March 15, 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 | >
March 15, 2011 NOTICE: This channel is no longer actively logged.
[00:02:56] <jbossbot> git [jboss-as] push master 9e49564.. David M. Lloyd Upgrade to JBoss Modules 1.0.0.Beta16[00:02:56] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/7a91d93...9e49564[00:09:43] <kkhan> bstansberry: https://github.com/kabir/jboss-as/commit/b73f974cf3cbec77b1f653b25abe6735166fa16c[00:09:43] <jbossbot> git [jboss-as] b73f974.. kabir Describe web services subsystem[00:11:34] <smarlow> does the user transaction bind into jndi work? I need a good example to follow for binding com.arjuna.ats.jbossatx.jta.TransactionManagerService.getTransactionManager()[00:12:02] <dmlloyd> it should work[00:12:19] <stuartdouglas> it was working at one point[00:12:29] <dmlloyd> you'll want to use a simple org.jboss.as.ee.component.ServiceBindingSourceDescription for TransactionManager though[00:15:09] <stuartdouglas> have a look in ComponentInstallProcessor, around line 234 for how to bind stuff[00:17:09] <dmlloyd> actually I thought we already had that one added[00:17:51] <Nihility> smarlow: How's jpa?[00:17:56] <dmlloyd> ah guess not but it should probably be added to org.jboss.as.ee.component.ResourceInjectionAnnotationParsingProcessor[00:18:07] <dmlloyd> that's where we add the other "global" comp bindings[00:18:27] <stuartdouglas> TransactionManager is not a comp binding[00:18:28] <dmlloyd> actually I stand corrected[00:18:34] <dmlloyd> ah, it's not?[00:18:35] <dmlloyd> how odd[00:18:54] <stuartdouglas> It is not standard, so you can put it wherever you want[00:20:31] *** smcgowan has joined #jboss-as7[00:20:42] <smarlow> we have a TransactionManagerService which gets a TransactionManager but it doesn't get bound to java:/TransactionManager[00:21:03] <stuartdouglas> traditionally JBoss uses java:/TransactionManager GF uses java:appserver/TransactionManager and Resin uses java:comp/TransactionManager[00:21:11] <dmlloyd> we don't bind anything to / anymore, smarlow[00:21:22] *** smcgowan has quit IRC[00:21:22] *** smcgowan has joined #jboss-as7[00:21:33] <dmlloyd> it goes in java:comp, java:module, java:app, or java:global[00:21:37] <stuartdouglas> websphere uses its own crappy UOWManager just to make life difficult[00:21:38] <dmlloyd> though we have talked about adding a java:jboss[00:21:40] <dmlloyd> or something[00:21:57] <dmlloyd> given the GF precedent maybe even more so now[00:22:00] <Nihility> yeah jboss ns for jboss stuff[00:22:07] <dmlloyd> maybe even using java:appserver is a good idae[00:22:19] <stuartdouglas> +1[00:22:23] <dmlloyd> the fewer things we put into comp, the faster deployment will be[00:22:28] <stuartdouglas> make it easier for people to switch from GF :-)[00:22:32] <dmlloyd> exactly :)[00:23:12] <bstansberry> seeing that NPE quite a bit now in the smoke tests; this time in the Stateless EJB one: http://pastebin.com/Afb2egFk[00:23:47] <bstansberry> seems JavaNamespaceSetup is getting a NamespaceSelectorService whose getValue() returns null[00:23:48] <smarlow> The hibernate integration is easier if we used the old style name but we probably could implement a custom Hibernate TransactionManagerLookup[00:23:52] <bstansberry> whatever that means :)[00:24:19] <dmlloyd> stuartdouglas: does that mean anything to you?[00:25:06] <stuartdouglas> kinda, JavaNamespaceSetup sets up the module context for ARQ tests[00:25:29] <smarlow> Nihility: I need to step back from getting java:/TransactionManager and just change the demo to use a non-transactional datasource. In the mean time, I think we need to bind a TM to jndi somewhere :)[00:25:47] <stuartdouglas> but if NamespaceSelectorService.getValue() returns null then there must be some filed/missing dep somewhere[00:26:01] * dmlloyd -> dinner, bbiab[00:26:35] <stuartdouglas> I will look into it[00:26:43] <smarlow> Nihility: hmm, actually, maybe were better off not trying non-tx datasources, that might be a deeper hole[00:27:27] <bstansberry> stuartdouglas: the NamespaceSelectorService is passed into JavaNamespaceSetup via it's constructor; it's not an injected dependency[00:27:43] <bstansberry> in ModuleContextProcessor[00:27:47] *** aslak has quit IRC[00:27:49] <stuartdouglas> yes, but the actual selector is set up in NamespaceSelectorService.strat[00:27:53] <stuartdouglas> start() even[00:28:10] <stuartdouglas> which means either the deps are failing/missing[00:28:14] <bstansberry> but if start() isn't called yet?[00:28:19] <stuartdouglas> or ARQ is running before the deployment is finished[00:28:23] <stuartdouglas> either one is bad[00:28:39] <stuartdouglas> but ARQ running before the deployment is finished has been a problem in the past[00:28:45] <bstansberry> i suspect that's it[00:28:58] <bstansberry> failing/missing deps wouldn't be intermittent[00:29:06] <Nihility> Hi[00:29:08] <stuartdouglas> damn[00:31:10] *** frainone_dinner is now known as frainone[00:31:44] <smarlow> dmlloyd, Nihility: I'll see if I can avoid the Hibernate jndi lookup of the transaction manager[00:33:19] <stuartdouglas> couldn't you implement TransactionManagerLookup and get the TM from TM service?[00:34:20] <Nihility> two approaches inject it into your service[00:34:25] <Nihility> And get it from there[00:34:34] *** asaldhan has left #jboss-as7[00:35:24] <Nihility> Or create the jboss namespace by copying the code for global into the ns subsystem and bind it there in tm startup[00:36:16] <Nihility> the former is less work[00:37:47] <Nihility> The latter one of us can help if you want, although I'm on and off for awhile tonight as I have to pack for tomorrow[00:39:31] <stuartdouglas> I think we are better off using java:appserver rather than java:jboss[00:40:31] <Nihility> How come?[00:40:52] <stuartdouglas> Then it is the same as GF, so it will be easier to port apps from GF[00:41:32] *** slaboure has quit IRC[00:41:41] <stuartdouglas> it is still going to break AS6 apps that rely on it, but at least that way we are compatible with one app server[00:41:53] <stuartdouglas> and code that looks in a number of locations is more likely to just work[00:42:12] <Nihility> Yeah although on the flip side everyone would expect a one to one mapping[00:42:59] <jpearlin> Nihility: just checking in if you need me to change anything with the deployment upload stuff[00:43:09] <stuartdouglas> I don't know, I think most people don't really expect anything except what the docs tell them about non-standard JNDI locations[00:43:35] <Nihility> The point of the ns is to be vendor specific, it seems like approver would be a emulated kind of thing,then we are better off havig it be just that[00:43:51] <smarlow> stuartdouglas, Nihility: yes, that is what I'm going to do (implement TransactionManagerLookup and get the TM from TM service)[00:44:44] <wolfc> got https://github.com/wolfc/jboss-as/commits/ejb3-ee with Cache & Pool impls, BMT Tx and injecting SessionContext[00:45:09] <wolfc> It is passing smoke-tests[00:45:17] <stuartdouglas> to be honest I am not sure exactly what else is in java:appserver on GF[00:45:34] <Nihility> jpearlin: nope, at this point i think it's fine. I made a small change to fix post on domain-API by giving upload it's own uri vs a query param[00:46:04] <jpearlin> ok...yeah, that makes more sense[00:46:10] <jpearlin> and easier to detect[00:46:39] <Nihility> btw one thing I thought of stuartdouglas is that hornetqnhas jndi bindingsni wonder if those still work haha[00:46:57] <stuartdouglas> bstansberry: were you seeing the failures with upstream master? I have not seen them yet[00:47:06] <stuartdouglas> Nihility: doubt it :-([00:47:07] <bstansberry> yes[00:47:16] <bstansberry> it happens a lot now[00:47:26] <wolfc> how many CPU cores do you have?[00:47:26] <stuartdouglas> hmm, I will try on my other computer[00:47:29] <bstansberry> to the point where i'm about to give up running the smole tests[00:47:49] <stuartdouglas> I have 4 + hyperthreading[00:47:57] <bstansberry> I have 2[00:48:04] <stuartdouglas> I'll try on my laptop[00:48:44] <Nihility> Strange i had no failures allnday[00:48:51] <stuartdouglas> same[00:49:18] <Nihility> Maybe i dont habe the most recent patches[00:49:47] <Nihility> smarlow: Was jndi the only issue remaining.[00:50:53] <smarlow> Nihility: its the only one I know of at this point.[00:51:11] <Nihility> wolfc: Nice! We might actually be able to tag tonight[00:51:17] <smarlow> but I don't know if there will be more[00:51:49] <wolfc> Nihility: two major pieces are missing, stateful concurrency and exception handling[00:53:11] <dmlloyd> Nihility: those bindings still work for now[00:53:20] <dmlloyd> Nihility: it's on my beta2 hit list[00:54:37] <wolfc> we also have another bit which needs thinking: entry into user code via injection and construction is not intercepted and thus no sec/tx/classloader context can be setup.[00:54:54] *** jpearlin1 has joined #jboss-as7[00:55:47] <baileyje> So where are we at?[00:55:47] <dmlloyd> rather think of it this way[00:56:01] <dmlloyd> we can set up sec/tx/tccl in more ways than just with interceptors[00:56:10] <dmlloyd> we don't have to be devoted to using interceptors for *everything*[00:56:17] <wolfc> true, that means duplicating stuff[00:56:56] <dmlloyd> not if you think of it in terms of, (a) have code which sets up those things and (b) have things which consume that code *including* interceptors[00:57:13] <dmlloyd> but also including lifecycle wrappers, injection handlers etc.[00:57:31] <wolfc> https://github.com/wolfc/jboss-as/commit/752359b1e8a3eb683c4bac0416caef6803cc6cf2#L8R49[00:57:32] <jbossbot> git [jboss-as] 752359b.. Carlo de Wolf Injecting SessionContext on a BMT bean and use UserTransaction[00:57:38] *** jpearlin has quit IRC[00:57:44] *** jwulf has joined #jboss-as7[01:00:03] <stuartdouglas> I have not seen any smoke test failures yet[01:00:08] *** jpederse has quit IRC[01:00:12] <wolfc> me neither[01:00:31] <wolfc> bstansberry, which JDK?[01:00:39] <dmlloyd> if you're running on linux, try running with one CPU: "taskset 0x1 mvn blah"[01:00:47] <bstansberry> apple[01:01:03] <jpearlin1> Nihility: I assume at some point I should add/update the testsuite to test the upload operation?[01:01:06] <dmlloyd> I find that more concurrency/race condition/sporadic issues seem to arise when running on *fewer* CPUs[01:01:13] <dmlloyd> probably because so many of us have multicore systems[01:01:19] <bstansberry> java version "1.6.0_24"[01:01:19] <bstansberry> Java(TM) SE Runtime Environment (build 1.6.0_24-b07-334-10M3326)[01:01:19] <bstansberry> Java HotSpot(TM) 64-Bit Server VM (build 19.1-b02-334, mixed mode)[01:01:48] <wolfc> just did two runs on dual core[01:02:01] <stuartdouglas> I am still on 1.6.0_22[01:02:06] <stuartdouglas> on my mac[01:02:20] <baileyje> smarlow: How is it going?[01:02:24] <bstansberry> back in a sec[01:02:38] <dmlloyd> maven is slloooowwww on one CPU[01:03:19] <smarlow> baileyje: I'm just going to hack getting the transactionmanager via our current TransactionManagerService and use a custom hibernate TM look class[01:03:21] *** bstansberry has quit IRC[01:03:33] <dmlloyd> that said I'm unable to reproduce any....[01:03:35] <dmlloyd> problems[01:04:15] <baileyje> smarlow: You need anything from me?[01:04:34] <stuartdouglas> I'll update to the same JDK[01:04:54] <smarlow> baileyje: not yet, when the TM shows up in jndi, I'll switch over to using that but no rush if this works[01:05:27] <baileyje> What is keeping it form JNDI?[01:05:54] <baileyje> Is it something I can fix?[01:05:57] *** bstansberry has joined #jboss-as7[01:05:57] *** ChanServ sets mode: +v bstansberry[01:06:42] * dmlloyd story time, be back shortly[01:06:43] <stuartdouglas> Odd, I can't see java in my updates list at the moment, I could have sworn it was there a day or two ago[01:06:50] <wolfc> doing a custom TM lookup into for the service shouldn't be that bad[01:07:52] <baileyje> yeah. It doesn't have to come from JNDI.[01:09:29] <jamezp> baileyje: Question for you on the @DependsOn processing if you have a moment.[01:09:38] <baileyje> jamezp: Sure[01:10:05] <jamezp> Here's what I have so far, but it's the addDependency() I'm not too sure about. https://github.com/jamezp/jboss-as/blob/depends-on/ejb3/src/main/java/org/jboss/as/ejb3/deployment/processors/EjbDependsOnAnnotationProcessor.java[01:11:11] <jamezp> From what I was understanding is we want to add a dependency on the Singleton annotated with @DependsOn.[01:11:56] <jamezp> What I'm not understanding is how to reference that dependency so that MSC? knows what dependency to create/find.[01:12:18] *** stuartdouglas has quit IRC[01:12:35] <wolfc> The value of the @DependsOn first needs to be resolved.[01:12:48] <wolfc> From ejb-name to MSC ServiceName.[01:12:59] <jamezp> Okay, I had a feeling it was something like that.[01:13:55] *** stuartdouglas has joined #jboss-as7[01:13:55] *** stuartdouglas has quit IRC[01:13:55] *** stuartdouglas has joined #jboss-as7[01:13:55] *** ChanServ sets mode: +v stuartdouglas[01:13:56] <jamezp> wolfc: Any idea where I could see an example of how to resolve dependencies :-)[01:14:58] <bstansberry> stuartdouglas: unfortunately, i think something was wrong on my machine[01:15:13] <bstansberry> it started behaving badly, so I rebooted; now the heisenbug is gone[01:15:29] <stuartdouglas> ah, I just rebooted mine[01:15:38] *** stuartdouglas_ has joined #jboss-as7[01:15:38] *** ChanServ sets mode: +v stuartdouglas_[01:15:38] <wolfc> jamezp, you need to use some magic created by baileyje. It should be very similar to @EJB stuff.[01:15:42] <stuartdouglas> to update to the same JDK you were using[01:16:03] <stuartdouglas> but I have been running it on an infinite loop on this one and it has not failed once[01:16:09] <baileyje> wolfc: we still haven't tackled the ejb-name with # in it yet[01:16:11] <bstansberry> so i suspect a concurrency issue that requires some magic sauce[01:16:34] *** stuartdouglas has quit IRC[01:16:35] *** stuartdouglas_ is now known as stuartdouglas[01:16:42] <jbossbot> git [jboss-as] push master a908b1e.. kabir Describe web services subsystem[01:16:43] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/9e49564...a908b1e[01:17:17] <jamezp> wolfc baileyje: Cool, I poke around.[01:17:23] <smarlow> baileyje: I'm fine for the moment, getting another exception, maybe its a bug in my demo code :)[01:17:29] <wolfc> jamezp, https://github.com/jamezp/jboss-as/blob/depends-on/ejb3/src/main/java/org/jboss/as/ejb3/deployment/processors/EjbResourceInjectionAnnotationProcessor.java#L129[01:18:48] <wolfc> I'm going to catch some Zzzz.... now.[01:19:09] <baileyje> wolfc: Time for 1 quick ?[01:19:21] <wolfc> yes[01:19:44] <wolfc> can't guarantee that I still make sense though :-)[01:19:52] <baileyje> When you have bean names like "../../test.jar#bean[01:19:54] <baileyje> "[01:20:17] <baileyje> WHat is the limit of the path structure?[01:20:31] <wolfc> the ear is the root of all things[01:20:43] <baileyje> Ok. Good. Just making sure.[01:20:56] <baileyje> Paths make no sense outside of the ear[01:21:31] <baileyje> wolfc: Thanks[01:21:33] <baileyje> get some sleep[01:21:47] <wolfc> the can make sense in terms of a global app server level[01:21:56] <wolfc> but it's impossible to intermix that[01:23:41] <kkhan> bstansberry: https://github.com/kabir/jboss-as/commit/ad25cb99dbdd929ae096ca6452fb5b3bc1701063[01:23:42] <jbossbot> git [jboss-as] ad25cb9.. kabir Descriptions for Tx subsystem[01:23:52] <kkhan> I'll do one more before bed[01:24:04] <bstansberry> :)[01:24:48] *** wolfc has quit IRC[01:24:51] <kkhan> We're missing osgi, jms, datasources, messaging, web, resource-adapter and security[01:25:47] <kkhan> When are you guys planning on tagging?[01:28:49] <alesj> smarlow, Nihility: I have Modules visualization done ;-)[01:29:14] <alesj> https://github.com/alesj/grapher[01:29:38] <smarlow> alesj: do you a screenshot?