[00:08:07] <statlor> bbl
[00:08:13] *** statlor has quit IRC
[00:17:31] <cowmix> hudson keeps requiring me to accept a SSL cert on a commit...
[00:17:37] <cowmix> but the checkout works fine
[00:20:17] <Lewisham> paste the output into pastebin
[00:30:25] <cowmix> ok.. here it is: http://hudson.pastebin.com/d1a482fe7
[00:33:27] <cowmix> it always wants to reconfirm the cert
[00:33:38] <cowmix> even after I manually do it from the command line.
[00:38:04] <Lewisham> when you say "manually do it" do you mean accept the cert permanently yourself?
[00:40:03] <cowmix> yeah
[00:40:07] <cowmix> after the error..
[00:40:12] <cowmix> I go to the workspace
[00:40:24] <Lewisham> are you running as the same user that runs the Hudson server?
[00:40:24] <cowmix> actually... it doesn't even ask me to do it..
[00:40:28] <cowmix> yeah
[00:40:35] <cowmix> its on ubuntu
[00:40:42] <cowmix> running as hudson
[00:40:49] <Lewisham> and you log in as hudson?
[00:40:52] <cowmix> you can see.. the 'checkout' always works fine
[00:40:54] <Lewisham> to the shell?
[00:40:56] <Lewisham> yeah
[00:41:04] <cowmix> i do a 'su - hudson'
[00:41:08] <Lewisham> hmmm
[00:41:16] <cowmix> yeah.. weird stuff
[00:41:31] <cowmix> does hudson use the command line svn client?
[00:41:38] * Lewisham wouldn't know
[00:41:52] <Lewisham> I know more about sys-adminy stuff than Hudson :)
[00:41:59] <Lewisham> it should do
[00:48:22] *** abayer has quit IRC
[00:55:26] <calculus> hudson uses svnkit (a java implementation of svn)
[01:02:39] <Lewisham> perhaps we have found our culprit?
[01:02:55] <Lewisham> but it should take the global svn config, you would think
[01:07:08] <cowmix> calculus: does svnkit store its cert info in a different place than the command line version?
[01:16:48] <cowmix> this is weird
[01:17:04] <cowmix> hudson is running as 'hudson' but it thinks its home dir is /root
[01:18:09] <calculus> cowmix: not sure where the cert would be stored, but it is likely that 'hudson' user does not have write privleges to /root
[01:19:13] <cowmix> calculus: yeah.. I'm trying to figure out why Hudson thinks: HOME='/root'
[01:23:23] <rtyler> kohsuke: thinking more on the chroot idea for executing, I think it might be an ideal solution for the notion of "anonymous slaves"
[01:23:37] <rtyler> i.e. "contribute your cycles to help build Project X"
[01:37:26] <calculus> we are trying to do something similiar here by using the swarm-slave plugin and auto-installing the tools (although we only need java and maven)
[01:42:11] *** brunomlopes has quit IRC
[01:46:38] <cowmix> manually setting 'HOME' fixed the problem
[01:49:45] *** BigAllan has quit IRC
[01:54:58] *** admc has quit IRC
[02:27:57] *** josh_k has quit IRC
[02:57:38] *** alexlod has quit IRC
[02:58:15] *** alexlod has joined #hudson
[02:59:16] *** alexlod has quit IRC
[03:00:18] *** alexlod has joined #hudson
[03:00:43] *** admc has joined #hudson
[03:10:20] *** rromanchuk has quit IRC
[03:17:59] *** alexlod has quit IRC
[03:18:05] *** rromanchuk has joined #hudson
[03:29:41] *** thomk has quit IRC
[03:30:46] *** thomk has joined #hudson
[03:31:34] *** abayer has joined #hudson
[03:32:59] *** abayer has joined #hudson
[03:36:54] *** rromanchuk has quit IRC
[04:00:26] *** j2sentient has joined #hudson
[04:07:03] *** rromanchuk has joined #hudson
[04:07:45] *** abayer has quit IRC
[04:08:46] *** abayer has joined #hudson
[04:21:01] *** j2sentient has quit IRC
[04:38:34] *** admc has quit IRC
[05:08:11] *** glass-eye has quit IRC
