(), true);
[19:12:27] <alesj>         moduleSpecification.addDependency(jsf);
[19:12:35] <Nihility> deployment imports api
[19:12:37] <alesj> this is from WarCLDP
[19:13:05] <dmlloyd> yeah I told stan he could do that temporarily until we can come up with a better solution (i.e. customized JSF API)
[19:13:26] <Nihility> if the api is META-INF/service based it should try TCCL and fall back to APIClass.getClassLoader()
[19:13:48] <dmlloyd> nah it has a FactoryFinder of its own
[19:13:55] <dmlloyd> not sure what it does
[19:14:01] <alesj> me neither
[19:14:02] <Nihility> ill look
[19:14:15] <alesj> IDEA doesn't know how to import sources … :-(
[19:14:34] <alesj> cannot find them ..
[19:14:51] <dmlloyd> yeah it does
[19:15:04] <dmlloyd> go to your maven window and click "get sources", it's the down arrow pointing into a box
[19:15:23] <alesj> yeah, that i know
[19:15:36] <alesj> but for some reason jsf sources are not accessible
[19:15:48] <alesj> do you get them from AS7 project?
[19:16:11] <dmlloyd> yeah it's in the spec JAR
[19:16:24] <dmlloyd> however I think that the different impls may require their own API JARs
[19:16:28] <dmlloyd> I don't recall the details
[19:16:52] <dmlloyd> I actually have the jboss spec project checked out in my AS7 project workspace so I guess that's cheating
[19:17:19] *** bstansberry is now known as bstans_afk
[19:18:47] <Nihility> ok i looked
[19:18:50] <Nihility> its wrong
[19:19:02] <alesj> ah, it picked up sources now … weird
[19:19:09] <Nihility> it looks in TCCL
[19:19:13] <alesj> i guess i just needed to be back home :-)
[19:19:14] <Nihility> does not fall back to the API CL
[19:19:40] <Nihility> why is it that all META-INF/services APIs are wrong :)
[19:20:11] <alesj> but I would expect the TCCL to be properly set
[19:20:23] <Nihility> it seems like anytime they run into a classloading problem they play "oh we guessed the wrong classloader"
[19:20:37] <alesj> since this happens at first request … or not … hmmm
[19:21:39] <alesj> it breaks at first request, but probably due to not being properly initialized at init time … not finding the factory
[19:22:01] <Nihility> alesj: is the JSF module exporting services
[19:22:23] <alesj> http://pastebin.com/yW9VEhYg
[19:22:32] <Nihility> nope its not
[19:22:59] <alesj> JSF module exporting service?
[19:23:42] <Nihility> dmlloyd: i forget are services exported by default?
[19:23:43] <Nihility> i thought not
[19:24:15] <alesj> i did change that boolean flag in WebCLDP
[19:24:19] <alesj> to export services
[19:24:23] <alesj> but no change
[19:24:27] <maxandersen> Nihility: you can enable autorefrseh for eclipse.
[19:24:31] <Nihility> yeah you dont need to export them in your deployment
[19:24:41] <maxandersen> Nihility: don't install m2eclipse wonky versions -> Main reason for crashes.
[19:24:58] <Nihility> maxandersen: i never could find a decent working m2e
[19:25:14] <Nihility> maxandersen: i will certainly agree that m2e is the source of 99% of the issues
[19:25:36] <Nihility> maxandersen: it should be called make-eclipse-slow-and-unstable
[19:25:49] <maxandersen> Nihility: Eclipse is the *fastest* IDE of all when it comes to feedback…everytime I try intellij/netbeans i'm speechless how long I have to wait to get an actual compile error/warning.
[19:26:12] <maxandersen> Nihility: that said - I completely understand the problems…
[19:26:33] <dmlloyd> Nihility: yes, everything is exported by default
[19:26:35] <maxandersen> Nihility: which version of m2e are you using ?
[19:26:41] <dmlloyd> Nihility: filtering happens at import by default
[19:27:16] <Nihility> maxandersen: 0.10.2
[19:27:44] <Nihility> services="import" should work
[19:27:56] <Nihility> hmm
[19:29:02] <Nihility> alesj: yeah your change should have done it
[19:29:12] <Nihility> alesj: however maybe jetty replaces the TCCL
[19:29:13] <alesj> who adds jsf-api to the apps modules?
[19:32:15] <Nihility> alesj: javaee.api
[19:32:33] <Nihility> which exports javax.faces.api
[19:35:25] <alesj> let me check what's the TCCL at the time we call that JSF' FF
[19:36:28] <alesj> ModuleClassLoader for Module "deployment.weld-permalink.war:main" from Service Module Loader
[19:36:37] <alesj> it looks the right one ...
[19:36:49] <Nihility> interesting
[19:37:10] <Nihility> if you step into that resource lookup does it succeed
[19:37:38] <alesj> META-INF/services/javax.faces.context.FacesContextFactory
[19:37:48] <alesj> lookup of this returns empty enumeration
[19:37:58] <alesj> hence no impl is found
[19:38:05] <dmlloyd> guys, kkhan also ran into an odd issue about service loading
[19:38:11] <dmlloyd> it might be related
[19:38:13] <Nihility> ok so somehow the module definition has to be wrong
[19:38:20] <dmlloyd> a services file, on the resource root, not being found
[19:38:34] <dmlloyd> which is quite odd because the resource roots should be visible
[19:38:48] <dmlloyd> unless someone installed a filter of some sort on it
[19:39:14] <Nihility> well in this case its an import
[19:39:35] <dmlloyd> yeah but I still have a hunch that the two problems are related
[19:40:50] <alesj> let me see how far i can debug the services lookup ...
[19:41:27] <Nihility> hmm
[19:41:33] <Nihility> maybe the modulespecprocessor is wrong
[19:42:27] <Nihility> hmm it looks right
[19:45:49] <alesj> I'm looking at Module "com.sun.jsf-impl:main" from local module loader @6490832e (roots: /Users/alesj/projects/cloud/as7/build/target/jboss-7.0.0.Alpha2/modules)
[19:45:50] <dmlloyd> jconsole may illuminate
[19:46:02] <dmlloyd> can you pastebin it alesj?
[19:46:07] <alesj> and I don't see META-INF/services/<actual-interface> exposed
[19:46:20] <dmlloyd> alesj, you'd only see META-INF/services, it's all directory-based
[19:46:45] *** frainone has quit IRC
[19:46:54] *** ChanServ changes topic to "JBoss AS 7 design discussion - to get started, see http://is.gd/0x6wkT - paste at http://fpaste.org - http://community.jboss.org/en/jbossas/dev/jboss_as7_development"
[19:46:55] <alesj> http://pastebin.com/1W2C1PXv
[19:46:58] <alesj> ah, ok
[19:47:12] *** bstans_afk is now known as bstansberry
[19:47:24] <alesj> let me check why it doesn't find actual iface file then ...
