[11:45] <Riddell> Ubuntu Desktop and Edubuntu being respun
[11:51] <bkerensa> :)
[11:51] <bkerensa> \o/ ISO Testing
[14:16] <brendand> jibel - ping
[14:41] <jibel> brendand, pong
[14:41] <brendand> jibel, do you know is there a way around the problem with creating i386 live images on an amd64 host?
[14:42] <brendand> jibel, i'd love to start iso testing, but my laptop is running amd64 now
[14:42] <brendand> and the only test systems i have are atom netbooks
[14:51] <jibel> brendand, what's the problem ?
[14:54] <brendand> jibel - when i create a usb stick on my laptup with usb-creator-gtk, even though the image is i386 i get the 'this is not a 64-bit host' error on the installation system. i guess syslinux is 64-bit
[14:54] <jibel> brendand, with virtualbox ?
[14:55] <brendand> jibel, no - bare metal
[14:57] <jibel> brendand, I never faced this bug. Did you report it or bug number you know ?
[14:57] <brendand> jibel - i'll try again. if it still happens i'll raise a bug
[14:58] <jibel> brendand, anything in your bios that could prevent booting x86 images ?
[14:59] <jibel> New Ubuntu images posted to the tracker
[14:59] <brendand> jibel - eh, the system i'm booting on has only an atom processor...
[15:00] <jibel> brendand, same here with an N450
[15:01] <jibel> it's booting an i386 image created from an x86_86 machine
[15:01] <jibel> x86_64 even
[15:04] <xdatap1> jibel, hello. I just received the email from the tracker. The download links still says "lucid"
[15:05] <jibel> xdatap1, hello. stgraber fixed it and we are waiting for a code push
[15:05] <jibel> which should happen any time soon
[15:05] <xdatap1> jibel, thanks. Starting the download anyway :)
[15:05] <brendand> jibel - should i be a bug? i mean usb-creator should install syslinux to match the architecture of the image, not the host right?
[15:05] <jibel> xdatap1, but you should see precise just below lucid ?
[15:06] <stgraber> it's actually been fixed weeks ago ;)
[15:06] <stgraber> and IS was supposed to do the rollout last week
[15:06] <stgraber> which they kind of did
[15:06] <stgraber> they just missed a final "bzr up" so we're still on the buggy code
[15:06] <jibel> brendand, yes please file a bug, I can't reproduce it locally.
[15:07] <xdatap1> jibel, I'm talking about this page: http://iso.qa.ubuntu.com/qatracker/milestones/208/builds/12594/downloads
[15:07] <stgraber> xdatap1: the duplicate (lucid + precise) there has been fixed in trunk for weeks ;)
[15:08] <stgraber> xdatap1: it's just a matter of having IS run the bzr up
[15:08] <xdatap1> stgraber, ok cool. No problem, just for letting you know ;)
[15:09] <brendand> jibel, ok. seems fine now! either i made an oopsie and accidentally selected an amd64 image, or usb-creator was momentarily bugged last week
[15:10] <jibel> brendand, cool. this bug was scared just talking about it :)
[15:11] <stgraber> jibel, xdatap1: I just poked IS directly. Someone is having a look now, so hopefully it should be all fixed in a few minutes
[15:12] <xdatap1> stgraber, great. You rock :)
[15:12] <xdatap1> stgraber, to be honest, you normaly rock, it's not just about the tracker :D
[15:13] <stgraber> xdatap1: fixed
[15:13] <xdatap1> stgraber, yep, it works, great
[16:59] <balloons> AlanBell, how'd that orca test end up?
[17:14] <AlanBell> orca got a bit broken with stuff landing for feature freeze I think, going to give it another go now, I want to make sure the test cases are somewhat realisticly do-able
[17:24] <balloons> kk, sounds good AlanBell
[17:42] <balloons> roadmr, cr3 quick question on checkbox test format.. what's the best way to list assumptions for the test case?
[17:45] <cr3> balloons: "assumptions" as in "dependencies on other tests" or "requirements on devices and/or packages"?