[05:08:11] *** huimies has quit IRC
[05:08:11] *** AgentIcarus has quit IRC
[05:08:11] *** cowmix has quit IRC
[05:09:04] *** cowmix has joined #hudson
[05:09:04] *** glass-eye has joined #hudson
[05:09:04] *** huimies has joined #hudson
[05:09:04] *** AgentIcarus has joined #hudson
[05:17:26] *** AgentIcarus has quit IRC
[05:17:26] *** huimies has quit IRC
[05:17:26] *** glass-eye has quit IRC
[05:17:26] *** cowmix has quit IRC
[05:18:23] *** AgentIcarus has joined #hudson
[05:18:23] *** huimies has joined #hudson
[05:18:23] *** glass-eye has joined #hudson
[05:18:23] *** cowmix has joined #hudson
[06:06:15] *** mindless has quit IRC
[06:42:23] *** thomk has left #hudson
[07:11:26] *** admc has joined #hudson
[07:57:02] *** khmarbaise has joined #hudson
[08:23:27] *** khmarbaise has quit IRC
[08:54:34] *** jieryn has quit IRC
[09:20:34] *** keshureddyp1 has joined #hudson
[09:21:16] *** keshureddyp has quit IRC
[10:47:10] *** dvrzalik has joined #hudson
[11:57:46] *** admc has quit IRC
[12:39:26] *** dvrzalik has quit IRC
[12:42:43] *** dvrzalik has joined #hudson
[13:27:16] *** velo has joined #hudson
[13:33:11] *** akostadinov has joined #hudson
[13:58:10] *** abayer_ has joined #hudson
[13:58:10] *** abayer has quit IRC
[14:03:50] *** jdolan has quit IRC
[14:20:38] *** jdolan has joined #hudson
[14:20:45] *** jieryn-w has joined #hudson
[14:34:26] *** jkramer has joined #hudson
[14:34:28] <jkramer> Hello
[14:35:31] <jkramer> I'm using the hudson plugin mail-ext and I'm trying to configure a custom mail body. However, I can't find a list of supported tokens in the wiki and the help buttons in the configuration UI just give me 404
[14:35:44] <jkramer> Does anyone know how to obtain a list of supported tokens?
[14:44:27] *** jdolan has quit IRC
[15:03:14] *** jdolan has joined #hudson
[15:03:53] *** abayer_ has quit IRC
[15:17:00] *** abayer has joined #hudson
[15:19:50] *** kekko has joined #hudson
[15:20:32] *** akostadinov has quit IRC
[15:21:17] <kekko> i'm doing a post-commit from svn to hudson which fail with exit code 1 "Cannot write to `notifyCommit?rev=57' (Permission denied). Any idea why?
[15:28:11] *** akostadinov has joined #hudson
[15:33:20] <mart> is it using wget?
[15:33:32] <mart> perhaps wget is trying to write the result somewhere.
[15:46:09] <kekko> mart: yes , wget is trying to write somewhere. Got the full link but not sure where is pointing
[15:50:43] *** jkramer has left #hudson
[15:54:56] <mart> kekko:  usually writes to current working directory.  could try getting it to write to stdout, then piping to /dev/null?
[16:00:24] *** j2sentient has joined #hudson
[16:03:37] *** j2sentient has quit IRC
[16:12:24] <kekko> mart: i've tried to add a &> /dev/null at the end of the line, but does not change the result
[16:14:11] <mart> kekko: and tried -O- with it?
[16:16:22] <kekko> mart: so it becames ?
[16:16:35] <mart> er, what've you got so far?
[16:18:48] <mart> how about: wget -o/dev/null -O/dev/null  http://YOURHOST/hudson/job/PROJECTNAME/build
[16:19:37] <kekko> mart: http://servername/hudson/subversion/$(svnlook uuid $REPOS)/notifyCommit?rev=$REV &>null
[16:20:30] <mart> not seeing the wget command there, but try using both -o/dev/null and -O/dev/null
[16:20:41] <mart> one is log, the other is where the output gets written
[16:28:16] <kekko> mart: this is my hook http://pastebin.ca/1505785 , permission 755 but still no difference
[16:30:27] <kekko> is it writing a log somewhere?