[19:47:25] <dmlloyd> yeah this should be correct
[19:47:37] <dmlloyd> assuming that [27] = {java.util.HashMap$Entry@4797} META-INF/services -> [org.jboss.modules.JarFileResourceLoader@70e284ab] is in fact the right JAR
[19:47:43] <dmlloyd> jconsole will be more descriptive
[19:47:58] <Nihility> you might want to look at the deployment modules path list
[19:48:14] <dmlloyd> if the deployment isn't *importing* services from jsf-impl it won't see it
[19:50:03] <dmlloyd> and I can see from the line you pasted, it isn't
[19:50:08] <dmlloyd> <alesj>         ModuleDependency jsf = new ModuleDependency(moduleLoader, JSF_IMPL, false, false, false);
[19:50:18] <dmlloyd> one of those falses has to be a true
[19:50:24] <dmlloyd> not sure which though outside of my IDE :)
[19:50:26] <Nihility> in his email he has it is true
[19:50:29] <dmlloyd> oh
[19:50:38] <Nihility> and i looked at module spec processor
[19:50:42] <Nihility> and it looks right
[19:51:02] <dmlloyd> ok not just:
[19:51:04] <dmlloyd> <alesj>         jsf.addImportFilter(PathFilters.getMetaInfFilter(), true);
[19:52:40] <alesj> i changed that as well
[19:52:45] <alesj> also didn't help
[19:52:57] <alesj> dmlloyd: i can try again ^^
[19:53:22] <dmlloyd> the deployment CL is where we need the META-INF/services path to appear
[19:53:59] <dmlloyd> well, the deployment module really
[19:54:05] <dmlloyd> the deployment CL will only include local paths of course
[19:54:26] * dmlloyd notes that the paste is the MCL paths, not the module paths
[19:56:00] <alesj> http://pastebin.com/YA2MxGuW
[19:56:03] <alesj> entry == null
[19:56:32] <dmlloyd> I'm not sure that means anything
[19:57:03] *** emuckenhuber has joined #jboss-as7
[19:57:03] *** emuckenhuber has joined #jboss-as7
[19:57:03] *** ChanServ sets mode: +v emuckenhuber
[19:58:51] <alesj> dmlloyd: why not?
[19:59:13] <dmlloyd> why would it
[19:59:19] <alesj> it goes into jsf-impl jar and doesn't find the right entry, the one we're actually looking
[19:59:25] <alesj> for
[19:59:46] <dmlloyd> well there's no context at all for your paste
[19:59:55] <dmlloyd> for all I know, it might just be lazily populating 'entry
[19:59:58] <dmlloyd> what ever "it" is
[20:00:07] <alesj> or where else should we find it if not in jsf-impl.jar
[20:00:10] <dmlloyd> is this some object class? some lexical scope
[20:00:11] <dmlloyd> ?
[20:00:20] <dmlloyd> just a bunch of variable names and values
[20:00:22] <dmlloyd> means nothing to me
[20:00:48] <alesj> #10 is the jar name
[20:01:24] <dmlloyd> yeah I get that
[20:01:28] <alesj> JarFileResourceLoader::getResource
[20:01:29] <dmlloyd> but that's irrelevant
[20:01:36] <dmlloyd> I mean are we in some code here
[20:01:48] *** fnasser has quit IRC
[20:01:57] <alesj> yeah, code == JarFileResourceLoader::getResource
[20:02:50] *** frainone has joined #jboss-as7
[20:02:51] *** ChanServ sets mode: +v frainone
[20:03:15] *** fnasser has joined #jboss-as7
[20:03:59] <alesj> http://pastebin.com/y1upjnfD
[20:04:21] <alesj> hmmm, looks like there is really no <iface file> in the jsf-impl jar
[20:07:17] <alesj> dmlloyd, Nihility: ^^
[20:07:37] <dmlloyd> well, that's the simpler explanation, isn't it
[20:08:10] <dmlloyd> I see META-INF/services/javax.servlet.ServletContainerInitializer and META-INF/services/com.sun.faces.spi.injectionprovider
[20:08:13] <dmlloyd> that is all
[20:11:09] <alesj> yeah
[20:13:21] <alesj> shouldn't impl have this services/<ifaces> setup?
[20:15:00] <alesj> dmlloyd, Nihility: ^
[20:15:07] <Nihility> maybe its not the lookuup thats failing
[20:15:10] <Nihility> THAT
[20:15:45] <dmlloyd> yeah that particular thing might be optional
[20:16:32] <Nihility> javax.faces.context.FacesContextFactory
[20:16:35] <Nihility> thats the one you care about
[20:17:21] <dmlloyd> that is the one which jsf-impl apparently hasn't got
[20:17:41] <dmlloyd> alesj, it's possible that stan has some extra magical piece which provides that
[20:18:08] <dmlloyd> like maybe he edited the JSF build or something
[20:18:09] <alesj> javax.faces.context.FacesContextFactory — this one is missing
[20:18:28] <alesj> ok, let me search for this
[20:18:44] <jpederse> jsf-api.jar ?
[20:19:12] <dmlloyd> it's not in there
[20:19:14] <alesj> nah, i mean for the actual impl setup
[20:19:27] <dmlloyd> at least, not in the spec JAR
[20:19:31] <alesj> as it needs to be set somehow, else i don't see how Stan makes FacesServlet work
[20:19:36] <dmlloyd> it might be in the specific mojarra API JAR
[20:19:58] <dmlloyd> back in the past you had to have the right API JAR for the impl you chose
[20:20:01] <dmlloyd> not sure if that's still true
[20:20:21] <alesj> afaik no
[20:21:07] <dmlloyd> I mean they meant to make it separate but in reality it just didn't work
[20:21:29] <alesj> http://community.jboss.org/wiki/JSFonAS7
[20:21:38] <alesj> Multi-JSF implementationsNoAS6 allowed you to plug in different JSF implementations.  Currently in AS7, only one implementation is allowed.
[20:21:45] <alesj> == No
[20:22:19] <alesj> i'm out of ideas on how this impl gets in ...
[20:22:47] <Nihility> let me brute force figure it out
[20:23:55] <Nihility> there must be a fallback or something
[20:26:28] <Nihility> i think we need stan
[20:27:36] <Nihility> i emailed him
[20:27:43] *** opalka has joined #jboss-as7
[20:27:43] *** opalka has joined #jboss-as7
[20:27:43] *** ChanServ sets mode: +v opalka
[20:28:28] <Nihility> i actually tested a jsf app
[20:28:34] <Nihility> to fix a bug he reported
[20:28:34] <alesj> and?
[20:28:39] <Nihility> and it worked
[20:28:53] <alesj> with FacesServlet?
[20:28:55] <Nihility> but maybe that was an older rev
[20:29:02] <Nihility> not sure it might not have had faces servlet
[20:29:16] <Nihility> hopefully he ges the email and pops in
[20:29:19] <alesj> if there was any .jsf then it probably was
[20:29:44] <Nihility> the only other thing i can think of to work around it is to edit the jf build and stick one in there
[20:29:47] <Nihility> i mean the jar
[20:30:11] <Nihility> hrm maybe its in the api jar
[20:30:23] <dmlloyd> are we using the jboss spec API JAR?