[17:45] <jibel> stgraber, when you have a chance, could you check ltsp-live. When I boot the client, I'm dropped to busybox
[17:46] <balloons> more like requirements cr3.. in other words, if I have a test talking about using a camera, I have an assumption that you have a camera :-)
[17:46] <stgraber> jibel: argh ... yeah will have a look, I actually think I know what's wrong
[17:46] <stgraber> at least that should be easy to fix
[17:46] <stgraber> (and shouldn't affect LTSP on alternates)
[17:46] <balloons> I have some dependencies on other tests inside tests as well, but generally I just run them in order.. I think that's fine for now
[17:47] <cr3> balloons: requires: device.category == 'CAPTURE'
[17:48] <balloons> ohh while I have you.. I have a number of these errors in my logs I'm trying to fix;  Exception: Template <stream> has a duplicate field 'plugin' with a new value 'manual'.
[17:48] <balloons> what could be wrong with my job if I'm getting that error?
[17:48] <cr3> balloons: don't necessarily assume tests will run in a particular order, if a test really makes more sense to run before or after another one, please make sure to specify depends: name_of_other_test
[17:49] <balloons> cr3, ohh cool.. in general I wasn't depending.. There was only one suite that has that dependency
[17:49] <cr3> balloons: probably a trailing space, make sure to separate your job definitions with an empty line, not a line with some spaces :)
[17:49] <balloons> it was lower priority to figure out. I'll add depends
[17:49] <balloons> ouch trailing spaces.. kk strippin
[17:49] <cr3> balloons: the parser could be more intelligent about that, I'll take a note for future improvement
[17:50] <cr3> ie, for the checkbox-core parser at very least
[18:29] <balloons> cr3, hmm that didn't seem to change anything.. still getting the same error
[18:47] <balloons> cr3, roadmr could you look at my branch again and see what might be up? I would appreciate it bzr branch lp:~nskaggs/+junk/checkbox-app-testing
[18:57] <bdmurray> jibel: can you tell me more about what happens with debconf in bug 940252?
[19:08] <roadmr> balloons: could you add a newline at the *end* of each of your job files and see if it helps fix the problem?
[19:09] <roadmr> balloons: basically what happens is that checkbox cats all the jobfiles together, if the result has duplicate keys within the same blankline-delimited block, then you'll get duplicate field errors
[19:15] <balloons> roadmr.. sure
[19:15] <balloons> makes sense
[19:16] <balloons> sweetness!
[19:16] <balloons> that looks like it did it
[19:16] <roadmr> balloons: we really need some kind of checkbox-lint tool, right? :(
[19:17] <roadmr> something that checks explains errors a bit better
[19:17] <balloons> well for now.. let me go add that to your wiki
[19:17] <balloons> trim trailing spaces, but a newline after each test.. end each file with a new line
[19:19] <balloons> woot! indeed no more errors in the log file
[19:21] <davmor2> AlanBell: there were a couple of at-spi2 lands that orca might need an update for
[19:24] <stgraber> jibel: do you have a bug number for the ltsp-live bug where you drop to busybox?
[19:24] <stgraber> jibel: if not that's fine, I just want to close any bug that's currently open with the upload
[19:29] <cr3> balloons: if you're still getting errors, could you provide me with something to look at? perhaps a branch where you could push your code...
[19:30] <balloons> cr3, i pasted my branch info above.. but roadmr solved it for me.. I didn't have newlines at the end of most of my files
[19:30] <balloons> everything is perfect now :-)
[19:30] <balloons> thanks to both of you
[19:30] <roadmr> balloons: glad to be of help, as always :)
[19:30] <balloons> i'm adding the info to the wiki here: https://wiki.ubuntu.com/Testing/Automation/Checkbox/Walkthrough
[19:39] <cr3> balloons: ah, because you were using cat to concatenate a bunch of job files. this might pose a problem if something creates a file~, ie with tilde in addition to the one you already have
[19:40] <cr3> balloons: if you're cat'ing multiple different files, you could also use filter_templates instead of cat, it'll take care of separating jobs properly so you don't need to remember to add a trailing empty line which is error prone
[20:52] <balloons> cr3, I'm confused by what you mean I'm catting files together
[20:52] <balloons> is there something I should/could be doing differently?
[20:55] <balloons> I was editing the wiki to say you needed to include the new lines, etc.. I'll leave it off for now
[20:57] <cr3> balloons: do you have a branch where all the code you're working on resides, just so that I can have a look?
[20:58] <balloons> yea, my current branch is here: bzr branch lp:~nskaggs/+junk/checkbox-app-testing
[21:01] <cr3> balloons: command: cat $CHECKBOX_APP_TESTING_SHARE/jobs/nautilus.txt*
[21:01] <balloons> ahh yes
[21:01] <cr3> balloons: that's what I was talking about, so you might have problems in the event something creates nautilus.txt.in~ for example
[21:01] <balloons> the local.txt.in file
[21:01] <balloons> how should i set it up?
