[00:08:41] <mrooney1> kohsuke: when it triggers the build it could tell it, couldn't it? Also I just want to know in the upstream what downstream failures were caused by the upstream
[00:09:10] <mrooney1> So as long as the upstream gets the results from the same build # that it triggered by remembering that, all is well I think
[00:17:11] <kohsuke> Like I said, there can be more upstream builds before the triggered downstream build runs
[00:21:30] * rtyler is digging the sectioned view plugin
[00:23:49] <mrooney1> kohsuke: but the upstream build knows what #s of downstream builds to aggregate, doesn't it? I guess I am guessing the way it works backwards
[00:24:08] <mrooney1> the downstream is giving the test result to the upstream, not the upstream pulling it?
[00:24:39] <kohsuke> My point is, just because upstream #100 has triggered downstream #110 doesn't mean D #110 is testing U #100.
[00:24:43] <kohsuke> It might be testing U #101.
[00:25:07] <kohsuke> So we can't rely on "who triggered what" to aggregate.
[00:28:24] *** mde_ has joined #hudson
[00:28:52] *** mde_ is now known as Guest76667
[00:30:12] *** mde has quit IRC
[00:33:53] <kohsuke> abayer: do you mind if I touch the incremental Maven build code a bit?
[00:33:56] <kohsuke> Or are you still working on it?
[00:34:04] <abayer> All yours. =)
[00:34:27] <kohsuke> thanks
[00:35:06] <abayer> and sorry 'bout the 1.316-SNAPSHOT POM thing - I'd tweaked it at one point and forgot about it when committing. I need to clean up my act.
[00:35:13] <rtyler> heh
[00:35:23] <kohsuke> Does this change have a test case?
[00:35:29] <rtyler> kohsuke: when do you think you'll context switch back to the pre-commit/concurrent-build project?
[00:35:33] <rtyler> post-318?
[00:35:58] <abayer> Not as such, no.
[00:36:01] <kohsuke> I was about to check the status of your bug report then got distracted.
[00:36:12] <kohsuke> you said you attached additional info to the bug?
[00:36:18] <rtyler> yeah, threaddump!
[00:36:19] <rtyler> wootz
[00:36:33] <abayer> I'm not entirely sure where unit testing for the maven plugin actually is.
[00:36:47] <rtyler> friday afternoon would be a good time to fuxor with the instance of Hudson we have, so let me know if you have debug stuff for me to try out
[00:37:00] <kohsuke> maven tests are actually all in main/test for historical reasons.
[00:37:06] <kohsuke> we should probably fix that.
[00:37:09] <kohsuke> rtyler: got it
[00:38:25] <abayer> 'k. I should probably write a bunch of tests at some point - doesn't look like there's much in the way of tests for the Maven plugin there.
[00:38:54] <kohsuke> abayer: Yes.
[00:39:03] <rtyler> mmm tests
[00:39:08] <kohsuke> think of this as a way of showing my guilt :-)
[00:39:12] <abayer> heehee
[00:40:47] <abayer> Give me some time to try to get comfortable with HudsonTestCase - I've looked at it a little in the past but not much. I'll try not to pick up any more random feature work until I've got a decent set of tests in place.
[00:41:12] <kohsuke> thanks
[00:41:16] <abayer> Also, I was wondering if we might want to move the Maven help HTML out of main/war/resources/help/maven and into the maven plugin itself - just for consistency?
[00:41:26] <kohsuke> Yes indeed
[00:45:59] <abayer> I'll get that done now, then, while I'm thinking of it.
[00:49:45] <abayer> Actually, that might be more pain than it's worth, now that I look at it - a lot of links to update, including the links for Maven in freestyle projects.
[01:02:19] <mrooney1> kohsuke: alright, so it sounds like I should just suck it up and use a parallel ant task? :)
[01:04:41] <abayer> ooo, kohsuke, I like those changes. I tend towards a "don't change existing code unless it's buggy" philosophy with code I didn't write, and it worked out here, 'cos your approach is much better than the vague ideas I'd had about generalizing the filtering. =)
[01:08:29] <rtyler> is there a one-time means of asking a job to build on a specific executor?
[01:32:38] <calculus> is there a way to have the hudson login get delegated to some other url that returns http 200 if successful and 401 if not?
[01:33:28] <calculus> once a return code of 200 is there, that user can be assigned to a hudson group that then has the read/build/delete/... permissions
[01:36:45] <calculus> or if the user is already part of hudson (e.g. with adminstrator privs) those privs still apply
[01:39:55] <calculus> inspiration is here: http://svn.sonatype.org/spice/trunk/plexus-security/security-realms/security-url-realm/README (plugin for nexus)
[01:45:37] <jieryn-w> i <3 hudson!!
[01:46:38] <abayer> It is pretty great.
[01:46:42] <calculus> hudson <3 you too!
[01:48:07] <jieryn-w> here's an interesting issue..
[01:48:25] <jieryn-w> i'm utilizing Hudson's per-job maven repository feature
[01:48:41] <jieryn-w> however, i also want to rm -rf $WORKSPACE/.repository on a nightly basis
[01:48:54] <abayer> Do you build on a nightly basis as well?
[01:48:58] <jieryn-w> i run with 5 executors
[01:49:06] <jieryn-w> abayer: i build on a per-checkin basis
[01:49:13] <abayer> Just checking. =)
[01:49:42] <jieryn-w> i want to make sure i don't clobber the .repo while another executor is using it
[01:50:09] <jieryn-w> any tips?
[01:50:11] <abayer> Locks!
[01:50:18] * jieryn-w checks
[01:50:19] <abayer> i.e., locks and latches.
[01:50:30] <jieryn-w> yah.. that plugin didn't blow my skirt up last i reviewed it
[01:50:50] <abayer> Have a nightly job that does nothing but nuke your repository, and have it share a lock with your real job.
[01:51:02] <abayer> Then the repo-nuker won't run when the real job is already running.
[01:51:04] <abayer> (and vice versa)
[01:52:29] <jieryn-w> can i have a job that locks out multiple other jobs?
[01:52:33] <jieryn-w> i'd rather just create one single cleanup job
[01:52:38] <abayer> Hmm.
[01:53:04] <jieryn-w> locks-and-latches hasn't been updated in like 6 months
[01:53:21] <abayer> I think you can add multiple locks. Lemme see.
[01:53:32] <abayer> and yeah, the "latches" aspect never happened. =)
[01:54:28] <abayer> Yes, you can add multiple locks - so that you can have the nuker job share a lock with real job A, share another lock with real job B, etc.
[01:56:37] <jieryn-w> on a heavily busy system, the locker may never obtain all the locks
[01:56:42] <abayer> True.
[01:56:47] <jieryn-w> it tries to obtain them all in one go, if it can't, it releases them.. i see in the code
[01:56:56] <jieryn-w> while i think that's good for deadlock..
[01:56:59] <abayer> Yeah.
[01:57:04] <abayer> I see the problem.
[01:57:09] <jieryn-w> i'm running 2-3 builds pretty much all the time
[01:57:34] * jieryn-w hms
[01:57:48] <jieryn-w> i may have to delete the local repo every stupid build.. blah
[01:57:54] <jieryn-w> OR, hm
[01:58:00] *** BigAllan has quit IRC
[01:58:29] <jieryn-w> i guess i could write a new maven plugin that runs dependency:purge-local-repository randomly
[01:58:59] <jieryn-w> OR, better yet.. :)
[01:59:21] <jieryn-w> enable a profile which runs dependency:purge-local-repository when we find a .purge-local-repository file in the workspace
[01:59:42] <jieryn-w> and then just have a job that runs @daily to touch all workspaces .purge-local-repository
[01:59:59] <abayer> That could work.
[02:00:03] <jieryn-w> i think it will :)
[02:00:09] <abayer> It's a bit hacky, but, well, life is a bit hacky. =)
[02:00:26] <jieryn-w> wonder if he have a tips and tricks section somewhere?
[02:00:48] <jieryn-w> i'll get my profile working, and then append it somewhere for someone to use.. hm
[02:29:22] *** statlor has joined #hudson
[02:30:48] <jieryn-w> ok, got it working
[02:44:20] <jieryn-w> http://dpaste.com/71105/
[02:46:02] <rtyler> ruh roh, tests failing in concurrent-build
[03:14:57] <jieryn-w> abayer: so, sorry - is 2736 fixed or no?
[03:15:03] <jieryn-w> it doesn't look like it's done yet.. curious
[03:20:02] <abayer> Which one's that?
[03:20:23] <abayer> Yes, it is.
[03:20:47] <abayer> I'll mark it as such.
[03:23:46] <jieryn-w> hotness
[03:24:02] <abayer> Indeed. I'm spending a lot of time hacking the Maven plugin these days. =)
[03:24:53] <jieryn-w> you're my hero
[03:25:05] <abayer> I try.
[03:26:08] <jieryn-w> i just started following your tweets
[03:27:09] <abayer> Be prepared to see me talk about my cat a lot.
[03:30:28] *** rromanchuk has quit IRC
[03:32:34] * jieryn-w grins :)
[03:37:11] * rtyler meows
[03:37:19] <abayer> You're not a cat.
[03:38:00] <abayer> Imposter.
[03:38:50] <rtyler> =^_^=
[03:39:00] <abayer> Now you're just anime.
[03:39:04] <rtyler> drats
[03:40:06] <jieryn-w> ok, i'm on 1.317
[03:40:16] <jieryn-w> hope to see some improvement in m2 downstream builds :)
[03:40:23] <abayer> Lemme know how it looks.
[04:05:25] *** davidstrauss has joined #hudson
[04:13:38] *** alexlod has quit IRC
[04:20:01] *** davidstrauss has quit IRC
[04:33:17] *** admc has quit IRC
[04:38:19] *** davidstrauss has joined #hudson
[04:52:39] <statlor> hm is there some feature or plugin that shows build stats, like top builds by frequency, by avg build time
[04:53:15] <statlor> might be something interesting to work on
[04:53:54] <statlor> maybe top causers of builds
[04:54:21] <statlor> (whether manually run or by their scm changes)
[05:00:23] <statlor> top failures by %age
[05:00:40] <statlor> replete with charts
[05:17:44] *** Lewisham has quit IRC
[05:18:58] <statlor> odd.. it said ssh-slaves plugin 0.6 is available, I upgraded successfully, and I'm still at 0.5.2
[05:19:25] <statlor> it hasn't picked up that 317 is available.. is there a way to force update the update-center.json file?
[05:21:40] <statlor> ha I knew it.. on the manage plugins advanced page
[05:22:07] <statlor> did I mention I love hudson
[05:23:39] *** |^JaMeS^| has joined #hudson
[05:29:35] *** dholst has joined #hudson
[05:32:28] <dholst> can anyone explain the reasoning behind not failing a build if there are test failures?
[05:58:58] *** abayer has quit IRC
[06:00:19] *** abayer has joined #hudson
[07:06:21] *** dholst has quit IRC
[07:29:20] *** Lewisham has joined #hudson
[07:42:06] <statlor> I could, but you left
[07:42:16] *** statlor has quit IRC
[08:27:36] *** davidstrauss has quit IRC
[08:43:46] *** khmarbaise has joined #hudson
[08:56:07] *** khmarbaise has quit IRC
[09:14:19] *** davidstrauss has joined #hudson
[09:32:36] *** davidstrauss has quit IRC
[09:34:41] *** Lewisham has quit IRC
[11:46:31] *** rromanchuk has joined #hudson
[12:10:50] *** jbriguet- has quit IRC
[12:29:04] *** khmarbaise has joined #hudson
[13:10:54] *** jbriguet has joined #hudson
[13:17:39] *** jbriguet has quit IRC
[13:20:32] *** jbriguet has joined #hudson
[13:40:13] *** jbriguet has quit IRC
[13:53:16] *** jbriguet has joined #hudson
[14:16:40] *** ppawel has joined #hudson
[15:10:39] *** akostadinov has joined #hudson
[15:27:44] *** khmarbaise has quit IRC
[16:10:34] *** khmarbaise_ has joined #hudson
[16:39:46] *** akostadinov has quit IRC
[16:57:03] *** jieryn-w has quit IRC
[17:34:38] *** |^JaMeS^| has quit IRC
[20:05:23] *** lhochet has joined #hudson
[20:54:17] *** lhochet has quit IRC
[21:12:18] *** Lewisham has joined #hudson
[22:24:01] *** resmo has joined #hudson
[22:42:13] *** Lewisham has quit IRC
[23:04:48] *** khmarbaise_ has quit IRC
[23:47:29] *** Lewisham has joined #hudson

top