[20:30:26] <dmlloyd> if so - it's not in there
[20:30:55] <Nihility> nope its the sun one
[20:30:58] <Nihility> and its not in there
[20:31:23] <dmlloyd> I wonder why we have a jboss one
[20:31:31] <dmlloyd> it's for 2.0
[20:31:39] <alesj> dmlloyd, Nihility: soccer time — Barca vs. Arsenal :-)
[20:31:39] <dmlloyd> so somebody must have been planning something
[20:31:52] <Nihility> we were creating ones for all ee apis
[20:32:00] <Nihility> to avoid non-OSS api jars
[20:32:08] <Nihility> alesj: oo
[20:32:24] <Nihility> alesj: sounds more fun than tracking down facesservlet
[20:32:24] <Nihility> :)
[20:33:09] <alesj> Nihility: this is 2nd leg game, the 1st one was one of the best games I ever saw
[20:33:18] <alesj> so, no brainer :-)
[20:34:17] <Nihility> damn conf call
[20:54:22] <opalka> alesj, Barca will win (my tip) ;)
[21:00:17] <alesj> opalka: we'l see ;-)
[21:01:08] <opalka> alesj, ;)
[21:02:31] <Jaikiran|Dinner> wolfc: you around?
[21:02:35] *** Jaikiran|Dinner is now known as Jaikiran
[21:03:24] <dmlloyd> good evening (morning?) Jaikiran
[21:03:35] <Jaikiran> dmlloyd: 1:31 in the morning :)
[21:03:43] <Jaikiran> almost time for bed
[21:03:48] <dmlloyd> yeah seems like you're on the border there :)
[21:03:50] <dmlloyd> everything going ok?
[21:03:55] <Jaikiran> i just pushed some of the @Local processing commits here https://github.com/jaikiran/jboss-as/commits/ejb3
[21:04:03] <dmlloyd> cool
[21:04:10] <Jaikiran> the problem i am running into is that the
[21:04:16] <Jaikiran> binding service doesn't start
[21:04:20] <Jaikiran> due to some missing dependency
[21:04:27] <Jaikiran> too late to figure out what's missing
[21:04:41] <Jaikiran> let me just pastebin the service dump if you can quickly see what's the issue
[21:04:46] <dmlloyd> okay
[21:05:05] <dmlloyd> someone (Jason I think) was telling me that some of our JNDI binding service names were wrong
[21:05:29] <Jaikiran> here's the missing dependency part http://pastebin.com/t7d7d8wV
[21:05:34] <Jaikiran> and here's the entire service dump
[21:05:54] <Jaikiran> http://pastebin.com/zurqXyVP
[21:06:14] <Jaikiran> i am guessing that the service named jboss.naming.context.java.global is missing
[21:06:20] <Jaikiran> since i can't find it in that dump
[21:06:25] <Jaikiran> but i might be wrong
[21:07:26] <dmlloyd> ah yes
[21:07:48] <dmlloyd> yeah that's a known problem, we'll get that sorted today perhaps...
[21:08:14] <Jaikiran> ok cool, so that's probably going to get the EJB3 part working too then
[21:08:16] <dmlloyd> Nihility: were you working on that, or was it stuartdouglas?
[21:08:21] <dmlloyd> great, Jaikiran!
[21:08:25] <dmlloyd> is there a demo already?
[21:08:40] <Jaikiran> yep, within the demos there the "ejb3" example
[21:08:43] <Jaikiran> which we are testing against
[21:08:53] <dmlloyd> if we can get the global naming context working and the demo works then I'll merge it all this evening
[21:08:55] <stuartdouglas> Nihility was working on it
[21:08:58] <Jaikiran> once this bindingservice is triggered, that demo should start working
[21:09:02] <dmlloyd> good morning, stuartdouglas
[21:09:10] <stuartdouglas> morning
[21:09:18] <Jaikiran> cool!
[21:09:40] <Jaikiran> dmlloyd: i'll get some sleep now. see you tomorrow. good night!
[21:09:44] <dmlloyd> good night Jaikiran
[21:09:50] *** Jaikiran has quit IRC
[21:09:50] * dmlloyd is getting confused by all these time zones
[21:12:00] *** kkhan has quit IRC
[21:12:33] <Nihility> i have a fix for that
[21:12:45] <Nihility> i just got off a call
[21:12:53] <Nihility> let me get power for my notebook
[21:12:59] <Nihility> and ill push the patch
[21:13:00] <dmlloyd> can you append it to https://github.com/jaikiran/jboss-as/tree/ejb3 ?
[21:13:13] <dmlloyd> well actually, they probably need a rebase anyway...
[21:13:20] <dmlloyd> so I guess it doesn't matter
[21:14:04] *** kkhan has joined #jboss-as7
[21:14:04] *** ChanServ sets mode: +v kkhan
[21:15:16] <Nihility> https://github.com/n1hility/jboss-as/commit/e584be3e5b83c069043634729dc00762438ef264
[21:15:17] <jbossbot> git [jboss-as] e584be3.. Jason T. Greene Initial supprot for @DataSourceDefinition
[21:15:26] <Nihility> dmlloyd: you want to review that?
[21:17:20] <dmlloyd> looks ok
[21:17:31] <jbossbot> git [jboss-as] push master e584be3.. Jason T. Greene Initial supprot for @DataSourceDefinition
[21:17:31] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/ed1eaae...e584be3
[21:17:43] <dmlloyd> I'm going to make a note though that we should look into having just one context service instead of four
[21:17:45] <Nihility> so that includes the fix for global
[21:17:59] <Nihility> the global one is "special"
[21:18:06] <Nihility> it doesnt use a selector
[21:18:25] <dmlloyd> yeah but later on we might need it to, so that app clients can import from the server
[21:18:34] <dmlloyd> but that's a problem for later
[21:18:42] <Nihility> right
[21:18:56] <dmlloyd> let me try rebasing jaikiran's branch and we'll see how that looks
[21:19:04] <Nihility> the big problem is that without it being this way, right now, then a managed bean cant do InitialContext().lookup
[21:19:48] <Nihility> (or any other random piece of in container code)
[21:19:54] <Nihility> we may want a special defauling selector
[21:20:07] <Nihility> but id rather do that when we tackle app client
[21:20:16] <Nihility> when we know how we want to do it
[21:20:17] <Nihility> :)
[21:21:17] <dmlloyd> yeah I mean we can have a global-only selector active when the component selector isn't, for example
[21:21:52] <dmlloyd> actually though managed beans *can* do new InitialContext().lookup, on global app and module without problems
[21:21:58] <dmlloyd> maybe even comp too
[21:22:01] <dmlloyd> it's just gonna be the caller's context
[21:22:08] <stuartdouglas> depends on the caller
[21:22:28] <Nihility> yeah in the case of the managedbean example
[21:22:31] <Nihility> its jmx
[21:22:44] <dmlloyd> right
[21:23:25] <Nihility> which has no context :/
[21:23:38] <dmlloyd> yes, but that's fine because only loonies like us would use JMX to test that
[21:24:04] <dmlloyd> really we should define some context for JMX and POJOs though
[21:25:08] <Nihility> not sure if you noticed
[21:25:12] <Nihility> but it uses the refleciton index
[21:25:19] <dmlloyd> ok so the merge is clean, testing now
[21:25:29] <dmlloyd> oh, great
[21:27:51] <Nihility> jpederse: fyi i added initial support for @DataSourceDefinition, it, however, does not use JCA to provide support because the contract for the annotation is kind of goofy. The user provides the jdbc DataSource (not Driver) impl class
[21:28:11] <jpederse> Nihility: k
[21:28:24] <jpederse> Nihility: should we leave it at that ?