[01:29:56] <jbossbot> git [jboss-as] push master 9432be9.. Emanuel Muckenhuber fix the read-children-names on host names[01:29:57] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/a908b1e...9432be9[01:30:01] <alesj> smarlow: it's not very nice, since I'm missing the layout library[01:30:16] <alesj> http://localhost:8080/grapher-web-2.0.0-SNAPSHOT/?extension=single-module&initial-value=org.jboss.msc[01:30:25] <alesj> http://localhost:8080/grapher-web-2.0.0-SNAPSHOT/?extension=this-module[01:30:33] <alesj> you can do this sort of commands[01:31:10] <smarlow> haha, I tried to follow your localhost link :)[01:34:02] <alesj> :-)[01:38:18] *** jwulf has quit IRC[01:38:18] <dmlloyd> baileyje: is that resolution logic tied tightly to @EJB or is it outside of that processing?[01:38:25] *** jwulf has joined #jboss-as7[01:38:30] <dmlloyd> if @EJB and @DependsOn needs it, odds are good that something else will too[01:39:16] *** alesj has quit IRC[01:41:18] <baileyje> dmlloyd: I am pretty sure it is EJB onlyl[01:41:53] <dmlloyd> EJB spec has at least @DependsOn(“b.jar#B”)[01:42:11] <dmlloyd> so we know we need it in those two places[01:46:13] <jamezp> dmlloyd: Not sure if you saw what I was doing, but if you did do I at least appear to be on the correct track now?[01:47:55] * dmlloyd looks back...[01:48:48] <dmlloyd> yeah it's the right idea - however the service name is wrong, of course[01:49:03] <dmlloyd> we need to resolve that service name the same way that @EJB does[01:49:09] <jamezp> :-) Yes, that's what I'm working on now. How to resolve that.[01:49:21] <dmlloyd> which is why I aksed baileyje if that code lives somewhere independent[01:49:50] <baileyje> dmlloyd: Right now the code live nowhere :)[01:49:54] <baileyje> well not all of it[01:50:19] <jamezp> Right, had a feeling. I'll just basically copy it for now, but note it should be replaced at some point.[01:51:13] <dmlloyd> okay[01:51:34] <baileyje> The in-module is there, but I am currently looking into the "...#..." version[01:51:38] <dmlloyd> one other thing is, I'm not sure that the code which EJB uses to resolve the EJB service name can live outside of the injection thing[01:52:06] <kkhan> bstansberry: I'm struggling a bit coming up with a description for things like: http://pastebin.com/ZvXzMMsn[01:52:16] <jamezp> I'll play around with it. Should just blow up on me if it doesn't work, right :-D[01:52:17] <kkhan> i.e. there is an attribute called properties[01:52:18] <smarlow> I got further, now trying to get the SFSB injection of the pc to work[01:52:45] <kkhan> which can contain more than one property where the value type is a string[01:52:48] <bstansberry> kkhan: does the xsd say anyting?[01:53:27] <bstansberry> noop[01:53:45] * smarlow will try to debug http://pastie.org/1672552 to see whats up :)[01:53:54] <kkhan> bstansberry: http://pastebin.com/cpG8zXig[01:54:15] <bstansberry> yeah, not too helpful[01:54:49] <kkhan> I think the type should perhaps be OBJECT with a value type of propery, but that misses the string value type[01:54:57] <smarlow> hmm, its that InjectedValue null value case again[01:55:50] <smarlow> we said the other day, that could be a missing dependency.[01:56:46] <bstansberry> kkhan: I think this is a good candidate for a TODO description[01:57:00] <bstansberry> let the osgi guys decribe it[01:57:09] <kkhan> ok, fine[02:00:18] <dmlloyd> smarlow: it almost always is.[02:01:30] <kkhan> I'm off to bed[02:05:30] <bstansberry> good night[02:05:38] <bstansberry> thanks for the help today![02:06:15] * bstansberry disables the ejb3 smoke tests on his checkouts -- failing again[02:06:46] <dmlloyd> that's very odd[02:07:46] <stuartdouglas> so did they just randomly start failing again?[02:07:55] <bstansberry> yep[02:08:04] <bstansberry> once[02:08:06] <jamezp> I've the smoke tests fail almost every time for me. I thought it was just me.[02:08:10] <bstansberry> and once is enough[02:08:15] *** kkhan has quit IRC[02:08:37] <stuartdouglas> jamezp: what JDK, OS and core count?[02:08:56] <dmlloyd> it's gotta be a missing dependency[02:09:17] <stuartdouglas> what tests does it fail on?[02:09:21] <dmlloyd> JavaNamespaceSetup is running before the NamespaceSelectorService comes up[02:09:26] <bstansberry> the 3 ejb3 ones[02:09:26] <jamezp> 1.6.0_24, OS X, duel core.[02:09:39] <dmlloyd> that is the same setup that brian has[02:09:44] <bstansberry> yes[02:10:05] <stuartdouglas> maybe they screwed the memory model in 1.6.0_24[02:17:29] <jbossbot> git [jboss-as] push master 14945a9.. bstansberry at jboss dot com Close the client on quit[02:17:30] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/9432be9...14945a9[02:18:21] <stuartdouglas> I am trying with 1.6.0_24, but I can't reproduce it, I have more cores though[02:18:47] *** rmaucher has quit IRC[02:23:33] <dmlloyd> there seems to be no command on mac to set CPU affinity for a process[02:23:37] <dmlloyd> +lame+[02:23:58] <jbossbot> git [jboss-as] push master 098af09.. kabir Descriptions for Tx subsystem[02:23:59] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/14945a9...098af09[02:24:08] <stuartdouglas> that is what I have been googling for[02:24:29] <stuartdouglas> I have tried using nice to slow things down a bit, but still not failures[02:24:35] <dmlloyd> and mac users always answer with a snobby "why would you want to do this"[02:24:40] <dmlloyd> idiots ;|[02:24:45] <jamezp> Here's the errors I'm getting, but I haven't looked into them at all.[02:24:47] <jamezp> testDeploymentStreamApi(org.jboss.as.test.surefire.servermodule.ServerInModuleStartupTestCase)[02:24:47] <jamezp> testDeploymentFileApi(org.jboss.as.test.surefire.servermodule.ServerInModuleStartupTestCase)[02:24:48] <jamezp> testFilesystemDeployment(org.jboss.as.test.surefire.servermodule.ServerInModuleStartupTestCase)[02:25:03] <stuartdouglas> jamezp: that one is a timing error[02:25:07] <dmlloyd> jamezp: can you pastebin a stack trace? then I can at least see if it's the same issue[02:25:19] <stuartdouglas> I have seen it before, different to bstansberry's error[02:25:25] <jamezp> Sure thing.[02:25:35] * dmlloyd prefers fpaste.org[02:25:44] <smarlow> dmlloyd: I must not be having the SFSB test bean depend on my service. That has to be it[02:26:31] <dmlloyd> smarlow, wouldn't you want to have it be the other way around?[02:26:45] <dmlloyd> you're expecting the value to be set in your service, so it should wait for the SFSB to appear[02:27:02] <jamezp> dmlloyd: THis is probably the most relevant bit. http://fpaste.org/mSyW/[02:27:41] <stuartdouglas> Those deployment tests have heaps of 'wait for a bit and hope that it works code'[02:27:41] <jamezp> I don't really get a stack trace as much as that failure just about every time.[02:27:42] *** jpearlin1 has left #jboss-as7[02:27:43] <smarlow> well, if the sfsb doesn't depend on my service, won't the sfsb start executing too soon?[02:27:45] <dmlloyd> jamezp: what about the output from testsuite/smoke/target/surefire-reports/test.Blah.txt[02:27:49] <bstansberry> jamezp: in testsuite/smoke/target/surefire-reports there are files for that test; those are what we need to see[02:27:55] <dmlloyd> stuartdouglas: I thought we fixed those :([02:28:02] * smarlow walking away for a min[02:28:28] <jamezp> Duh, didn't think of that. Let me have a look.[02:28:28] <stuartdouglas> I did for the demo based tests, this is for filesystem based deployment, which is kinda hard to do any other way :-([02:28:37] <bstansberry> yeah[02:28:39] <dmlloyd> ah[02:29:08] <bstansberry> the test gets a notification of the mbean in the deployment coming up, but it has to wait a bit longer for the scanner to complete its work[02:29:51] <stuartdouglas> I notice that the scan interval is 1000, and the tests seem to wait for 100 lots of 10ms[02:30:02] <stuartdouglas> that does not seem to leave much room for error[02:30:15] <bstansberry> it doesn't need to wait for the scan interval[02:30:58] <bstansberry> it needs to wait the time between when the mbean sends a notification and the deployment thread returns and finishes with the marker files[02:31:07] <bstansberry> should be some few ms[02:32:40] <bstansberry> no argument it's ugly though[02:32:52] <smarlow> dmlloyd: is there a way to have the sfsb injection wait until my service is ready? I think you said before, that if my injector service depends on the target, the target will wait for me to have my value set?[02:32:58] <jamezp> My errors could be my fault. I had the server running.[02:33:18] <bstansberry> all right! welcome to the club![02:33:22] <dmlloyd> smarlow: yeah that's how dependencies work. if A depends on B, A will not start until B is fully started[02:33:33] <dmlloyd> smarlow: it's all callback-driven, so no threads block waiting for anything[02:34:38] <dmlloyd> smarlow: when you add an entry to a component's dependency set, or its resource injection set (with the dependency flag set to true), you're adding another external service/component that has to be up before your component will be started[02:35:00] <dmlloyd> this means that when your component *does* start, everything you depend on is guaranteed to be up, and to stay up, until after your component has stopped again[02:41:22] <smarlow> dmlloyd: in http://fpaste.org/QOnq/ , how do I add a dependency on the SFSB?[02:41:48] <smarlow> getInjectorServiceName() is where I add the injector service[02:43:03] <dmlloyd> I don't even really know what the semantics of the two annotations even are[02:43:31] <dmlloyd> are you sure you want to add a dependency on the SFSB? seems likw what you really want is to add one _from_ the SFSB to another service?[02:43:59] <baileyje> from SFSB to the PU service[02:44:16] <smarlow> dmlloyd: from a dependency point of view, it seems like the SFSB is not directly connected (dependency wise) with the PU[02:44:27] <dmlloyd> okay in that case you have to do: bindingDescription.setDependency(true)[02:44:29] <smarlow> the injector depends on the PU[02:44:30] <dmlloyd> everywhere you create it[02:44:52] <dmlloyd> check the javadoc for org.jboss.as.ee.component.BindingDescription#setDependency()[02:46:53] <baileyje> smarlow: Is your latest changes pushed. I can take a quick look.[02:47:14] <dmlloyd> baileyje: http://fpaste.org/QOnq/ is his main processor[02:47:16] <dmlloyd> if tha thelps[02:47:34] <baileyje> Yeah. I was messing with that before.[02:49:01] <smarlow> I just set the bindingDescription to true and will run again. I haven't committed a few changes since I started hacking up the access to the TM.[02:50:13] <dmlloyd> hey stuartdouglas - if a TLD has a class-path ref defined on it, we're treating that as a dependency on a sibling TLD right?[02:50:57] <stuartdouglas> yes, although just on the module service for that deployment[02:51:15] <dmlloyd> okay, and what if the module isn't there yet at the time of deployment? does it set up an unsatisfied dep?[02:51:22] <dmlloyd> MSC dep I mean?[02:51:39] <stuartdouglas> it will block till it is[02:51:53] <dmlloyd> using a service dependency though, right?[02:51:59] <stuartdouglas> yea[02:52:06] *** rawbdor has quit IRC[02:52:19] <dmlloyd> so what causes: Class Path entry foo.jar in "/content/blah.war" does not point to a valid jar for a Class-Path reference.[02:53:03] <dmlloyd> the "/content/" I can't explain, it doesn't appear in other messages[02:53:48] <stuartdouglas> a bug? I think the /content comes from VFS, but that looks like the message that you get when a class-path entry points somewhere inside the deployment[02:54:15] <dmlloyd> okay, so maybe something at some level thinks that the deployment root is /content or something...[02:54:53] <stuartdouglas> what is the actual structure of the deployment? I assume it is a war, but what does the cp ref say?[02:56:25] <dmlloyd> CP ref says just "foo.jar" in this case[02:56:36] <dmlloyd> or actually, "derby.jar" which is deployed as a standalone JAR[02:57:00] <stuartdouglas> that will be interpreted as foo.jar in the root of the war[02:57:12] <dmlloyd> really, not as a sibling deployment?[02:57:18] <dmlloyd> guess that explains it...[02:57:36] <dmlloyd> does that make sense though?[02:57:39] <stuartdouglas> It makes a lot more sense for ears[02:57:59] <stuartdouglas> also the spec does not require sibling deployments to work[02:58:13] <dmlloyd> yeah I get that[02:58:33] <dmlloyd> so what's the sibling syntax then? ../derby.jar?[02:58:38] <stuartdouglas> yes[02:58:52] <stuartdouglas> and for absolute ones you can use a /[02:59:00] <dmlloyd> but that's not really consistent with things inside the deployment[02:59:10] <stuartdouglas> ?[02:59:13] <dmlloyd> like something in web-inf/lib would refer to a sibling without .. correct?[02:59:37] <dmlloyd> and of course on the command line (java -classpath) you wouldn't use .. either[02:59:44] <stuartdouglas> no, the cp refs are resolved relative to the resource root that declares them[02:59:58] <dmlloyd> otoh those things don't have a syntax to see "inside" themselves at all[03:01:04] <stuartdouglas> The referenced .jar file or directory is named using a URL relative to the URL of the referencing JAR file[03:01:10] <stuartdouglas> EE 8.2.1[03:01:33] <dmlloyd> yeah but relative means: "bar" relative to "file:///foo" is "file:///bar", not "file:///foo/bar"[03:01:55] <smarlow> at least the app is stuck now rather than failing.[03:02:11] <dmlloyd> stuck is good, smarlow! it just means that your dep isn't there :)[03:02:23] <dmlloyd> and a stuck app can still be undeployed cleanly[03:02:35] <dmlloyd> (the beauty of callback-driven dependencies)[03:02:57] <stuartdouglas> hmm, actually you are probably right[03:03:18] <stuartdouglas> foo.jar should be a sibling deployment[03:04:59] <jamezp> dmlloyd: I'm still not sure this is right, but I think it's at least close. https://github.com/jamezp/jboss-as/blob/depends-on/ejb3/src/main/java/org/jboss/as/ejb3/deployment/processors/EjbDependsOnAnnotationProcessor.java[03:05:07] * dmlloyd bets that had he finished his MSC visualizer for jconsole, people would really grok dependencies much better right now[03:05:47] <jamezp> The name resolution is my biggest concern.[03:05:54] <dmlloyd> jamezp: yeah this looks reasonable. baileyje what do you think about this as a first pass[03:07:00] <baileyje> Yeah. Looks good.[03:07:07] <jamezp> My main concern is I'm not appending the beanInterface name to the ServiceName. I'm not sure how to get it :-)[03:07:34] <baileyje> well.. hmmm. YOu have a component name correct?[03:07:38] <baileyje> You aren't injecting.[03:07:52] <baileyje> So you could just depend on the install service of the component.[03:07:52] <baileyje>[03:07:53] <baileyje>[03:08:18] <jamezp> All I get is what would be equivalent to EJB.beanName().[03:09:37] <dmlloyd> hm yeah should depend on START, not VIEW[03:09:51] <jamezp> Ah, that's where that comes in then.[03:09:53] <baileyje> Yeah. For sure. That should be good enough for now[03:09:58] <dmlloyd> then you don't need the interface name[03:10:03] <dmlloyd> I'm... 80% sure that's right :)[03:10:15] <jamezp> Sweet. That's a change I can handle :-)[03:10:21] <dmlloyd> non-@Startup singletons might flake out but we'll deal with those later[03:10:22] <baileyje> Then I will trick someone else into parsing and resolving the # version of the bean names.[03:10:43] <jamezp> Hehe, sounds good.[03:11:24] <baileyje> dmlloyd: Are we attaching the DU to the nested roots for sub-deployments?[03:11:50] <smarlow> hmm, I'm missing the component: jboss.naming.context.java.comp.jpa-example.jpa-example.SimpleStatefulSessionBean[03:12:11] <baileyje> smarlow: You are missing some commits on upstream.[03:12:20] <baileyje> I fixed an issue earlier with the context not being added.[03:12:29] <baileyje> try stashing your changes and rebasing upstream[03:12:43] <dmlloyd> baileyje: hmm, that code changed recently...[03:12:43] <smarlow> cool, I'll sync up[03:12:57] <jamezp> dmlloyd: Changed and pushed.