[02:02] <wallyworld> sinzui: around?
[02:14] <wallyworld> wgrant: StevenK: on the sharing details page, we show bugs and branches in the one table. there's a (3rd party) sorttable.js bug wrt row colspans which means sorting is broken. i can work on fixing the bug but i think we perhaps should be showing branches and bugs in different tables in any case. you agree?
[02:15] <lifeless> wallyworld: why different tables?
[02:16] <lifeless> wallyworld: we have lots of artifact types, having one table per is really just a manual pre-defined facet search.
[02:16] <wallyworld> lifeless: like we do for milestones with blueprints and bugs i think it is
[02:16] <wgrant> wallyworld: Well, you know how I feel about sorttable on that page at all :)
[02:16] <wallyworld> different things therefore different ables
[02:16] <wallyworld> i can keep everything in the one table
[02:17] <wallyworld> i'll fix the sorttable bug
[02:18] <wallyworld> but i'm not sure how much sense sorting makes if we are sorting on bug nr and branch name
[02:18] <wallyworld> that's why i was thinking separate tables, unless we think sorting on bug nr, branch name is silly
[02:25] <StevenK> wgrant: I wonder if the >20 minute tests branch I approved had a wildly different time in ec2.
[02:28] <wgrant> StevenK: Quite possibly. Those tests are rather DB-heavy.
[02:31] <lifeless> wallyworld: why have js sorting anyway? db can sort more efficiently
[02:31] <wallyworld> sure, but why enforce a round trip just to sort 20 rows by information type for example
[02:32] <wallyworld> i think js sorting definitely has a place
[02:33] <lifeless> wallyworld: if the page doesn't show /all/ the content
[02:33] <wallyworld> then js sorting should be disabled
[02:33] <wallyworld> it's not currently, and i don't think we do that elsewhere, but i think we should
[04:31]  * StevenK tries to get his head around YUIXHR tests and feels his brain slowly dribbling out of his ears.
[04:35] <StevenK> loltpg
[07:58] <adeuring> good morning
[08:00] <czajkowski> aloha
[08:15] <RfADdlS> I was upload the package and get error on build "Can't exec "qmake": No such file or directory at /usr/share/perl5/Debian/Debhelper/Dh_Lib.pm line 215." How to fix it? Full log http://goo.gl/nxYFz
[08:33] <StevenK> RfADdlS: We don't support PPAs in this channel, but you're missing a Build-Depends.
[08:39] <RfADdlS> StevenK: where should I write? I pointed out in debian/control all depends.
[09:24] <jelmer> RfADdlS: #ubuntu-packaging is the right channel for questions about packaging
[09:33] <RfADdlS> StevenK: jelmer: thank you
[10:56] <jml> blooop
[10:56] <jml> http://paste.ubuntu.com/938096/
[11:11] <wgrant> Heh
[11:11] <czajkowski> thats definately a blooop and not a blip
[11:13] <wgrant> jml: Is that really all of them?
[11:13] <wgrant> It's excluding tests obviously
[11:13] <wgrant> But I would have expected more
[11:13] <jml> wgrant: excluding tests and doc#@$!s
[11:14] <jml> bzr grep -n 'removeSecurityProxy\(' |grep -v test |grep -v '\.txt'
[11:14] <jml> also excludes 'testing'
[11:14] <jml> which accounts for about 100
[11:14] <wgrant> Ah
[11:14] <wgrant> The lack of testing would do it, indeed.
[12:04] <jml> gary_poster: https://code.launchpad.net/~jml/subunit/filter-tags/+merge/102840 up for review and I'm happy for it to land.
[12:05] <jml> gary_poster: jelmer or vila are valid reviewers.
[12:10] <gary_poster> jml, fantastic thanks.  jelmer or vila, if you have time to review soon it would help us a lot.
[12:11] <gary_poster> jml. did benji sufficiently address https://code.launchpad.net/~benji/testrepository/add-worker-id-tagging/+merge/102574 for you to approve it?