[21:28:36] <Nihility> jpederse: so theres not much we could do other than implement some features like connection-tracking
[21:29:03] <jpederse> Nihility: community contribution imho
[21:29:29] <jpederse> Nihility: nobody in their right mind will use that feature anyway
[21:29:47] <Nihility> jpederse: yeah i figure its a development only kind of feature
[21:30:01] <dmlloyd> I'm ok with keeping it minimal
[21:30:06] <dmlloyd> they get what they pay for
[21:30:08] <Nihility> im fine with leaving it as is
[21:30:09] <jpederse> Nihility: while you are here - did you get a chance to look at the jandex classloader issue ?
[21:30:31] <Nihility> note yet i can do that now
[21:30:54] <jpederse> Nihility: thanks - there should be a smoke test ready - just need to uncomment
[21:31:49] <jpederse> Nihility: I found a problem with our security integration - do you know when Marcus have planned the entire integration ?
[21:32:14] <Nihility> we certainly need it by CR
[21:32:30] <jpederse> Nihility: yup ;)  I'll check with him once I see him
[21:32:32] <Nihility> but yeah i dont know when they expected to be finished
[21:32:56] <dmlloyd> darn, test failure :)
[21:33:06] * dmlloyd thought we were just about to have EJBs!
[21:33:06] <jpederse> Nihility: the bug won't show up until a security domain is used for a datasource
[21:33:17] <dmlloyd> looks like another JNDI problem though...
[21:33:27] <Nihility> JBAS-8924 looks familiar
[21:33:28] <jbossbot> jira [JBAS-8924] Port connector subsystems to IronJacamar Beta4 [Open (Unresolved) Task, Major, Stefano Maestri] https://issues.jboss.org/browse/JBAS-8924
[21:33:48] <Nihility> stuartdouglas: hey is JBAS-8924 what you and smarlow were running into
[21:33:49] <Nihility> ?
[21:33:49] <jbossbot> jira [JBAS-8924] Port connector subsystems to IronJacamar Beta4 [Open (Unresolved) Task, Major, Stefano Maestri] https://issues.jboss.org/browse/JBAS-8924
[21:33:57] <jpederse> Nihility: JBAS-8925
[21:33:59] <jbossbot> jira [JBAS-8925] Failure in the rar demo smoke test [Open (Unresolved) Bug, Blocker, Jason Greene] https://issues.jboss.org/browse/JBAS-8925
[21:34:07] <Nihility> oops typo
[21:34:22] <stuartdouglas> not sure
[21:34:37] <Nihility> ok ill dig deeper
[21:34:40] <jpederse> Nihility: getting an annotation instance from jandex based on classloader and reflection
[21:34:40] <stuartdouglas> It was saying that the driver was wrong for some reason
[21:34:52] <stuartdouglas> I would need to debug into it
[21:34:59] <jpederse> stuartdouglas: there are classloading issues
[21:35:11] <jpederse> maybe JBAS-8925 is just the first one we hit
[21:35:12] <jbossbot> jira [JBAS-8925] Failure in the rar demo smoke test [Open (Unresolved) Bug, Blocker, Jason Greene] https://issues.jboss.org/browse/JBAS-8925
[21:36:04] <Nihility> yeah the interesting thing
[21:36:11] <Nihility> is that it looks like classloading isnt available
[21:36:19] <Nihility> but it is in the right phase for it
[21:36:23] <Nihility> (INSTALL)
[21:36:28] <stuartdouglas> which seems very odd
[21:36:41] <Nihility> its a similar setup to web
[21:36:46] <jpederse> Nihility: something changed between alpha1 and now - since this did work in the release
[21:36:50] <Nihility> where they expect classes to be avilable to interpret annotaitons
[21:37:33] <Nihility> jpederse: yeah the deployment service lifecycle is 100% different
[21:37:42] <jpederse> Nihility: the SPI impl is under connector/ so feel free to change
[21:37:47] <dmlloyd> maybe it's OSGi meddling again
[21:38:10] <Nihility> jpederse: yeah we need to change that eventually, but what you guys are doing should work, and that bothers me
[21:38:11] <Nihility> :)
[21:38:17] <jpederse> :)
[21:38:36] <stuartdouglas> If it was osgi I think it would be a NPE, as module would be null
[21:39:11] <jpederse> stuartdouglas: maeste tried it with/without osgi - same result
[21:40:16] <Nihility> time to debug
[21:40:17] <Nihility> :)
[21:40:29] <Nihility> jpederse: is that in upstream
[21:40:37] <Nihility> or maeste's tree
[21:41:04] <jpederse> Nihility: bstansberry merged it to master
[21:41:12] <Nihility> cool
[21:41:50] <Nihility> oh isnt john back today
[21:41:58] <dmlloyd> hmm, ejb doesn't seem to be binding anything
[21:42:04] <stuartdouglas> javax/resource/Referencable is missing
[21:42:16] <dmlloyd> Nihility: yeah tonight, though I wager he'll probably be a bit tired to do anything other than catch up email
[21:42:38] <stuartdouglas> I will add it to the deployment and see if that works
[21:42:59] <jpederse> stuartdouglas: javax.resource.api
[21:43:25] <stuartdouglas> It doesn't look like you have a dependency processor there already?
[21:44:05] <jpederse> stuartdouglas: all the processors are under connector/src/main/java/org/jboss/as/connector/deployers/processors/
[21:44:19] *** mbg has quit IRC
[21:44:26] <Nihility> well that would explain it
[21:46:07] <Nihility> RaConfigProcessor
[21:46:21] <stuartdouglas> editing it now...
[21:47:21] <Nihility> speaking of, we should fix this one at some point:
[21:47:22] <Nihility>  moduleSpecification.addDependency(new ModuleDependency(moduleLoader, IRON_JACAMAR_IMPL_ID, false, true, false));
[21:47:44] <stuartdouglas> yea
[21:48:00] <dmlloyd> yeah we have a couple of those
[21:48:04] <dmlloyd> I see those as an M3 task
[21:48:41] <Nihility> jpederse: is anything in impl used directly by the user's deployment classes
[21:49:15] <jpederse> Nihility: not directly
[21:49:39] <Nihility> ok how do you locate your impl
[21:49:40] <jpederse> Nihility: well, define directly ;)
[21:49:55] <jpederse> Nihility: the container is injected into the resource adapter of course
[21:50:27] <Nihility> well by difectly i mean does a users xml or java code ever do loadClass(name of an impl class)
[21:51:28] <jpederse> Nihility: no, the container is created by the deployer chain and then injected into the resource adapter - the resource adapter can only use javax.resource.api and org.jboss.ironjacamar.api
[21:51:45] <Nihility> oh ok
[21:51:59] <Nihility> so does your api directly link to the impl or do you do service discovery
[21:52:10] <Nihility> e.g. FactoryFinder
[21:52:29] * dmlloyd is more of a direct linkage fan :)
[21:52:49] <jpederse> Nihility: reflection
[21:53:15] <Nihility> ah ok so do you look for a META-INF/service/icronjavamar or something
[21:53:16] <Nihility> ?
