[00:11:35] *** kenfinnigan has quit IRC
[00:39:39] *** lightguard_jp is now known as lightguard_jp_aw
[00:49:20] *** michaelschuetz has quit IRC
[01:08:19] *** ldimaggi has joined #jbosstesting
[01:29:47] *** Tashtego has quit IRC
[01:57:30] *** aslak has quit IRC
[02:38:32] *** yendor has joined #jbosstesting
[02:48:20] *** yendor has quit IRC
[04:05:28] *** ldimaggi_ has joined #jbosstesting
[04:06:11] *** ldimaggi has quit IRC
[04:08:03] *** ldimaggi_ has quit IRC
[04:17:12] *** bleathem_ has joined #jbosstesting
[04:18:32] *** bleathem_ is now known as bleathem
[04:18:42] *** bleathem has quit IRC
[06:49:41] *** PeteRoyle has left #jbosstesting
[06:49:44] *** lightguard_jp_aw is now known as lightguard_jp
[07:06:53] *** pmuir has quit IRC
[07:09:01] *** mgoldmann has joined #jbosstesting
[07:09:50] *** mgoldmann has quit IRC
[07:11:19] *** mgoldmann has joined #jbosstesting
[08:02:19] *** jharting has joined #jbosstesting
[08:09:37] *** tdiesler has joined #jbosstesting
[08:14:10] *** alesj has joined #jbosstesting
[08:22:56] *** ge0ffrey has joined #jbosstesting
[08:25:55] *** lightguard_jp has quit IRC
[08:27:44] *** lightguard_jp has joined #jbosstesting
[08:36:48] *** oskutka has joined #jbosstesting
[08:49:21] *** kpiwko has joined #jbosstesting
[09:02:19] *** aslak has joined #jbosstesting
[09:27:31] *** michaelschuetz has joined #jbosstesting
[09:38:54] <kpiwko> hey aslak
[09:40:27] *** lfryc has joined #jbosstesting
[09:40:41] <tdiesler> aslak, my container does not get loaded anymore - any obvious reason?
[09:41:39] <aslak> tdiesler, yea, spi changes. got the fixes locally here, but i'm in trouble else where. the jmx-bean is never installed. any obvious reason ?
[09:41:39] <tdiesler> aslak, there is no arquillian.xml . AFAIK it was loaded by the ServiceLoader API
[09:42:19] <aslak> i see when running osgi embedded in eclipse it can't find the arquillian-osgi-bundle, but not sure if it runs there at all
[09:42:46] <tdiesler> aslak, you can leave the osgi container to me. I updated all the group/artifact ids already
[09:43:45] <tdiesler> aslak, I suggest you commit your stuff to some branch where I can pick it up. So we can have a tight feedback loop and try to get this fixed together
[09:45:56] <aslak> yea, working on it
[09:46:43] <aslak> tdiesler, https://github.com/arquillian/arquillian-container-osgi/tree/ARQ-361
[09:46:45] <jbossbot> jira [ARQ-361] Split up Arquillian repository [Coding In Progress (Unresolved) Feature Request, Major, Aslak Knutsen] https://issues.jboss.org/browse/ARQ-361
[09:54:39] <tdiesler> aslak, you can leave the osgi container to me. To get the arq bundle to work is not trivial and prob a waste of your time. Perhaps we can take this step by step and first find out why the DeployableContainer does not get loaded any more. I pushed my WIP to https://github.com/tdiesler/arquillian-container-osgi/tree/arq429
[09:55:37] <tdiesler> The associated core changes are here: https://github.com/tdiesler/arquillian-core/tree/arq429
[09:59:50] <tdiesler> aslak, I can also use your brach and take it from there if you think that works best
[10:00:59] <aslak> tdiesler, the arquillian-container-osgi:ARQ-361 branch has the required changes to match arquillian-core:master
[10:01:01] <jbossbot> jira [ARQ-361] Split up Arquillian repository [Coding In Progress (Unresolved) Feature Request, Major, Aslak Knutsen] https://issues.jboss.org/browse/ARQ-361
[10:01:26] <aslak> tdiesler, so the container is loading and starting, but it does not install the osgi-jmx bundle for some reason
[10:01:26] <tdiesler> aslak, ok. I'll take it from there
[10:06:24] <tdiesler> aslak, have a look at target/test.log The ARQ bundle cannot get started because of http://pastie.org/1842527
[10:06:27] <tdiesler> I'll fix it
[10:07:14] <aslak> tdiesler, aa ha.. missed that. it's in arquillian-test-spi
[10:10:31] *** lfryc has quit IRC
[10:22:37] *** wolfc has joined #jbosstesting
[10:29:43] <tdiesler> aslak, back to you - I now see http://pastie.org/1842577 You can pull from here https://github.com/tdiesler/arquillian-container-osgi/tree/ARQ-361
[10:29:45] <jbossbot> jira [ARQ-361] Split up Arquillian repository [Coding In Progress (Unresolved) Feature Request, Major, Aslak Knutsen] https://issues.jboss.org/browse/ARQ-361
[10:36:29] <aslak> tdiesler, it's suppose to run from eclipse right?