[16:31:30] *** akostadinov has quit IRC
[16:32:07] <mart> that's what -o is specifying.
[16:32:29] <mart> but it shouldn't be trying to write the file if you've specified -O/dev/null
[16:33:04] <kekko> mart: i think it does not write it now , but in this way hudson does not start
[16:48:34] *** rromanchuk has quit IRC
[16:58:00] *** mindless has joined #hudson
[17:00:57] *** mdelaney has joined #hudson
[17:01:28] <mdelaney> Does anyone know of a good way to update a jobs config.xml from within a plug-in?
[17:17:47] *** mdelaney has quit IRC
[17:23:04] *** mdelaney has joined #hudson
[17:34:15] *** khmarbaise has joined #hudson
[17:36:59] *** mindless has quit IRC
[17:58:24] *** admc has joined #hudson
[18:00:45] *** jdolan has quit IRC
[18:13:00] *** dvrzalik has quit IRC
[18:35:30] *** SteveC_ has joined #hudson
[18:40:57] *** abayer_ has joined #hudson
[18:42:36] *** kekko has quit IRC
[18:53:59] <kohsuke> mdelaney: better to do it programmatically, instead of at XML level.
[18:54:23] <kohsuke> AbstractProject class exposes a lot of methods that let you mutate the project setting
[18:54:29] *** abayer has quit IRC
[18:57:28] *** alexlod has joined #hudson
[18:59:50] *** abayer_ has quit IRC
[19:14:06] *** SteveC_ has quit IRC
[19:15:39] *** mdelaney has quit IRC
[19:22:50] *** jtong has joined #hudson
[19:28:46] *** timp_ has quit IRC
[19:46:56] *** statlor has joined #hudson
[19:49:33] *** abayer has joined #hudson
[19:50:14] <abayer> Well, I think I just figured out doing incremental Maven builds through Hudson.
[20:07:13] * AgentIcarus dances and waves flags
[20:07:25] <mrooney1> kohsuke: okay, let me make sure I get this release right. My plug-in in the hudson plug-in dashboard doesn't have a description. Is there something in the pom that controls that?
[20:07:45] <mrooney1> under "Available" plug-ins, I am not sure what supplies that description
[20:07:56] <abayer> (a) You need to have the wiki page specified in your pom.
[20:08:06] <abayer> (b) The wiki page needs to start with an {excerpt}.
[20:08:13] <abayer> I believe that does the trick.
[20:08:37] <mrooney1> I have both of those, I think
[20:08:45] <abayer> Which plugin is this?
[20:08:50] <abayer> ncover?
[20:09:12] <mrooney1> yeah, I have "<url>http://wiki.hudson-ci.org/display/HUDSON/NCover+Plugin</url>"
[20:09:22] <mrooney1> abayer: can you confirm that it doesn't have a description?
[20:09:31] <abayer> Checking...
[20:09:45] <mrooney1> thanks!
[20:09:47] <abayer> (and comparing against the Clearcase plugin, which I know works)
[20:12:05] <statlor> I don't see a description of the ncover plugin in my available list
[20:12:13] <statlor> 0.2.5 is what it says is available
[20:12:28] <abayer> I'm checking the guts to see what, if anything, differs...
[20:12:33] <statlor> though I don't see one for Clearcase plugin either
[20:13:23] <abayer> Ok, I'm a little confused. More poking is needed. =)
[20:17:41] <abayer> ah. I was looking at the Installed list, not the Available. And Installed gets its description from the index.jelly for the plugin. Still trying to figure out the available.
[20:18:58] <abayer> Ok, this might be as silly as "don't have a new line right after the opening {excerpt}".
[20:19:33] <abayer> ...yes, I think it is.
[20:19:40] <abayer> That is kinda silly.
[20:22:59] <statlor> yeah the body of my except is on its own line (AccuRev plugin)
[20:23:05] <statlor> and it has a description in the avail plugins list
[20:23:28] <statlor> that text is also in the index.jelly under src/main/resources for the plugin
[20:23:33] <statlor> that must be for somewhere else
[20:23:59] <abayer> yeah, my text (for the ClearCase plugin) in index.jelly is different from the excerpt text on the wiki.