[12:11] <jml> gary_poster: haven't looked. will do so now.
[12:12] <gary_poster> thank you
[12:38] <jml> merged.
[12:38] <jml> there's definitely some releasing that needs to be done
[12:45] <jml> gary_poster: once you guys are actually using this new stuff in anger and getting value from it, let me know, and I'll release the bits I can.
[13:00] <gary_poster> awesome thanks jml.  We already have the tagging stuff in place for our ec2 parallel tests, thanks to our ppa, so those parts are being used successfully in anger.  this has given us subunit results that we can now analyze with your subunit-filter changes.
[13:01] <jml> gary_poster: cool. I guess once you've done such an analysis and learnt something, I'll do a release.
[13:01] <jml> and maybe I'll do a testtools release today anyway.
[13:02] <gary_poster> cool.  I intend/hope to have such that analysis today; will ping you about it one medium or another when I do
[13:03] <rick_h> abentley: hey, can I bug you pre stand up for a call to help with this bug stuff?
[13:04] <abentley> rick_h: Sure.  Bug me with your bug stuff.
[13:04] <rick_h> awesome, normal stand up hangout?
[13:05] <abentley> rick_h: Okay.
[14:14] <jam> flacoste: are you around? Are we still chatting today?
[14:15] <abentley> adeuring: The lazr.jobrunner release HOWTO says "Test the generated source distribution in dist/".  How do I do that?
[14:17] <adeuring> abentley: good question... I don't know. The text stems from creating a template for the buildout machinery...
[14:17] <abentley> adeuring: Oh, okay.  Maybe we'll remove that text for 0.4, then :-).  I wondered why it didn't mention anything about Launchpad...
[14:18] <flacoste> jam: yes
[14:18] <adeuring> abentley: what you can do: create the tar ball, unpack it somewhere and run "make develop; make check"
[14:18] <flacoste> jam: hang out?
[14:18] <flacoste> jam: let me get you an invite
[14:19] <jam> flacoste: google hangouts makes it sound like I'm talking under water according to everyone I've tested with.
[14:19] <jam> but I'll try it again
[14:19] <flacoste> jam: let's skype then
[14:19] <flacoste> jam: or voip
[14:19] <flacoste> either works for me
[14:19] <jam> JohnArbashMeinel on skype
[14:20] <flacoste> jam: yes, let me switch input
[14:20] <jam> flacoste: ok
[14:21] <flacoste> jam: i need to hang up
[14:21] <abentley> adeuring: Makefile isn't included in the tarball.  Should it be?
[14:21] <deryck> hi flacoste.  wb!
[14:22] <abentley> adeuring: Neither is bootstrap.py.
[14:22] <adeuring> abentley: hmmm... the usual "python -S bootstrap.py; bin/buidout" should be enough, but sure, bootstrap.py should be included...
[14:23] <abentley> adeuring: And buildout.cfg?
[14:23] <adeuring> abentley: yes, of course...
[14:23] <abentley> adeuring: I guess we can fix that for 0.4, too.
[14:24] <adeuring> abentley: right
[14:24] <gary_poster> jml, hey.  we think the new subunit-filter might be broken.  Are we maybe doing something wrong?  Examples:
[14:25] <gary_poster> cat ~/Downloads/testrepo-0.txt | env PYTHONPATH=python ./filters/subunit-filter -s --with-tag='worker-0' | head -n 300
[14:25] <gary_poster> this gives us all tests irrespective of worker-0
[14:25] <gary_poster> cat ~/Downloads/testrepo-0.txt | env PYTHONPATH=python ./filters/subunit-filter -s --with-tag='worker-0' --without-tag='zope:layer' | head -n 300
[14:25] <gary_poster> this gives us...something weird.  Will pastebin, one sec
[14:27] <gary_poster> http://pastebin.ubuntu.com/938332/
[14:27] <gary_poster> oh, easier to read there
[14:27] <gary_poster> without-tag may have worked
[14:28] <gary_poster> but it did not for benji
[14:28] <gary_poster> he tried --without-tag worker-3
[14:28] <gary_poster> and got worker-3
[14:33] <abentley> adeuring: Can you please authorize me to upload packages for lazr.jobrunner to pypi?