[03:13:04] <smarlow> walking away for a min[03:14:43] <dmlloyd> jamezp: can you squash it down to one commit?[03:15:28] <jamezp> I can try :-) I would guess just switch to master and make the changes?[03:16:03] <stuartdouglas> git fetch origin && git rebase -i origin/master[03:16:20] <dmlloyd> yeah then follow the directions[03:16:23] <stuartdouglas> and then when the editor opens squash the commits into one[03:16:35] <jamezp> Sweet. Thanks![03:16:54] * jamezp is still a git newbie[03:18:45] <Nihility> stuartdouglas: dmlloyd: yes sibling is the correct answer[03:18:59] <stuartdouglas> I will change[03:19:02] <stuartdouglas> it[03:19:07] <Nihility> class-path in EE just follows the JAR spec[03:19:16] <dmlloyd> was gonna say, nah we like you just how you are stuartdouglas[03:19:18] <dmlloyd> :)[03:19:24] <Nihility> ROFL[03:19:45] <Nihility> he only codes in a certain outfit[03:19:53] <dmlloyd> haha[03:20:10] <dmlloyd> a purple tie and a fez cap[03:20:25] <dmlloyd> and bunny slippers[03:20:28] <Nihility> yes and a shirt that says IZ L33TZ[03:20:50] <Nihility> lol reminds of real genius[03:21:40] <dmlloyd> so is he val kilmer or the dude who lives in the basement?[03:22:36] <Nihility> well the bunny slippers definitely val kilmer[03:22:41] <Nihility> if it was pjs, dude in the basement[03:23:01] <dmlloyd> then I'm the kid who just starts freaking out while studying in the library[03:23:09] <Nihility> stuartdouglas: all you need is a big spinning mirror[03:23:34] * stuartdouglas has no idea what is going on[03:24:12] <dmlloyd> what, you haven't seen "real genius"? did australia just skip over the 80s? :)[03:24:22] <Nihility> and you can make a hellavu lot of popcorn[03:25:17] * stuartdouglas was a good 4 months old when real genius came out[03:25:23] <dmlloyd> ouch[03:25:26] <dmlloyd> just.... ouch[03:25:37] <dmlloyd> :D[03:27:38] <stuartdouglas> so looking at it inside the deployment foo.jar is treated as a sibling, but I have screwed up handling siblings for top level deployments[03:28:27] <Nihility> http://www.youtube.com/watch?v=wibqG9OCNj0[03:28:33] <Nihility> bunny slippers @ 47 seconds[03:29:50] <Nihility> guy screaming at 2:03[03:32:28] <stuartdouglas> why were they building a laser? Was it part of a plan for world domnation?/[03:32:48] <Nihility> in a way yes[03:32:57] <Nihility> but i dont want to give away too much :)[03:33:14] <Nihility> they are on a special research team[03:33:24] <Nihility> that is trying to come up with a 5 mw laser[03:33:36] <dmlloyd> megawatt, not milliwatt :)[03:33:43] *** JimMa has joined #jboss-as7[03:33:49] <Nihility> yeah Mw[03:33:50] <dmlloyd> or jiggawatt![03:33:59] <Nihility> its an awesome movie[03:34:24] <Nihility> a lot of funny college stunts[03:34:38] <Nihility> like turning the dorm into an ice skating rink[03:34:44] <dmlloyd> a lot of good 80s stuff in there[03:40:09] <Nihility> so can we consider 3, 4, 7 and 8 done:[03:40:13] <Nihility> http://community.jboss.org/wiki/Beta1MUSTHAVEList[03:40:41] <stuartdouglas> Is 8 really done?[03:41:00] *** pgier has joined #jboss-as7[03:41:00] *** ChanServ sets mode: +v pgier[03:41:03] <dmlloyd> I didn't do it[03:41:16] <Nihility> oh i thought someone did[03:41:19] <Nihility> do we need it?[03:41:25] <dmlloyd> 1.3.1 isn't even on the map for beta1 imo[03:41:34] <dmlloyd> beta 3 maybe[03:41:39] <Nihility> carlo committed didnt he?[03:41:46] <Nihility> 1.3[03:41:48] <dmlloyd> dunno, I thought we agreed on fixed impls[03:41:55] <dmlloyd> not configurable caches or pools[03:42:21] <dmlloyd> oh I released modules btw[03:42:52] <Nihility> [00:44:44] <wolfc> got https://github.com/wolfc/jboss-as/commits/ejb3-ee with Cache & Pool impls, BMT Tx and injecting SessionContext[03:42:52] <Nihility> [00:45:09] <wolfc> It is passing smoke-tests[03:43:05] <dmlloyd> oh ok cool[03:43:19] <Nihility> i guess we do not have CMT[03:43:23] * smarlow synced up with the latest changes and almost ready to try again[03:43:26] <Nihility> unless he just left that out[03:43:43] <Nihility> stuartdouglas: how is ejb and weld?[03:44:01] <stuartdouglas> it works for some definition of 'works'[03:44:10] <Nihility> also this:[03:44:11] <Nihility> <wolfc> https://github.com/wolfc/jboss-as/commit/752359b1e8a3eb683c4bac0416caef6803cc6cf2#L8R49[03:44:11] <Nihility> [00:57:32] <jbossbot> git [jboss-as] 752359b.. Carlo de Wolf Injecting SessionContext on a BMT bean and use UserTransaction[03:44:13] <jbossbot> git [jboss-as] 752359b.. Carlo de Wolf Injecting SessionContext on a BMT bean and use UserTransaction[03:45:02] <stuartdouglas> Weld is made a aware of the EJB's and can instantiate them[03:45:16] <jamezp> dmlloyd: Done, https://github.com/jamezp/jboss-as[03:45:34] <dmlloyd> great, thanks jamezp[03:45:37] <stuartdouglas> no real interceptor support yet, although I should have a really hacky way of doing it up and running soon[03:46:07] <stuartdouglas> jpa support is mostly ready to go, but waiting in a topic branch until it gets integrated[03:46:13] <jamezp> Sorry it took so long. I'll have my wife give a hard punch to the kidney so I won't do it again ;-)[03:46:32] <stuartdouglas> also none of this stuff has really been tested past the basic 'it works' stage[03:49:39] <jamezp> dmlloyd: I'm going offline. If something doesn't look right, feel free to email me.[03:50:11] <frainone> dmlloyd: I'm assuming that it is alright if the @WebServiceRef implementation does not support @Addressing,@HandlerChain, @MTOM, etc for now?[03:51:06] <dmlloyd> jamezp: okay, thanks.[03:51:16] <dmlloyd> frainone: yeah taht's fine.[03:51:22] * dmlloyd didn't even know about any of that stuff :)[03:51:30] <Nihility> frainone: all of that stuff seems like it should be handled by the ws framework[03:51:31] <jamezp> Thank you. I really appreciate all the help.[03:51:48] <dmlloyd> jamezp: no prob. come back again for more fun stuff! :)[03:52:08] <Nihility> @HandlerChain for example just specifies afile[03:52:28] <jamezp> Oh I will. I'll be checking in Wednesday probably. Last day off tomorrow and I better spend it with the wife :-)[03:52:37] <Nihility> so yeah looks like carlo's stuff does not include CMT[03:52:51] <stuartdouglas> are we really going to release today?[03:53:13] <Nihility> looks like its not happening[03:53:13] <frainone> Nihility: as far as I understand, it won't be handled unless I treat it along with @WebServiceRef[03:54:42] <Nihility> we dont have a ton of stuff on this must have list[03:54:49] <Nihility> so either we need to decide its not releavant[03:54:56] <Nihility> or we block[03:55:31] *** jamezp has quit IRC[03:55:36] <Nihility> if by some luck we get jpa[03:55:41] <stuartdouglas> the weld ejb integration will take quite a bit to get working properly, imho whatever is working once everything else is ready is what we go with[03:55:57] <stuartdouglas> jpa is probably worth blocking for[03:56:04] <Nihility> and carlos branch merges ok[03:56:27] <frainone> Nihility: how do I test if a org.jboss.jandex.Type is assignable from a class? Is there a way to do it?[03:56:28] <Nihility> then i think we should tag[03:56:45] <dmlloyd> I already merged carlo's stuff earlier[03:56:47] <dmlloyd> afaik I have everything[03:57:22] <frainone> and I'm almost done with a bare implementation of @WebServiceRef[03:57:31] <stuartdouglas> frainone: walk the classes superclasses/interfaces looking for the Type[03:57:47] <frainone> stuartdouglas: thanks![03:58:08] <Nihility> frainone: Type has a kind() and a name() on it[03:59:00] <Nihility> ah i misunderstood the question[03:59:27] <stuartdouglas> The other way round does not work, as you cannot be sure that Type's superclass/interfaces are present in the Index[04:00:32] <frainone> stuartdouglas, Nihility: I don't get it. I have a Type. I need to know if it is assignable from Service.class. I can't find a way to do it[04:01:04] <Nihility> Type.name() gives you the class name[04:01:28] <Nihility> index lets you look up a ClassInfo from that[04:01:42] <Nihility> the info containes a super class name[04:01:56] <Nihility> thts the walking part[04:02:14] <Nihility> there is an index per jar[04:02:14] <stuartdouglas> so you need to know if the Type implements Service ?[04:02:53] <Nihility> we have a composite index[04:02:56] <frainone> exactly[04:02:59] <Nihility> that encompasses a deployment[04:03:29] <Nihility> the class info contains a list of interfaces[04:03:30] <dmlloyd> hmm guess I missed four of carlo's commits....[04:03:33] <Nihility> you crawl those[04:04:20] *** rawbdor has joined #jboss-as7[04:04:40] <stuartdouglas> the problem is that there can in theory be situations where you have Type extends SuperType, SuperType implements Service, where supertype is not in the index[04:05:01] <stuartdouglas> although this is very unlikely[04:06:24] <Nihility> yeah it depends on the rules of the annotaiton[04:07:52] <Nihility> to truly know you have to go up the type try[04:08:00] <Nihility> and for each tree get interfaces[04:08:04] <Nihility> merge them into a set[04:08:19] <Nihility> then for all entries in the set crawl for super types[04:08:44] <frainone> yeah, I'm doing this recursively[04:09:17] <Nihility> although note that if you are doing this just for validation you can also choose to delay processing of it[04:10:15] <Nihility> like at some point the value class must be instantiated[04:10:20] <Nihility> and at that time you could fail[04:11:46] <Nihility> stuartdouglas: so that patebin demo requires all of the ejb int stuff to finish right[04:12:19] <stuartdouglas> yea, I think part of the aim of the demo was to show off ejb and CDI working together[04:12:53] <Nihility> so is there a demo that works well enough that we could point to working on AS7[04:13:10] <stuartdouglas> numbergess, princess-rescue and permalink[04:13:24] <stuartdouglas> or at least, they worked a few days ago[04:13:35] <Nihility> thats something[04:13:39] <Nihility> :)[04:13:54] <stuartdouglas> so are you considering releasing without JPA?[04:14:00] <Nihility> no[04:14:20] <Nihility> just trying to get an idea of where we are[04:14:32] <Nihility> i am out for the rest of the week[04:14:43] <Nihility> but was going to be online for doing the binary release after the tag is done[04:16:27] <Nihility> technically the due date (after being shifted 2 weeks) is tommorow[04:16:46] <Nihility> brian or david is going to be doing the tag since i wont be here[04:17:18] <Nihility> smarlow: i think its safe to say that we wont have jpa today right?[04:17:19] <smarlow> I'm close to having my little (works at all) non-transactional demo to pass[04:18:33] <jbossbot> git [jboss-as] push master d0e4cdd.. James Perkins Initial commit to handle @DependsOn annotation.[04:18:33] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/098af09...d0e4cdd[04:21:08] <smarlow> Nihility: I'm just testing the non-tx path so far. The next test would be tx (maybe with bmt)[04:21:45] *** smcgowan has quit IRC[04:23:38] *** mbg has joined #jboss-as7[04:27:16] <frainone> Nihility: how do I retrieve the jandex index? I thought I could do a du.getAttachment(Attachments.ANNOTATION_INDEX), but that returns null[04:27:31] <dmlloyd> you may be doing it too early[04:27:35] <dmlloyd> (before the index is built)[04:27:38] <stuartdouglas> Attachments.COMPOSITE_ANNOTATION_INDEX[04:27:45] <dmlloyd> ah, or that ;)[04:27:47] <stuartdouglas> Attachments.ANNOTATION_INDEX is attached to resource roots[04:28:00] <stuartdouglas> not the DeploymentUnit[04:28:08] *** pgier has quit IRC[04:28:31] <baileyje> Yeha. Yo would want the composite index if you ware working against the DU[04:28:42] <Nihility> can i get an opinion[04:28:52] <baileyje> Attachments.COMPOSITE_ANNOTATION_INDEX[04:28:56] <Nihility> who thinks we need CMT on EJBs for beta1[04:28:59] <Nihility> ?[04:29:12] <Nihility> im doing an imprecise tally here[04:29:41] * dmlloyd thinks probably yes[04:29:48] * stuartdouglas agrees[04:29:55] <dmlloyd> not 100% tho[04:30:06] <stuartdouglas> 'We have ejb's, they just don't do anything that you would expect ejb's to do'[04:32:51] <frainone> +1, I wouldn't call it beta[04:35:07] *** Jaikiran has joined #jboss-as7[04:35:07] *** ChanServ sets mode: +v Jaikiran[04:36:01] <Jaikiran> dmlloyd: haven't completely booted up, but is there something that i can help with in the next few min/hours if Beta1 is still in progress?[04:37:39] <Nihility> speak of the devil[04:37:40] <Nihility> :)[04:38:03] <Nihility> it looks like the release will not be tagged tonight[04:38:48] <Nihility> Jaikiran: dmlloyd was running into conflicts with wolfc's patch and your changes[04:39:12] <Jaikiran> Nihility: i see. i can take a look at those and see what's causing it[04:39:27] <Jaikiran> oh just saw your mail too[04:39:49] <dmlloyd> it's carlo's ejb-ee branch, I think he just didn't get a chance to rebase maybe[04:40:06] <Jaikiran> i'll check with him then[04:40:25] <Nihility> yeah everyone dont stay up too late[04:40:31] <dmlloyd> or maybe I just grabbed the wrong branch, or something[04:41:16] <Nihility> we still have another day and the release will be better[04:42:02] <dmlloyd> okay[04:42:11] <dmlloyd> I kinda feel crappy anyway, I think I'm going to head to bed[04:42:15] <dmlloyd> unless someone needs something[04:42:44] <Jaikiran> good night. see you later[04:42:51] <Nihility> i am still blown away at the number of commits today[04:42:52] <Nihility> :)[04:42:56] <Nihility> its a new record[04:43:36] <Jaikiran> :)[04:44:27] <frainone> dmlloyd: in that case, I can commit the initial version of @WebServiceRef tomorrow[04:44:35] <stuartdouglas> I also have a few commits ready to go[04:44:41] <stuartdouglas> or nearly ready to go[04:44:52] <dmlloyd> okay, frainone - just leave a message here whenever you like, I'm sure someone will be around to review[04:45:10] <dmlloyd> if it's already reviewed by someone we can just merge it straightaway in the morning[04:45:12] <Nihility> stuartdouglas: anything you want in right now[04:45:19] <frainone> dmlloyd: ok[04:45:46] <stuartdouglas> nothing that can't wait[04:47:04] <dmlloyd> all right folks, have a good night[04:47:30] <frainone> you too[04:54:41] <Nihility> ok guys im taking off[04:54:45] <Nihility> have a good rest of the week[04:54:50] <Nihility> will see you later[04:55:38] <smarlow> i'm good (just fixing some transaction management code...)[04:55:55] <smarlow> dmlloyd: ^ I meant for you.[04:56:22] <smarlow> Nihility: have a great week, looks like I have about 4 minutes left to today :)[04:56:39] <Nihility> dont sweat it[04:56:50] <Nihility> maybe a fresh start will helpo[04:57:04] <smarlow> probably will[04:59:53] <frainone> Nihility: see you[05:00:11] <Nihility> frainone: have a good evening thanks![05:00:29] <jbossbot> git [jboss-as] push master ed8093d.. bstansberry at jboss dot com Move domain-client code to controller-client[05:00:29] <jbossbot> git [jboss-as] push master dd7d422.. bstansberry at jboss dot com Remove legacy protocol stuff from DomainClient[05:00:29] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/d0e4cdd...dd7d422[05:08:25] *** Nihility has quit IRC[05:17:43] <smarlow> cool, got my first success from the jpa demo. It also looked like some CMT is working as well[05:17:59] <bstansberry> :-)[05:18:30] <smarlow> should be able to actually create some data instead of just search for a row, that will not exist.[05:35:13] *** mbg has quit IRC[05:41:53] *** irooskov has quit IRC[05:43:52] <bstansberry> g'night all[05:43:59] *** bstansberry is now known as bstans_zzz[05:47:50] <stuartdouglas> night[06:02:30] <smarlow> got a basic jpa demo (insert/select) going. Fixed a few bugs along the way. :)[06:03:07] <smarlow> thanks for the help today dmlloyd, baileyje, stuartdouglas[06:03:23] * smarlow night[06:03:29] *** smarlow has quit IRC[06:19:23] *** ALR has quit IRC[06:30:59] <frainone> here is my @WebServiceRef commit for review: https://github.com/fl4via/jboss-as/commit/53d56b24a1593efd7deea42856765b657db7e5fd[06:31:00] <jbossbot> git [jboss-as] 53d56b2.. Flavia Rainone Add initial support for @WebServiceRef[06:31:32] <frainone> there is only basic functionality implemented and there is no support yet for @WebServiceRef for classes, only for fields and methods[06:32:34] *** frainone is now known as frainone_afk[06:50:06] <stuartdouglas> Weld TCK is now over 90%: Tests run: 847, Failures: 79, Errors: 0, Skipped: 0[06:55:36] *** ALR has joined #jboss-as7[06:55:36] *** ChanServ sets mode: +v ALR[07:08:55] *** ALR has quit IRC[08:06:49] *** jfclere has joined #jboss-as7[08:06:49] *** ChanServ sets mode: +v jfclere[08:13:22] *** tdiesler has joined #jboss-as7[08:13:22] *** ChanServ sets mode: +v tdiesler[08:31:31] *** emuckenhuber has quit IRC[08:32:20] *** miclorb has quit IRC[08:38:00] *** dimitris_ has joined #jboss-as7[08:38:00] *** dimitris_ has joined #jboss-as7[08:38:00] *** ChanServ sets mode: +v dimitris_[08:52:08] *** miclorb has joined #jboss-as7[08:55:58] *** rmaucher has joined #jboss-as7[08:56:21] *** aslak has joined #jboss-as7[08:56:24] *** aslak has joined #jboss-as7[08:56:24] *** ChanServ sets mode: +v aslak[08:56:29] *** emuckenhuber has joined #jboss-as7[08:56:29] *** emuckenhuber has joined #jboss-as7[08:56:29] *** ChanServ sets mode: +v emuckenhuber[08:59:18] *** darranl has joined #jboss-as7[08:59:18] *** darranl has joined #jboss-as7[08:59:18] *** ChanServ sets mode: +v darranl[09:08:13] *** ALR has joined #jboss-as7[09:08:13] *** ChanServ sets mode: +v ALR[09:12:26] *** aslak has quit IRC[09:12:45] *** aslak has joined #jboss-as7[09:12:45] *** ChanServ sets mode: +v aslak[09:15:01] *** adietisheim has joined #jboss-as7[09:18:00] *** pil-afk is now known as pilhuhn[09:36:53] *** opalka has joined #jboss-as7[09:36:53] *** ChanServ sets mode: +v opalka[09:45:55] *** epbernard has joined #jboss-as7[09:45:55] *** epbernard is now known as emmanuel[09:45:55] *** ChanServ sets mode: +v emmanuel[09:47:25] *** alesj has joined #jboss-as7[09:49:40] *** torben has joined #jboss-as7[09:49:40] *** torben has joined #jboss-as7[09:49:40] *** ChanServ sets mode: +v torben[10:05:04] *** slaboure has joined #jboss-as7[10:10:00] *** baileyje has quit IRC[10:10:17] *** baileyje has joined #jboss-as7[10:10:17] *** ChanServ sets mode: +v baileyje[10:14:52] *** jhalliday has quit IRC[10:23:06] *** darranl has quit IRC[10:27:46] *** jhalliday has joined #jboss-as7[10:28:02] *** jhalliday has quit IRC[10:33:50] *** miclorb has quit IRC[10:42:46] *** wolfc has joined #jboss-as7[10:42:46] *** ChanServ sets mode: +v wolfc[10:47:44] *** AndyTaylor has joined #jboss-as7[10:47:44] *** ChanServ sets mode: +v AndyTaylor[10:50:59] *** maeste has joined #jboss-as7[10:50:59] *** ChanServ sets mode: +v maeste[10:56:54] *** jhalliday has joined #jboss-as7[10:59:52] *** jcosta has joined #jboss-as7[10:59:52] *** ChanServ sets mode: +v jcosta[11:10:46] *** miclorb has joined #jboss-as7[11:16:04] *** pmuir has joined #jboss-as7[11:16:04] *** pmuir has joined #jboss-as7[11:16:04] *** ChanServ sets mode: +v pmuir[11:30:17] *** kkhan has joined #jboss-as7[11:30:17] *** ChanServ sets mode: +v kkhan[11:36:52] *** JimMa has quit IRC[11:58:55] <kkhan> maeste: Emanuel and I are addind descriptions to the subsystems as outlined here: http://community.jboss.org/wiki/DetypedDescriptionoftheAS7ManagementModel. TIme is getting short before the release, would you be able to look at datasources and resourceadpaters? I can do the connector subsystem, that one seems simple enough[11:59:04] <kkhan> s/addind/adding[12:01:37] <maeste> kkhan: let me finish an open task before. I'll ping you in an hour or so[12:01:46] <kkhan> Thanks :-)[12:13:15] *** stansilvert has quit IRC[12:17:36] *** emmanuel has quit IRC[12:23:59] *** smcgowan has joined #jboss-as7[12:30:06] *** epbernard has joined #jboss-as7[12:30:06] *** epbernard is now known as emmanuel[12:30:06] *** ChanServ sets mode: +v emmanuel[12:32:35] *** miclorb has quit IRC[12:40:49] *** aloubyansky has joined #jboss-as7[12:55:41] *** ALR has quit IRC[13:00:20] *** stansilvert has joined #jboss-as7[13:09:59] *** jpederse has joined #jboss-as7[13:10:06] *** jpederse has quit IRC[13:10:06] *** jpederse has joined #jboss-as7[13:10:43] *** jhalliday is now known as jhalli_lunch[13:20:34] *** tdiesler has quit IRC[13:21:58] *** smcgowan has quit IRC[13:36:52] *** smarlow has joined #jboss-as7[13:36:52] *** ChanServ sets mode: +v smarlow[13:43:07] *** mmoyses has joined #jboss-as7[13:43:07] *** ChanServ sets mode: +v mmoyses[13:44:04] *** darranl has joined #jboss-as7[13:44:04] *** darranl has joined #jboss-as7[13:44:04] *** ChanServ sets mode: +v darranl[13:50:53] *** balunasj has joined #jboss-as7[13:50:53] *** balunasj has joined #jboss-as7[13:50:54] *** ChanServ sets mode: +v balunasj[13:57:25] *** fnasser has joined #jboss-as7[13:57:25] *** ChanServ sets mode: +v fnasser[13:58:21] *** frainone_afk has quit IRC[13:58:34] *** frainone has joined #jboss-as7[13:58:34] *** ChanServ sets mode: +v frainone[14:01:38] *** frainone has quit IRC[14:02:16] *** frainone has joined #jboss-as7[14:02:16] *** ChanServ sets mode: +v frainone[14:06:50] *** balunasj has quit IRC[14:09:26] *** jhalli_lunch is now known as jhalliday[14:31:43] <darranl> does anyone else manage to attach a debugger to AS7? For some reason the JVM always crashes for me[14:31:53] <dmlloyd> yeah it works ok for me[14:31:57] <dmlloyd> sun jvm, linux 64-bit[14:32:08] *** AndyTaylor has quit IRC[14:32:14] <dmlloyd> it might be bad if you attach a 32-bit debugger to a 64-bit jvm, I"ve seen that fail before[14:32:33] <dmlloyd> I've also seen failure when attaching, say, a linux debugger to a JVM running on solaris[14:33:00] <darranl> everything for me is 32bit running on the same machine - hardly ever had problems like this on previous AS versions[14:33:36] <darranl> just checking if OpenJDK and Sun are getting mixed up[14:34:07] <dmlloyd> wolfc: give me a ping when you're ready to merge. I saw you had some stuff on ejb3-ee which I thought you wanted to merge but I guess it wasn't ready yet[14:34:42] <wolfc> dmlloyd, it's passing smoke-tests[14:35:12] <dmlloyd> when I merged it last night it didn't merge cleanly[14:35:18] <dmlloyd> so it probably at least needs a rebase[14:35:28] <Jaikiran> wolfc: can you pull in my latest changes on my branch (@Lock interceptor change)[14:35:28] <wolfc> weird, I did a clean rebase this morning[14:35:40] <wolfc> okay hold on[14:36:04] <dmlloyd> ok this would be about 10 hours ago[14:36:08] <dmlloyd> 10 or more[14:36:31] <Jaikiran> dmlloyd: wolfc: looking at the current code, i think we don't have system level interceptor facility for lifecycles callbacks on component instances, do we?[14:36:48] <Jaikiran> for ex: @PostConstruct of singleton beans need CMT tx interceptors[14:36:52] <wolfc> no we don't[14:36:59] *** bstans_zzz is now known as bstansberry[14:37:09] <Jaikiran> ok, so that needs doing then[14:37:14] <wolfc> it is still something that needs to be coded, but currently the interceptor setup is becoming to complex to handle[14:37:14] *** torben has quit IRC[14:37:29] <wolfc> especially once we start to mix ordering in[14:38:15] <dmlloyd> so you're saying we need to intercept lifecycle methods?[14:38:16] <Jaikiran> we have numerous levels of interceptors, yes. but wouldn't ordering only apply at the EJB level?[14:38:23] <dmlloyd> or are you saying we need our own lifecycle methods[14:38:24] <Jaikiran> dmlloyd: yep[14:38:38] <Jaikiran> dmlloyd: we need to interceptor lifecycle methods on the bean instances[14:38:49] <dmlloyd> with around-invoke interceptors?[14:39:00] <dmlloyd> oh, its this again[14:39:01] <bstansberry> darranl: are you trying to attach to a standalone server or a host controller?[14:39:11] <darranl> bstansberry, to a host controller[14:39:14] <Jaikiran> not really. @PostConstruct isn't an around-invoke.[14:39:26] <bstansberry> yeah, that always fails for kkhan and I[14:39:35] <dmlloyd> like I said last night - yes we need a way to *provide* txn/security/TCCL contexts to lifecycle methods[14:39:36] <bstansberry> our greatest bane[14:39:49] <dmlloyd> that does NOT mean that we need interceptors on them per se[14:39:56] <darranl> the one area I want to get a better feeling of with the debugger ;-([14:40:01] <bstansberry> illegal memory access[14:40:15] <dmlloyd> really bstansberry, that's very interesting[14:40:18] <Jaikiran> yeah, interceptors is one of those ways. there might be other which i probably missed[14:40:23] <maeste> kkhan: ping[14:40:41] <dmlloyd> Jaikiran: yeah well the way we were doing it before was simply to manually set those things up before invoking lifecycle methods[14:40:41] <bstansberry> yep. it's greatly slowing us down[14:41:06] <dmlloyd> bstansberry: any clues where the crash comes from?[14:41:21] <wolfc> dmlloyd, you're talking about instantiation & injection[14:41:27] <wolfc> we're talking about lifecycle callbacks[14:41:35] <dmlloyd> wolfc: and also lifecycle methods, it can be done the same way[14:41:38] <darranl> just applied the latest sun version and still the same crash ;-([14:41:58] <bstansberry> i'd have to do it again[14:41:59] <wolfc> right now we don't lifecycle callbacks at system or user level on interceptors[14:42:04] <bstansberry> it happens immediately[14:42:10] <darranl> on a positive note at least it proves the host controller re starts automatically ;-)[14:42:16] <bstansberry> haha, true![14:42:27] <darranl> for me it is happening as soon as it hits a breakpoint[14:42:56] <bstansberry> you have better luck than I do :)[14:43:09] <darranl> bstansberry, are you on a Mac or Fedora?[14:43:14] <bstansberry> mac[14:43:43] <darranl> and so it kkhan - I am on Fedora so Fedora gets you a little further[14:43:58] <bstansberry> FOSS for the win![14:44:10] <bstansberry> well, maybe...[14:44:22] <darranl> dmlloyd, and your successful debugging is that a domain instance or standalone?[14:44:22] <bstansberry> FOSS for the slightly less humiliating loss![14:44:35] <bstansberry> i can do standalone no problem[14:44:57] <bstansberry> but the HC is spawned by the PC, not from the command line[14:45:16] <dmlloyd> wolfc: we did before. They stopped working when we started changing from using org.jboss.as.ee.component.ComponentLifecycle to Interceptor for lifecycle methods[14:45:34] <darranl> yeah I have to attach the debugger twice - once to the PC and then once to the HC[14:45:34] <dmlloyd> darranl: it was always standalone[14:46:03] <darranl> dmlloyd, ok so it could be possible it will fail for you as well[14:46:08] <dmlloyd> darranl: what crashes, the HC or the debugger?[14:46:11] <opalka> darranl I was facing these debugging problems too. Are U using Eclipse, right?[14:46:12] <Jaikiran> dmlloyd: wolfc: we probably would need more than that, because we also have to take into account EJB specific exception handling semantics[14:46:18] <darranl> the HC crashes[14:46:28] *** jwulf has quit IRC[14:46:36] <opalka> darranl, Cleaning up the workspace, closing eclipse, opening eclipse, updating eclipse usually helped me to workaround it[14:46:37] <darranl> I was on Eclipse so was blaming eclipse but same behaviour with InteliJ[14:46:56] <dmlloyd> how are you guys configuring the HC to be debugged?[14:47:31] <darranl> I am just uncommenting the "JPDA settings for remote socket debuging" line in domain.conf[14:47:46] <darranl> and setting to suspend=y[14:48:04] <dmlloyd> ah, that's why you have to attach multiple times[14:48:13] <dmlloyd> are you using a different port # for each?[14:48:38] <darranl> no same port - after the first one is connected the second is free to listen on that port again[14:48:49] <dmlloyd> interesting.[14:49:07] <darranl> that was the reason I needed supend=y to have a chance to allow the first port to be used and freed[14:49:37] <dmlloyd> no crash for me[14:49:42] <dmlloyd> I'm attached to the PC and the HC[14:49:52] <wolfc> rebased onto upstream/master https://github.com/wolfc/jboss-as/commits/ejb3-ee[14:50:02] <dmlloyd> granted I'm still at d0e4cdd[14:50:06] <dmlloyd> let me update and retry[14:50:18] <darranl> dmlloyd, this has been happening for me for a while now[14:50:48] <dmlloyd> I'm 64-bit[14:51:05] *** asaldhan has joined #jboss-as7[14:51:05] *** ChanServ sets mode: +v asaldhan[14:51:24] <darranl> are you hitting a breakpoint? for me I can connect but it crashes when the breakpoint is hit[14:51:33] <dmlloyd> ah, no breakpoint[14:51:35] <kkhan> maeste pong[14:51:44] <dmlloyd> I wonder if it's related to 2x debugger on the same classes though[14:51:54] <dmlloyd> when I'm done rebuilding I'll test again[14:51:59] <dmlloyd> I'm using 64 bit if that helps[14:52:10] <maeste> kkhan: I've taken a look to the wiki[14:52:28] <maeste> kkhan: are there any subsystem already full documented to have a look?[14:52:48] <bstansberry> dmlloyd: i don't have the 2x debugger issue; I only configure debugging on the HC process[14:52:49] <maeste> kkhan: or can you say what basically I have to do to don't forget anything[14:53:11] <kkhan> Lots of them are: transactions, threads, logging are some more advanced examples[14:53:26] <dmlloyd> how do you manage that bstansberry?[14:53:43] <kkhan> When registering the subsystem operation handlers you provide it with a descriptionProvider[14:54:06] <maeste> kkhan: yup and we need to complete the properties file[14:54:20] <kkhan> To try it out start up AS in domain mode and run the cli[14:54:21] <kkhan> ./build/target/jboss-7.0.0.Alpha2/bin/jboss-admin.sh[14:54:26] <kkhan> Then type connect[14:54:28] <maeste> kkhan: localDescriptions if I well remember[14:54:31] <dmlloyd> where are you setting your breakpoint darranl?[14:54:32] <bstansberry> darranl: http://fpaste.org/F91U/[14:54:34] <kkhan> and to look at the subsystem[14:54:38] <kkhan> you execute[14:54:46] <dmlloyd> ah of course[14:54:47] <bstansberry> basically I stick the config in the correct place in domain.sh[14:54:51] <kkhan> you execute /profile=default/subsystem=security:read-resource(recursive=1, operations=1)[14:55:04] <darranl> dmlloyd, my breakpoint is on the first line of the execute method of ServerStartHandler[14:55:07] <dmlloyd> bstansberry: one difference is I'm using suspend=y[14:55:17] <kkhan> or replace 'security' with datasources or whatever[14:55:19] *** torben has joined #jboss-as7[14:55:19] *** torben has joined #jboss-as7[14:55:19] *** ChanServ sets mode: +v torben[14:55:21] <darranl> I then call ./host=local/server/server-config=server-one:start using the CLI to hit it[14:55:40] <maeste> kkhan: oki trying[14:56:07] <kkhan> To see less initially (since you get all the inherited operations) you can also just do /profile=default/subsystem=security:read-resource(recursive=1)[14:56:17] <dmlloyd> darranl: you mean like like 73 of org.jboss.as.host.controller.operations.ServerStartHandler ?[14:56:33] <darranl> dmlloyd, yes that's the line[14:56:51] *** stansilvert has quit IRC[14:56:54] <dmlloyd> okay how do I start up the CLI?[14:56:57] <dmlloyd> :)[14:57:07] <bstansberry> ./jboss-admin.sh[14:57:21] <bstansberry> then[14:57:24] <bstansberry> connect[14:57:39] *** hbraun has joined #jboss-as7[14:57:49] <alesj> dmlloyd: is there a get my hands on MSC' ServiceRegistry from Servlet?[14:58:01] <dmlloyd> well it hit the breakpoint darranl[14:58:19] <dmlloyd> alesj: that's a sure sign that "you're doing it wrong" :) no not really[14:58:20] <darranl> ;-( Looks like we all need 64 bit[14:58:43] <dmlloyd> darranl: but then it crashed[14:58:57] <alesj> dmlloyd: well, it's not like I will write nuclear power plant on top of that code or new NASA space shuttle ;-)[14:58:59] <darranl> ok so it does crash for us all[14:59:18] <alesj> dmlloyd: i just need it to hack my Grapher stuff :-)[14:59:20] <dmlloyd> it's the JDWP thread which crashed[14:59:35] <dmlloyd> alesj: maybe grab the platform MBeanServer instead[14:59:40] <dmlloyd> use the mbeans[14:59:49] <dmlloyd> they have the most complete info[14:59:59] <dmlloyd> somewhat simpler to code to[15:00:22] <alesj> which MBean gets me ServiceReg?[15:01:04] <maeste> kkhan: the subsystem already documented you have mentioned are on master?[15:01:16] <kkhan> maeste: Yes[15:02:07] <dmlloyd> alesj: jboss.msc:container=jboss-as[15:02:21] <dmlloyd> alesj: use jconsole to explore the available operations[15:02:21] <maeste> kkhan: and so is it a correcct output? https://gist.github.com/870742[15:02:26] <alesj> which returns which MBean class?[15:02:32] <alesj> dmlloyd: ^[15:02:39] <dmlloyd> you realize you could answer all your questions just by starting jconsole :)[15:02:53] <dmlloyd> it's ServiceContainerMBean or something like that[15:02:57] <dmlloyd> in the msc management pkg[15:03:01] <alesj> and AS7, and Grapher, ...[15:03:11] <alesj> while I'm actually working on EAP51 ...[15:03:47] <alesj> btw: what's the plan for JBAS-7882 in AS7?[15:03:49] <jbossbot> jira [JBAS-7882] Wrong provider code base for security provider included in packed ear [Open (Unresolved) Bug, Major, John Bailey] https://issues.jboss.org/browse/JBAS-7882[15:03:54] <kkhan> maeste: Sorry I gave you the wrong command, 'read-resource' gives you what is actually in the model, use 'read-resource-description' instead with the same guidelines regarding 'operations=1'[15:04:09] <kkhan> r-r-d gives you the description[15:04:45] <hbraun> dmlloyd: can i ask you something related to classloaders within modules?[15:05:05] <dmlloyd> hbraun: go for it[15:05:15] <maeste> kkhan: oki...a moment rebuildin my AS...I was thinking I haven't rebuilded after rebase[15:05:23] <hbraun> dmlloyd: i am integrating the console with the http api module: http://bit.ly/h0mp4O[15:05:25] <kkhan> It should look something like this https://gist.github.com/870742#comments[15:05:47] <hbraun> it basically servres the console resources from a jar file through the embedded httpd[15:06:11] <hbraun> but it requires a proper classloader to get hold of the resources[15:06:27] <dmlloyd> what JAR file?[15:06:34] <kkhan> maeste: It looks worse than it is :-) In the operations part everything apart from the 'add' operation is inherited[15:06:36] <dmlloyd> the console itself?[15:06:41] <hbraun> a jar file with the GWT application contents[15:06:43] <hbraun> yes[15:06:51] <darranl> going to try OpenJDK and see if that has similar issues[15:06:52] <dmlloyd> then just use ConsoleHandler.class.getClassLoader()[15:07:06] <dmlloyd> put it in a static final field even[15:07:36] <hbraun> ok, i tried context classloader and it didn't work[15:07:47] <kkhan> maeste: Added another comment to the gist where operations=1 is left out[15:07:48] <hbraun> i'll try what you suggest[15:07:50] <hbraun> makes sense[15:08:11] <dmlloyd> hbraun, unfortunately it's relatively inconvenient to find out your own initiating class loader so it often doesn't occur to people to try.[15:08:15] <alesj> hbraun: seen my email on Modules visualization?[15:08:24] *** kkhan is now known as kabir_lunch[15:08:35] <hbraun> alesj: nope[15:08:46] <alesj> it's on jboss-as7-dev ml[15:08:48] <hbraun> alesj: why do you ask?[15:08:59] <alesj> hbraun: as it needs your involvement ;-)[15:09:16] <hbraun> ;)[15:09:24] <darranl> No OpenJDK crashes in what looks like exactly the same way[15:09:26] <hbraun> oh well, then I better take alook ;)[15:09:37] <hbraun> alesj: tnx for the notice[15:09:38] <alesj> ;-)[15:09:41] <alesj> np[15:10:45] <maeste> kabir_lunch: so connector is already fine: https://gist.github.com/870757 have you already did something there?[15:11:30] <frainone> dmlloyd: when I finished the initial commit last night, there was nobody around to review it: https://github.com/fl4via/jboss-as/commit/53d56b24a1593efd7deea42856765b657db7e5fd[15:11:31] <jbossbot> git [jboss-as] 53d56b2.. Flavia Rainone Add initial support for @WebServiceRef[15:11:57] <frainone> dmlloyd: there is still a lot to be done, this is just basic functionality, and it is not implemented for @WebServiceRef on classes yet[15:12:52] <dmlloyd> okay I'll check it out shortly[15:14:23] <maeste> kabir_lunch: or is just because I've written LocalDescriptions.properties during my development[15:15:13] <hbraun> dmlloyd: works like charm[15:15:37] <Jaikiran> rmaucher: does the current AS7 upstream support web deployments containing web.xml?[15:15:42] <Jaikiran> i.e. servlets mapped in web.xml[15:15:43] <smarlow> dmlloyd: what do you think about merging https://github.com/scottmarlow/jboss-as in? I have more changes that I want to make but I could pause for a bit. :) Should I squash it down to one commit?[15:15:49] <hbraun> dmlloyd: great, now the console is served from a simple jar file that's inlcuded with the httpapi module[15:16:19] <dmlloyd> smarlow: yeah that's probably a good idea. it'll make it easier to rebase[15:17:25] <rmaucher> Jaikiran well, yes, it's the basics[15:17:51] <Jaikiran> rmaucher: ok, i am looking at a tck stuff where i don't see @EJB being injected into such web classes[15:18:06] <Jaikiran> so wanted to make sure those are currently supported before looking into it further[15:18:21] <rmaucher> ah ok, but the servlets will be there no problem[15:18:37] <Jaikiran> ok[15:23:51] *** mmoyses has quit IRC[15:26:25] *** rawblem has joined #jboss-as7[15:31:37] *** AndyTaylor has joined #jboss-as7[15:31:37] *** ChanServ sets mode: +v AndyTaylor[15:32:46] *** rawbdor has quit IRC[15:33:12] *** hbraun has quit IRC[15:33:25] <bstansberry> emuckenhuber: you want your "add web subsystem descriptions" commit pulled?[15:34:01] <emuckenhuber> bstansberry: hmm, yes.[15:34:24] <bstansberry> what domain demos have you tested?[15:36:02] <emuckenhuber> bstansberry: i think i did not look at the DS ones - the others used to work for me[15:36:18] <emuckenhuber> i can look at them - in case something is not working?[15:36:55] <bstansberry> used to when? yesterday when you were working on them?[15:37:05] <bstansberry> :)[15:37:21] <bstansberry> i'm just trying to get a sense of what exactly has been tested[15:38:36] <smarlow> dmlloyd: I squashed https://github.com/scottmarlow/jboss-as down after sync with the latest as7 bits.[15:38:49] <emuckenhuber> bstansberry: i ran them yesterday, but mostly did the config and servers stuff - so i will run them again and get back to you[15:38:51] <dmlloyd> okay, thanks. You're #3 in the queue, smarlow :)[15:39:01] <bstansberry> thanks![15:39:23] *** bgeorges has joined #jboss-as7[15:39:26] <dmlloyd> wolfc: build is failing for me - saying there's no trove:trove:jar:2.1.1 in the jboss repository.[15:39:48] <dmlloyd> (do we _really_ need trove?)[15:40:25] * wolfc is digging[15:40:42] <bstansberry> emuckenhuber: btw, IIRC the DS demos basically work by enabling H2, which is now enabled by default[15:41:25] <bstansberry> so 1) are they really doing anything and 2) do we want to go back to not enabling H2 by default?[15:42:06] <bstansberry> I enabled it as part of the detyped conversion, just so that code would be getting exercised without me having to change the file[15:43:27] <wolfc> dmlloyd, no. The persistence store hasn't been updated to jbmar yet.[15:43:43] <emuckenhuber> bstansberry: hmm, the DS one does not work for me[15:43:45] <emuckenhuber> the others do[15:44:40] <baileyje> The DS demo only works if you enable the H2 driver by installing it.[15:45:14] <baileyje> unless something changed late yesterday[15:45:23] <bstansberry> ah, ic. so we have the config ready, but nothing happens because the driver isn't deployed?[15:45:48] <baileyje> RIght. If you enable the DS in the config it will have a dep issue on server start.[15:45:58] <baileyje> Which is resolved when you install the driver[15:46:09] <bstansberry> ok, then for sure we should go back to commenting out that config[15:46:11] <baileyje> and then it will work. And continue to work on server start[15:46:27] <bstansberry> having a config that errors on start is not right :)[15:46:28] <baileyje> bstansberry: I just set the "enabled" attribute to false[15:46:53] <baileyje> and change the subsystem to actually adhere to it[15:46:59] <bstansberry> hmm, but we have no op to change that after boot[15:47:06] <bstansberry> unless you added one!! :)[15:47:11] <baileyje> bstansberry: Not yet[15:48:41] <bstansberry> so lets go back to commenting out the config; then the demos can do what they used to, but with the step of deploying the driver added[15:49:04] <dmlloyd> wolfc: would you (and/or Jaikiran) mind also updating the demos pom so that it lists the new demo(s) that you guys have introduced?[15:49:21] <Jaikiran> dmlloyd: it already has ejb3 demo AFAIK[15:49:24] <Jaikiran> let me check though[15:49:33] <wolfc> dmlloyd, everybody mucks into that one demo[15:49:40] <dmlloyd> oh ok[15:49:54] <dmlloyd> I thought maybe there were more of them[15:49:59] <wolfc> we really need to do stuff in testsuite[15:50:18] <Jaikiran> except for those basic demos, i have added some tests in smoke-tests[15:50:53] *** kabir_lunch is now known as kkhan[15:51:00] <Jaikiran> the "integration" tests would probably be a better place, but i haven't yet looked at those or run them successfully[15:51:41] <kkhan> maeste: Yes, connector looks fine. For some reason I thought it had not been done last night[15:52:13] <maeste> kkhan: yep, I've checked I've done it at the beginning[15:52:26] <maeste> kkhan: anyway I'm going to take care of other 2 subsystem[15:52:39] <kkhan> Cool, thanks a lot[15:54:02] <dmlloyd> Jaikiran: when I do "mvn package" the EJB3 demo isn't listed[15:54:16] * Jaikiran checks[15:55:21] <jbossbot> git [jboss-as] push master 53d56b2.. Flavia Rainone Add initial support for @WebServiceRef[15:55:22] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/dd7d422...53d56b2[15:55:41] <Jaikiran> ah ok, i now see what you mean[15:55:43] <Jaikiran> let me add it[15:56:13] <smarlow> do we have a place to list "how to run a demo" steps for JPA? These are basically here http://pastie.org/1674635[15:58:18] *** dimitris_ has quit IRC[16:00:08] <kkhan> smarlow: You could make the demo ExamleRunner output that info (I think that is what was done for datasources in the past)[16:00:49] *** balunasj has joined #jboss-as7[16:00:49] *** balunasj has joined #jboss-as7[16:00:49] *** ChanServ sets mode: +v balunasj[16:01:02] <smarlow> kkhan: great idea, thanks![16:01:53] <dmlloyd> smarlow: looks pretty good - you didn't even have to do the XML bits for this release TBH.[16:02:09] <dmlloyd> but assuming everything tests out OK, I'll take it :)[16:02:43] <Jaikiran> dmlloyd: https://github.com/jaikiran/jboss-as/commits/ejb3 done[16:02:50] <smarlow> dmlloyd: great, thanks! :)[16:02:53] <dmlloyd> bah, forgot to shut down the server again.[16:05:16] *** mbg has joined #jboss-as7[16:05:16] *** ChanServ sets mode: +v mbg[16:05:35] *** bgeorges has quit IRC[16:05:47] <dmlloyd> I'm seeing a lot of "10:02:59,818 WARN [org.jboss.as.ejb3.component.EJBComponent] (pool-9-thread-100) EJBTHREE-2120: deprecated getTransactionAttributeType method called (dev problem)"[16:05:50] <jbossbot> jira [EJBTHREE-2120] Tx logic doesn't take method-intf type into acount [Open (Unresolved) Bug, Major, Carlo de Wolf] https://issues.jboss.org/browse/EJBTHREE-2120[16:05:59] <dmlloyd> is this (a) a problem and (b) strictly necessary?[16:06:27] <wolfc> it's a bug that must be fixed soon[16:07:28] <wolfc> I'm still have issues where a CL picks up a service/* file, but can't find the class or package itself.[16:07:49] <jbossbot> git [jboss-as] push master 5bfb3f4.. Scott Marlow keep changes linear...[16:07:49] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/53d56b2...5bfb3f4[16:08:22] <dmlloyd> wolfc: ok we're gonna tag in, like, 4-5 hours[16:08:25] <dmlloyd> at the outside[16:09:09] <wolfc> dmlloyd, what are the bits you absolutely need?[16:09:37] <wolfc> Jaikiran, rmaucher, I still suspect that @EJB is not working for servlets.[16:09:40] <dmlloyd> well, your earlier patch would be good[16:09:50] <Jaikiran> wolfc: it's not working for any web artifacts[16:10:12] <Jaikiran> wolfc: i looked briefly at it and it appears web artifacts aren't converted to componentconfigurations and such[16:10:12] <wolfc> dmlloyd, I'll put trove on the ban list and clean up a bit[16:10:19] <Jaikiran> just bypassing the injection and other stuff[16:10:26] <dmlloyd> okay, I'll grab in Jaikiran's patch here[16:10:43] <dmlloyd> ping me when yours is ready[16:10:56] <dmlloyd> would be nice to at least hide that warning too, if it can't be fixed[16:11:58] <baileyje> where can I help?[16:12:23] *** kkhan_ has joined #jboss-as7[16:12:46] *** kkhan has quit IRC[16:12:47] <jbossbot> git [jboss-as] push master 1c0ba3f.. jaikiran Include ejb3 in the demo usage instructions[16:12:48] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/5bfb3f4...1c0ba3f[16:13:03] *** kkhan_ has quit IRC[16:13:24] <dmlloyd> wolfc: after your patch is it safe to say that we'll have working "Cache" and "Pool" behavior?[16:13:28] *** kkhan has joined #jboss-as7[16:13:28] *** ChanServ sets mode: +v kkhan[16:13:36] <dmlloyd> and EJB TX[16:14:06] <wolfc> dmlloyd, yes, but they are not configurable. EJB TX is already working. SFSB synchronization is coming with the patch. Exception handling is not yet implemented.[16:14:27] <rmaucher> wolfc, well, I never said @EJB worked for servlets ;)[16:15:09] <dmlloyd> we can get that working though[16:15:23] <dmlloyd> with the work by john and stuart we should be like 90% there[16:15:32] <dmlloyd> it's not necessary for this release IMO[16:15:42] <dmlloyd> but if someone wants to get it working in the next hour or two, go for it :)[16:18:23] <kkhan> asaldhan: Seems to be a mismatch between jboss-security.xsd and standalone.xml. xsd mentions 'application-policy' as a child of jaas, while 'security-domain' is used in practice[16:19:42] <dmlloyd> looks like the JPA demo is hanging, smarlow, when looking up an EJB[16:19:47] *** frainone has quit IRC[16:19:48] * dmlloyd will look in jconsole for more info[16:20:35] *** jfd has joined #jboss-as7[16:20:35] *** ChanServ sets mode: +v jfd[16:22:14] <baileyje> we aren't currently managing servlets as components right?[16:23:08] <dmlloyd> stuart started that code baileyje[16:23:12] <dmlloyd> but I don't know what the current status is[16:23:57] <baileyje> Ok. Servlets will inherently get EJB injection once they have a component, since I did not limit which kinds of components can have @EJB injection.[16:23:57] <baileyje>[16:25:44] <darranl> Interestingly attaching JDB directly does not crash the HostController[16:26:44] <smarlow> dmlloyd: I'll update my master branch and try the same. There are steps to avoid hanging :)[16:28:16] <dmlloyd> Service "jboss.naming.context.java.comp.jpa-example.jpa-example.SimpleStatefulSessionBean.env/persistence/entityManager" (class org.jboss.as.naming.service.BinderService) mode ACTIVE state DOWN (parent: jboss.deployment.unit."jpa-example.jar".INSTALL) (dependencies: jboss.naming.context.java.comp.jpa-example.jpa-example.SimpleStatefulSessionBean, jboss.jpa.persistenceunit."jpa-example.jar#H2DS"."javax.persistence.EntityManager org.jboss.as.demos[16:28:16] <dmlloyd> .jpa.archive.SimpleStatefulSessionBean.entityManager", jboss.deployment.unit."jpa-example.jar".component.SimpleStatefulSessionBean.CREATE) (has missing dependency)[16:28:37] <dmlloyd> that's a pretty complex one[16:29:28] <jbossbot> git [jboss-as] push master f7d2014.. Emanuel Muckenhuber add proxies operation param to read-resource[16:29:28] <jbossbot> git [jboss-as] push master 83174f7.. Emanuel Muckenhuber add web subsystem descriptions[16:29:29] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/1c0ba3f...83174f7[16:29:32] <smarlow> I need to switch some internal names to more simple unique names :)[16:29:55] <smarlow> especially like ^ that leak outside in plain view :)[16:30:04] <wolfc> dmlloyd, rebased onto upstream/master smoke-tests passing, https://github.com/wolfc/jboss-as/commits/ejb3-ee[16:30:16] <dmlloyd> in parciular - jboss.jpa.persistenceunit."jpa-example.jar#H2DS"."javax.persistence.EntityManager org.jboss.as.demos.jpa.archive.SimpleStatefulSessionBean.entityManager" does not look right[16:30:20] <dmlloyd> ok thanks wolfc[16:36:07] <smarlow> dmlloyd: works after following the steps http://pastie.org/1674635[16:36:22] <dmlloyd> ah interesting.[16:36:28] <smarlow> I'll print usage on the ^ steps and fix the ugly name ^[16:36:52] <dmlloyd> okay, not saying the name is ugly it just stood out as a possible issue since it doesn't look like how we name other things.[16:38:23] <smarlow> well, I keep clean it up in the next go around, with everything else. :)[16:39:47] <wolfc> smarlow, I'm missing classes around JPAMockSFSBContainerTestCase[16:40:14] <dmlloyd> maybe for beta2 we'll actually print nice missing dep error messages :)[16:40:16] <wolfc> Employee, SFSBInterface and SFSB[16:40:16] <smarlow> dmlloyd: might be nice to take https://github.com/scottmarlow/jboss-as/commit/9ed007d5c06431b9ca3de3fc1b70c4057d06f8c6[16:40:17] <jbossbot> git [jboss-as] 9ed007d.. Scott Marlow show usage info in JPA demo[16:40:57] <smarlow> wolfc: I decided to mock things up differently and killed some of the mock classes. I'll clean that up[16:41:13] <smarlow> wolfc: I couldn't build integration yesterday, is that fixed today?[16:41:55] <wolfc> smarlow, I can build it. Trying to get something to run in my IDE, but Arq & Modules don't agree upon class loading bits.[16:42:33] <dmlloyd> one last test build[16:42:50] <asaldhan> kkhan: let me ping marcus on that.[16:42:58] <kkhan> Cool[16:43:27] <asaldhan> kkhan: as7, we want to use security-domain[16:43:45] <kkhan> I'll change the schema then[16:45:07] *** emuckenhuber has quit IRC[16:46:07] <smarlow> wolfc, dmlloyd: I have two minor changes, show jpa demo usage steps and compiler errors fixed in JPAMockSFSBContainerTestCase. Should I squash them to one commit and do you want to pull them in?[16:46:32] <dmlloyd> I already grabbed your usage one[16:46:43] <dmlloyd> so I'd push the other one separately[16:46:58] <jbossbot> git [jboss-as] push master 79837ef.. Carlo de Wolf Setup BMT interceptors[16:46:59] <jbossbot> git [jboss-as] push master 5926977.. Carlo de Wolf Using real Cache and Pool implementations[16:46:59] <jbossbot> git [jboss-as] push master 7bbf4b2.. Carlo de Wolf Fixed NPE for BMT beans[16:46:59] <jbossbot> git [jboss-as] push master 55d3abc.. Carlo de Wolf Injecting SessionContext on a BMT bean and use UserTransaction[16:46:59] <jbossbot> git [jboss-as] push master b023d67.. Carlo de Wolf Created AbstractEJBInterceptor[16:46:59] <jbossbot> git [jboss-as] push master 35b5e21.. Carlo de Wolf Added StatefulSessionSynchronizationInterceptor[16:46:59] <jbossbot> git [jboss-as] push master d70e82f.. jaikiran Include ear name (a.k.a appName) for EJB bindings when deployed through .ear files. Also added tests for these bindings[16:47:00] <jbossbot> git [jboss-as] push master a3b2212.. jaikiran Change the @Lock interceptor to be at ComponentInstance level[16:47:00] <jbossbot> git [jboss-as] push master 974f7ff.. Carlo de Wolf Only warn once for EJBTHREE-2120[16:47:01] <jbossbot> jira [EJBTHREE-2120] Tx logic doesn't take method-intf type into acount [Open (Unresolved) Bug, Major, Carlo de Wolf] https://issues.jboss.org/browse/EJBTHREE-2120[16:47:01] <jbossbot> git [jboss-as] push master 1276c90.. Carlo de Wolf Banned trove:trove[16:47:01] <dmlloyd> better yet, just rebase on upstream[16:47:02] <jbossbot> git [jboss-as] push master 20f8922.. Scott Marlow show usage info in JPA demo[16:47:02] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/83174f7...20f8922[16:47:07] <dmlloyd> got a couple things in there :)[16:47:28] <smarlow> the other one is https://github.com/scottmarlow/jboss-as/commit/85b49f940df99879ec1861d50515846e1567c993[16:47:29] <jbossbot> git [jboss-as] 85b49f9.. Scott Marlow comment out references to deleted classes[16:47:56] <smarlow> oh, I'll rebase[16:48:01] <bstansberry> kkhan: can you compare domain.xml to standalone.xml for missing subsystems? I see that jax-rs and jpa aren't there[16:48:54] <asaldhan> kkhan: marcus: anil: for AS7 we will use picketbox-security-domain-configuration_4_0.xsd and not jboss-security.xsd[16:48:55] *** mmoyses has joined #jboss-as7[16:48:56] *** ChanServ sets mode: +v mmoyses[16:49:14] <asaldhan> kkhan: marcus: anil: for AS7 we will use picketbox-security-domain-configuration_4_0.xsd and not jboss-security.xsd (mmoyses)[16:49:22] <asaldhan> mmoyses: ^[16:50:03] <kkhan> asaldhan mmoyses Then that should be added to the AS 7 sources[16:50:52] <jbossbot> git [jboss-as] push master 533eae0.. Scott Marlow comment out references to deleted classes[16:50:52] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/20f8922...533eae0[16:51:35] <dmlloyd> wolfc: for that CL issue, you could probably just add the EE API module to ..protocol/jmx/main/module.xml[16:52:00] <mmoyses> kkhan: i will update the xsd in AS7. we were talking about the xsd used for standalone security[16:53:12] <wolfc> dmlloyd, but that's probably not what we want, or do we?[16:53:29] <dmlloyd> that's the CL it's using (I *think*)[16:53:37] <dmlloyd> as to whether that's the CL it *should* be using, I don't know[16:53:40] <wolfc> it's the test case that is using those classes[16:53:44] <wolfc> not arq-jmx[16:54:02] <dmlloyd> TCCL would be the only way to hand it a specific CL to use[16:54:19] <dmlloyd> I'm not sure that'll work though[16:54:28] <dmlloyd> for now that'll get you past the error though[16:54:38] <dmlloyd> unless of course there are other classes missing too :)[16:54:43] <dmlloyd> in which case, you're screwed![16:54:54] <dmlloyd> you might try TCCL though[16:55:18] <dmlloyd> ah yes, it's using OIS so TCCL should definitely work[16:55:57] <wolfc> OIS?[16:56:09] <dmlloyd> ObjectInputStream[16:56:15] <dmlloyd> java serialization[16:56:25] <wolfc> ah okay, I didn't even look that deep into the problem[16:57:30] *** pferraro has joined #jboss-as7[16:57:30] *** ChanServ sets mode: +v pferraro[16:58:01] <wolfc> I found another bug in Arq. I thought it was Modules, but Arq has an anti-pattern.[17:00:43] *** hbraun has joined #jboss-as7[17:01:17] <hbraun> does anybody else have problems with the nexus login?[17:01:31] <dmlloyd> not as of yesterday[17:02:16] <dmlloyd> if your jboss.org username changed you might want to try the new one[17:04:02] <hbraun> it changed several weeks ago[17:04:11] <hbraun> and used to be working[17:04:25] <hbraun> just want to make sure it's not just me (for whatever reason)[17:04:49] <dmlloyd> it's just you[17:05:00] <dmlloyd> I'm logged in (using my new jboss.org username though)[17:05:17] <hbraun> new as of stripped the email suffix?[17:05:31] <dmlloyd> yeah, I think that's how it went for most people[17:05:42] <dmlloyd> I had them change mine to "dmlloyd" though[17:05:49] <hbraun> yeah, I've verified jboss.org. It works. nexus doesn't[17:05:54] *** jfd is now known as jfd_away[17:06:04] <hbraun> nevermind. i'll figure it out[17:06:06] <dmlloyd> they sync up in a goofy way[17:06:13] <dmlloyd> it's possible that only your new one works on nexus[17:06:25] <dmlloyd> (now)[17:09:49] *** emuckenhuber has joined #jboss-as7[17:09:49] *** emuckenhuber has joined #jboss-as7[17:09:49] *** ChanServ sets mode: +v emuckenhuber[17:10:54] <mmoyses> kkhan: https://github.com/mmoyses/jboss-as/commit/81e28e65e57ad007ea77c46aa4f8b5854c157063[17:10:55] <jbossbot> git [jboss-as] 81e28e6.. Marcus Moyses change application-policy to security-domain in the xsd[17:11:33] <kkhan> Cool, bstansberry ^^[17:12:14] <bstansberry> thanks, mmoyses[17:12:18] *** hbraun has quit IRC[17:12:55] * smarlow intellij is still complaining about not finding org.jboss.as.ee.component.Attachments. I synced the project manually + reimported maven. I forget where the option to detect externally changed files is...[17:12:58] <wolfc> https://github.com/wolfc/jboss-as/commit/740f4a8fa04f1066722220f2e07958999ee9711c[17:12:59] <jbossbot> git [jboss-as] 740f4a8.. Carlo de Wolf Added web-ejb test[17:13:12] <wolfc> the good news, it is passing... hmm, there is bad news. :-)[17:13:20] <wolfc> is no bad news[17:13:34] <wolfc> I suspect that Weld does the injection.[17:13:59] *** balunasj is now known as balunasj_mtg[17:14:49] <rmaucher> nice, no bad news ;)[17:15:01] <Jaikiran> hmm, why would weld come in picture for that example?[17:15:31] <darranl> Is anyone currently working on the host registration with the domain controller?[17:17:20] <dmlloyd> wolfc: shall I pull it?[17:17:35] <wolfc> dmlloyd, I would say pull it all :-)[17:17:52] <dmlloyd> wolfc: okay, would you mind rebasing though[17:17:53] <wolfc> I've raised the problem here https://issues.jboss.org/browse/SHRINKWRAP-260 for the moment.[17:17:55] <jbossbot> jira [SHRINKWRAP-260] Anti-pattern use of getPackage [Open (Unresolved) Bug, Major, Unassigned] https://issues.jboss.org/browse/SHRINKWRAP-260[17:18:51] <jbossbot> git [jboss-as] push master 81e28e6.. Marcus Moyses change application-policy to security-domain in the xsd[17:18:51] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/533eae0...81e28e6[17:19:03] <bstansberry> dmlloyd: fyi ^^^[17:19:20] <bstansberry> i.e. we're having push races today ;)[17:19:31] <dmlloyd> got it[17:19:41] <dmlloyd> yeah it's been interesting :)[17:20:08] *** opalka has quit IRC[17:20:53] <kkhan> bstansberry: https://github.com/kabir/jboss-as/commit/598f03608712dd2eff8dfd4a2acf42009772a655 This is only a partial description missing out the security domain children. Maybe mmoyses or asaldhan can pick it up? (I'm going on vacation in an hour or so)[17:20:54] <jbossbot> git [jboss-as] 598f036.. kabir Partly describe the security subsystem[17:20:54] *** pmuir has quit IRC[17:21:37] <bstansberry> that's fine for the beta; no need to fill in more today[17:22:45] <aslak> wolfc, the issue has been reported to Stan previously and he is working on a fix. Alpha-12 of ShrinkWrap also supports adding packages as strings[17:22:46] <wolfc> please pick https://github.com/wolfc/jboss-as/commit/c9d1d33ecdfd4bf10388650cc90364fe5e156dec[17:22:47] <jbossbot> git [jboss-as] c9d1d33.. Carlo de Wolf Added web-ejb test[17:23:04] <dmlloyd> building now[17:23:08] <wolfc> aslak, that won't solve the issue[17:23:52] <aslak> wolfc, no, but it opens for a alternative[17:25:56] <emuckenhuber> bstansberry: i started on some messaging descriptions as well: https://github.com/emuckenhuber/jboss-as/commit/2cdf6fe29349af5cb3d2a47446825a1f44fb6bfb[17:25:57] <jbossbot> git [jboss-as] 2cdf6fe.. Emanuel Muckenhuber basic messaging subsystem descriptions[17:25:58] <mmoyses> kkhan: sure. i can take over[17:26:01] <wolfc> aslak, did you see my mail to jboss-as7-dev?[17:26:08] <kkhan> mmoyses: Thanks[17:27:33] <emuckenhuber> bstansberry: not sure if it makes sense to add much more, since i think quite a few things should be updated there[17:27:47] <bstansberry> emuckenhuber: no more please :)[17:28:14] <aslak> wolfc, aa, we have similar issues with arq, but that is when we are client server. with as7 you are all incontainer so this is a classloader issues, kkhan ?[17:28:28] <emuckenhuber> bstansberry: do we want the server-group rename for the tag?[17:28:45] <bstansberry> emuckenhuber: no, but can you look at the DS demo thing?[17:28:52] <emuckenhuber> bstansberry: for the domain?[17:28:58] <bstansberry> in general[17:29:12] <bstansberry> the issue is we are shipping a config that will log an error[17:29:39] <bstansberry> but if we comment out H2, then demos that need it will have to add it[17:29:58] <bstansberry> hmm, maybe better to just leave it as is[17:30:22] <emuckenhuber> hmm, we could automatically deploy the driver as part of the demos?[17:30:29] <emuckenhuber> not sure if that makes much sense though[17:30:29] <kkhan> bstansberry: https://github.com/kabir/jboss-as/commit/d4e187d402030e7583dbcb5c9341cb3c48703041 - I did not add the deployment scanner since I think that is a standalone only thing. :read-resource-description(recursive=1,operations=1) works[17:30:30] <jbossbot> git [jboss-as] d4e187d.. kabir Make sure domain.xml contains the same subsystems as standalone.xml and reorder for easier comparison[17:30:30] <jbossbot> git [jboss-as] push master 00e22f8.. kabir Partly describe the security subsystem[17:30:31] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/81e28e6...00e22f8[17:30:31] <smarlow> kkhan, bstansberry: I think that I missed added jpa completely to domain.xml. Missed: <subsystem xmlns="urn:jboss:domain:jpa:1.0"/>[17:30:50] <bstansberry> smarlow: kkhan is chasing that[17:30:56] <kkhan> smarlow: Added in my latest commit[17:31:00] <bstansberry> oh, he fixed it :)[17:31:08] <jbossbot> git [jboss-as] push master e5d9303.. Carlo de Wolf Added web-ejb test[17:31:08] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/00e22f8...e5d9303[17:31:19] <smarlow> bstansberry, kkhan: thanks, I saw the comments but wasn't sure. Thank you! :-)[17:32:07] <bstansberry> emuckenhuber: a demo that uses the DS will have to deploy the driver[17:32:22] <bstansberry> i figure if smarlow wrote a jpa demo that he's doing that?[17:33:33] <kkhan> aslak: It is not necessarily all incontainer, we have some as_client tests too. But to get the package don't you normally do MyClass.class.getPackage() anyway, so I don't get the problem in SHRINKWRAP-260[17:33:41] <jbossbot> jira [SHRINKWRAP-260] Anti-pattern use of getPackage [Open (Unresolved) Bug, Major, Unassigned] https://issues.jboss.org/browse/SHRINKWRAP-260[17:33:42] <smarlow> bstansberry: I printed usage instructions but didn't do that. Clearly, we found that not following the steps, results in a hanging experience[17:34:49] <aslak> kkhan, sorry, that was refering to wolfc email on jbossas7-dev about ClassLoader issues when deserializing the exception[17:34:56] <bstansberry> smarlow: ok thanks[17:36:22] <aslak> kkhan, subject: Missing Java EE API classes in Arq (via JMX)[17:36:29] <kkhan> aslak: Yes, that is something I have seen too. Forgot to mention it before[17:36:59] <dmlloyd> bstansberry: I think in the end we're going to want to have a <jdbc-driver module="org.h2"/> kind of thing in the appropriate subsys[17:37:21] <dmlloyd> it's gonna be the only real practical approach for distributing drivers[17:37:25] <aslak> kkhan, looking at this as a general solution: ARQ-299[17:37:26] <jbossbot> jira [ARQ-299] Create a IN_CONTAINER Exception Proxy [Open (Unresolved) Feature Request, Major, Andy Gibson] https://issues.jboss.org/browse/ARQ-299[17:37:45] <kkhan> The problem is that JMXExecutor uses its own classloader to deserialize the result. If they are application or other classes they are not visible from the org.jboss.arquillian.protocol.jmx module[17:37:45] <bstansberry> dmlloyd: makes sense[17:38:19] <aslak> kkhan, basically deserialize what we have available.. if not, the stack will have a 'undeserializable exception: bla bla bla"[17:38:45] <aslak> kkhan, to avoid having to have the world on the classpath.[17:39:01] <kkhan> If it instead used the test class classloader instead the classes would be (more likely to be) visible[17:39:37] <baileyje> bstansberry, dmlloyd: I still think it would be nice to allow the DS subsystem to deploy driver services as well[17:39:47] <dmlloyd> baileyje: yeah definitely[17:39:52] <dmlloyd> we need both[17:40:03] <dmlloyd> deployments are our only distribution mechanism right now[17:40:12] <baileyje> You want me to put something together?[17:40:19] <dmlloyd> sure, that'd be great[17:40:34] <baileyje> ok. When are you tagging?[17:40:43] <dmlloyd> in a few hours[17:41:10] <baileyje> dmlloyd: Ok. No problem. I will get something working[17:41:14] <aslak> kkhan, in your case yes. since the deserialization is incontianer.[17:41:20] <kkhan> baileyje / bstansberry / dmlloyd : If you want to do that as part of the smoke tests as well you could probably put some extra annotations on the test classes and modify the arquiilian container to deploy the ds stuff[17:41:39] <kkhan> I'd do it if I didn't have to leave very soon :-)[17:41:52] <dmlloyd> well my point is, if we're distributing a module for H2/derby/whatever, we should NOT also make it a deployment[17:41:56] <dmlloyd> that'll just screw things up[17:41:59] <baileyje> kkhan: I think it would be better to just have the provided DS drivers installed by the bootstramp.[17:42:26] <dmlloyd> if we have a subsystem element for it as well then you can define our bundled drivers in domain.xml and deploy user drivers as deployments.[17:42:52] <baileyje> dmlloyd: That is my plan..[17:43:13] <dmlloyd> baileyje: the deployment part is already done, so you'd just have to do the subsystem part[17:43:19] <dmlloyd> should be pretty easy though[17:43:21] <baileyje> RIght.[17:43:31] <baileyje> I have to run for an hour, but I will still have time to get it done[17:43:38] <dmlloyd> okay.[17:43:38] <baileyje> before the tag.[17:43:44] <bstansberry> baileyje: make sure any extension includes a "describe" operation handler[17:43:53] <bstansberry> otherwise it won't work on the domain[17:44:08] <dmlloyd> and we can then make that part of our default configuration, which will make demos work a bit better[17:44:10] <baileyje> ok. I will figure out what that means :)[17:44:17] <baileyje> dmlloyd: For sure[17:44:22] <bstansberry> there's one for every subsystem[17:44:41] <baileyje> bstansberry: Yeah. I see kkhan has been kncking them[17:44:42] <baileyje> out[17:44:44] <kkhan> baileyje: ok, if it is a module/subsystem that is better[17:45:07] <bstansberry> baileyje: no, what kabir has been doing is not what I mean[17:45:24] <baileyje> ok. I will track it down either way[17:45:35] <bstansberry> there is an operation called "describe" that every subsystem has[17:45:58] <kkhan> bstansberry: org.jboss.as.domain.controller.operation.ProfileDescribeHandler handles describe at domain level and forwards to each subsystem's describe handler[17:46:08] <baileyje> Ok. I will see what the other subsystems are doing and follow that.[17:46:11] <bstansberry> yeah[17:46:22] <kkhan> e.g. LoggingDescribeHandler[17:46:54] <kkhan> bstansberry: Anything tiny I could look at before going?[17:47:50] <bstansberry> you can just play a bit, see what breaks[17:47:59] <kkhan> ok[17:48:03] <bstansberry> although TBH I'm not sure I want to hear about breakage!!![17:48:19] *** Jaikiran has quit IRC[17:48:21] <kkhan> I'm not sure I want that to be the last thing on my mind before going either :-)[17:48:50] <baileyje> ok. back in a bit[17:50:15] *** frainone has joined #jboss-as7[17:50:15] *** ChanServ sets mode: +v frainone[17:50:41] *** imebcc2000 has joined #jboss-as7[17:51:08] *** imebcc2000 has quit IRC[17:51:38] *** smarlow has quit IRC[17:51:45] *** marcus has joined #jboss-as7[17:54:12] *** mmoyses has quit IRC[17:54:40] *** marcus has quit IRC[17:54:50] *** mmoyses has joined #jboss-as7[17:54:50] *** ChanServ sets mode: +v mmoyses[17:58:51] *** jfd_away is now known as jfd[17:59:13] <kkhan> ok, I'm off - good luck with the tagging and see you in 2 weeks[18:00:01] <dmlloyd> have a pleasant holiday, kkhan[18:00:07] <kkhan> Thanks[18:00:13] <alesj> kkhan: have fun ;-)[18:01:53] *** kkhan has quit IRC[18:05:23] *** frainone has quit IRC[18:07:08] <alesj> dmlloyd: i see CL magic is gonna be revealed again at JUDCon :-)[18:07:11] *** torben has quit IRC[18:07:26] <dmlloyd> yeah I'm going to try to update the class loading talk from previous ASes[18:07:51] <dmlloyd> and explain modules in simple terms[18:10:46] *** jcosta has quit IRC[18:16:37] *** pilhuhn is now known as pil-afk-dinner[18:17:54] *** pil has joined #jboss-as7[18:17:54] *** pil has joined #jboss-as7[18:17:54] *** ChanServ sets mode: +v pil[18:17:54] *** pil-afk-dinner has quit IRC[18:25:13] *** aloubyansky has quit IRC[18:26:03] *** jhalliday has quit IRC[18:29:26] <jbossbot> git [jboss-as] push master e6bb99e.. Emanuel Muckenhuber basic messaging subsystem descriptions[18:29:26] <jbossbot> git [jboss-as] push master 8c42592.. Emanuel Muckenhuber add missing describe handlers[18:29:26] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/e5d9303...8c42592[18:30:17] *** mmoyses has quit IRC[18:36:22] *** jfclere has quit IRC[18:36:51] *** mmoyses has joined #jboss-as7[18:36:51] *** ChanServ sets mode: +v mmoyses[18:38:32] *** adietisheim has quit IRC[18:41:18] *** frainone has joined #jboss-as7[18:41:24] *** ChanServ sets mode: +v frainone[18:43:37] <darranl> hi bstansberry, do you think you would have some time to talk tomorrow?[18:43:56] <bstansberry> sure[18:44:23] <bstansberry> i might start a bit later than normal, but tomorrow should be low stress[18:44:28] * bstansberry knocks on wood[18:45:16] <darranl> bstansberry, ok, I have another task I can start on to begin with anyway so will ping when I see you - just need to go over how we can configure items that live in the HostController as at the moment we have nothing really configurable there[18:46:17] <bstansberry> good topic[18:46:39] <darranl> hence trying to get the debugger to work to see what was really happening in a HC ;-)[18:47:30] *** smcgowan has joined #jboss-as7[18:47:30] *** ChanServ sets mode: +v smcgowan[18:47:30] <bstansberry> i'll gladly move to intellij if it works there[18:47:38] <bstansberry> although that would be odd[18:48:03] <darranl> no I have fallen back to jdb , that works without any problems[18:49:38] <darranl> I think the IDEs when debugging pro-actively retrieve a lot more over the connection and it is something they are retrieving that is causing the crash[18:51:23] *** stansilvert has joined #jboss-as7[19:02:42] *** stansilvert has quit IRC[19:04:12] *** darranl has quit IRC[19:12:04] <dmlloyd> can someone look over http://github.com/dmlloyd/jboss-as/commit/4661aa8 please[19:12:05] <jbossbot> git [jboss-as] 4661aa8.. David M. Lloyd Support signed deployments[19:12:46] <bstansberry> looks good[19:13:50] <jbossbot> git [jboss-as] push master b8d1238.. David M. Lloyd Support signed deployments[19:13:50] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/8c42592...b8d1238[19:14:11] <dmlloyd> was easier than I thought :)[19:19:19] *** balunasj_mtg has quit IRC[19:21:01] <dmlloyd> looks like JBAS-7882 still exists under AS 7[19:21:02] <jbossbot> jira [JBAS-7882] Wrong provider code base for security provider included in packed ear [Open (Unresolved) Bug, Major, John Bailey] https://issues.jboss.org/browse/JBAS-7882[19:21:08] <dmlloyd> oh forgot to rebuild[19:21:09] * dmlloyd blush[19:21:57] <dmlloyd> if it still exists then the problem is either that javax.crypto.Cipher uses the cert somehow to get the code base rather than the CL, or that VFS doesn't propagate the CL info correctly, or a bug in modules[19:22:05] <dmlloyd> if this fixes it then the problem is jboss-cl[19:22:16] <dmlloyd> (not using the code source from vfs)[19:22:44] *** ALR has joined #jboss-as7[19:22:44] *** ChanServ sets mode: +v ALR[19:23:04] <dmlloyd> I really am liking the FS deployer btw[19:23:07] <dmlloyd> it seems logical[19:23:09] <dmlloyd> makes sense[19:23:13] <dmlloyd> and there's a README.txt :)[19:25:25] <dmlloyd> yeah doesn't work[19:25:41] <dmlloyd> it won't authenticate a VFS code source URL for some reason[19:25:44] *** pmuir has joined #jboss-as7[19:25:44] *** pmuir has joined #jboss-as7[19:25:44] *** ChanServ sets mode: +v pmuir[19:26:21] <dmlloyd> it wants to be able to read the JAR itself[19:26:23] <dmlloyd> via a JAR URL[19:26:35] <dmlloyd> pretty serious problem[19:28:07] <dmlloyd> gonna try to look at the code in openJDK but I'm not confident I'll find anything[19:28:18] <dmlloyd> otherwise this JCE stuff is all sourceless[19:29:35] <baileyje> man. WHy does the DsParser have to be in ironjacamar..[19:35:45] <dmlloyd> hahaha[19:36:00] <dmlloyd> the parser for the domain config?[19:36:11] <dmlloyd> that's got no business being in ironjacamar[19:36:12] <dmlloyd> none at all[19:36:45] <baileyje> Yeah. It is there. In order to get it functional with additional DS subsystem elements, I have to pull it in..[19:36:59] <dmlloyd> so looks like this JCE problem doesn't appear in openJDK (which I have sources for), just sun JDK (which I do not)[19:37:00] <dmlloyd> :|[19:37:07] <dmlloyd> guess it's gonna have to wait for beta3 or so[19:37:24] <dmlloyd> baileyje: I say go for it, copy it in[19:37:34] <dmlloyd> it doesn't belong in ironjacamar[19:37:38] <jpederse> baileyje: the parser is in IJ because it is pure StAX[19:37:47] <jpederse> baileyje: all the domain stuff is in AS 7[19:38:01] <baileyje> jpederse: Do a degree you are correct[19:38:04] <dmlloyd> ...except for data source parsing[19:38:06] <jpederse> baileyje: and you shouldn't need to extend the XSD[19:38:37] <baileyje> The issue it is assuming it is getting full control of the subsystem.[19:39:03] <baileyje> Sine the Extension passes the full subsystem in, not just the datasources element[19:39:16] <jpederse> because it has - the model is read-only[19:39:25] <jpederse> the StAX model that is[19:39:29] <baileyje> So there is no ability to add anything else to the subystem[19:39:48] <jpederse> the XSD is not controlled in AS[19:39:55] *** alesj has quit IRC[19:40:03] <baileyje> If it instead allowed you to parse just the datasources element and below it would be ok, and the rest of the subsystem could be extended[19:40:07] <dmlloyd> the XSD for domain things is controlled by, and shipped with AS[19:40:37] <dmlloyd> for deployments there's more flexibility[19:40:42] <jpederse> baileyje: what do you need to add ?[19:41:06] <baileyje> jpederse: If there was a third method that said parseDataSources(reader) it would allow the datasource to remain controlled by the xsd, but allow other elements to go in the subsystem[19:41:23] <jpederse> baileyje: the subsystem implementation is in AS[19:41:25] <baileyje> jpederse: We are going to add support for getting jdbc drivers deployed[19:41:32] <baileyje> Right. Take a look at how it works through[19:41:35] <baileyje> though[19:41:53] <baileyje> It finds the SUBSYSTEM element, and passes directly into the DsParser[19:42:08] <jpederse> baileyje: the <module> element allows you to specify the JDBC driver[19:42:22] <jpederse> baileyje: it is up to the deployer chain to link them together[19:42:23] <baileyje> That isn't going be enough.[19:43:47] <jpederse> baileyje: why[19:44:30] <baileyje> one second..[20:04:31] <baileyje> jpederse: We need to add and additional element to the datasources subsysyetem. The datasources element can remain the same[20:06:24] <jpederse> baileyje: sure, that part should be done in AS[20:06:45] <baileyje> Right. What I am trying to say is the way the DsParser is written it can not be done[20:07:29] <baileyje> Since it is assuming it is being called with "subsystem" currently element.[20:07:52] <jpederse> baileyje: because there is a one-to-one mapping atm between the datasources.xsd and the subsystem[20:07:58] <dmlloyd> the domain parser has to be in AS[20:08:05] <dmlloyd> we just can't do it another way[20:08:10] <jpederse> dmlloyd: it is[20:08:50] <baileyje> http://pastebin.com/cng9XcsW[20:08:51] <stuartdouglas> I have some stuff in my branch that should be ready to be merged[20:08:58] <jpederse> baileyje: so an XSD is needed in AS for the stuff you are thinking about[20:08:58] <dmlloyd> stuartdouglas: okay[20:09:16] <stuartdouglas> with it the weld tck is above 90%[20:09:20] <baileyje> basically a peer element to the datasources element[20:09:20] <dmlloyd> nice![20:09:31] <jpederse> baileyje: correct, that is the one-to-one mapping[20:09:56] <jpederse> baileyje: what is the use-case ?[20:09:56] <baileyje> which is the one-to-one, the datasources element?[20:10:07] <dmlloyd> the use case is, I said so[20:10:14] <dmlloyd> we need jdbc-driver as an element[20:10:16] <baileyje> To allow drivers to be installed from a module.[20:10:19] <dmlloyd> DS subsystem is the most logical place for that[20:10:27] <dmlloyd> there is zero reason for this to be in IJ, nor for it to be an IJ schema[20:10:37] <jpederse> baileyje: since it'll be "case: AS_DATASOURCE_SUBSYSTEM_1_0"[20:10:37] <dmlloyd> it should be a domain schema, and the parser wholly contained in AS[20:11:17] <jpederse> dmlloyd: IJ is the subsystem atm[20:12:03] <jpederse> baileyje: yes, which was agreed with Nihility that we should allow a classloading spi for the resource adapter[20:12:35] <jpederse> baileyje: we don't need a new XSD for that[20:12:49] <jpederse> baileyje: that is an implementation detail in the deployer chain[20:12:56] <dmlloyd> if IJ is the subsystem then why isn't in the AS code?[20:12:59] <baileyje> THis is unrelated[20:13:16] <jpederse> dmlloyd: all the AS stuff is in AS[20:13:26] <baileyje> this is allow JDBC drivers to be installed at server boot, not as user deployments[20:14:15] <dmlloyd> ...except the DS parser[20:15:16] <jpederse> baileyje: that is an implementation detail too - dmlloyd wrote that service[20:15:32] <jpederse> baileyje: what is needed to link those with their deployments[20:15:46] <jpederse> baileyje: using JBoss Modules sematics[20:16:37] <baileyje> jpederse: I can get done what I need. I will circle back when complete.[20:16:40] <jpederse> dmlloyd: correct, since we havn't found an element that is specific to an AS deployment[20:17:32] <jpederse> baileyje: ok - maeste is working on the classloading spi, so we can hook that up in the deployer chain at the next drop[20:34:07] <stuartdouglas> I just pushed weld/jpa integration into my branch as well[20:34:34] <stuartdouglas> but now I am going to the gym, I will be back in about an hour and a half[20:36:20] *** mbg is now known as mbg|away[20:38:54] *** nickboldt has joined #jboss-as7[20:38:54] *** ChanServ sets mode: +v nickboldt[20:40:40] <dmlloyd> okay I'll check it out shortly[20:53:42] *** manstis has joined #jboss-as7[20:54:28] *** asaldhan has quit IRC[21:00:52] *** emmanuel has quit IRC[21:03:01] *** mbg|away is now known as mbg[21:06:18] *** smcgowan has quit IRC[21:09:44] *** nickboldt has quit IRC[21:10:16] *** asaldhan has joined #jboss-as7[21:10:31] *** asaldhan has quit IRC[21:10:32] *** asaldhan has joined #jboss-as7[21:31:31] <dmlloyd> stuartdouglas: what is the purpose of the viewServices map?[21:31:46] <dmlloyd> the view service name can always be determined from the component name and view class name[21:31:50] <dmlloyd> unless you've changed that[21:35:14] *** pil has quit IRC[21:38:32] <dmlloyd> also adding a session ID concept to component doesn't make a lot of sense to me[21:38:54] * dmlloyd wonders if he's reading the wrong branch[21:42:06] *** rmaucher has quit IRC[21:42:09] *** slaboure has quit IRC[21:42:35] *** fnasser has quit IRC[21:45:48] *** maxandersen has joined #jboss-as7[21:45:48] *** ChanServ sets mode: +v maxandersen[21:47:26] <maxandersen> bstansberry: is there actually a way to deploy via the CLI/http api today or does it still require writing a java client using the native api to do ?[21:48:27] <bstansberry> a contributor added a way to upload content via a form post[21:48:32] <dmlloyd> I believe the CLI uses the native API[21:48:37] <bstansberry> but you'd be better off using a java client[21:49:12] <bstansberry> actually you can use the CLI if you can create a URL that the server can access[21:50:08] <bstansberry> there's an op for that, but... I can explain more tomorrow after our deadline ;)[21:53:18] <maxandersen> bstansberry: okey :)[21:53:49] <maxandersen> dmlloyd: yes, but i still haven't found a way to deploy anything with it ;)[21:54:17] <baileyje> bstansberry: I am getting an error adding a model op... "because parent resource does not have child "[21:54:41] <baileyje> I have update the parents DescriptionProvider to list the child.. What else am I missing?