[21:02] <cr3> balloons: good question, nautilus.txt* was intended to express "either nautilus.txt *or* nautilus.txt.in", not "both nautilus.txt.in *and* nautilus.txt.in~"
[21:04] <balloons> yes.. I was just copying the example I had
[21:04] <balloons> I wasn't sure why the syntax existed like that
[21:04] <balloons> so again.. happy to adopt whatever best practice for this you thin
[21:04] <balloons> *think
[21:12] <cr3> balloons: out of curiosity, what editor are you using?
[21:12] <balloons> meh.. normally geany
[21:12] <balloons> today just gedit to quick fix things
[21:13] <cr3> balloons: ok, I wasn't expecting the editor to thwart checkbox with its temporary swap files.
[21:13] <cr3> balloons: so, an alternate filename matching expression might be:
[21:13] <cr3> command: cat $CHECKBOX_APP_TESTING_SHARE/jobs/nautilus.txt[\.in]*
[21:14] <cr3> this doesn't accurately express "either nautilus.txt *or* nautilus.txt.in", but it's close enough with the limited pattern matching at our disposal
[21:15] <balloons> yea.. not a fan of those
[21:15] <balloons> did they get into the branch
[21:15] <balloons> they should be in the ignore file
[21:16] <balloons> cr3, but that still requires the trailing blank line for files... as long as I'm using cat
[21:20] <bdmurray> Does anybody know if there is a bug about ubuntu slideshow display a paw print of an ocelot?
[21:21] <cr3> balloons: don't worry, your branch looks fine and none of the tilde files were committed.
[21:22] <cr3> balloons: however, it's still a pain that you have to close your editor every time you want to run checkbox, so that's where [\.in]* might be useful
[21:22] <balloons> i will update.. thanks cr3
[21:22] <cr3> balloons: or, of course, remember to add a trailing blank line at the end of your files at which point your files will be opened twice, once with the tilde and once without it because there are still two files matching nautilus.txt*
[21:23] <cr3> balloons: I might as well do the same in the base checkbox package :)
[21:23] <balloons> hehe
[21:23] <balloons> ohh
[21:23] <balloons> while i've got you
[21:24] <balloons> I was wondering on what you thought about a "file-a-bug" button on tests for when they fail?
[21:24] <balloons> ideally we'd like to be able to hit a button on the test, have it file a comment with the new bug number, and have it launch ubuntu-bug SOURCEPACKAGE -t TAG
[21:24] <cr3> balloons: if apport is enabled, a failing test should already cause apport to popup. if that's not the case, please let me know
[21:25] <cr3> balloons: fyi, apport should be enabled by default during the development cycle and only disabled by default 7 days before release
[21:25] <balloons> cr3, yes but apport won't appear if nothing actually crashed :-)
[21:25] <cr3> balloons: even if you just answered "no", apport should popup
[21:25] <balloons> ohh really?
[21:25] <cr3> balloons: if not, let me know and I could have a quick look
[21:25] <balloons> doesn't happen on my branch, so I'm not doing something correct
[21:26] <balloons> can i specify a tag as well?
[21:26] <balloons> I don't specify source packages anywhere, so I'm assuming that's the issue
[21:26] <cr3> balloons: and you're sure apport is enabled on your system? for example, you're running precise?
[21:26] <balloons> yes, running precise
[21:26] <balloons> apport is working fine
[21:27] <cr3> I think checkbox does a best effort of feeding apport with an appropriate tag, probably based on the requires line
[21:28] <balloons> hmm.. ok, I'm missing requires
[21:28] <cr3> balloons: just to be sure, can you show me the output of this command from your system: grep -i enabled /etc/default/apport
[21:28] <balloons> enabled=1
[21:28] <cr3> balloons: ok, try specifying a package then, let me know if that solves the problem
[21:29] <balloons> cr3, that's my guess
[21:29] <cr3> # Do not report a bug if no package nor symptom is defined
[21:29] <balloons> trying real quick
[21:29] <cr3> balloons: that's what checkbox says, so it must be right :)
[21:29] <balloons> has to be!
