March 20, 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 20, 2011 NOTICE: This channel is no longer actively logged.
[00:55:11] *** bobmcw has joined #jboss-as7[00:55:11] *** ChanServ sets mode: +v bobmcw[01:04:54] *** bobmcw has quit IRC[02:34:19] *** mbg is now known as mbg|away[02:37:35] *** bobmcw has joined #jboss-as7[02:37:35] *** ChanServ sets mode: +v bobmcw[02:44:27] *** bobmcw_ has joined #jboss-as7[02:48:52] *** bobmcw has quit IRC[03:27:51] <smarlow> bstansberry: here is another crazy thought on filesystem action driven hot deployment :) Eliminate the tags from the file system, just keep the current application deployment state internally. Then, we could let users express the same actions, that the tags represent, through the admin console.[03:30:15] <smarlow> at least, that would help if people don't like the tag approach[03:37:49] <bstansberry> tag == marker?[03:39:34] <smarlow> yup, marker[03:43:23] <smarlow> I don't see a perfect answer though, filesystem deployment will always be somewhat primitive[03:44:02] <smarlow> having to use the admin console, to alter your file system deployments, is also a bit clumsy[03:44:22] <bstansberry> we do keep the deployment state is kept internally[03:44:47] <bstansberry> sorry; "we do keep the deployment state internally"[03:45:27] <smarlow> one thing that might help, is if we log our intentions and actions taken, in response to what the user does with markers in the file system[03:46:44] <bstansberry> so users could use the admin console to trigger some things; and probably eventually everything -- once we rework a bit what state we keep internally[03:47:08] *** aslak has quit IRC[03:48:15] <bstansberry> I think what people are asking for is focused on allowing certain kinds of users have to think or understand much[03:48:19] <bstansberry> which i can understand[03:50:14] <smarlow> externalizing the file system scanner, could also help. Then users could launch, with whatever behaviour they want. :)[03:52:43] <smarlow> I posted to the thread about that idea. Basically, have support for AS4 (old school) scanning (no markers) and support for the new approach. We could also scan from other things besides the file system (version control systems).[04:00:47] *** smarlow has quit IRC[06:03:55] *** bstansberry has quit IRC[06:18:50] *** bstansberry has joined #jboss-as7[06:18:50] *** ChanServ sets mode: +v bstansberry[07:07:08] *** bstansberry has quit IRC[07:12:59] *** baileyje has joined #jboss-as7[07:12:59] *** ChanServ sets mode: +v baileyje[07:47:58] *** bobmcw_ has quit IRC[09:48:28] *** JimMa has joined #jboss-as7[11:16:46] *** aslak has joined #jboss-as7[11:16:46] *** aslak has joined #jboss-as7[11:16:46] *** ChanServ sets mode: +v aslak[11:35:39] *** ALR has joined #jboss-as7[11:35:39] *** ChanServ sets mode: +v ALR[12:36:53] *** aslak has quit IRC[13:28:30] *** sannegrinovero has joined #jboss-as7[13:28:30] *** sannegrinovero has joined #jboss-as7[13:28:30] *** ChanServ sets mode: +v sannegrinovero[13:32:05] *** pretec has joined #jboss-as7[13:34:46] *** pretec has left #jboss-as7[14:03:11] *** smarlow has joined #jboss-as7[14:03:12] *** ChanServ sets mode: +v smarlow[14:04:04] *** aslak has joined #jboss-as7[14:04:04] *** ChanServ sets mode: +v aslak[14:16:47] <smarlow> dmlloyd: I'm squashing again[14:22:18] <smarlow> dmlloyd: could you please commit https://github.com/scottmarlow/jboss-as/commit/5eedc8b69fd76a6c1c3cdd9f16366e24290655c2 when you can. Thanks[14:22:20] <jbossbot> git [jboss-as] 5eedc8b.. Scott Marlow switch to using DataSource injection, flipped the loop in JPAAnnotationParseProcessor, improved service names and addressed JBAS-9034 (allow JPA operations in deployments that use JPA annotations that reference external applications persistence.xml),delete duplicate module org.slf4j.slf4j-api, only pass validator if validation is enabled, switch to LazyValidatorFactory, JBAS-9065 Invoked without a tx[14:22:28] <jbossbot> jira [JBAS-9034] Applications should be able to inject a persistence context/unit from a separate deployment by using a fully scoped (appname#puname) reference [Resolved (Done) Sub-task, Major, Scott Marlow] https://issues.jboss.org/browse/JBAS-9034[14:22:29] <jbossbot> jira [JBAS-9065] Invoked without a tx, transaction scoped entity mananger managed entities should be detached at