[21:54:59] <bstansberry> that means whatever the thing you are adding too needs to have an undefined child of the type you are adding[21:55:25] <bstansberry> so if the address is /foo=a/bar=b and you are adding bar=b[21:55:46] <bstansberry> the foo node needs to have foo.get("bar")[21:55:59] *** mmoyses has quit IRC[21:56:13] <bstansberry> foo.get("bar") should have been done when foo was added[21:58:13] <baileyje> bstansberry: http://pastebin.com/VNbPpebn[21:58:25] <baileyje> I can see the Threads subsystem doing the same thing.[21:58:36] <baileyje> but I may be screqing it up still.[21:58:40] <baileyje> screwing[21:58:44] <bstansberry> it's not the operation[21:58:54] <bstansberry> it's the SubsystemAdd for the DS subsystem[21:59:08] <bstansberry> i assume you are adding a child to that subystem>[21:59:09] <bstansberry> ?[22:00:12] <baileyje> I am attempting to create a child address[22:00:25] <baileyje> I think I am messing it up. Let me just dig a bit more[22:00:56] *** irooskov has joined #jboss-as7[22:02:23] <bstansberry> baileyje: http://pastebin.com/HQMdjxnT[22:02:42] <bstansberry> in the DataSourcesSubsystemAdd[22:03:22] *** pferraro has quit IRC[22:05:43] *** nickboldt has joined #jboss-as7[22:05:43] *** ChanServ sets mode: +v nickboldt[22:05:45] <stuartdouglas> dmlloyd: There is no way of getting the view classes from Component (only ComponentDescription) atm, and having the service names availible in a map seemed better that building them in lots of different places[22:06:19] *** fnasser has joined #jboss-as7[22:06:32] <dmlloyd> ah so you're trying to enumerate them basically[22:06:36] <dmlloyd> at runtime?[22:08:21] <stuartdouglas> yea, although to be honest I could just build the view name and hope that it is there[22:10:57] <stuartdouglas> building services names all over the place kinda scares me a bit though, it seems like it is not very refactor friendly[22:11:16] <dmlloyd> we should have a single spot for building each "type" of service name[22:11:31] <stuartdouglas> yea[22:18:22] *** epbernard has joined #jboss-as7[22:18:22] *** epbernard is now known as emmanuel[22:18:22] *** ChanServ sets mode: +v emmanuel[22:19:04] *** emmanuel has quit IRC[22:19:14] <dmlloyd> I only am concerned because Component has become rather heavyweight[22:19:43] <stuartdouglas> I will change it to use view service name builders[22:20:31] *** jfd has quit IRC[22:20:41] *** jpederse has quit IRC[22:21:55] *** manstis has quit IRC[22:26:15] <dmlloyd> the main thing I'm concerned about is the session ID thing[22:28:01] <stuartdouglas> yea, that is going to have to change, but at the moment I can't really see any other way to get different views for the same SFSB[22:28:49] <dmlloyd> where is this being used from?[22:29:12] <dmlloyd> I assume it's related to weld integration?[22:29:24] * dmlloyd is looking at the change via https://github.com/stuartwdouglas/jboss-as/compare/master[22:30:40] <stuartdouglas> yea, in SessionObjectReferenceImpl, basically if you have a session scope SFSB with two different views, you need to be able to get at the same bean instance for both of those views[22:30:57] <stuartdouglas> or any other normal scoped sfsb[22:31:48] <dmlloyd> TODO: This is a massive hack[22:31:51] <dmlloyd> :)[22:32:38] *** emmanuel has joined #jboss-as7[22:32:38] *** ChanServ sets mode: +v emmanuel[22:32:41] *** nickboldt has quit IRC[22:34:43] <stuartdouglas> yea :-) This is all just 'get it done for the release' type stuff, so we have ejb's at least partly integrated with CDI[22:36:44] <dmlloyd> okay as long as I have your promise! :)[22:39:32] <stuartdouglas> It's definitely going to change in the next few weeks :-)[22:40:14] <dmlloyd> still need a release code name, for some stupid reason[22:42:33] <dmlloyd> could use "Floccinaucinihilipilification"[22:43:55] <baileyje> dmlloyd: I have it working. Just need to fix this parsing test in smoke tests. I must not be writing it correctly.[22:44:06] *** balunasj has joined #jboss-as7[22:44:06] *** balunasj has joined #jboss-as7[22:44:06] *** ChanServ sets mode: +v balunasj[22:44:12] <baileyje> I still need to javadoc, but that can wait till after the tag[22:44:27] *** balunasj is now known as balunasj_away[22:44:29] <dmlloyd> okay, baileyje - I've got a big commit from stuart, so you might want to rebase on that in a second[22:44:34] <baileyje> Also, I really need to fix the way DS are being deployed. I will fix that after tag as well.[22:45:04] <stuartdouglas> I fixed the integration tests as well, the main problem was a transitive arquillian-jsfunit dep[22:45:07] <baileyje> sorry for the bad time estimate. I didn't really know what all went into the description pieces.[22:45:12] *** pgier has joined #jboss-as7[22:45:12] *** ChanServ sets mode: +v pgier[22:45:22] *** pgier has left #jboss-as7[22:45:39] <jbossbot> git [jboss-as] push master 4b8a7e6.. Stuart Douglas Class-Path issues[22:45:39] <jbossbot> git [jboss-as] push master fdb2080.. Stuart Douglas fix npe[22:45:39] <jbossbot> git [jboss-as] push master e80efcd.. Stuart Douglas missing dep[22:45:39] <jbossbot> git [jboss-as] push master 05206c3.. Stuart Douglas resolve module name conflicts[22:45:39] <jbossbot> git [jboss-as] push master 4070d43.. Stuart Douglas Add basic EJBInjectionServices[22:45:40] <jbossbot> git [jboss-as] push master 07fb011.. Stuart Douglas inital weld-EJB integration hack[22:45:40] <jbossbot> git [jboss-as] push master 4f0de99.. Stuart Douglas Add basic support for multiple views for SFSB's[22:45:40] <jbossbot> git [jboss-as] push master 99e327e.. Stuart Douglas allow access to META-INF between deployment modules[22:45:41] <jbossbot> git [jboss-as] push master e85aaf4.. Stuart Douglas Composite index generation bug[22:45:41] <jbossbot> git [jboss-as] push master 2285727.. Stuart Douglas Change to a singleton provider that is aware of the deployment class loader structure[22:45:42] <jbossbot> git [jboss-as] push master 03905e6.. Stuart Douglas module deps[22:45:43] <jbossbot> git [jboss-as] push master fd56965.. Stuart Douglas Fix integration tests[22:45:43] <jbossbot> git [jboss-as] push master d586f7d.. Stuart Douglas Inital weld JPA integration[22:45:43] <jbossbot> git [jboss-as] push master 7a7ba43.. Stuart Douglas Bug fix with component views[22:45:44] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/b8d1238...7a7ba43[22:46:12] <dmlloyd> there it is[22:46:15] *** nickboldt1 has joined #jboss-as7[22:49:47] <wolfc> dmlloyd, stuartdouglas, there are two points where we need to have a view on a specific SFSB instance, Weld's weird home-interface and SessionContext.getBusinessObject(Class<?> viewClass);[22:50:26] <wolfc> I'm more worried about https://github.com/stuartwdouglas/jboss-as/commit/4f0de996aa8deac4ba5a64865c40ec5d3cbdee9a#L7R56 though[22:50:27] <jbossbot> git [jboss-as] 4f0de99.. Stuart Douglas Add basic support for multiple views for SFSB's[22:50:50] <wolfc> It's a doubled checked lock[22:52:19] <stuartdouglas> what wrong with double checked locking? It's fixed in java 5+[22:52:57] *** alesj has joined #jboss-as7[22:53:14] <stuartdouglas> as long as the variable is volatile, or the object in question is immutable[22:53:29] <dmlloyd> it's not a "double checked lock" in the dangerous sense if the variable is volatile[22:53:39] <dmlloyd> then it's just unlocked reads[22:53:46] <dmlloyd> aka lock-on-write[22:53:48] <wolfc> ah okay[22:54:44] <dmlloyd> I think as far as getting an instance from a proxy we should use information attached to the proxy itself or its invocation handler[22:54:55] <dmlloyd> for instances anyway...[22:55:00] *** ALR has quit IRC[22:55:21] <dmlloyd> it may be that we can't avoid having a map from class to... something - that'd be unfortunate though[22:55:25] <wolfc> where would we want to get instances from a proxy?[22:55:30] <dmlloyd> but not the end of the world[22:55:38] <dmlloyd> dunno, that's what I figured the weld stuff is doing[22:56:13] <wolfc> no, they work the other way around. SessionObjectReference is basically the instance itself. They want proxies on top of it.[22:56:30] <stuartdouglas> yea[22:57:16] <wolfc> It's the same as SessionContext.getBusinessObject(Class<?> businessInterface); but then with some added features for explicit lifecycle management.[22:57:32] <stuartdouglas> I'm going to be offline for a bit soon, I'm going up to the office[22:57:55] <wolfc> I'm off to bed. I'll implement getBO tomorrow using your stuff.[22:58:56] <wolfc> the SessionContext one that is. :-)[23:00:08] <dmlloyd> okay I'm going to tag today for sure[23:00:12] <dmlloyd> probably once baileyje's stuff is in[23:00:30] <wolfc> Woohoo, tag it and bag it[23:00:59] * stuartdouglas runs to test the weld JPA integration[23:01:17] *** wolfc has quit IRC[23:06:45] *** miclorb_ has joined #jboss-as7[23:08:31] *** pmuir has quit IRC[23:09:03] <baileyje> dmlloyd: doing final run after the rebase.[23:11:27] <stuartdouglas> gah, jpa fails[23:12:04] <dmlloyd> hm I didn't retest JPA[23:12:11] <dmlloyd> just saw that the smoke tests pass[23:12:49] <stuartdouglas> no, with weld[23:13:12] <stuartdouglas> it can't find a Component for the CDI bean with @PersistenceContext, so it errors out[23:13:15] *** pmuir has joined #jboss-as7[23:13:15] *** pmuir has joined #jboss-as7[23:13:15] *** ChanServ sets mode: +v pmuir[23:14:39] <stuartdouglas> to me the logic is round the wrong way, it should look through components for @PersistenceContext, not look for @PersistenceContext and then try and find a component[23:15:39] <dmlloyd> seems reasonable[23:15:44] <dmlloyd> I think that's how we do other things[23:16:05] <stuartdouglas> how long till you do the release?[23:16:10] <dmlloyd> depends[23:16:15] <dmlloyd> how long until you fix whatever you're fixing :)[23:16:24] <dmlloyd> I need baileyje's change anyway[23:16:56] <dmlloyd> gonna start dinner soon but it takes a while to boil 2 gallons of water so there's time[23:17:06] <baileyje> dmlloyd: I thought I was done. Got a couple sporadic test fails[23:17:09] <dmlloyd> and I need a *&#*( release code name[23:17:34] <baileyje> blurg[23:18:59] <baileyje> I rebased and now org.jboss.as.test.embedded.demos.ejb3.StatelessBeanTestCase is hung.[23:19:58] <dmlloyd> hm I"ll rerun all my tests to be sure[23:20:51] <frainone> dmlloyd: what about WebServiceRef?[23:21:04] <dmlloyd> frainone: I think I merged it already[23:21:07] <dmlloyd> your branch[23:21:19] <dmlloyd> unless you have more changes[23:21:45] <frainone> yeah, I did some small changes[23:25:33] <stuartdouglas> I have jpa weld int fixed in my trunk[23:25:37] * frainone is rebasing to provide an updated commit link[23:25:58] <stuartdouglas> I will re-write the processor properly on the train, it does not matter for the release[23:26:24] <stuartdouglas> but the demo passed with a CDI injected EM[23:27:28] <baileyje> dmlloyd: https://github.com/baileyje/jboss-as/commit/b27d7237935dc53f877aa24f4d2e109882523a2a[23:27:30] <jbossbot> git [jboss-as] b27d723.. John E. Bailey Allow JDBC drivers to be installed based on domain model.[23:28:09] <stuartdouglas> now I have to run to make the train[23:28:26] <baileyje> dmlloyd: i Have to take the dog for a walk, be on in a little bit[23:28:35] <bstansberry> dmlloyd: I am about to push in my stuff[23:28:51] <bstansberry> i have to bail for a couple hours :([23:29:29] <frainone> dmlloyd: I rebased, but I can wait for all this to be pushed in, so I can rebase again, run tests and give you a new commit link[23:30:42] *** pmuir has quit IRC[23:31:50] <dmlloyd> okay so I'll do baileyje's, stuartdouglas, then frainone's.[23:31:57] <dmlloyd> I'll rebase stuartdouglas' work.[23:32:12] <dmlloyd> let me just get dinner started...[23:39:04] *** jpearlin has joined #jboss-as7[23:39:18] *** stuartdouglas_ has joined #jboss-as7[23:40:15] <jbossbot> git [jboss-as] push master 026ca62.. bstansberry at jboss dot com Initial work on converting domain.interactive demo to detyped API[23:40:15] <jbossbot> git [jboss-as] push master 719f41b.. bstansberry at jboss dot com Remove multi-deployment-set capability from domain deployment plan API[23:40:15] <jbossbot> git [jboss-as] push master 7e2457d.. bstansberry at jboss dot com DomainDeploymentManager executes the deployment plan[23:40:15] <jbossbot> git [jboss-as] push master 50c9c33.. Emanuel Muckenhuber addJMSQueue operation[23:40:15] <jbossbot> git [jboss-as] push master 88ad5e5.. Emanuel Muckenhuber fix connection-factory describe action[23:40:15] <jbossbot> git [jboss-as] push master a7704d7.. bstansberry at jboss dot com Fix wrong method; add missing class[23:40:15] <jbossbot> git [jboss-as] push master d9ae20a.. bstansberry at jboss dot com Use correct address for server lifecycle ops[23:40:16] <jbossbot> git [jboss-as] push master d1f8840.. bstansberry at jboss dot com Add missing file[23:40:16] <jbossbot> git [jboss-as] push master b4089ff.. bstansberry at jboss dot com Update results use ModelNode, not <R>[23:40:17] <jbossbot> git [jboss-as] push master 2d4d6ac.. Emanuel Muckenhuber demo addServer()[23:40:17] <jbossbot> git [jboss-as] push master ee37449.. bstansberry at jboss dot com Turn on deployment plan[23:40:18] <jbossbot> git [jboss-as] push master 1ecc882.. Emanuel Muckenhuber use the hostModel on the DC to start servers[23:40:18] <jbossbot> git [jboss-as] push master 8f9c4c0.. bstansberry at jboss dot com ServerRemove should remove, not add[23:40:19] <jbossbot> git [jboss-as] push master 9c56af3.. Emanuel Muckenhuber remove system.out[23:40:19] <jbossbot> git [jboss-as] push master b84c166.. bstansberry at jboss dot com Disable domain deployment part of demo[23:40:20] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/7a7ba43...b84c166[23:40:44] *** bstansberry is now known as bstans_afk[23:42:37] *** stuartdouglas_ has quit IRC[23:49:04] *** maxandersen has quit IRC[23:50:57] <dmlloyd> JPA demo seems...(?)...to work without having to run the special instructions[23:51:06] <dmlloyd> I guess that's the whole point, so... one down, two to go[23:51:19] <jbossbot> git [jboss-as] push master d9ada2a.. John E. Bailey Allow JDBC drivers to be installed based on domain model.[23:51:19] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/b84c166...d9ada2a[23:51:38] *** maxandersen has joined #jboss-as7[23:51:38] *** ChanServ sets mode: +v maxandersen[23:52:50] <dmlloyd> frainone: your stuff is unlikely to conflict with this other stuff so I can do the rebase and a simple test run for it[23:52:56] <dmlloyd> then I'll just run thru all the demos[23:53:21] <dmlloyd> there is a POM validation warning about one of the WS deps though, it seems to be listed twice in the parent pom.xml[23:53:23] <frainone> dmlloyd: I rebased anyway[23:53:32] <frainone> oh, really?[23:56:09] <dmlloyd> yeah you can see it if you build with maven 3.0.3[23:56:41] <dmlloyd> [WARNING] 'dependencyManagement.dependencies.dependency.(groupId:artifactId:type:classifier)' must be unique: org.apache.cxf:cxf-rt-ws-security:jar -> duplicate declaration of version ${version.org.apache.cxf} @ org.jboss.as:jboss-as-parent:7.0.0.Alpha2-SNAPSHOT, /home/david/src/java/as7/jboss-as/pom.xml, line 1084, column 25[23:57:35] <frainone> dmlloyd: I can't build the server, I'm getting a failure at EarDeploymentEjbJndiBindingTestCase[23:57:59] <frainone> Caused by: java.lang.RuntimeException: java.lang.NullPointerException[23:57:59] <frainone> at org.jboss.arquillian.context.ContextManager.setup(ContextManager.java:74)[23:59:01] <jbossbot> git [jboss-as] push master ea971e1.. Stuart Douglas Change JPAAnnotationParseProcessor to allow for CDI @PersistenceContext injection[23:59:01] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/d9ada2a...ea971e1[23:59:08] <dmlloyd> hmm, interesting[23:59:12] <dmlloyd> I get no failure locall[23:59:13] <dmlloyd> y[23:59:29] <dmlloyd> try a rebase, clean build, etc?[23:59:48] <frainone> yeah, exactly, that's what I tried