[10:37:01] *** jeand has joined #jbosstesting
[10:37:03] <tdiesler> aslak, I use mvn -Dmaven.surefire.debug=true -Dtest=EmbeddedFrameworkTestCase clean instal
[10:37:12] <aslak> ok
[10:37:14] <tdiesler> and connect eclipse to 5005
[10:37:35] <tdiesler> aslak, ultimately - yes it should run from eclipse
[10:37:55] *** pmuir has joined #jbosstesting
[10:38:03] *** pmuir has quit IRC
[10:38:03] *** pmuir has joined #jbosstesting
[10:38:13] <tdiesler> aslak, actually its mvn -Dmaven.surefire.debug=true -Dtest=OSGiEmbeddedFrameworkTestCase clean instal
[10:45:50] <tdiesler> aslak, I verified that the JMXMethodExecutor returns the TestResult, which is later null and causes the NPE in Arquillian
[10:46:16] <aslak> hmm, even stranger.. looking
[10:47:14] <aslak> tdiesler, aa.. it's from incontainer
[10:54:06] <aslak> tdiesler, do you just ignore the auxilliary archives ?
[10:54:31] <tdiesler> aslak, yes
[10:56:54] *** alesj has quit IRC
[10:56:57] <tdiesler> aslak, I can deploy bundes. Have a look at ARQ-338 and the EmbeddedDeployableContainer on how it deploys the support bundles on demand
[10:56:59] <jbossbot> jira [ARQ-338] Add test case to specify a list of support bundles [Open (Unresolved) Task, Major, Thomas Diesler] https://issues.jboss.org/browse/ARQ-338
[10:57:02] <aslak> tdiesler, extension like the ContainerTestExtension use the AuxillaryArchiveAppender SPI to create a 'new' of it self runtime.
[10:57:24] <aslak> v. of it self
[10:58:20] <aslak> ContainerTestExtension is loaded on client, runtime creates a ContainerTestRemoteExtension that is passed to the server and reloaded there. slightly different in behavior.
[11:01:20] *** lfryc has joined #jbosstesting
[11:06:01] *** alesj has joined #jbosstesting
[11:14:42] <tdiesler> aslak, are you still fixing or do I need to do stuff?
[11:15:30] *** tdiesler has quit IRC
[11:17:26] *** tdiesler has joined #jbosstesting
[11:28:36] <aslak> tdiesler, i'm wondering how this has worked before
[11:29:32] *** lightguard_jp has quit IRC
[11:29:34] <tdiesler> aslak, why? The method executor returned the TestResult, which was deserialized on the client side.
[11:29:45] <aslak> tdiesler, that's not the problem
[11:29:57] <aslak> tdiesler, the protocol works fine
[11:30:35] <aslak> tdiesler, the problem is there is no LocalTestExecutor that is registered when running incontainer, so TestMethodExecturor is never called and TestResult not returned
[11:33:47] <aslak> tdiesler, just to clearify, you don't modify anything else besides the MANIFEST.fm for visibility etc in the user defined deployment.. everything else is in the osgi-bundle that is installed straight from maven.. nothing dynamic going on there?