end of EntityManager call [Resolved (Done) Sub-task, Major, Scott Marlow] https://issues.jboss.org/browse/JBAS-9065[15:05:53] *** mbg|away is now known as mbg[15:08:13] *** JimMa has quit IRC[15:23:19] *** ALR has quit IRC[15:24:23] *** mbg is now known as mbg|away[15:37:58] *** bstansberry has joined #jboss-as7[15:37:58] *** ChanServ sets mode: +v bstansberry[16:18:32] *** baileyje has quit IRC[17:10:43] *** mbg|away is now known as mbg[17:45:33] <jbossbot> git [jboss-as] push master 90e8d1e.. Scott Marlow switch to using DataSource injection, flipped the loop in JPAAnnotationParseProcessor, improved service names and addressed JBAS-9034 (allow JPA operations in deployments that use JPA annotations that reference external applications persistence.xml),delete duplicate module org.slf4j.slf4j-api[17:45:35] <jbossbot> jira [JBAS-9034] Applications should be able to inject a persistence context/unit from a separate deployment by using a fully scoped (appname#puname) reference [Resolved (Done) Sub-task, Major, Scott Marlow] https://issues.jboss.org/browse/JBAS-9034[17:45:35] <jbossbot> git [jboss-as] push master 1395108.. Marcus Moyses removing commented out code[17:45:35] <jbossbot> git [jboss-as] push master 52efe7b.. Marcus Moyses JBAS-8981: changing java:/jaas to java:jboss/jaas[17:45:36] <jbossbot> jira [JBAS-8981] Move java:/jaas into appserver-wide global context. [Resolved (Done) Task, Major, Marcus Moyses] https://issues.jboss.org/browse/JBAS-8981[17:45:36] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/a627097...52efe7b[17:45:55] <dmlloyd> smarlow: I had merged it previously, just forgotten to push it...[18:10:29] *** jamezp has joined #jboss-as7[18:14:10] *** aslak has quit IRC[18:14:34] *** aslak has joined #jboss-as7[18:14:34] *** ChanServ sets mode: +v aslak[18:27:44] *** aslak has quit IRC[18:28:00] <jamezp> dmlloyd: Looking at https://issues.jboss.org/browse/JBAS-8991 I'm not sure I will be much good on it. I don't have any access to RHEL and I can't get it to fail on Fedora.[18:28:01] <jbossbot> jira [JBAS-8991] Cannot start beta1 server on RHEL6 with OpenJDK 1.6 [Open (Unresolved) Bug, Critical, James Perkins] https://issues.jboss.org/browse/JBAS-8991[18:53:14] *** jpearlin has joined #jboss-as7[19:01:20] *** ccrouch has quit IRC[19:14:33] <dmlloyd> jamezp: just out of curiosity what JVM are you using on fedora?[19:14:56] <jamezp> It's OpenJDK, but it's a newer release.[19:15:11] <jamezp> It's 1.6.0_20 instead of _17[19:15:20] <dmlloyd> I had a copy of the JVM which exhibited the problem but I stupidly erased it[19:15:39] <dmlloyd> which would have told us whether it's the JVM or other environmental issue[19:16:04] <jamezp> Right… …maybe I'll just look around for the JVM then.[20:19:26] *** aslak has joined #jboss-as7[20:19:26] *** ChanServ sets mode: +v aslak[20:19:49] *** jpearlin has quit IRC[20:30:46] *** jwulf has quit IRC[20:59:23] <smarlow> dmlloyd: It seems that I have squashed some commits that you already merged (I shouldn't of squashed after asking for a merge).[21:00:33] <dmlloyd> yeah. Easiest way out of that is to try and locate your pre-squash commit ID (maybe in your console log scrollback) and reset --hard to it[21:00:45] <dmlloyd> or create a branch from that ID if you dont' want to lose your current work[21:01:00] <dmlloyd> i.e.: git branch temp 0123456[21:01:11] <dmlloyd> (note you can use a partial commit ID as well)[21:01:26] * dmlloyd has done this exact thing on more than one occasion :)[21:01:53] <smarlow> probably best to just go back in history (to before my bad squash) and reapply changes :)[21:02:22] <dmlloyd> if you have them sure[21:02:32] <dmlloyd> the nice thing about git is that you never really lose anything[21:02:47] <dmlloyd> the exact state of your checkout is always a unique commit ID[21:11:53] <smarlow> looks like the commit id scrolled off the screen.[21:12:40] <dmlloyd> ah, damn[21:13:15] <smarlow> can we just go forward somehow, forget about my changes. I can then diff and create new commits?[21:13:28] <dmlloyd> you could try git fsck --lost-found[21:13:41] <dmlloyd> yeah you can, just trying to save you the hassle[21:13:49] <dmlloyd> it doesn't matter from my perspective[21:14:46] <dmlloyd> the git prompt stuff can help too[21:14:54] <dmlloyd> because you know your commit ID every time you hit enter[21:15:24] <dmlloyd> that has saved me a lot of trouble and also makes me a lot more comfortable with being in the notorious "detached head" state, where you're basically committing with no branch[21:16:01] <dmlloyd> I will often do stuff like check out a remote ref (with no local branch), rebase it on to master, note the new commit ID, check out master, then merge the commit ID[21:22:57] *** bobmcw has joined #jboss-as7[21:22:57] *** ChanServ sets mode: +v bobmcw[21:22:59] <smarlow> I already pushed another change to my branch also, could you pull https://github.com/scottmarlow/jboss-as/commit/befc451a85ac0a521e69eaeceb8ef15cb6e14fd0 Then, I'll see what the remaining diff is and either give you a patch or drop/reclone my branch. does that work?[21:23:01] <jbossbot> git [jboss-as] befc451.. Scott Marlow JBAS-8962 start adding XPC support[21:23:05] <jbossbot> jira [JBAS-8962] add XPC support [Open (Unresolved) Sub-task, Major, Scott Marlow] https://issues.jboss.org/browse/JBAS-8962[21:25:10] *** bobmcw_ has joined #jboss-as7[21:27:18] <smarlow> dmlloyd: if you apply the above, I should just be able to truth up by generating a new commit for the diff. No patch should be needed.[21:27:32] *** bobmcw_ has quit IRC[21:28:09] *** bobmcw has quit IRC[21:31:09] <dmlloyd> yeah that's fine with me[21:31:39] <dmlloyd> just make sure you don't include the two commits after yours in the diff image[21:32:26] <smarlow> I won't :)[21:33:01] * smarlow is getting called to grill :)[21:33:27] <dmlloyd> must be nice. It's cold & rainy here[21:34:09] <jamezp> Cold and rainy here too, of course that's typical for Oregon.[21:34:59] <jbossbot> git [jboss-as] push master 999ef17.. Scott Marlow JBAS-8962 start adding XPC support[21:35:00] <jbossbot> jira [JBAS-8962] add XPC support [Open (Unresolved) Sub-task, Major, Scott Marlow] https://issues.jboss.org/browse/JBAS-8962[21:35:00] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/52efe7b...999ef17[21:52:38] *** ccrouch has joined #jboss-as7[21:52:38] *** ChanServ sets mode: +v ccrouch[22:03:10] *** irooskov has joined #jboss-as7[22:13:58] <smarlow> its like 40 degrees here but dry today[22:20:25] *** mbg has quit IRC[22:30:53] <jamezp> Alright, confirmed that https://issues.jboss.org/browse/JBAS-8991 is an OpenJDK 1.6.0_17 issue. Now to figure out why :-)[22:30:55] <jbossbot> jira [JBAS-8991] Cannot start beta1 server on RHEL6 with OpenJDK 1.6 [Open (Unresolved) Bug, Critical, James Perkins] https://issues.jboss.org/browse/JBAS-8991[22:31:15] <dmlloyd> nice[22:31:40] <dmlloyd> a good start might be to run the modules & msc tests on that JVM[22:31:44] <dmlloyd> see if they start screwing up[22:32:08] <jamezp> CentOS and VirtualBox to the rescue.[22:32:39] <dmlloyd> ah, centos, good idea[22:32:46] <jamezp> Good idea![22:33:03] <jamezp> I figured it's the closest I'll get for free.[22:45:00] <smarlow> dmlloyd: https://github.com/scottmarlow/jboss-as/commit/c8912bcdbeafab6a0f2658743cb0f83f8b7bb9da contains the missing part. Could you please merge it in[22:45:01] <jbossbot> git [jboss-as] c8912bc.. Scott Marlow use BeanValidationAttachments.VALIDATOR_FACTORY, detach entities after nontx invocation (trans EM only)[22:46:31] *** miclorb_ has joined #jboss-as7[22:51:03] <dmlloyd> smarlow: can you confirm that this is what you want it to look like: http://github.com/dmlloyd/jboss-as/commit/be7f2e3[22:51:03] <jbossbot> git [jboss-as] be7f2e3.. Scott Marlow use BeanValidationAttachments.VALIDATOR_FACTORY, detach entities after nontx invocation (trans EM only)[22:51:15] <dmlloyd> seemed to merge without an issue, just making sure[22:56:48] *** irooskov has quit IRC[23:33:22] * smarlow checking[23:34:03] <smarlow> dmlloyd: looks correct, thanks! :)[23:34:14] <dmlloyd> cool[23:34:25] <jbossbot> git [jboss-as] push master be7f2e3.. Scott Marlow use BeanValidationAttachments.VALIDATOR_FACTORY, detach entities after nontx invocation (trans EM only)[23:34:26] <jbossbot> git [jboss-as] push master URL: http://github.com/jbossas/jboss-as/compare/999ef17...be7f2e3[23:43:00] <smarlow> dmlloyd: Since all of my changes are merged in that I want, I assume its safe to "reset --hard" to a few days ago now?[23:43:54] <dmlloyd> yeah. remember though that nothing is actually lost as long as you know the commit IDs.[23:44:00] <dmlloyd> unless you do "git gc" :)[23:44:17] <smarlow> cool