March 12, 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 12, 2011 NOTICE: This channel is no longer actively logged.
[00:01:31] <jbossbot> git [jboss-as] push master 531880e.. bstansberry at jboss dot com Log OperationFailedException at debug[00:01:31] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/f1046ca...531880e[00:06:35] <wolfc> anything you guys need before I go to bed?[00:07:06] * dmlloyd can't think of anything[00:07:46] <wolfc> In that case I bid you all a good night[00:07:48] *** mbg has joined #jboss-as7[00:07:48] *** ChanServ sets mode: +v mbg[00:08:11] *** wolfc has quit IRC[00:13:19] *** aslak has quit IRC[00:14:14] *** emmanuel has joined #jboss-as7[00:14:14] *** ChanServ sets mode: +v emmanuel[00:15:41] *** jpederse has quit IRC[00:21:33] <bstansberry> geez; the FS deployment forum thread has 27,000 views[00:23:26] <dmlloyd> and yet so little factual information :)[00:23:39] <emuckenhuber> i'm sure max was checking this post every minute![00:24:34] *** smarlow has quit IRC[00:35:58] *** ALR has quit IRC[00:42:07] *** pferraro has quit IRC[00:42:58] *** mbg is now known as mbg|away[00:47:35] *** mbg|away is now known as mbg[00:48:51] *** jpearlin has joined #jboss-as7[00:53:27] *** jwulf has joined #jboss-as7[00:59:23] *** jpearlin has quit IRC[00:59:29] *** jpearlin has joined #jboss-as7[01:11:49] *** rawbdor has quit IRC[01:15:25] *** smcgowan has quit IRC[01:19:32] *** emmanuel has quit IRC[01:23:07] *** rmaucher has quit IRC[01:24:05] *** smcgowan has joined #jboss-as7[01:24:05] *** ChanServ sets mode: +v smcgowan[01:28:39] *** frainone has joined #jboss-as7[01:28:39] *** ChanServ sets mode: +v frainone[01:28:56] *** aamonten has joined #jboss-as7[01:33:57] *** rawbdor has joined #jboss-as7[01:38:22] *** rawbdor has quit IRC[01:53:40] *** adietisheim has joined #jboss-as7[01:53:59] *** adietisheim has quit IRC[01:54:48] *** rawbdor has joined #jboss-as7[01:59:04] <stuartdouglas> dmlloyd: Can you merge my master? It should fix the extension-list stuff and the WebServiceContext injection[01:59:33] <dmlloyd> ok will grab it[01:59:47] <dmlloyd> thanks for taking time to do that[02:00:06] *** asaldhan has left #jboss-as7[02:00:22] <stuartdouglas> np, sorry I could not get to it earlier, I did not have much time at NHO[02:04:44] <jbossbot> git [jboss-as] push master a092685.. Stuart Douglas Add support for WebServiceContext @Resource injection[02:04:44] <jbossbot> git [jboss-as] push master 6ce608c.. Stuart Douglas Fix extension-list war bug[02:04:44] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/531880e...6ce608c[02:06:02] <stuartdouglas> I also noticed that the web services module is installing DUP's that are not listed in Phase[02:06:25] <stuartdouglas> instead they are doing a priority++ type thing to register them, not sure if this is a problem[02:06:57] <dmlloyd> ugh[02:07:05] <dmlloyd> I must have missed that on reviews[02:07:10] <dmlloyd> yeah it's definitely a problem[02:09:02] <stuartdouglas> thought so[02:11:35] *** mbg is now known as mbg|away[02:20:29] *** jamezp has left #jboss-as7[02:41:33] <smcgowan> stuartdouglas: with your commit: http://pastebin.test.redhat.com/43782 hmm[02:41:58] *** fnasser has quit IRC[02:44:06] <stuartdouglas> hmm, that is another issue[02:44:37] <stuartdouglas> basically, what classes to we make availible to extension-list items[02:46:56] <stuartdouglas> we probably should make all the Java EE api's available[02:47:08] <stuartdouglas> and possibly all as7 classes as well[02:52:56] <stuartdouglas> smcgowan: can you try with my master?[02:54:28] <smcgowan> stuartdouglas: this fix has improved the ability to deploy the test archives is other cases so does seem like a different issue[02:55:03] <smcgowan> stuartdouglas: you have another fix to try in your master?[02:55:08] <stuartdouglas> yes[02:55:50] <stuartdouglas> the issue is that the extension list items are loaded as seperate modules, so they can be shared between modules, without having multiple definitions of the same class[02:56:05] <stuartdouglas> but at the moment we don't have any real way of configuring dependencies for them[02:56:44] <stuartdouglas> manifest based dependencies are not ideal, as in many cases this would involve opening up a third party jar to modify the manifest[02:57:19] *** miclorb has joined #jboss-as7[02:57:47] <smcgowan> ya - i'll can try your master - probably in the a.m. though, i was just about to sign off to take a break[02:59:22] <stuartdouglas> ok, I will hopefully have the weld tck mostly passing again by then[03:00:05] <smcgowan> and I saw your message about that, i'll set that up as well[03:10:29] <smcgowan> stuartdouglas: i'll send you an email after i test with your master - g' night ...[03:10:53] *** smcgowan has quit IRC[03:17:50] *** rawbdor has quit IRC[03:21:12] *** smarlow has joined #jboss-as7[03:21:13] *** ChanServ sets mode: +v smarlow[03:25:22] *** pferraro has joined #jboss-as7[03:25:22] *** ChanServ sets mode: +v pferraro[03:30:16] *** rawbdor has joined #jboss-as7[03:47:33] <Nihility> man i am at a loss for how i can fix the jca jndi to use the propper classloader[03:48:03] <Nihility> hmm[03:48:53] <Nihility> oh no im making progress[03:56:52] *** aamonten has quit IRC[04:03:06] *** aamonten has joined #jboss-as7[04:11:40] *** miclorb has quit IRC[04:12:58] *** miclorb has joined #jboss-as7[04:16:20] *** jpearlin has left #jboss-as7[04:27:47] <Nihility> damn[04:27:54] <Nihility> im going to have to patch jca to make this work[04:34:10] <Nihility> i guess i could stuff all referenced jdbc drivers in one classloader[04:34:15] <Nihility> which is really damn gross[04:34:31] <Nihility> we MUST get off this jdbc-local.rar design[04:36:39] *** miclorb has quit IRC[04:39:17] *** rawbdor has quit IRC[04:41:31] <jbossbot> git [jboss-as] push master b6613b1.. John E. Bailey File System deployment API updates. Initial work at marker file driven API.[04:41:31] <jbossbot> git [jboss-as] push master 6c956e7.. bstansberry at jboss dot com Tests of FileSystemDeploymentService[04:41:32] <jbossbot> git [jboss-as] push master cddff1a.. bstansberry at jboss dot com Add in-progress markers[04:41:32] <jbossbot> git [jboss-as] push master 81d2fa8.. bstansberry at jboss dot com Get FS deployment scanning smoke test working[04:41:32] <jbossbot> git [jboss-as] push master 904e52b.. Emanuel Muckenhuber exploded deployment handling[04:41:32] <jbossbot> git [jboss-as] push master b9032f7.. bstansberry at jboss dot com Get the test to compile[04:41:32] <jbossbot> git [jboss-as] push master 7121c2e.. bstansberry at jboss dot com Port rename of ExecutionContext to Operation[04:41:33] <jbossbot> git [jboss-as] push master a123b3a.. bstansberry at jboss dot com Add README...[04:41:33] <jbossbot> git [jboss-as] push master 4f74ac8.. bstansberry at jboss dot com Make FS deployment test more robust[04:41:34] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/6ce608c...4f74ac8[04:42:58] <smarlow> one classloader for all referenced jdbc drivers would be cleaner than one global classloader for all jdbc drivers[04:43:09] <Nihility> both are stupid[04:43:20] <Nihility> but yeah[04:43:27] <Nihility> actually we cant do the latter anyway[04:44:19] <Nihility> (you can have multiple versions of the same jdbc driver deployed)[04:46:28] <smarlow> I would want to have a classloader visualizer, that showed which versions were accessible in different deployments, otherwise you would be unsure if the right version was used in every place. Or maybe other developers are more trusting than me :)[04:47:10] <Nihility> whats annoying about stuff like this[04:47:20] <Nihility> is when you go through all the trouble to have a modular cl architecture[04:48:01] <Nihility> and then some framework decides that the world should be on one classloader[04:49:01] <Nihility> smarlow: the ds definitnio specifies which version of which driver to use[04:49:21] <Nihility> they get mapped to different jndi names so its pretty clear what the intention is[04:51:35] <smarlow> I still think Ales should write the cl visualizer :)[05:01:47] <Nihility> actrually i dont think we can even do that[05:01:54] <Nihility> its a chicken egg problem[05:02:01] <Nihility> the driver might not exist yet[05:02:38] <Nihility> so we cant create an aggregate module to represent it[05:02:54] <Nihility> im totally painted into a corner here[05:05:31] <smarlow> did we consider multiple jdbc driver groups that sit below the system classloader but above our module classloading[05:06:05] <smarlow> kind of a hack on the global/static jdbc lib folder[05:07:01] <smarlow> so at least there is some flexibility for different versions of the global jars via different groups.[05:09:37] <stuartdouglas> So what is/was the plan for loading jdbc drivers? I have not really been following it so I am not 100% sure what the issue is[05:18:31] *** pgier has quit IRC[05:20:07] <baileyje> Nihility: Yeah. You ran into what I did..[05:20:16] <baileyje> It needs a JCA patch.[05:21:04] <baileyje> stuartdouglas: The issue is the JDBC drivers are being installed as deployments.[05:21:29] <baileyje> and the other deployments that lookup datasources need accesss to the jdbc driver jars[05:21:51] <baileyje> and right now it is using the TCCL to lookup the driver[05:21:56] <stuartdouglas> how are we enforcing the deployment order?[05:22:11] <stuartdouglas> service dependencies?[05:22:15] <baileyje> The connection factory binding has a dep o the driver[05:22:22] <baileyje> so yes service deps[05:23:10] <baileyje> The best solution would be to have the connection factory be bound using a BinderService with the correct deps and have the CF just use the driver class that is already loaded by the driver service[05:24:19] *** rawbdor has joined #jboss-as7[05:28:47] *** rawbdor has quit IRC[05:40:36] *** frainone has quit IRC[05:53:59] *** rawbdor has joined #jboss-as7[05:59:22] <Nihility> yeah its a mess[06:01:11] *** rawbdor has quit IRC[06:01:54] *** smarlow has quit IRC[06:04:15] <baileyje> Nihility: I am half tempted to git rid of the RAR this weekend[06:05:59] *** rawbdor has joined #jboss-as7[06:07:03] <Nihility> http://anonsvn.jboss.org/repos/jbossas/projects/jboss-jca/trunk/adapters/src/main/java/org/jboss/jca/adapters/jdbc/local/LocalManagedConnectionFactory.java[06:07:07] <Nihility> see getDriver[06:07:22] <baileyje> yeah. I was looking at that earlier[06:07:44] <baileyje> You would need to plug in a custom ManagedConnectionFactory[06:07:54] <baileyje> to override it[06:08:36] <baileyje> I have an idea of how..[06:09:09] <Nihility> amazing that this code still exists[06:09:16] <baileyje> hahahahaha[06:09:30] <baileyje> all to create the CF and bind it in jndi[06:10:38] <Nihility> hey look you can serialize a CF[06:10:42] <Nihility> that looks like a good idea[06:13:33] <Nihility> note that there are many factories[06:14:52] <Nihility> well i guess there is two that would need changes[06:15:13] <Nihility> the problem is though[06:15:27] <Nihility> it doesnt matter if we fix it[06:15:30] <baileyje> the classes are all in the RAR[06:15:41] <Nihility> because the deps wont be correct[06:16:03] <baileyje> why not[06:16:55] <Nihility> the rar gets deployed before the driver does[06:17:27] <Nihility> i guess it wont bind the cf[06:17:35] <baileyje> It should add the DataSourceDeploymentService with a dep on the driver service[06:17:53] <baileyje> which it does. So the driver is required before the DataSourceDeploymentService is started.[06:18:01] <baileyje> so order shouldn't matter[06:18:40] <Nihility> nice so it waits for all drivers to be deployed?[06:19:32] <Nihility> oh no one per instance[06:19:35] <Nihility> thats progress[06:19:58] <baileyje> yeah. For sure[06:20:17] <baileyje> Say the word and we will have RARless DS deployment by monday. :)[06:20:35] <Nihility> haha would you just import the classes into AS[06:21:00] <baileyje> for the most part. Just bring the CF classes and bind them using a BinderService with the correct deps[06:21:12] <baileyje> and fix the CL in the MCF[06:21:31] *** aamonten has left #jboss-as7[06:21:34] <baileyje> It should be like 5 classes[06:24:48] <Nihility> the big problem is it still has to go through the jca layer though[06:25:15] <Nihility> e.g. AbstractDsDeployer[06:25:25] <Nihility> but you are thinking a special CF impl then[06:26:07] <Nihility> that would probably work[06:26:43] <baileyje> Yeah. I haven't fully speced it, but I am pretty sure it would be fairly easy[06:26:59] <baileyje> it all comes down to the CFs[06:27:03] <Nihility> yeah you are right i think the issue is lifecycle though[06:27:14] <Nihility> they are instantiated by the resource stuff iirc[06:27:36] <baileyje> I think the DS services should be added no matter what with a dep on the driver[06:27:38] <Nihility> but i guess you could have a map or something[06:27:52] <baileyje> right now all the DSs wait for the RAR to install[06:28:15] <baileyje> by wait I mean they are installed as part of the RAR deployment[06:28:31] <baileyje> I think they should be added as part of the DS subsystem start[06:28:42] <baileyje> with deps on the drivers.[06:28:55] <baileyje> If the drivers are not installed it will just have a service dep issue[06:29:06] <Nihility> oh look at this[06:29:16] <baileyje> once the driver is added it would start.[06:29:27] <Nihility> we do get to control initialization[06:30:40] <Nihility> uh[06:30:46] <Nihility> take a look at util.injection[06:30:52] <Nihility> Injection[06:31:34] <baileyje> yeah..,..[06:32:45] <Nihility> would be nicer if this was a service[06:32:56] <Nihility> instead of a javabean injection thing[06:33:08] <Nihility> but yeah we could stuff the driver in there[06:33:25] <baileyje> Anyway, do you mind if I hack at it this weekend?[06:33:31] <Nihility> not at all[06:33:49] <baileyje> I will start by just making it work with the driver, but maybe I will get rid of the RAR completely.[06:34:01] <Nihility> it looks like our jdbc driver approach needs to be updated to handle XA[06:34:25] <baileyje> The main issue with the RAR is there is no way to cleanly depend on it from a AS codebase.[06:34:44] <baileyje> So we can't just extend LocalManagedConnectionFactory, which would be the easier.[06:35:54] <baileyje> Ok. Well I am going to head to bed. You should stay up and fix the jdbc driver issue with XA :)[06:35:55] <Nihility> local-tx uses Driver[06:36:07] <Nihility> but XA needs to use DataSource[06:36:40] <Nihility> although it could do Driver.getClass().getClassLoader().loadClass(xa-datasource-name)[06:37:11] <Nihility> thats probably the only realistic way to do it[06:37:39] <Nihility> baileyje: so what was happening with JPA[06:38:05] <baileyje> well, it is somewhat dependent on this working.[06:38:18] <baileyje> I have the JPA PU starting, but it fails to get a DS connection[06:38:31] <Nihility> ah ok[06:38:43] <Nihility> yeah i thought you mentioned something[06:38:46] <baileyje> part of the PU starting requires a connection.[06:39:17] <baileyje> Once it starts, then I will continue to the @PU and @PC working, which they likely already do.[06:39:42] <baileyje> then there are the listeners carlo is doing.[06:39:45] <baileyje> Not sure about those..[06:40:10] <Nihility> ok well i will be online tommorow evening as well[06:40:20] <Nihility> maybe sometime during hte day[06:40:23] <baileyje> ok. I will see where I am by then..[06:40:38] <baileyje> I have to get o bed now, butI will be working tomorrow afternoon/night[06:40:45] <Nihility> ok cool talk to you later[06:40:49] <baileyje> later man[06:57:00] *** pferraro has quit IRC[07:08:39] *** Nihility has quit IRC[07:08:51] *** Nihility has joined #jboss-as7[07:08:51] *** Nihility has joined #jboss-as7[07:08:51] *** ChanServ sets mode: +v Nihility[08:17:13] *** mbg|away is now known as mbg[08:29:06] *** mbg has quit IRC[08:35:33] *** rawbdor has quit IRC[08:48:19] *** rawbdor has joined #jboss-as7[08:54:27] *** jfclere has joined #jboss-as7[09:01:41] *** miclorb_ has joined #jboss-as7[10:22:09] *** jfclere has quit IRC[10:22:40] *** rawbdor has quit IRC[10:27:33] *** maeste has quit IRC[10:31:18] *** dimitris_ has joined #jboss-as7[10:31:18] *** dimitris_ has joined #jboss-as7[10:31:18] *** ChanServ sets mode: +v dimitris_[10:32:27] *** alesj has joined #jboss-as7[10:35:20] *** rawbdor has joined #jboss-as7[10:47:35] *** maxandersen has joined #jboss-as7[10:47:36] *** ChanServ sets mode: +v maxandersen[11:29:19] *** kabir_afk is now known as kkhan[11:41:34] *** Darran_L has joined #jboss-as7[11:48:16] *** aslak has joined #jboss-as7[11:48:16] *** ChanServ sets mode: +v aslak[11:54:01] *** Darran_L has quit IRC[11:54:01] *** rawbdor has quit IRC[11:59:27] *** dimitris_ has quit IRC[12:01:44] *** alesj has quit IRC[12:04:02] *** jfclere has joined #jboss-as7[12:44:25] *** epbernard has joined #jboss-as7[12:44:25] *** epbernard is now known as emmanuel[12:44:25] *** ChanServ sets mode: +v emmanuel[12:46:29] *** emmanuel has quit IRC[12:53:34] *** jfclere has quit IRC[13:07:24] *** miclorb_ has quit IRC[13:37:09] *** kkhan has quit IRC[14:41:19] *** smcgowan has joined #jboss-as7[14:41:19] *** ChanServ sets mode: +v smcgowan[14:57:03] *** jfclere has joined #jboss-as7[15:12:56] *** jfclere has quit IRC[15:13:52] *** dimitris_ has joined #jboss-as7[15:13:52] *** dimitris_ has joined #jboss-as7[15:13:52] *** ChanServ sets mode: +v dimitris_[15:21:17] *** jwulf has quit IRC[15:27:44] *** aslak has quit IRC[15:55:24] *** dimitris_ has quit IRC[16:12:27] *** jpearlin has joined #jboss-as7[16:14:47] *** smarlow has joined #jboss-as7[16:14:48] *** ChanServ sets mode: +v smarlow[16:32:24] *** maxandersen has quit IRC[16:59:51] <smarlow> Nihility, baileyje: I committed some more fixes to my branch last night (some bug fixes and implemented some missing JPA code). I'll fix any other holes I find today in JPA and continue building it up. I'll fill in the transaction support next. Bailey, if there is a simple way for you to give me a patch or merge your changes with mine, I would like to work on top of your fixes :-)[17:00:45] <baileyje> smarlow: My changes are pretty minor. I would move forward without them for now. They may not be strictly necessary.[17:01:13] <smarlow> baileyje: sounds good, thanks again for your help! :)[17:01:19] <baileyje> np..[18:19:24] *** smarlow has quit IRC[18:19:45] *** smarlow has joined #jboss-as7[18:19:45] *** ChanServ sets mode: +v smarlow[18:35:58] *** smcgowan has quit IRC[18:39:24] *** smarlow has quit IRC[18:48:34] *** dimitris_ has joined #jboss-as7[18:48:34] *** dimitris_ has joined #jboss-as7[18:48:34] *** ChanServ sets mode: +v dimitris_[18:50:19] *** pferraro has joined #jboss-as7[18:50:19] *** ChanServ sets mode: +v pferraro[18:51:54] *** cs02rm0 has joined #jboss-as7[19:32:48] *** cs02rm0 has quit IRC[19:45:13] *** cs02rm0 has joined #jboss-as7[19:47:16] *** pferraro has quit IRC[19:47:41] *** pferraro has joined #jboss-as7[19:47:41] *** ChanServ sets mode: +v pferraro[20:02:33] *** dimitris_ has quit IRC[20:30:38] *** pferraro has quit IRC[20:53:47] *** cs02rm0 has quit IRC[20:54:12] *** dimitris_ has joined #jboss-as7[20:54:12] *** dimitris_ has joined #jboss-as7[20:54:12] *** ChanServ sets mode: +v dimitris_[20:54:42] *** alesj has joined #jboss-as7[21:11:17] *** pferraro has joined #jboss-as7[21:11:17] *** ChanServ sets mode: +v pferraro[21:14:06] <jpearlin> Nihility: I was able to merge all of the changes with my upload private branch...let me know if you need me to make any changes[21:32:45] <alesj> Nihility, dmlloyd: what's the Beta1 date?[21:56:38] *** pferraro has quit IRC[22:42:52] <Nihility> alesj: we are going to try tagging monday evening us hours[22:43:07] <bstansberry> he's gone[22:43:51] <Nihility> jpearlin: ill take a look at merging tonight thanks.[22:44:03] <Nihility> bstansberry: Ah[22:44:18] <bstansberry> oh, no he's not[22:44:22] <bstansberry> sorry, i'm blind[22:45:38] <Nihility> Im at an overloaded kids play area at the mall with parents that cant red the minimun heighth sign[22:47:16] <bstansberry> scary[22:47:23] <bstansberry> but good[22:47:34] <Nihility> baileyje: hows it going? Ill be coding at around 7 so ill provablt miss you. Leave a message if you want me to help with that change[22:48:03] <baileyje> I am just getting into it.[22:48:14] <baileyje> been busy. Will be on most of the night[22:48:25] <Nihility> cool[22:49:21] <Nihility> bstansberry: Yeah 13 year olds junping between play stuctures over infants[22:50:30] <Nihility> but madeline LOVEs it[22:50:34] <bstansberry> I remember those days; the fear toughened me up[22:50:39] <bstansberry> exactly[22:50:41] <jpearlin> Nihility: I am tracking down one issue with it...[22:50:52] <jpearlin> I hope to have it resolved tonight, but if not, I will definitely have it fixed tomorrow[22:51:11] <Nihility> whars the issue?[22:51:24] <jpearlin> looks like my boundary for the header is not right in all cases[22:51:32] <jpearlin> I am trying an upload with two hidden fields and the file[22:51:47] <jpearlin> and its choking, so I just have to figure out why I am not reading the body right[22:51:54] <Nihility> ah[22:52:05] <jpearlin> but as it stands now, it handles the simple case (just the file in the post)[22:52:14] <jpearlin> and removes the header from the .war written to disk[22:52:20] <jpearlin> or whatever the uploaded file is[22:54:23] <Nihility> ok when i merge ill see if i notice anything[22:54:30] <Nihility> bbl[22:54:40] <jpearlin> yeah...I'm not in love with how I am finding the body, but we can clean that up later I suppose[23:05:12] *** aslak has joined #jboss-as7[23:05:12] *** aslak has joined #jboss-as7[23:05:12] *** ChanServ sets mode: +v aslak[23:39:27] <alesj> Nihility: ah, ok, thought there was more time[23:43:08] <alesj> anyway, I'll impl something against the Beta1 tag[23:43:19] <alesj> and see how soon I can make something work[23:43:41] <alesj> e.g. reusing my MC grapher should speed this up[23:56:29] *** alesj has quit IRC