[11:35:45] <tdiesler> I add the test case. Have a look at OSGiApplicationArchiveProcessor
[11:36:32] <tdiesler> aslak, the test bundle is supposed to import packages that the arq-bundle exports
[11:36:40] <aslak> right
[11:37:26] <tdiesler> aslak, it is guarantied that there is no arq stuff coming from the syscl. Everything comes from the arq-bundle
[11:38:45] <tdiesler> aslak, so its a client side issue, right?
[11:39:53] <aslak> tdiesler, not really. the client side works, it's the container side that seems to be setup incorrectly when running in the osgu-bundle
[11:41:03] *** maschmid has joined #jbosstesting
[11:41:10] <tdiesler> aslak, in that case we want the arq-bundle fail to start
[11:41:46] <aslak> tdiesler, well, it's not really known at bundle start time.
[11:42:15] <aslak> tdiesler, invoking the TestRunner results in the loading of the extension
[11:42:50] <tdiesler> aslak, can this not be done when the TestRunner gets created?
[11:43:04] <tdiesler> aslak, or maybe explicitly from the bundle activator
[11:43:52] <aslak> tdiesler, and good new i think.. the internal extension loading is now using the CL from where LoadableExtension(core-impl-base) is loaded from, so i think the JUnitBundleTestRunner is not needed anymore (atleast not for arquillian)
[11:45:47] <tdiesler> aslak, thats good. Second best is that the first test that runs fails with a cause that shows the init failure
[11:46:37] <tdiesler> aslak, perhaps its something trivial like a missing jar in the arq-bundle. You can crack it open to have a look
[11:48:02] <aslak> tdiesler, the problem is knowing programatically what is suppose to be there. it's all just event listeners.. but Arquillian test framework integration could check the result alla: result == null throw new Exception("No result was returned from Test invoker, setup issues.") just to avoid the NullPointer
[11:49:24] <tdiesler> aslak, that doesn't really help I don't think. We want to know why the result is null
[11:50:24] <aslak> "most likely no TestMethodExecutor registered"
[11:50:41] <tdiesler> aslak, the bundle only has one service (i.e. TestRunner) Is that right?
[11:54:00] <aslak> tdiesler, the rest of the services are in the bundled jars
[11:54:11] <aslak> and all should be of type, LoadableExtension
[11:54:34] *** michaelschuetz has quit IRC
[11:55:36] <tdiesler> aslak, let me run this in a remote framework. Then we can look at what extensions actually get loaded
[11:57:23] <aslak> you can see it if you put on some FINE logging of LoadableExtensionLoader
[11:58:10] <aslak> tdiesler, testing something here now..  bundling the standalone junit v. instead of the container v. it's won't be correct, but might get us passed this
[12:00:04] <tdiesler> nothing, obvious http://pastie.org/1842769 the bundle install/starts ok. Running the test does not cause log messages
[12:07:12] <tdiesler> aslak, I see what you're saying the NPE already happens in the framework and then gets (correctly) transported to the client
[12:07:21] <aslak> tdiesler, only things in bundle/META-INF is loaded, not bundle/*.jar/META-INF?
[12:09:40] <tdiesler> alsak, I don't think so. The osgi testenricher would not have worked if that was the case
[12:10:07] <tdiesler> aslak, unless there is a different discovery algorithm
[12:10:58] <tdiesler> aslak, what API do you use to load the services?
[12:11:07] <aslak> no, my bad
[12:12:25] <aslak> tdiesler, https://github.com/arquillian/arquillian-core/blob/master/core/impl-base/src/main/java/org/jboss/arquillian/core/impl/loadable/JavaSPIServiceLoader.java
[12:12:25] *** Tashtego has joined #jbosstesting
[12:18:08] *** aslak has quit IRC
[12:18:50] *** aslak has joined #jbosstesting
[12:29:35] <tdiesler> aslak, ClassLoader.getResources() should give you all resources in the wired class space. That would include the ones from nested jars
[12:55:59] <aslak> tdiesler, yea, that's what's being used.  but i didn't include the spi i thought i did so.. :)
[13:04:02] *** michaelschuetz has joined #jbosstesting
[13:24:49] <aslak> tdiesler, i'm wondering if creating the osgi-bundle dynamically based on the AuxiliaryArchives and installing it on first deploy might work better
[13:30:50] <tdiesler> have a look at the pom. Its not trivial to create the bundle. There are many variables involved
[13:31:13] <tdiesler> aslak, could you find what was causing the NPE?