[20:25:29] <mrooney1> so I just need to remove that line?
[20:25:52] <abayer> Should do. Though I'm also going to change the pattern it uses to grab the excerpt to ignore leading newlines.
[20:27:33] <mrooney1> cool
[20:27:55] <mrooney1> when is loaded? per view or at startup or something?
[20:28:24] <abayer> when the update-center.js gets rebuilt - so whenever that cronjob of kohsuke's kicks off.
[20:28:36] <mrooney1> oh I see
[20:29:37] *** jbriguet- has joined #hudson
[20:46:15] *** jbriguet has quit IRC
[20:52:58] <kohsuke> and that's 4 times a day
[21:00:09] <abayer> fyi, I just committed the change to EXCERPT_PATTERN so that it should pull "excerpt text" from "{excerpt}[new line]excerpt text{excerpt}" anyway.
[21:05:09] *** admc has quit IRC
[21:06:58] <mrooney1> Hm, I wonder how mature the Clover coverage plug-in is
[21:07:17] <mrooney1> It seems to be a little buggy and lacking in information
[21:08:42] <bobo_> yeh i got abit disapointed
[21:08:59] <bobo_> seems cobertura is the best code coverage plugin
[21:09:44] <mrooney1> where do the "downstream" test results come from? the "build other projects" option?
[21:10:10] <bobo_> haventr tried it
[21:10:16] <bobo_> but sounds probable
[21:10:27] <abayer> Is that the clover plugin talking about downstream test results?
[21:17:56] *** jdolan has joined #hudson
[21:18:02] <mrooney1> abayer: those were unrelated questions
[21:18:13] <abayer> Wasn't sure. =)
[21:18:22] <mrooney1> I was just trying to figure out how to have one project execute as a result of another but show the actions in the parent
[21:18:37] <statlor> 317 going out today isn't it?
[21:19:43] <statlor> make sure that 4045 gets a big red dot beside it :)
[21:19:51] <statlor> the memory leak issue
[21:20:04] <statlor> there will be much rejoicing
[21:21:24] <abayer> Argh. Found a problem with my incremental Maven build in Hudson stuff. MavenBuildProxy confuses me.
[21:25:16] <mrooney1> Anyone know off-hand if I list multiple projects to build as downstream, do they just get thrown in the queue so they can execute in parallel?
[21:25:30] <abayer> I believe so.
[21:27:16] <mrooney1> how excellent, I assume I just comma separate them?
[21:27:39] <abayer> Ahyup.
[21:27:40] *** alexlod has quit IRC
[21:35:34] <mrooney1> hm I must be doing something wrong with this aggregating, it isn't working :[
[21:36:30] <abayer> I haven't done much with aggregating - are your downstream builds not firing right?
[21:38:55] <mrooney1> it is running and shows a test result with no failures
[21:39:11] <mrooney1> I click the upstream project link from it, and it shows no test result
[21:39:31] <abayer> Hrm.
[21:39:45] <mrooney1> and I enabled fingerprinting for both of them
[21:47:06] <mrooney1> kohsuke: is there any guide for setting up downstream projects? I have "/job/trunk-nightly/lastBuild/aggregatedTestReport/" telling me "trunk-browsertests                   (test result not available)"
[21:47:36] <mrooney1> But if I click the trunk-browsertests link, it shows me "Latest Test Result       (no failures)" just fine, and shows trunk-nightly as the upstream project.
[21:56:15] *** rromanchuk has joined #hudson
[22:03:38] *** Lewisham has quit IRC
[22:05:46] <kohsuke> abayer: I think the change to update-center is failing to compile.
[22:05:55] <abayer> Argh. Lemme check.
[22:06:05] <kohsuke> illegal escape character, javac says.
[22:06:11] <abayer> arrrrgh. My bad. Gimme a sec.
[22:06:18] <abayer> I checked in the wrong version.
[22:07:07] <abayer> Fix committed.
[22:07:10] <abayer> Sorry about that.
[22:10:18] <kohsuke> no problem. I do it all the time.
[22:10:46] <abayer> that's what I get for tossing that fix off while I was focused on the incremental build Maven stuff. =)
[22:12:09] <mrooney1> do I have to explicitly archive the test result xml file in my downstream project to get the upstream to show it?
[22:12:45] <mrooney1> to get "Aggregate downstream test results" to work
[22:12:55] <kohsuke> Yes.
[22:13:01] <kohsuke> No, not archiving
[22:13:07] <kohsuke> but you have to record them as JUnit reports
[22:15:06] <mrooney1> kohsuke: yeah I have my downstream set to "Publish JUnit test result report", and record fingerprints
[22:15:38] <mrooney1> Oh my fingerprint list is blank I guess I have to specify explicitly the test file
[22:16:05] <kohsuke> Not the fingerprints of the test files. Fingerprint is to track what artifacts from upstream is used in which build of the downstream
[22:16:12] <kohsuke> So you fingerprint the binary you grabbed from the upstream.
[22:16:29] <mrooney1> Grabbing a binary from upstream? What?
[22:17:23] <mrooney1> It seems like if A builds B and B has test results, and I tell A to aggregate downstream test results, isn't that all the info it needs?
[22:21:04] *** alexlod has joined #hudson
[22:27:28] *** Lewisham has joined #hudson
[22:31:14] <mrooney1> kohsuke: am I missing something there?
[22:34:29] <mrooney1> I can hack around it by archiving that file and wget'ing it in the upstream though that seems a bit silly
[22:40:25] *** admc has joined #hudson
[22:42:53] <mrooney1> Looks like others are confused too, there is http://www.nabble.com/Can-someone-explain-how-%22Aggregate-downstream-test-results%22-works--tt20947020.html#a20947020 which is unanswered
[22:43:17] <mrooney1> haha also http://www.nabble.com/Aggregate-Test-Results-td18699694.html
[22:51:41] <rtyler> USE THE SOURCE LUKE!
[22:52:20] <mrooney1> I'm reading https://hudson.dev.java.net/fingerprint.html but not understanding it
[22:53:35] <mrooney1> rtyler: you are familiar with downstream projects right? you have like hundrends? :)
[22:53:47] <mrooney1> In my downstream do I fingerprint the test result xml file?
[22:55:48] <rtyler> yeah, we have a lot
[22:56:54] <rtyler> admc: impart some wisdom damnit
[22:57:32] *** jdolan has quit IRC
[22:57:44] <admc> ah hmm
[22:58:24] <admc> were not really aggregating, we just have downstream jobs and those jobs produce junit XML and have their own report
[22:58:35] <admc> having one downstream fail doesn't make the initial job fail
[22:58:51] <admc> unless you are talking about the heirarchy internal to the tests
[22:59:02] <admc> so I don't really know that I can help with that use case
[22:59:16] <admc> although I do see the value, im sure I will want it at some point
[23:15:10] *** khmarbaise has quit IRC
[23:18:07] <mrooney1> admc: yeah I want to aggregate the test result of a bunch of downstream projects with junit XML, that is what you are doing?
[23:18:14] <mrooney1> I just can't figure out how to configure it to work
[23:18:56] <admc> so each job generates a file of junit output
[23:19:02] <admc> and I just check the box on the job for Publish JUnit test result report
[23:19:16] <admc> and tell it where to find it
[23:19:18] <admc>  Test report XMLs
[23:19:36] <admc> but I imagine what you could do, is generate the junit file on the file system
[23:19:41] <admc> and have all the down stream jovs append to it
[23:19:51] <admc> that way your main job could just read that one big file
[23:19:59] <admc> which would include data for all of the downstream projects
[23:20:10] <admc> we aren't doing the big aggregation piece yet
[23:20:26] <admc> I actually really like having it broken down by job for generating graphcs etc
[23:21:01] <rtyler> christ that's ugly
[23:23:51] <admc> yeah I wouldnt be surprised if there was a MUCH better way to do it
[23:23:53] <admc> I just don't know about it
[23:24:15] <admc> I bet I know who would know though
[23:24:27] <admc> the hudson big cheese, you around kohsuke ?
[23:24:38] <kohsuke> at your service, master.
[23:24:42] <admc> hahah
[23:24:48] <rtyler> big cheese?
[23:25:07] <rtyler> http://icanhascheezburger.files.wordpress.com/2009/03/funny-pictures-your-cat-is-disappointed-in-you.jpg
[23:25:08] <admc> you have any ideas about a nicer way to aggregrate all your down stream jobs into one set of results?
[23:25:28] <kohsuke> reading the back log now
[23:27:06] <kohsuke> I guess mrooney1 misunderstands the test report aggregation
[23:28:01] <mrooney1> oh, can you give me the understanding? :)
[23:28:57] *** statlor has quit IRC
[23:29:03] <admc> rtyler: that picture makes me feel guilty and I didnt even do anything :/
[23:29:38] * rtyler has a lolcat for every occasion
[23:30:56] <kohsuke> Having the upstream triggering the downstream build is not good enough
[23:31:20] <kohsuke> It actually doesn't give you any guarantee that the downstream is testing what the upstream produced
[23:31:37] <kohsuke> For example, there can be more builds of the upstream before the triggered downstream started.
[23:31:42] <kohsuke> That's where fingerprint kicks in.
[23:32:02] <kohsuke> If the downstream is testing something from upstream, record fingerprint of that something on both upstream and downstream
[23:32:09] <kohsuke> That allows Hudson to create a link between them.
[23:32:40] <kohsuke> Hudson should probably diagnose the situation better to explain this to you.
[23:32:52] <kohsuke> But that's why you are failing to aggregate the results.
[23:34:03] <mrooney1> kohsuke: okay, maybe I am using it wrong. Technically they could execute in parallel, the downstream doesn't need anything from the upstream
[23:34:36] <kohsuke> if it doesn't depend on anything from the upstream, how is it testing the upstream?
[23:34:40] <mrooney1> I am running phpunit tests and have two projects for javascript and selenium tests, which are testing against the php in a live environment. So I was thinking those could be separate projects which build with the upstream and the test results show together
[23:35:12] *** BigAllan has joined #hudson
[23:35:15] <mrooney1> it isn't really, I just thought it was a way to have smaller, cohesive/modular projects
[23:35:45] <mrooney1> I could have a large ant file and run the tasks in parallel but I'd rather have the logic in Hudson I think?
[23:36:26] <kohsuke> You can certainly have them as different projects, but then I don't understand the motivation for aggregating the results
[23:40:10] <calculus> probably just one place to get an overall view of how the testing is going
[23:40:30] <mrooney1> yeah, I have a trunk-nightly which should show phpunit, javascript, and selenium test results
[23:41:19] <mrooney1> but I don't necessarily want a monster project to maintain, and it would be cool to run the little ones themselves
[23:41:58] <abayer> Thanks for the reminder re: changelog, kohsuke - I always forget that I have that in a separate directory from the rest of my hudson tree and so my commits don't pick it up.
[23:42:53] <mrooney1> kohsuke: like I said I could have a large build.xml, though I'd rather move some of that logic to Hudson where I have a shiny UI and can more easily run the things separately if I desire :)
[23:44:26] <kohsuke> If you just want to get the "overall view of how the testing is going", maybe you can use some view plugins that can show graphs from multiple projects in one page.
[23:44:47] <kohsuke> Test result aggregation is for keeping track of what chanegs in the upstream broke what tests in the downstream.
[23:49:11] <mrooney1> technically that might still be what I want?
[23:49:36] <mrooney1> breakages in the JS or Selenium are caused by changes in the PHP code of the upstream
[23:52:13] <admc> yeah I totally see what your use case is, I haven't seen anyone running jobs like that
[23:52:31] <admc> makes sense though.. maybe you should write a hudson plugin that works like that :)
[23:55:23] <kohsuke> mrooney1: but for Hudson to track that, it needs to know exactly which build of the upstream a given downstream build is testing
[23:55:29] <kohsuke> That's only possible with fingerprinting.
[23:57:16] *** rromanchuk has quit IRC
[23:58:50] *** rromanchuk has joined #hudson

top