[21:53:27] <dmlloyd> icron javamar++
[21:53:28] <jpederse> Nihility: mhmm, actually let me check that I may have removed it
[21:53:48] <stuartdouglas> odd, the rar test never actually hits the deployment processor
[21:53:50] <Nihility> i cant even blame the ipad this time
[21:53:59] *** emmanuel has quit IRC
[21:54:27] <jpederse> Nihility: not atm - we do overrides with system properties - e.g. things we will expose in the subsystem in AS 7
[21:54:34] <jpederse> Nihility: closer to CR
[21:54:40] <dmlloyd> I love discovering new and interesting words via typo
[21:55:25] <Nihility> jpederse: its fine to do impl discover
[21:55:49] <Nihility> jpederse: im only asking so i can know how to recommend we fix the inclusion of the impl import
[21:55:57] <jpederse> Nihility: all that should be gone
[21:56:14] <Nihility> the lookup pattern if you want to discovery should be:
[21:56:19] <Nihility> 1) check TCCL
[21:56:24] <jpederse> Nihility: class discovery is done in org.jboss.ironjacamar.impl through the passed classloader
[21:56:28] <Nihility> 2) if fail check apiClass.getCL
[21:56:44] <jpederse> Nihility: there is no TCCL in IronJacamar ;)
[21:56:45] <Nihility> oh
[21:56:48] <Nihility> even better
[21:57:20] <jpederse> Nihility: the wildest stuff we have is WeakReference<ClassLoader>
[21:59:03] <Nihility> ah what do you use that for
[21:59:15] <Nihility> (just wondering)
[21:59:43] <jpederse> Nihility: hehe, I nuked that too
[21:59:57] <jpederse> Nihility: so only direct classloader references during startup
[22:00:32] <jpederse> Nihility: its now
[22:01:02] <jpederse> Nihility: http://pastebin.com/9HV7h7DU
[22:01:14] *** irooskov has joined #jboss-as7
[22:01:28] <Nihility> ah nice
[22:01:36] <jpederse> Nihility: impl class
[22:02:04] <jpederse> Nihility: but users shouldn't mess with that - so plugins aren't really an issue
[22:02:21] *** wolfc1 has joined #jboss-as7
[22:02:21] *** ChanServ sets mode: +v wolfc1
[22:02:44] *** rmaucher has quit IRC
[22:02:53] <wolfc1> dmlloyd, darn, test failure :) ***dmlloyd thought we were just about to have EJBs! ?
[22:03:00] *** wolfc has quit IRC
[22:03:03] <dmlloyd> yeah
[22:03:05] *** wolfc1 is now known as wolfc
[22:03:11] <wolfc> is that in jaikiran/ejb3?
[22:03:12] <dmlloyd> the demo isn't finding the global binding
[22:03:14] <dmlloyd> yes
[22:03:16] <wolfc> ah
[22:03:22] <dmlloyd> looking to see now if it's something with the test
[22:03:39] <Nihility> you see the problem is you have test cases
[22:03:42] <asoldano> Nihility, dmlloyd: I have 8 regressions in jbossws testsuite due to today's commits on AS7 upstream master, basically some war deployments failures cause 2 NPE in AbstractComponentDescription and ComponentInstallProcessor
[22:03:50] <Nihility> if you didnt have them you wouldnt know they existed
[22:03:55] <dmlloyd> asoldano, we should probably have smoke tests
[22:03:55] <wolfc> :-)
[22:04:08] <wolfc> gitk crashed my display when I tried to look into that branch
[22:04:22] <asoldano> dmlloyd, well... Jim is working on something like that
[22:04:34] <asoldano> dmlloyd, but covering everything is probably too much
[22:04:44] <dmlloyd> yeah just enough to be sure that the basics aren't broken
[22:04:49] <asoldano> dmlloyd, what the details of the exceptions?
[22:04:53] <asoldano> s/what/want
[22:04:59] <Nihility> sure
[22:05:13] <Nihility> ill fix that other issue you saw
[22:05:14] <wolfc> dmlloyd, do you know if the local view annotation processor is engaged?
[22:05:40] <asoldano> Nihility, dmlloyd : first one: http://fpaste.org/OWww/
[22:06:24] <asoldano> Nihility, dmlloyd : 2nd one: http://fpaste.org/6P0k/
[22:06:44] <asoldano> Nihility, dmlloyd : I can provide the war files to reproduce if you want
[22:06:47] <Nihility> its all stuart's fault
[22:06:49] <Nihility> hahahahahaha
[22:07:07] <stuartdouglas> whats all my fault?
[22:07:08] <wolfc> the demo was passing some time ago, so I'm assuming my test code is correct :-P
[22:08:01] <dmlloyd> wolfc: yeah I'm looking at the services right now and TBH they all look correct
[22:08:12] <dmlloyd> that's why I'm thinking it might be a test env issue
[22:08:24] <Nihility> stuartdouglas: im kidding, asoldano saw web issues and since you are the web component ee integration maintainer.......:)
[22:08:24] <dmlloyd> since we just did some JNDI changes
[22:08:56] <dmlloyd> well I saw one thing, the @Local BindingDescriptions do not need dependency set to true, that just slows down deployment a little bit
[22:08:59] <dmlloyd> at least that I can see
[22:09:05] <dmlloyd> but that wouldn't cause a problem
[22:09:15] <stuartdouglas> I was not actually following, although the rar demo does actually kinda look like my fault :-)
[22:09:24] <wolfc> the managed bean test doesn't use JNDI, does it?
[22:09:31] <dmlloyd> it does, wolfc
[22:09:37] <wolfc> I'm still waiting for my machine to wake up properly
[22:09:38] <dmlloyd> but not java:global
[22:09:41] <asoldano> stuartdouglas, http://fpaste.org/OWww/ and http://fpaste.org/6P0k/
[22:09:44] <Nihility> stuartdouglas: ah so i dint have to feel bad you are debugging it instead of me then :)
[22:09:56] <dmlloyd> OWww++
[22:10:06] <stuartdouglas> You need to put classes in the rar into nestled jars now
[22:10:17] <stuartdouglas> where as before it would just let you do whatever
[22:10:29] <Nihility> dmlloyd: with less than a week to have ejb3 implemented, i really think you should spend some time implementing a karma module for jbossbot
[22:10:30] <Nihility> :)
[22:10:52] <stuartdouglas> the ws stuff probably is my fault to :-(
[22:11:12] <dmlloyd> hey jbossbot is gonna be hosted by AS 7 eventually
[22:11:32] <wolfc> only if we can get a rar to inflow it to a MDB :-)
[22:12:25] <asoldano> stuartdouglas, let me know if you want the war files for reproducing the problem
[22:12:36] <stuartdouglas> that would be good
[22:13:45] <asoldano> stuartdouglas, sent them
[22:13:51] <stuartdouglas> thx
[22:14:14] <stuartdouglas> now I have new and different failures in the rar demo :-)
[22:14:43] <jpederse> stuartdouglas: if you are looking at it - or fix the issue, email maeste
[22:14:47] <jpederse> stuartdouglas: cool :)
[22:15:41] <Nihility> private final TransformerFactory actual;
[22:15:41] <Nihility> private final SAXTransformerFactory saxtual; // Snicker
[22:16:00] <asoldano> Nihility, ;-)
[22:16:13] <stuartdouglas> actually, I think it may be a jndi issue now
[22:18:09] <jpederse> stuartdouglas: our SPI implementation connector/src/main/java/org/jboss/as/connector/jndi/AS7JndiStrategy.java
[22:19:36] <stuartdouglas> That looks like a problem
[22:22:58] <stuartdouglas> none of the methods on AS7JndiStrategy are called
[22:23:04] <jpederse> stuartdouglas: you can change that class
[22:23:19] <stuartdouglas> nothing is actually calling it :-)
[22:23:20] <jpederse> stuartdouglas: hmm, that means the deployer isn't kicking in
[22:23:45] <jpederse> stuartdouglas: AbstractResourceAdapterDeployer in the IronJacamar projects does
[22:24:26] *** slaboure has quit IRC
[22:24:45] <stuartdouglas> well, it nevel hits my breakpoint
[22:25:08] <jpederse> stuartdouglas: the jndi name is defined in testsuite/smoke/src/test/resources/demos/rar-example.rar/META-INF/ironjacamar.xml
[22:26:04] *** davidbos has joined #jboss-as7
[22:26:07] <stuartdouglas> ok, actually it is probably a packaging issue
[22:27:00] <jpederse> stuartdouglas: both .rar/org/... and .rar/myjar.jar must be supported
[22:27:18] <stuartdouglas> according to the spec that is wrong though
[22:27:19] <stuartdouglas> The Java interfaces, implementation, and utility classes required by the resource
[22:27:20] <stuartdouglas> adapter must be packaged as one or more JAR files as part of the resource adapter
[22:27:20] <stuartdouglas> module. A JAR file must use the .jar file extension.