[21:30] <cr3> that's an obscur side effect: no package requirement, no apport. I can understand how that makes sense but the relation is subtle at best
[21:30] <balloons> yes.. I knew there was no way to run apport.. since I didn't give a package
[21:30] <balloons> but lol, I didn't realize it would
[21:32] <balloons> hmm.. not sure I've got the require syntax correct
[21:33] <balloons> what happens if I do a requires and I don't have it? it just skips the test right?
[21:33] <balloons> and what goes on the requires line?
[21:34] <balloons> for instance, requires: nautilus
[21:37] <bdmurray> okay well I reported bug 942262 about it
[21:43] <cr3> requires: package.name == 'nautilus'
[21:44] <cr3> balloons: ^^^ try that and, to answer your question, a test is indeed just skipped if the requires expression evaluates to false
[21:44] <balloons> cr3, k trying now
[21:45] <balloons> cool.. test popped up this time
[21:45] <balloons> ok, I hit no, but apport didn't pop
[21:47] <balloons> hmm hmm
[21:47] <balloons> gonna try something
[21:57] <cr3> balloons: how so?
[21:58] <cr3> balloons: err, nevermind :)
[21:58] <balloons> trying it now..
[21:58] <balloons> i'm going to try and use your upstream branch with my tests
[21:58] <balloons> instead of the bin on my box
[22:02] <cr3> balloons: fyi, I created bug #942273 about the filename matching expression in checkbox core and already submitted a branch with a merge request
[22:03] <balloons> cr3, nice!
[22:04] <balloons> thanks
[22:04] <cr3> balloons: thanks to you :)
[22:36] <balloons> sidetracked for a few mins.. will get back to trying your upstream later.. however, if you have any other ideas cr3 on why apport isn't popping up on a 'no', let me know now :-)
[22:38] <cr3> balloons: my changes haven't been merged into upstream trunk yet, but that shouldn't affect the problem though. it's mostly a nice-to-have when editing jobs with gedit
[22:38] <cr3> balloons: as for apport not appearing, do you see anything suspicious in the logs? I can't try it right now, only tomorrow
[22:40] <balloons> ahh.. lol, I keep forgetting those logs!
[22:40] <balloons> so I tried to just drop my jobs into the upstream branch bzr+ssh://bazaar.launchpad.net/+branch/checkbox/ and then run checkbox-qt.. nothing shows up ;-)
[22:41] <balloons> same deal for checkbox-gtk ;-) obviously needs a bit more than that
[22:41] <balloons> checking logs on run now
[22:44] <balloons> yep sure enough
[22:44] <balloons> Traceback (most recent call last):
[22:44] <balloons>   File "/usr/lib/python2.7/dist-packages/checkbox/reactor.py", line 74, in fire
[22:44] <balloons>     results.append(handler(*args, **kwargs))
[22:44] <balloons>   File "/usr/share/checkbox/plugins/apport_prompt.py", line 229, in prompt_test
[22:44] <balloons>     'suite:' : test["suite"]}
[22:44] <balloons> KeyError: 'suite'
[22:51] <cr3> balloons: aha, that looks like the culprit!
[22:51] <cr3> balloons: I really need to jet but we could pickup from there tomorrow
[22:52] <cr3> balloons: until then, what that means is that it seems that the test is not part of a local job
[22:52] <balloons> cr3, sounds good.. enjoy your evening, chat with you tomorrow.. you've been a big help
[22:52] <cr3> balloons: in other words, the test selection window should only contain local jobs, ie they should have the + to expand the tests within the local job
[22:53] <balloons> yes, the test selection window contains a list with =
[22:53] <cr3> balloons: if you see any jobs in the test selection window without a + at the root node, that's a problem
[22:53] <balloons> a list with +'s
[22:53] <balloons> yes, I had that problem for a long time
[22:53] <cr3> balloons: it should only have +'s
[22:54] <balloons> yep, it does
[22:54] <balloons> actually they are > arrows, but :-)
[22:54] <cr3> balloons: hm, in that case, we could have a closer look tomorrow
[22:54] <cr3> balloons: heh, you get the idea though :)
[22:54] <cr3> cheerio!
[23:08] <skaet> A full set of rebuilds has just been kicked off to pick up the fixes from the weekend and today - they should be emerging on the iso tracker over the next couple of hours.
[23:10] <skaet> ubuntu alternate,  kubuntu alternate posted.
[23:14] <skaet> xubuntu alternate posted
[23:45] <bkerensa> skaet: getting ready to test them now