[13:32:23] <tdiesler> aslak, what is right for the first test bundle might not be right for subsequent tests
[13:34:05] <tdiesler> aslak, so IMHO the arq-bundle should be well defined infrastructure that test bundles can use. We can look at possibilities to extend the infrastructure on a per-test-bases e.g. the test provides an enricher
[13:34:26] <tdiesler> aslak, if needed
[13:35:48] <aslak> tdiesler, i know what is causing it, the 'Container/Remote' side is being dynamically generated.. and osgi-bundle is halfway based on the static client side
[13:36:45] <aslak> aa.. now i know why it worked before
[13:38:51] <aslak> previously a Extension had a getClientProfile vs getContainerProfile based on how the TestRunner was started. but the Client Profile would still have to bundle it up dynamiccaly etc to be shipped of to the remote container. so even tho packaged i nthe bundle, 'parts' of the ContainerProfile was reachable based on the Client jar.
[13:39:37] <aslak> so with a little luck that would/could work
[13:40:54] <aslak> in this v. tho the client/container side of the profile has been removed, and is all based on how the AuxillaryArchiveAppenders generate the Container side. since this was needed anyway, the dual side thing seemed pointless and confusing
[14:27:08] *** ldimaggi has joined #jbosstesting
[14:29:43] <tdiesler> aslak, from this I understand that you'd provide a fix, right?
[14:30:43] <aslak> tdiesler, well, the fix is what we discussed on the last call. we need a dynamic infrastruct for the Modular (osgi/msc) runners
[14:32:09] <tdiesler> aslak, ok we can talk about this again in a little while
[14:32:18] <aslak> yeah
[14:33:14] *** Tashtego is now known as Fiatego
[14:36:56] *** Fiatego is now known as Tashtego
[14:55:13] *** nickarls has quit IRC
[14:56:13] *** echelog-2` has joined #jbosstesting
[14:56:44] *** echelog-2` is now known as echelog-2
[14:56:46] *** dblevins_ has joined #jbosstesting
[14:58:00] *** dblevins has quit IRC
[14:58:01] *** jc3 has quit IRC
[14:58:03] *** nickarls has quit IRC
[15:00:53] *** jc3` is now known as jc3
[15:13:44] *** michaelschuetz1 has joined #jbosstesting
[15:15:40] *** michaelschuetz has quit IRC
[15:54:47] *** oskutka has quit IRC
[16:00:17] *** tdiesler has quit IRC
[16:09:07] *** jharting has quit IRC
[16:58:11] *** michaelschuetz has joined #jbosstesting
[17:00:06] *** michaelschuetz1 has quit IRC
[17:01:20] *** rruss has joined #jbosstesting
[17:09:08] *** kpiwko has quit IRC
[17:26:39] *** dblevins_ has quit IRC
[17:26:47] *** dblevins has joined #jbosstesting
[17:38:19] *** bobmcw has quit IRC
[17:42:53] *** alesj has quit IRC
[17:45:34] *** maschmid has quit IRC
[17:51:36] *** vtunka has quit IRC
[17:55:28] *** lfryc has quit IRC
[18:02:13] *** bobmcw has joined #jbosstesting
[18:09:58] *** michaelschuetz1 has joined #jbosstesting
[18:13:32] *** michaelschuetz has quit IRC
[18:15:35] *** rruss has quit IRC
[18:18:19] *** rruss has joined #jbosstesting
[18:22:59] *** rruss has quit IRC
[19:10:55] <aslak> ALR, you've been around the MavenResolver code a little, is there a way to tell it to only resolve the given cordinates?
[19:11:11] <ALR> At present only w/ filtering
[19:11:22] <aslak> StrictFilter only applies to after resolve
[19:11:28] <aslak> right
[19:11:35] <ALR> Though the new refactorings suggest a new syntax
[19:11:44] <ALR> resolveWithDeps and resolve or something similarly-named
[19:12:06] <aslak> yea, but will that change how it is resolved ? or is it just adding the filters ?
[19:14:24] <ALR> Umm, looking
[19:14:24] <ALR> http://community.jboss.org/message/596588
[19:14:41] <ALR> No impl notes there I don't believe
[19:14:50] <ALR> Though it shouldn't require resolution I wouldn't think
[19:15:16] <ALR> aslak: Do you know the command to make ARQ tests in Forge shell?
[19:15:24] <ALR> aslak: It's not listed in tab-view
[19:15:35] <aslak> you have to install it
[19:16:16] <aslak> there was something funky there..  which i fixed up in my local v.
[19:16:23] <aslak> let's see
[19:16:28] <ALR> Calling Lincoln
[19:20:09] <ALR> Oooh, there's not JBOSS_AS7 container for Arquillian Forge Plugin now
[19:23:27] <aslak> have it locally..
[19:23:48] <ALR> And my WAR generated by Forge doesn't deploy in AS7
[19:23:48] <ALR> Caused by: java.lang.NoClassDefFoundError: org/springframework/context/MessageSource
[19:24:17] <aslak> springframework?
[19:24:24] <ALR> Guess so.
[19:25:12] <ALR> Hmm, might take a different tack.
[19:25:21] <ALR> And make the EJB3 Examples run on AS7 instead
[19:26:32] <ALR> Book Examples that is.
[20:09:30] *** nickarls_ is now known as nickarls
[20:12:31] *** rruss has joined #jbosstesting
[20:17:25] *** rruss has quit IRC
[20:19:50] *** rruss has joined #jbosstesting
[20:27:57] <aslak> ALR, https://github.com/aslakknutsen/plugin-arquillian/commits/jbossas7 if interested
[20:39:48] *** lightguard_jp has joined #jbosstesting
[20:46:38] *** Tashtego has quit IRC
[20:54:06] *** lightguard_jp has quit IRC
[20:54:29] *** lightguard_jp has joined #jbosstesting
[20:55:00] *** ALR1 has joined #jbosstesting
[20:55:03] *** ALR has quit IRC
[21:02:38] *** aslak has quit IRC
[21:09:42] *** pmuir has quit IRC
[21:22:01] *** michaelschuetz1 has quit IRC
[21:22:30] *** alesj has joined #jbosstesting
[21:26:49] *** bobmcw has quit IRC
[21:27:14] *** bobmcw has joined #jbosstesting
[21:34:17] *** michaelschuetz has joined #jbosstesting
[21:50:16] *** jeand has quit IRC
[21:56:29] *** ldimaggi has quit IRC
[21:57:40] *** ge0ffrey has quit IRC
[21:58:15] *** bobmcw has quit IRC
[21:58:26] *** lightguard_jp has quit IRC
[21:58:28] *** bobmcw has joined #jbosstesting
[21:58:52] *** lightguard_jp has joined #jbosstesting
[22:10:27] *** lightguard_jp has quit IRC
[22:14:14] *** lightguard_jp has joined #jbosstesting
[22:14:14] *** rruss has quit IRC
[22:14:55] *** alesj has left #jbosstesting
[22:22:58] *** Tashtego has joined #jbosstesting
[22:26:02] *** lightguard_jp has quit IRC
[22:26:04] *** mgoldmann has quit IRC
[22:38:32] *** pmuir has joined #jbosstesting
[22:38:32] *** pmuir has quit IRC
[22:38:32] *** pmuir has joined #jbosstesting
[22:38:40] *** lightguard_jp has joined #jbosstesting
[22:40:56] *** lightguard_jp has quit IRC
[22:41:15] *** wolfc has quit IRC
[22:43:07] *** kpiwko has joined #jbosstesting
[22:45:44] *** Tashtego has quit IRC
[22:49:13] *** oskutka1 has joined #jbosstesting
[22:51:41] *** lightguard_jp has joined #jbosstesting
[22:52:04] *** pmuir has quit IRC
[23:03:24] *** lightguard_jp has quit IRC
[23:14:36] *** lightguard_jp has joined #jbosstesting
[23:15:07] *** oskutka1 has quit IRC
[23:22:52] *** lightguard_jp has quit IRC
[23:27:56] *** kpiwko has quit IRC
[23:56:03] *** michaelschuetz has quit IRC
[23:59:45] *** lightguard_jp has joined #jbosstesting





top