[22:27:41] <jpederse> stuartdouglas: correct, but tell that to the 3rd party vendors
[22:27:43] <dmlloyd> http://github.com/dmlloyd/jboss-as/commit/d265b25
[22:27:44] <jbossbot> git [jboss-as] d265b25.. David M. Lloyd Dependency is not necessary
[22:28:30] <Nihility> jpederse: how many do you know of that do this? (just wondering)
[22:28:47] <jpederse> Nihility: ~50% of the .rar I see
[22:28:50] <dmlloyd> still no test though, gonna have to bust out the debugger
[22:29:21] <Nihility> jpederse: thats a strong argument
[22:29:33] *** davidbos has quit IRC
[22:29:51] <jpederse> Nihility: mostly older .rar's though... but there is some old EIS' out there ;)
[22:30:12] <jpederse> Nihility: we will have a validator warning in CR
[22:30:14] <wolfc> dmlloyd, I might be cross-eyed, but the MB tests do not appear to use JNDI
[22:30:30] <dmlloyd> ah not directly, the MB tests deploy a SAR which uses a service which does the JNDI lookup
[22:32:04] <wolfc> dmlloyd, the service uses MSC
[22:32:20] <dmlloyd> really?
[22:32:21] * dmlloyd looks
[22:32:36] <Nihility> yeah LookupService
[22:32:43] <wolfc> oh we are both right and the thing is more wicked :-)
[22:32:47] <dmlloyd> ah yeah it uses a service
[22:32:51] <dmlloyd> but the service yields a context
[22:33:06] <Nihility> why does it do a poll loop like that
[22:33:21] <Nihility> i wondered that this weekend
[22:33:29] <dmlloyd> it lets it look right into someone else's context
[22:33:45] <dmlloyd> as for the poll loop I think it was because it didn't do deps right
[22:33:45] <Nihility> oh
[22:33:47] <dmlloyd> but that should have been fixed
[22:35:43] <Nihility> stuartdouglas: you remember how you were looking for ctrl-o in eclipse
[22:35:46] <Nihility> in idea
[22:35:46] <dmlloyd> yeah I don't see any poll loop any more
[22:35:59] <Nihility> oh maybe it was using wait and notify
[22:36:22] <Nihility> stuartdouglas: i found it in idea :)
[22:36:50] <stuartdouglas> I just set idea to import automatically :-)
[22:37:01] <stuartdouglas> so have you moved to idea?
[22:37:17] <Nihility> no the heirarchy view
[22:37:23] <Nihility> not package organize
[22:37:32] <stuartdouglas> ah
[22:37:44] <Nihility> its ctrl/apple-command - F12
[22:37:54] <Nihility> yeah i loaded it today
[22:38:02] <Nihility> eclipse pissed me off for the last time
[22:38:08] <Nihility> im never using it again
[22:38:25] <stuartdouglas> I just realised that I don't actually know the key shortkeys enough to actually talk about them, I just hit them without thinking :-)
[22:38:40] <stuartdouglas> I am never going back
[22:38:56] <Nihility> i must have wasted around a man week of time for the life of the as7 project
[22:39:09] <Nihility> just dealing with eclipse flaking out on a daily basis
[22:41:06] <wolfc> dmlloyd, the ComponentInstallProcessor puts in a dependency on jboss.naming.context.java.global which is not a service at all
[22:41:52] <dmlloyd> wolfc: you need the latest commits
[22:42:06] <wolfc> where from?
[22:43:11] <dmlloyd> https://github.com/dmlloyd/jboss-as/tree/master is jaikiran's tree/ejb3 rebased on upstream/master plus one commit
[22:45:14] <wolfc> hmm, I thought I had already tried that
[22:45:50] <dmlloyd> jason just pushed it an hour or two ago
[22:47:03] <wolfc> oh I didn't have the master commits yet
[22:48:06] *** kkhan has quit IRC
[22:51:25] <stuartdouglas> jpederse In my master I have made a lot of fixes to the rar demo
[22:51:38] <stuartdouglas> actually only a few fixes
[22:51:54] <stuartdouglas> but now I am at: Caused by: org.jboss.jca.deployers.common.DeployException: TODO: We need security context service to implement it!
[22:53:14] <stuartdouglas> Nihility: would you be able to merge my master when you have a few minutes?
[22:54:24] <Nihility> stuartdouglas: whats in it
[22:54:57] <jpederse> stuartdouglas: 2 sec
[22:56:11] *** smcgowan has joined #jboss-as7
[22:56:11] *** ChanServ sets mode: +v smcgowan
[22:56:53] <jpederse> stuartdouglas: in connector/src/main/java/org/jboss/as/connector/metadata/deployment/ResourceAdapterXmlDeploymentService.java and connector/src/main/java/org/jboss/as/connector/metadata/deployment/ResourceAdapterDeploymentService.java
[22:57:12] <jpederse> stuartdouglas: change "getSubjectFactory" to return null
[22:57:18] <jpederse> stuartdouglas: that should do it
[22:57:26] <darranl> hi dmlloyd do you have time for a quick question?
[22:59:18] <dmlloyd> go ahead
[22:59:27] <dmlloyd> just helping the boy with some math homework...
[22:59:55] <jpederse> stuartdouglas: I have to get going - meetup with Dan Allen - send an email to maeste about your findings
[23:00:02] <jpederse> stuartdouglas: thanks for looking into it
[23:00:14] *** miclorb_ has joined #jboss-as7
[23:00:17] <darranl> dmlloyd, should only be quick - the SAML mechanisms in Remoting, are these implemented yet?  Really thinking about the digest mechanism here
[23:00:19] <stuartdouglas> ok
[23:01:04] <darranl> sorry SASL not SAML
[23:01:11] <alesj> jpederse: ping Dan in my name, saying I have some email for him ;-)
[23:01:20] <darranl> keep making that typo today ;-)
[23:01:26] <jpederse> alesj: will do :)
[23:01:39] *** jpederse has quit IRC
[23:01:48] <alesj> jpederse: tell him I need to pimp-up Permalink example
[23:03:07] *** pferraro has quit IRC
[23:03:12] <wolfc> dmlloyd, I found the issue. The EJB name in the EJBComponentConfiguration is wrong
[23:03:45] <stuartdouglas> Nihility: The rar demo is fixed in my master
[23:04:03] <wolfc> Nihility, can the reflection index provide Class.getSimpleName() ?
