[00:05:07] <BlueFire_> kohsuke: I think I have narrowed down a reasonable test case
[00:05:20] <kohsuke> thanks!
[00:05:54] <BlueFire_> maybe not minimal, but reasonable.
[00:06:11] <BlueFire_> looks like if I remove the cobertura from the pom it works, otherwise not
[00:06:24] <BlueFire_> I can zip up the project and put it somewhere
[00:06:30] <kohsuke> It doesn't have to be very small.
[00:06:36] <kohsuke> Yes, please.
[00:08:52] <mrooney> kohsuke: I am mcrooney I believe, is that all you need?
[00:09:28] <kohsuke> yes, and I normally want to know what parts of Hudson you are interested in hacking
[00:10:25] <rtyler|opera> kohsuke: I believe mrooney was volunteering to fix core so one job could be run more than once at a time
[00:10:42] <mrooney> hmmm....
[00:10:56] <kohsuke> OK. For that I think we need to talk a bit more
[00:11:02] <mrooney> kohsuke: haha I assume that was a joke
[00:11:17] <mrooney> I am currently just interested in adding my NCover plugin
[00:11:22] <kohsuke> Oh, OK
[00:11:25] <kohsuke> That's easier
[00:11:38] *** admc has quit IRC
[00:11:51] <kohsuke> A part of the reasons I ask is because core changes require a bit of paperwork
[00:11:57] <kohsuke> Your commit access should be all set now
[00:12:07] <mrooney> excellent, thanks!
[00:12:22] *** admc has joined #hudson
[00:12:56] <mrooney> oh I have one question about enhancing the plugin, I thought I read somewhere there was a way to give a stylesheet to a page so it is "embedded" in Hudson such as the project page, instead of being a separate page which you have to keep hitting back until you get back to Hudson
[00:13:09] <mrooney> is something like that possible?
[00:13:56] <kohsuke> it's not really a stylesheet, but there are a set of tags
[00:14:04] <mrooney> right now I am using a DirectoryBrowserSupport (thanks for the setIndexFileName in 1.312), some sort of way to wrap that in Hudson would be neat
[00:14:13] *** brunomlopes has quit IRC
[00:14:13] <mrooney> oh okay, what do you mean by tags?
[00:14:20] <kohsuke> Jelly tags
[00:14:32] <mrooney> oh okay, are there any plugins which do this as an example?
[00:14:35] <kohsuke> No, those aren't designed to take arbitrary HTML and wrap that into Hudson's L&F
[00:14:46] <kohsuke> ... assuming that's what you are asking
[00:14:57] <mrooney> yeah, basically
[00:15:18] <mrooney> or a little bar at the top which has a hudson picture, which a link back to hudson
[00:15:22] <mrooney> so you can always get right back
[00:15:44] <kohsuke> Because we have jtidy in core, I suppose conceivably we can parse HTML on the fly and inject layout in the right place
[00:15:58] <kohsuke> maybe your next hacking project, mrooney :-)
[00:16:31] <mrooney> I'd be perfectly happy with just being able to supply an address and say, put this in an iframe within hudson or something
[00:16:34] <mrooney> haha, maybe :)
[00:16:53] <mrooney> EmbeddedDirectoryBrowserSupport :)
[00:21:48] <BlueFire_> kohsuke: https://hudson.dev.java.net/issues/show_bug.cgi?id=3911
[00:21:53] <BlueFire_> good luck ;)
[00:22:03] <kohsuke> thanks
[00:28:00] <calculus> kohsuke: did you get any of the PATH separators worked out? (sorry I had to catch the train yesterday when we were discussing this)
[00:43:39] <rtyler|opera> kohsuke: http://twitter.com/tehfailwhale/status/2301344671 any idea who this guy is or what he's talking about?
[00:44:21] <kohsuke> no clue
[00:46:38] <rtyler|opera> ditto
[00:46:39] * rtyler|opera shrugs
[00:47:32] <kohsuke> too much information in one tweet
[00:48:37] <rtyler|opera> heh
[00:48:46] <rtyler|opera> kohsuke: FWIW, just stumbled across this as well http://github.com/stephenh/git-central/blob/master/server/post-receive-hudson
[00:48:51] <rtyler|opera> haven't had a chance to deploy it here
[00:48:55] <rtyler|opera> but looks snazzy ;D
[00:50:08] <kohsuke> We should link from Wiki
[00:51:11] <kohsuke> Added as a comment
[00:52:13] <mrooney> okay so is "<version>1.0-SNAPSHOT</version>" where my plugin version should go?
[00:52:19] <mrooney> in the pom.xml
[00:53:01] <rtyler|opera> kohsuke: to the Git plugin page?
[00:53:38] <kohsuke> yes
[00:53:47] <kohsuke> mrooney: that's your plugin version
[00:54:06] <mrooney> okay so I can change that to just say, 0.1
[00:54:10] <rtyler|opera> you may or may not car, but stephenh is also the gentleman responsible for the git2 plugin that caused such a stir ;)
[00:54:13] <rtyler|opera> care*
[00:55:47] *** BlueFire_ has quit IRC
[00:57:19] <kohsuke> I forgot the technical details of the issue
[00:57:54] * rtyler|opera shrugs
[00:58:05] <rtyler|opera> doesn't really matter, it was an enjoyable argument nonetheless ;)
[00:59:16] <mrooney> How do I know if my plugin is "Maven2 Project Type Compatbile"?
[00:59:53] <kohsuke> I think that label lost its meaning
[00:59:57] <kohsuke> I should probably remove it
[01:01:07] <kohsuke> OK, removed.
[01:05:21] *** jdolan has quit IRC
[01:25:55] <calculus> I remember seeing a place in hudson to upload custom webpages/content
[01:26:20] <calculus> something like http://myhudson/personalContent
[01:28:17] <calculus> where is this location and I can I POST new data to it?
[01:33:47] <calculus> got it, it is http://myhudson/userContent
[01:34:10] <calculus> now the question is if I can POST new files to that url to upload them
[01:35:54] *** soominlee has joined #hudson
[01:37:33] *** jdolan has joined #hudson
[01:43:36] <mrooney> alright I've made a wiki page (http://wiki.hudson-ci.org/display/HUDSON/NCover+Plugin), now to push it to svn
[01:45:11] <mrooney> should I be doing or doing the mvn:release, I'm a bit confused
[01:53:10] <mrooney> kohsuke: any tips on that? do I need to manually add the stuff to svn AND do the mvn:release?
[01:55:23] *** jdolan has quit IRC
[02:04:42] *** jodell has joined #hudson
[02:34:57] <kohsuke> mrooney: http://wiki.hudson-ci.org/display/HUDSON/Release+Plugin
[02:35:04] <kohsuke> sorry
[02:35:05] <kohsuke> wrong link
[02:35:24] <kohsuke> see http://wiki.hudson-ci.org/display/HUDSON/Hosting+plugins
[02:36:20] <mrooney> kohsuke: that's where I am
[02:36:58] <mrooney> I don't understand if committing to svn and mvn release tasks are mutually exclusive or what the difference is though
[02:37:19] *** rromanchuk has quit IRC
[02:37:30] <kohsuke> svn commits and releases are independent
[02:39:19] <mrooney> ah okay, so I do one and then the other? Commits don't trigger a release until I tell it the next version is ready/stable by using mvn release?
[02:49:40] *** admc has quit IRC
[02:49:59] *** admc has joined #hudson
[02:56:32] *** dragon3_away is now known as dragon3
[02:59:27] <calculus> how does Project based Matrix Security differ from plain Matrix Security (they have the same description from the help dialog)?
[03:15:06] <calculus> how can I delete users from hudson's own database?
[03:15:18] <calculus> I created some test user accounts that I would like to get rid of
[03:17:14] <mrooney> kohsuke: okay, I requested a role for maven2-repository, I assume I need one to deploy?
[03:17:38] <kohsuke> you already have it
[03:29:33] <jodell> is anyone using the twitter plugin?
[04:10:57] *** jieryn-w has quit IRC
[04:16:15] <mrooney> excellent okay, thanks
[04:16:52] <mrooney> kohsuke: I can't figure out what "Missing required setting: scm connection or developerConnection must be specified." means then, I've added all the necessary things to my pom.xml as far as I can tell
[04:17:01] <mrooney> on doing mvn release:prepare
[04:22:51] *** davidstrauss has joined #hudson
[04:27:50] *** admc has quit IRC
[04:44:50] *** wsmoak has joined #hudson
[04:56:45] <jodell> quit
[04:56:47] *** jodell has quit IRC
[05:13:25] *** mrooney has quit IRC
[05:31:48] *** davidstrauss has quit IRC
[05:31:48] *** soominlee has quit IRC
[05:32:29] *** davidstrauss has joined #hudson
[05:32:29] *** soominlee has joined #hudson
[05:32:52] *** mirage` has quit IRC
[05:32:52] *** btQuark has quit IRC
[05:32:52] *** kohsuke has quit IRC
[05:32:52] *** maxb_ has quit IRC
[05:32:53] *** calculus has quit IRC
[05:32:53] *** usv has quit IRC
[05:33:24] *** mirage` has joined #hudson
[05:33:24] *** btQuark has joined #hudson
[05:33:26] *** kohsuke has joined #hudson
[05:33:26] *** maxb_ has joined #hudson
[05:33:26] *** calculus has joined #hudson
[05:33:26] *** usv has joined #hudson
[05:43:40] *** mindless has quit IRC
[06:01:57] *** admc has joined #hudson
[06:14:12] *** akostadinov has joined #hudson
[06:20:09] *** dmulter has joined #hudson
[06:20:42] *** dmulter has quit IRC
[06:22:35] *** decriptor_ has joined #hudson
[06:30:48] *** admc has quit IRC
[06:35:02] *** akostadinov has quit IRC
[06:46:39] *** waz has quit IRC
[06:57:41] *** wsmoak has quit IRC
[07:00:34] *** akostadinov has joined #hudson
[07:09:27] *** decriptor_ has quit IRC
[07:11:33] *** akostadinov1 has joined #hudson
[07:28:01] *** akostadinov has quit IRC
[07:35:14] *** khmarbaise has joined #hudson
[07:37:25] *** khmarbaise has quit IRC
[07:38:56] *** akostadinov1 has quit IRC
[09:01:10] *** Denogini1er is now known as Denoginizer
[09:34:30] *** khmarbaise has joined #hudson
[10:06:02] *** arthru has joined #hudson
[10:07:39] <arthru> hi
[10:07:53] <arthru> I found a bug in hudson
[10:08:04] <arthru> i copyed the log here : http://hudson.pastebin.com/m27eb3504
[10:08:25] <arthru> the error happens when trying to deploy a maven build to a repository
[10:08:36] <arthru> it happens since Hudson 1.311
[10:08:53] <arthru> i tried 1.312 this morning and the problem is still hapening
[10:15:13] *** maxb_ is now known as maxb
[10:55:20] <usv> oh, 1.312 is out
[11:01:38] <usv> I posted this in the users list regarding hudson, svn and p12 : http://www.nabble.com/Problem-with-subversion-authentication-with-p12-key-td24170558.html
[11:02:33] <usv> I'm currently getting hudson trunk to try to identify the problem; it might have something to do with the following
[11:02:36] <usv> X509KeyManager passed to SSLContext.init():  need an
[11:02:39] <usv> X509ExtendedKeyManager for SSLEngine use
[11:12:50] *** khmarbaise has quit IRC
[11:19:04] <usv> maybe I should try 1.312 first and then post to issue tracker
[12:07:27] *** gits1 has joined #hudson
[12:36:45] <usv> posted as issue 3913: https://hudson.dev.java.net/issues/show_bug.cgi?id=3913
[12:46:15] *** timp has joined #hudson
[12:55:25] *** dragon3 is now known as dragon3_away
[13:14:50] *** gits1 has quit IRC
[13:17:52] *** gits1 has joined #hudson
[13:56:56] *** mirage` has quit IRC
[13:56:56] *** btQuark has quit IRC
[13:57:14] *** mirage` has joined #hudson
[13:57:14] *** btQuark has joined #hudson
[14:01:00] *** gits1 has quit IRC
[14:03:17] *** akostadinov has joined #hudson
[14:03:41] *** gits1 has joined #hudson
[14:05:58] *** waz has joined #hudson
[14:08:33] *** jieryn-w has joined #hudson
[14:10:29] *** mirage` has quit IRC
[14:10:29] *** btQuark has quit IRC
[14:10:45] *** mirage` has joined #hudson
[14:10:45] *** btQuark has joined #hudson
[14:49:43] *** khmarbaise has joined #hudson
[14:53:22] *** wsmoak has joined #hudson
[15:05:28] *** jdolan_ has joined #hudson
[15:11:25] *** khmarbaise has quit IRC
[15:22:09] *** gits1 has quit IRC
[15:23:29] *** wsmoak has quit IRC
[15:24:19] *** oxi has joined #hudson
[15:25:11] *** oxi has quit IRC
[15:34:43] *** oxi has joined #hudson
[15:35:43] *** oxi has quit IRC
[15:49:24] *** kstreith has joined #hudson
[15:50:26] *** timp has quit IRC
[15:53:45] *** waz has quit IRC
[15:54:58] *** waz has joined #hudson
[16:22:56] *** mindless has joined #hudson
[16:35:32] *** davidstrauss has quit IRC
[16:50:08] *** davidstrauss has joined #hudson
[17:10:11] <glass-eye> can someone take a look at this http://pastehere.com/?mkpxqp and maybe give me some pointers on figuring our why the "dead" task is there?
[17:24:20] *** rromanchuk has joined #hudson
[18:27:55] *** khmarbaise has joined #hudson
[18:29:14] *** wsmoak has joined #hudson
[18:31:41] *** gits1 has joined #hudson
[18:33:26] *** davidstrauss has quit IRC
[18:49:31] *** PiPoCidE has joined #hudson
[18:50:24] <PiPoCidE> anyone have problem with svn & symlinks?
[18:55:33] <usv> I have problems with svn + https
[18:55:46] <usv> debugging for 3 days now
[18:56:25] <PiPoCidE> im getting regular files instead of symlinks when it check's out the source for the project
[18:56:54] <PiPoCidE> upgraded from 1.304 to 1.312
[19:00:42] <rtyler|opera> PiPoCidE: wait, why do you expect symlinks? o_O
[19:01:22] <PiPoCidE> the project has some prebuilt lib's + symlink in the tree
[19:01:26] *** rtyler|opera has left #hudson
[19:01:30] *** rtyler|opera has joined #hudson
[19:01:33] <rtyler|opera> goddamnit
[19:01:40] <PiPoCidE> :)
[19:01:46] <rtyler|opera> keep forgetting Ctrl+W in Opera isn't "back one word"
[19:02:03] <rtyler|opera> PiPoCidE: you have symlinks in your SVN tree?
[19:02:16] <PiPoCidE> yes and some prebuilt libraries
[19:02:21] <rtyler|opera> glass-eye: do you have that sidebar link plugin installed?
[19:02:33] <rtyler|opera> PiPoCidE: I'm not entirely sure if SVNKit supports symlinks
[19:02:41] <rtyler|opera> I can't imagine what they're supposed to be on windows
[19:03:11] <PiPoCidE> i would imagine just real copies of the file
[19:03:33] <rtyler|opera> I would imagine that's asking a lot of SVNKit ;)
[19:04:37] *** akostadinov has quit IRC
[19:04:56] *** mrooney has joined #hudson
[19:05:19] <rtyler|opera> gooood mooooorning mrooooney
[19:05:43] <mrooney> good morning!
[19:06:06] <PiPoCidE> looks like svnkit supports it
[19:06:10] <mrooney> do you know anything about plug-in licensing, are plug-ins in the repo required to be MIT?
[19:06:11] <PiPoCidE> but im guessing its disabled
[19:06:21] <rtyler|opera> mrooney: it's generally appreciated
[19:06:40] <rtyler|opera> mrooney: I don't think we can deal with GPL, but LGPL works
[19:06:42] <rtyler|opera> or BSD
[19:06:50] <rtyler|opera> or Beerware license
[19:07:02] <mrooney> rtyler|opera: okay, for some reason the NCover guys said "please tell me this is not released as GPL or LGPL or something like that" so I don't know what their concerns are
[19:07:13] <mrooney> as far as I understand it, MIT is more liberal than GPL
[19:07:14] <rtyler|opera> sketchy
[19:07:18] <rtyler|opera> way more liberal
[19:07:23] <rtyler|opera> BSD 2-clause ~= MIT
[19:07:33] <mrooney> they said if it is something other than GPL it should be good but I wonder why
[19:07:42] <rtyler|opera> GPL is infectious
[19:07:52] <rtyler|opera> particularly poorly suited for a plugin
[19:07:55] <mrooney> the only thing that would be bad is GPL Affero for a plug-in like this I would think
[19:08:02] <mrooney> I don't see why GPL is harmful
[19:08:08] <mrooney> but I don't know that much after all :)
[19:08:12] <rtyler|opera> Affero is teh lame too
[19:08:23] <mrooney> but it closes an important loophole! :)
[19:08:53] <rtyler|opera> the useful one? :D
[19:09:48] <mrooney> haha
[19:11:12] <rtyler|opera> mrooney: the general problem that most folks have with GPL is that it is restrictive on other developers' freedoms, so in this case, if you GPL'd the plugin, kohsuke and Snoracle couldn't commercially package Hudson and include (AFAIK) unless they had copyright
[19:11:34] <mrooney> yeah that is annoying in this case
[19:11:39] * rtyler|opera is generally a fan of BSD 3-clause and MIT
[19:11:55] <mrooney> for my person projects I tend to prefer GPL because I don't want someone commercially packaging my stuff without my permission
[19:12:21] <kohsuke> mrooney: unless you have reasons to choose something else, I generally recommend sticking to MIT in Hudson
[19:12:26] <mrooney> but for this kind of stuff MIT sounds good
[19:12:33] <mrooney> nope I don't think, I was mostly just curious
[19:12:35] <kohsuke> It simplifies things for users and other contribuers to have a fewer licenses.
[19:13:08] <mrooney> I was worried about it being too restrictive which I don't think is a problem with MIT
[19:15:12] <rtyler|opera> did you know, MIT licensed code just smells better too!
[19:16:04] *** rtyler|opera has left #hudson
[19:16:10] *** rtyler|opera has joined #hudson
[19:16:12] <rtyler|opera> gahhhhhhhhhhhh
[19:20:58] <mrooney> is there any infrastructure set up for multiple instances of a plug-in?
[19:21:17] <mrooney> like if I wanted to show two coverage reports, it would be cool to be able to just enable two instances of my plugin
[19:21:27] *** abayer has joined #hudson
[19:21:27] <PiPoCidE> anyway to modify svnkit's default behavior?
[19:26:07] <glass-eye> rtyler|opera: let me look
[19:26:37] <glass-eye> rtyler|opera: yes, the sidebar link plug-in is installed
[19:27:46] <rtyler|opera> glass-eye: that just looks like a broken sidebar link
[19:27:53] <rtyler|opera> or like you have it installed ,but none configured
[19:27:54] <glass-eye> rtyler|opera: ok, I'll take a look, thanks
[19:28:09] <glass-eye> not a huge deal...I was just wondering what was up :)
[19:35:18] *** matt-wynne has quit IRC
[19:37:47] *** davidstrauss has joined #hudson
[19:38:41] *** rtyler|opera has left #hudson
[19:43:04] <glass-eye> rtyler: I removed the sidebar link plugin and it went away, I wasn't using it anyway so no harm
[19:43:05] <glass-eye> thanks
[19:58:04] *** rasilon has joined #hudson
[19:59:02] <rasilon> hi
[20:00:47] <rasilon> I have the need to split a job into smaller jobs. The user needs to enter some parameters for job #1 and I need jobs #2, #3.... to inherit the same parameters
[20:00:52] <rasilon> is this possible ?
[20:05:21] <rasilon> ah, just found the page saying you can't do this. apologies for the intrusion ;)
[20:10:38] *** velo has joined #hudson
[20:11:22] <mrooney> hm I've lost my plugin checkout, how embarrassing
[20:12:49] <velo> kohsuke, hi, FWIW, I found a way to workarund the package filtering.... right now is a negative filter.... everything on classpath (maven CP doesn't include rt.jar) won't be checked ... I will create a possitive checking, so users will be able to specify witch jar is in check... if not specified will behave like it is right now, otherwise will use this positive filter
[20:13:13] <velo> kohsuke, not sure when I will have time to do, but at least I have an idea to work with
[20:13:19] <velo> kohsuke, what do you think about it?
[20:14:00] <kohsuke> that information should be a part of the signature file
[20:16:12] *** arthru has left #hudson
[20:16:29] *** rasilon has quit IRC
[20:28:23] <rtyler> kohsuke: so I'm thinking more today about highly concurrent job runs
[20:28:43] <rtyler> keeping external jobs and some external pool manager in my back pocket as a last resort
[20:29:03] <rtyler> I'm assuming there's no current push to allow for a job to have two executions running at the same time? :/
[20:30:24] <kohsuke> I suppose you mean if anyone is planning to work on it right now
[20:30:29] <kohsuke> and if so, not that I know of.
[20:30:41] <kohsuke> Of course there's a push for any feature (with different degree of urgency)
[20:33:49] <rtyler> well, so here's my thinking right now
[20:33:52] <rtyler> I want continuous deployment
[20:33:54] <rtyler> I want it bad
[20:34:17] <rtyler> but we can't get that because our test suite takes 20 minutes and only one job can run at a time
[20:34:53] <rtyler> so pusher A queues the release job, then pusher B does, and by the time pusher B's job is done, it's almost been an hour
[20:35:45] <rtyler> my backup plan is to write a queue manager in Python that will queue external jobs
[20:35:59] <rtyler> err, starts jobs externally (not hudson external jobs)
[20:36:06] <rtyler> which will be configured to report to the external job
[20:36:09] <abayer> I'm still totally confused as to what you're trying to do, but hey. =)
[20:36:28] <rtyler> abayer: think of it like this, 1 commit, 1 job queued and run
[20:37:02] <rtyler> abayer: we have ~150-200 commits merged into trunk a day, with Hudson as it currently stands, there's no way we could have that level of atomikc job/commit running
[20:37:06] <rtyler> just not enough time
[20:37:35] <mrooney> hm, how do I provide a default for a plug-in setting? does that happen in the config.jelly or the class?
[20:37:39] <abayer> Longer polling window doesn't do the trick?
[20:38:23] <rtyler> abayer: then we get a bunch of commits in one job, makes it harder to isolate the reason for failures
[20:38:43] <rtyler> we also don't do polling at all for the trunk job, whenever somebody does a git push in that branch, it's queueued
[20:38:48] <rtyler> queueueueueueueud
[20:38:57] <abayer> Ok, that's totally lisp-like spelling.
[20:45:54] <rtyler> abayer: do you at least understand what I'm going for?
[20:46:34] <abayer> Running
[20:46:41] <abayer> multiple instances of the same build at the same time.
[20:46:42] <abayer> yes?
[20:46:51] <rtyler> pretty much
[20:47:02] <abayer> That's a weird one.
[20:47:11] <rtyler> why?
[20:47:23] <rtyler> if your build takes 1 hour to complete
[20:47:36] <rtyler> do you want to max out at 8-10 builds during a workday?
[20:47:48] <abayer> Yes.
[20:48:03] <abayer> Because having multiple builds running at once is potentially confusing.
[20:48:05] <rtyler> why? you have far more than 8-10 commits?
[20:48:33] <abayer> I think it's simpler to deal with multiple commits in a given build.
[20:48:52] <rtyler> I disagree wholeheartedly
[20:49:10] <rtyler> <o3 (<-- ex)
[20:49:36] <abayer> *shrug*
[20:49:37] <rtyler> if your commit history is linear, which is sure as shit should be
[20:49:47] <rtyler> what does it matter if 12 builds are running at the same time?
[20:49:56] <rtyler> they'd still very likely complete linearly
[20:50:04] <abayer> Load, external dependencies, easily confused developers. =)
[20:50:18] <rtyler> abayer: I stopped worrying about confused developers ages ago
[20:50:43] <rtyler> Load: slaves, dependencies: shouldn't matter IMHO
[20:51:04] <rtyler> I can't get those confused developers to write unit tests, so fuk em :P
[20:51:20] <abayer> See, I'd say that if your build takes 1+ hour and you really want to have an isolated build per commit, then the build needs to get split up into multiple components with shorter builds, etc.
[20:52:05] <abayer> Something just doesn't quite seem right about what you're talking about. *shrug*
[20:52:19] <rtyler> in our case, "build" really means a 5 minute deploy, and then regressing the fuxor out of it with our test suite
[20:52:30] <abayer> Same here.
[20:52:56] <abayer> Well, 15 minute compile, a couple hours of tests.
[20:53:02] <abayer> But yeah, same basic thing.
[20:53:10] <rtyler> abayer: how often are you guys releasing?
[20:53:18] <abayer> Nowhere near as often as you are.
[20:53:34] <rtyler> heh
[20:54:09] <rtyler> we get a good 8-20 a day
[20:54:16] *** glass-eye has quit IRC
[20:54:51] <rtyler> but we only get about 5-8 actual executions of the trunk job :/
[20:54:54] <abayer> Well, if nothing else, you're a distinctly weird case - when the window between releases is shorter than the build time, that's kinda odd.
[20:55:38] <abayer> Not a situation that I've ever seen before, that's for sure.
[20:55:53] <rtyler> I don't think concurrent jobs is just a solution for that
[20:56:00] <rtyler> IMHO that's how things "should work"
[20:56:17] <rtyler> having a job only execute one at a time is odd to me conceptually
[20:56:25] <abayer> I vehemently disagree. For example, that just straight up won't work with ClearCase, at all.
[20:56:42] <abayer> Or any other case where you're updating a workspace rather than recreating it every time.
[20:59:16] <abayer> At the very least, there's no way that concurrent instances of a single job should be default behavior.
[21:02:51] <rtyler> hmm
[21:06:37] <abayer> While concurrent instances may seem like how things should work by default to you, it's not how most build processes have historically worked, so yeah, a lot of processes may not work properly in that context. If your SCM/tests/process are designed with the assumption that only one build can run at a time, there could be a LOT of problems if suddenly multiple builds could runa t once.
[21:07:53] <rtyler> well, first things first ;)
[21:08:01] <rtyler> it's not even possible right now :P
[21:08:14] *** brunomlopes has joined #hudson
[21:09:24] <abayer> true. But don't be shocked if there's not a lot of interest in concurrent instances of a build as a feature.
[21:11:24] <rtyler> yeah, I don't see people recognizing the value of if unless they're in CVS/SVN/Git/Hg land, and having an active enough code base that has these sorts of choke points
[21:11:48] <abayer> And an active enough release schedule that it would matter.
[21:24:59] *** khmarbaise has quit IRC
[21:27:45] *** waz has quit IRC
[21:28:11] *** waz has joined #hudson
[21:31:58] <mrooney> Hooray I finally figured out how to use locks, now I can have more than 1 executor!
[21:32:54] <rtyler> beware!
[21:34:41] <mrooney> of?
[21:34:55] <rtyler> being eaten by a grue?
[21:35:05] <mrooney> but I love grues
[21:35:07] <rtyler> you're referring to the locks and latches plugin right?
[21:35:11] <mrooney> yes indeed
[21:35:18] *** admc has joined #hudson
[21:35:20] <mrooney> though there aren't latches are there
[21:35:34] <mrooney> more of just a "locks" plugin, which a very ambiguous changelog also
[21:35:36] <mrooney> does anyone maintain it?
[21:35:50] <rtyler> not sure, it's "quirky"
[21:36:03] <mrooney> oh no.
[21:36:03] <rtyler> admc can probably tell you more about using locks and latches more successfully than I
[21:36:08] *** kstreith has quit IRC
[21:36:17] <rtyler> admc: drop some knowledge son
[21:36:24] <mrooney> by the way do you know how to set a default value for a plugin?
[21:36:39] <mrooney> do I do that in the config.jelly, or in the class somewhere
[21:36:42] <rtyler> another admc question!
[21:36:51] <mrooney> oh really?
[21:36:57] <mrooney> I just mean generically
[21:37:00] <admc> mrooney: what specifically are you trying to do
[21:37:09] <rtyler> heh
[21:37:34] <mrooney> admc: so for a post-build plugin I have some options in the configuration UI, I want a textbox to default value, instead of being blank
[21:37:49] <admc> as for default values in the config.jelly you can, the way to do is different depending on the kind of form element
[21:37:58] <mrooney> I guess I should look at the cobertura perhaps
[21:38:00] <mrooney> just a string
[21:38:51] <admc> well, it appears the way that works is that you can set a "value" property on the jelly tag
[21:38:55] <admc> which will stick that value in there
[21:39:06] <admc> however reading it back out becomes an issue
[21:39:28] <admc> I did a bunch of reserach on this because I need to default a checkbox to true
[21:39:52] <admc> the problem is that if I set it to true, it doesn't receive the state again after someone changes it
[21:39:57] <mrooney> right
[21:40:00] <admc> its super annoying but the docs are really unclear
[21:40:28] <admc> so my guess is that you have to do it with code
[21:40:56] <admc> what are you trying to do with locks?
[21:41:19] <mrooney> oh nothing that doesn't work, just the super simplest use case
[21:41:42] <mrooney> I have a few projects and only one can execute at a time but all other projects are fine, I gave those few the same lock and it seems to be working fine
[21:41:53] <admc> your just trying to keep jobs from stepping on each other when running on a shared slave?
[21:42:16] <mrooney> yeah we have like 6 projects, three of which are branches of the same thing and can't run at the same time
[21:42:28] <admc> right, yeah locks make good sense for that
[21:43:17] <mrooney> yes indeed
[21:43:23] <mrooney> now for the default, there must be plugins that do it
[21:43:37] <admc> if you find it, please let me know
[21:44:11] <mrooney> cobertura seems to have some defaults
[21:44:35] <mrooney> also "Aggregate downstream results" has a checkbox which is checked by default
[21:44:37] <mrooney> so you could look at that for yours
[21:44:55] <mrooney> I don't know what plugin that is though or it is one
[21:45:40] <admc> yeah ill have to spend some time this afternoon digging back into that
[21:49:41] <mrooney> wow I keep getting confused between this channel and the "Mr Hudson" jabber user which IMs me build information
[21:53:20] <rtyler> hah
[21:53:26] <rtyler> we use a jabber conference
[21:54:14] <mrooney> admc: looks like something to do with getDefaultTargets on a Descriptor? https://svn.dev.java.net/svn/hudson/trunk/hudson/plugins/cobertura/src/main/java/hudson/plugins/cobertura/CoberturaPublisher.java
[21:54:47] *** PiPoCidE has quit IRC
[21:55:48] <admc> erm
[21:56:01] <admc> the way that is done is so unintuitive for me
[21:56:04] <admc> :/
[21:57:34] *** brunomlopes has quit IRC
[21:59:34] <mrooney> unintuitive > impossible
[21:59:47] <mrooney> also kohsuke can probably give us some tips, I bet that is deprecated and there is a better way
[21:59:47] <admc> hehe true
[22:04:38] <admc> that would be awesome
[22:11:30] *** brunomlopes has joined #hudson
[22:13:55] *** admc_ has joined #hudson
[22:19:25] *** brunomlopes has quit IRC
[22:25:13] *** admc has quit IRC
[22:38:11] <calculus> rtyler: I have a similar process, our tests are 30-50 minutes (depending on which slave gets the job) and would like a test per commit per job
[22:38:34] <calculus> but many times there are 4-5 commits for a build
[22:50:20] *** admc_ has quit IRC
[23:07:40] <mrooney> hmm isn't there a hudson stylesheet somewhere that I can use?
[23:09:16] *** admc has joined #hudson
[23:18:47] *** jdolan_ has quit IRC
[23:23:56] <mrooney> kohsuke: any tips on having a default value for a checkbox or text field of a plugin option?
[23:24:39] <kohsuke> default="true"
[23:24:41] <kohsuke> etc.
[23:34:39] *** abayer has quit IRC
[23:45:05] *** davidstrauss has quit IRC
[23:47:23] <mrooney> oh how simple :)
[23:48:19] <mrooney> now if I can figure out how to return a boolean so the checkbox understands it
[23:48:23] <mrooney> it doesn
[23:48:28] <mrooney> 't seem to want a boolean type

top