[14:34] <adeuring> abentley: if you can tell me to do that, sure ;)
[14:35] <adeuring> abentley: found it. what is your pypi user name?
[14:35] <abentley> adeuring: abentley
[14:36] <adeuring> abentley: done
[14:36] <abentley> adeuring: Thanks, works.
[14:39]  * deryck does the post-update reboot
[14:49] <jml> gary_poster: ok, that sucks. Will try adding more tests.
[14:50] <jml> gary_poster: can you give me a small snippet that has worker-0 tags, zope:layer tags and tests with neither?
[14:50] <gary_poster> jml, we are futzing with it within timebox.  10 minutes left.  tag leakage appears to be the problem.
[14:51] <gary_poster> jml, will send after 10 minutes :-)
[14:51] <jml> gary_poster: tag leakage should be fixed as of testtools r254 :(
[14:52] <gary_poster> jml, yes, different tag leakage
[14:52] <jml> \o/
[14:52] <gary_poster> our output does not have tag leakage
[14:52] <gary_poster> from the testrun
[14:52] <gary_poster> but when the tag predicate compares with-tags with existing tags
[14:52] <gary_poster> the existing tags gradually grows
[14:53] <gary_poster> within the logic of subunit-filter
[14:54] <jml> harumph.
[14:55] <jml> gary_poster: which revno of filter-tags are you using?
[14:55] <gary_poster> jml 184
[14:56] <jml> gary_poster: that's the correct one... dammit.
[15:00] <gary_poster> jml, ok, our futzing time's up.  benji is putting up a pastebin of a small output you can use
[15:01] <benji> jml: http://paste.ubuntu.com/938374/
[15:01] <jml> \o/
[15:01] <gary_poster> jml, that is a testrepository run, not LP, but that means it is complete, fwiw, and he duped my experiences
[15:02] <gary_poster> --with-tag='worker-0' or --without-tag='worker-0'
[15:04] <jml> intriguing.
[15:07] <jml> subunit-filter trunk (and 0.0.7) both don't seem to parse that correctly
[15:08] <jml> even when I strip out tags.
[15:10] <jml> :(
[15:11] <abentley> adeuring: r=me
[15:11] <adeuring> abentley: thanks!
[15:13] <jml> benji: when I download that, it has dos line endings. Is that the case for you?
[15:14] <benji> jml: I don't think it did on my end, I'll check in a second
[15:25] <benji> jml: I just emailed you the stream as an attachment so we can be sure it isn't being mangled by the pastebin
[15:39] <jml> ok.
[15:39] <jml> I've figured it out
[15:40] <jml> I learned about the local / global distinction partway through implementing this branch and messed up the tag tracking in _PredicateFilter
[15:40] <jml> :(
[15:41] <jml> What I wouldn't give for a solid fortnight and a reviewer to hand.
[16:21]  * jml twirls his pen, clicks it twice, stands up with both arms high in the air and shouts "I am invincible!"
[16:22] <jml> benji, gary_poster: r186 seems to work for me.
[16:22] <jml> the rot in tags goes deep
[16:22] <jml> this is the first time that it's actually been used for anything, afaict.
[16:25] <jml> I've got to head now. Keep in touch if there's any problems or any successes.
[16:26] <benji> thanks jml; I'll see if I'm not using r186 or something
[17:18] <gary_poster> thank you jml
[19:05] <jcsackett> weird. bin/py -t buildmailman is blowing up in my face on `make run` now. anyone seen this before?
[19:16] <sinzui> jcsackett, no.
[19:16] <jcsackett> sinzui: yeah, i thought i had just trashed a config, but stashing some changesets shows it's actually some code i changed in publisher.
[19:16] <jcsackett> which is weird.
[19:19] <jcsackett> sinzui: and solved. import of `Breadcrumb` apparently fubared things. i'm guessing circular import issues, since moving the import to the clause where i need it works.
[19:21] <sinzui> jcsackett, I agree with your assessment
[19:25] <jcsackett> sinzui: weird though that would break the mailman build step though...or is that more tightly coupled than i thought?
[19:26] <sinzui> jcsackett, it is very coupled because we import Lpism in moneypatches
[19:26] <jcsackett> aaah.
[19:26] <jcsackett> ok.
[19:26] <jcsackett> well then, as long as i'm not nuts. :-)
[19:26] <sinzui> Mailman 3 promises better decoupling
[20:39] <gary_poster> abentley, hi.  hallyn from the server team would really like some assistance getting qemu-kvm and libvirt bzr trees auto-importing again.  He was hoping that someone would be at UDS to help him, but I will have to tell him no.  Would you be able to help him sometime in the next few days or weeks, or could you give another suggestion for him to talk to?
[20:40] <abentley> gary_poster: Let me have a quick look.
[20:40] <gary_poster> thanks abentley
[20:44] <abentley> gary_poster: The problem is that these git branches use submodules, and Bazaar does not have an equivalent feature.
[20:44] <gary_poster> abentley, ah :-(
[20:45] <gary_poster> abentley, so this will not be fixed anytime soon, if ever, given the current state of things, is what I should report, right?
[20:46] <abentley> gary_poster: More complicated than that.
[20:46] <gary_poster> ok
[20:46] <abentley> gary_poster: If the git trees don't actually use submodules and just have some revisions in the past that did, then there's a workaround.
[20:47] <abentley> gary_poster: Although I'm not sure if it's accessible to users.  Probably not, now that I think of it.
[20:47] <gary_poster> abentley, but in that scenario, you, or someone else with the proper direction, would be able to help him?
[20:48] <abentley> gary_poster: I think we'd have to get webops hand-upgrading branches in the code import branch store.  I don't know if that would fly.
[20:49] <gary_poster> abentley, ah, ok.  it's a remote possibility at best then, I guess.  thank you very much for looking into it, abentley.
[20:49] <abentley> gary_poster: np.
[20:52] <jelmer> bug 402841 is (IIRC) the relevant bug
[20:52] <_mup_> Bug #402841: file-roller crashed with SIGSEGV in memmove() <apport-crash> <apport-failed-retrace> <i386> <file-roller (Ubuntu):New for desktop-bugs> < https://launchpad.net/bugs/402841 >
[20:52] <jelmer> hmm no, bug 402814 ?
[20:52] <_mup_> Bug #402814: Importing revisions with submodules is not supported <lp-code> <udd> <Bazaar:Triaged> <Bazaar Git Plugin:Fix Released> <cloudfoundry:Confirmed> <Launchpad itself:Triaged> < https://launchpad.net/bugs/402814 >
[20:55] <abentley> jelmer: I tried it locally, and it died with this: http://pastebin.ubuntu.com/938863/
[20:55] <jelmer> ah, so it's more than just bug 402814
[20:55] <_mup_> Bug #402814: Importing revisions with submodules is not supported <lp-code> <udd> <Bazaar:Triaged> <Bazaar Git Plugin:Fix Released> <cloudfoundry:Confirmed> <Launchpad itself:Triaged> < https://launchpad.net/bugs/402814 >
[20:56] <jelmer> abentley: bug 963525
[20:56] <_mup_> Bug #963525: signed tag support <Bazaar Git Plugin:Triaged> <Dulwich:Triaged by jelmer> <Launchpad itself:Triaged> < https://launchpad.net/bugs/963525 >
[20:57] <abentley> gary_poster: So even if we were lucky on the first issue, there's a second bug: 963525 which definitely requires code to fix.
[20:59] <gary_poster> abentley, ack. passing along.  thank you.