[23:06:47] <Nihility> wolfc: it looks like it doesnt but you can do classIndex.getIndexedClass.getSimpleName()
[23:07:42] <wolfc> ah crap, I meant ClassInfo from jandex, not reflection index
[23:08:40] <dmlloyd> darranl: SASL you mean? yeah the JDK provides them
[23:09:23] *** davidbos has joined #jboss-as7
[23:09:37] <Nihility> wolfc: well sort of
[23:09:53] <Nihility> wolfc: if you pull the name off of a ClassInfo, it will have a componentized dotname
[23:10:01] <darranl> dmlloyd, yeah keep making that typo today - and google isn't helping as it keeps swapping my acronyms for something it thinks I mean
[23:10:06] <Nihility> on a componentized dotname you can call local()
[23:10:36] <darranl> dmlloyd, thanks for the info - I will have a look at what the JDK provides
[23:12:40] *** davidbos has quit IRC
[23:16:30] <wolfc> dmlloyd, https://github.com/wolfc/jboss-as/commit/ff7da7c2907c4be66597c9e76c7102d9d92c7434
[23:16:30] <jbossbot> git [jboss-as] ff7da7c.. Carlo de Wolf Set the proper EJB name in ComponentDescription
[23:16:38] <wolfc> It makes the binding come up properly
[23:16:42] <dmlloyd> nice
[23:17:08] *** mbg has joined #jboss-as7
[23:17:09] *** ChanServ sets mode: +v mbg
[23:19:25] *** mbg has quit IRC
[23:19:40] <Nihility> woa
[23:19:49] <Nihility> i love the way idea does most things already
[23:20:10] <dmlloyd> looks like instance association is failing
[23:20:24] <dmlloyd> the component interceptor isn't doing it
[23:21:46] <stuartdouglas> the only shortcut key I changed was to make command + 1 the quick fix command
[23:21:51] <darranl> Nihility, so worth the effort switching?
[23:21:59] <stuartdouglas> the rest of the commands I got used to pretty quick
[23:22:13] <dmlloyd> good idea stuartdouglas, I usually go through the alt-enter menu maze
[23:22:21] <wolfc> dmlloyd, the setting up of component interceptors has changed, so I'm redoing that bit now
[23:22:25] <dmlloyd> ok great
[23:22:30] <dmlloyd> this is easier than I thought :)
[23:23:05] <stuartdouglas> I was just to used to doing it with my left hand, alt + enter felt kinda awkward
[23:23:16] <Nihility> darranl: totally
[23:23:44] <darranl> Nihility, will get it installed shortly ;-)
[23:24:00] <Nihility> darranl: yeah and its defaults are mostly right
[23:24:21] <Nihility> darranl: ironically it seems to understand maven better than m2e
[23:24:30] <dmlloyd> yeah the maven integration is pretty good
[23:24:37] <darranl> Nihility, I had head that from someone else before
[23:24:43] <Nihility> like when i kicked off a test case
[23:24:45] <dmlloyd> I kinda wish it would respect <exclude> a little better though
[23:24:50] <Nihility> i didnt have to add resource imports
[23:24:58] <dmlloyd> doesn't cause any adverse behavior, just lots of extra useless deps
[23:25:07] <dmlloyd> useless when coding anyway
[23:25:22] <Nihility> they are very F happy with key bindings
[23:25:24] <dmlloyd> at test time since it's maven it's a flat classpath with all transitives, which also sucks but for different reasons
[23:26:11] <dmlloyd> the thing about the maven integration I really don't like though is that every maven dep becomes a project library
[23:26:18] <dmlloyd> so you end up with 900000000000 project libraries
[23:26:29] <stuartdouglas> It is still better than m2eclipse, which most of the time just plain does not work :-)
[23:26:34] <dmlloyd> and they never get cleaned up
[23:26:40] <dmlloyd> so every now and then you just have to nuke the whole list
[23:26:40] <alesj> this tests fails for me:
[23:26:41] <alesj> Tests in error:
[23:26:42] <alesj>   org.jboss.as.test.embedded.demos.managedbean.ManagedBeanTestCase
[23:26:45] <dmlloyd> then re-import all maven stuff
[23:26:56] <stuartdouglas> also it hooks up maven libraries source automatically when debugging
[23:26:57] <dmlloyd> alesj: every time?  or just sometimes
[23:27:10] <dmlloyd> alesj: there's a sporadic MSC bug which is actually fixed in upstream that makes the tests fail sometimes
[23:27:30] <Nihility> dmlloyd: what bug
[23:27:31] <dmlloyd> stuartdouglas, yeah that's great BUT it does get kind of confused when you have many versions of something
[23:27:44] <Nihility> oh one problem with IDEA's defaults
[23:27:47] <dmlloyd> Nihility: it's a state machine glitch on install which causes an ISE
[23:27:48] <Nihility> is that the warnings are subdued
[23:28:09] <Nihility> dmlloyd: did you just find it today?
[23:28:11] <dmlloyd> I'll do releases of everything before we tag
[23:28:16] <dmlloyd> no last week
[23:28:39] <stuartdouglas> can someone merge my master? It fixes the rar demo
[23:29:04] <dmlloyd> so if you get an ISE on doInstall() in MSC, just retry it and it'll probably work
[23:29:46] <alesj> dmlloyd: let me try it once again
[23:30:07] *** ALR has quit IRC
[23:30:16] <dmlloyd> I'll look at it stuartdouglas
[23:32:08] <asoldano> stuartdouglas, I'm heading to bed... should I create a jira for those war deployment issues, or you have all you need and are going to work on that "soon" ?
[23:32:42] <stuartdouglas> asoldano: I'll probably work on them on the plane/train in a few hours
[23:32:52] <asoldano> stuartdouglas, cool, thanks
[23:33:30] * asoldano is trying to minimize waste of time due to timezones...
[23:34:13] <wolfc> dmlloyd, demo is working again: https://github.com/wolfc/jboss-as/commits/ejb3-ee
[23:34:33] *** darranl is now known as darranl_afk
[23:37:07] <wolfc> The DummyComponentInterceptor is a temporary measure. It'll be refactored once the real pool and cache impls go in.
[23:37:43] <wolfc> In fact I think I want to hide it all behind a SessionManager which delegates to either a pool or a cache or a single instance.
[23:39:01] <wolfc> Now I'm going to catch some Zzzz...
[23:39:12] * wolfc is shutting down
[23:40:04] <dmlloyd> cool, thanks wolfc
[23:40:41] <wolfc> likewise, good night
[23:41:16] *** wolfc has quit IRC
[23:44:21] <jbossbot> git [jboss-as] push master 5fc838c.. Stuart Douglas Fix Class-Path entries that point to other deployments
[23:44:21] <jbossbot> git [jboss-as] push master c7e0eaa.. Stuart Douglas Work on rar demo
[23:44:22] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/e584be3...c7e0eaa
[23:45:42] <dmlloyd> the build seems somewhat more verbose these days
[23:46:52] <stuartdouglas> also slower, those extra smoke tests slow things down a fair bit
[23:48:32] <alesj> Nihility, dmlloyd: this is how JSF gets its impl
[23:48:39] <alesj> Nihility, dmlloyd: http://pastebin.com/R8TgMNPc
[23:49:30] <dmlloyd> ah
[23:51:03] <jbossbot> git [jboss-as] push master 97e78ed.. David M. Lloyd Give every component a namespace selector, even if it is not gonna use it
[23:51:03] <jbossbot> git [jboss-as] push master 41d0450.. David M. Lloyd Fixed extension imports
[23:51:04] <jbossbot> git [jboss-as] push master 930bf63.. David M. Lloyd Initial work on ejb-jar.xml parsing deployment unit processor - jaikiran, carlo
[23:51:04] <jbossbot> git [jboss-as] push master 57acfa9.. David M. Lloyd Moved jboss-metadata-ejb version to dependency management - jaikiran
[23:51:04] <jbossbot> git [jboss-as] push master 15bf384.. Carlo de Wolf Created ejb3 demo
[23:51:04] <jbossbot> git [jboss-as] push master d9cfa67.. David M. Lloyd Enabled EjbJarParsingDeploymentUnitProcessor - wolfc
[23:51:04] <jbossbot> git [jboss-as] push master 2aa8239.. jaikiran Change the version of xml to 1.1 to prevent the javax.xml.stream.XMLStreamException: ParseError at [row,col]:[1,39] Message: Premature end of file.
[23:51:05] <jbossbot> git [jboss-as] push master e123374.. Carlo de Wolf Cleanup module dependencies
[23:51:05] <jbossbot> git [jboss-as] push master 4c41d6d.. David M. Lloyd Initial attempt at StatelessEJBComponent for deploying/processing SLSBs - ed. by dmlloyd, remove the EJB-specific processor for future integration with the EE component processor
[23:51:06] <jbossbot> git [jboss-as] push master 933caa9.. Carlo de Wolf Add a dependency on jboss-modules
[23:51:06] <jbossbot> git [jboss-as] push master 09d011c.. Carlo de Wolf Configured the modules to allow EjbComponentDeploymentUnitProcessor to function
[23:51:06] <jbossbot> git [jboss-as] push master cffe94e.. jaikiran Bring in the latest from upstream and fix the compilation issues
[23:51:07] <jbossbot> git [jboss-as] push master 31454c3.. jaikiran Move the EjbComponentDeploymentUnitProcessor to PARSE phase
[23:51:08] <jbossbot> git [jboss-as] push master 5422a27.. jaikiran Move away from Managed Bean based EJB components....
[23:51:08] <jbossbot> git [jboss-as] push master c1c64c3.. jaikiran Add DEPENDENCIES_EJB phase
[23:51:09] <jbossbot> git [jboss-as] push master 3369bce.. jaikiran Setup correct module dependenices
[23:51:09] <jbossbot> git [jboss-as] push master f107f57.. jaikiran Include EjbDependencyDeploymentUnitProcessor in the Ejb3Subsytem
[23:51:10] <jbossbot> git [jboss-as] push master 9c5b9c9.. jaikiran Setup the system interceptor through the component configuration
[23:51:10] <jbossbot> git [jboss-as] push master 3117cd0.. Carlo de Wolf Setup java:global binding (wip)
[23:51:11] <jbossbot> git [jboss-as] push master a4b30ab.. Carlo de Wolf Properly bind local business views (wip)
[23:51:11] <jbossbot> git [jboss-as] push master 694ac98.. jaikiran Final few changes to get the ejb3 demo working
[23:51:12] <jbossbot> git [jboss-as] push master b80aa08.. Carlo de Wolf Initial work on interceptors (wip)
[23:51:12] <jbossbot> git [jboss-as] push master aa36a83.. Carlo de Wolf Initial work on stateful session beans (wip)
[23:51:13] <jbossbot> git [jboss-as] push master 7a83884.. David M. Lloyd Fix merge error on closing tag
[23:51:13] <jbossbot> git [jboss-as] push master 153893b.. Carlo de Wolf Make it compilable (smoketests are failing) (wip)
[23:51:14] <jbossbot> git [jboss-as] push master c456659.. Carlo de Wolf Removed module dependency org.jboss.as.domain (smoketests pass)
[23:51:14] <jbossbot> git [jboss-as] push master 85f699a.. jaikiran Fix compilation errors
[23:51:15] <jbossbot> git [jboss-as] push master 0b99723.. Carlo de Wolf Refactored demo to use annotations instead of descriptor
[23:51:15] <jbossbot> git [jboss-as] push master 13ffc51.. Carlo de Wolf Added TransactionManagementAnnotationProcessor
[23:51:16] <jbossbot> git [jboss-as] push master 66cd45e.. jaikiran Initial work on processing EJB annotations - Merge the tx annotation processor changes introduced by Carlo...
[23:51:16] <jbossbot> git [jboss-as] push master 2850fc9.. Carlo de Wolf Fire up the correct bean class
[23:51:16] <jbossbot> git [jboss-as] push master 28fbd6f.. Carlo de Wolf Engage dependencies if it is an EJB deployment
[23:51:17] <jbossbot> git [jboss-as] push master d7b8745.. jaikiran EjbAnnotationProcessor refactoring
[23:51:18] <jbossbot> git [jboss-as] push master 767b102.. jaikiran Handle component creation in EJB component configuration
[23:51:18] <jbossbot> git [jboss-as] push master 04649d5.. jaikiran @Local EJB view processing
[23:51:19] <jbossbot> git [jboss-as] push master afbdf16.. David M. Lloyd Dependency is not necessary
[23:51:19] <jbossbot> git [jboss-as] push master 855bb83.. Carlo de Wolf Set the proper EJB name in ComponentDescription
[23:51:20] <jbossbot> git [jboss-as] push master 58c5f63.. Carlo de Wolf Setup a temporary instance association interceptor for Stateless
[23:51:20] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/c7e0eaa...58c5f63
[23:52:21] <dmlloyd> there it is
[23:52:22] <dmlloyd> EJBs
[23:52:30] <dmlloyd> only local views supported
[23:52:37] <dmlloyd> only annotation-driven deployment
[23:52:42] <stuartdouglas> quick, release before we find something doesn't work :-)
[23:53:17] <dmlloyd> may we have no more topic branches before beta1 :)
[23:53:44] <smarlow> nice!  :)  I'll sync up with it.
[23:54:02] *** Darran_L has joined #jboss-as7
[23:54:21] <dmlloyd> I'm kind of amazed that jbossbot didn't get booted after that one :)
[23:56:38] <stuartdouglas> has anyone seen this: http://chriswanstrath.com/hub/
[23:56:41] <smarlow> dmlloyd:  I still have an objectFactory question from the other day.  I looked at ResourceInjectionAnnotationParsingProcessor and based JPAAnnotationParseProcessor on that, but didn't really grok how to plug in an object factory.  I hacked in the service name that represents the JPA persistence unit definition (persistence.xml) but I'm not sure how to plug in an actual injector for what I want.
[23:57:27] *** irooskov has quit IRC
[23:57:33] <dmlloyd> nice stuartdouglas
[23:57:52] <dmlloyd> smarlow: why are you trying to plug in an ObjectFactory?  Is something in JPA providing you with that?
[23:58:47] <smarlow> I don't actually need an object factory, but just looking for the equivalent.
[23:59:13] <dmlloyd> smarlow: the equivalent is... org.jboss.as.naming.ManagedReferenceFactory

   March 8, 2011  
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | >