[00:03] <poolie> sinzui, hi, i don't really understand your comment on bug 801242
[00:03] <_mup_> Bug #801242: no way to request membership in a private team - can only be added <privacy> <teams> <Launchpad itself:Triaged> < https://launchpad.net/bugs/801242 >
[00:04] <cjwatson> How come the QA bot hasn't got round to doing its thing on bug 899972 yet?
[00:04] <_mup_> Bug #899972: cron.germinate is very slow <Launchpad itself:In Progress by cjwatson> < https://launchpad.net/bugs/899972 >
[00:05] <StevenK> cjwatson: Which revision?
[00:05] <lifeless> cjwatson: qastaging isn't running it yet
[00:06] <lifeless> cjwatson: look at qastaging, bottom right corner.
[00:07] <cjwatson> Ah
[00:07] <cjwatson> StevenK: r14516
[00:28] <StevenK> cjwatson: That revision hasn't passed buildbot, so it's not in stable.
[00:29] <cjwatson> Ah, I keep forgetting how many stages are involved :-)
[00:30] <StevenK> The QA tagger should get it's grubby hands on the bug in about 3-3.5 hours
[00:32] <cjwatson> I've just posted my QA notes in the bug.  In the unlikely event that the deployment pipeline gets that far between the tagger getting to it and me noticing that, it can be marked qa-ok.
[00:32] <StevenK> I still need to sort out the deployment report
[00:32] <StevenK> It's still hung up on r14489 being bad
[00:59]  * StevenK tries to understand bzrlib.gpg and goes cross-eyed
[01:02]  * StevenK finds the right option, finally.
[02:44]  * wallyworld_ goes for a coffee run
[02:59]  * nigelb realizes he hasn't written LP code in a while.
[03:12] <StevenK> nigelb: Fix it!
[03:13] <nigelb> StevenK: Yeah, I need to find an interesting bug to fix. Updating my tree now.
[03:31]  * StevenK kicks the qa-tagger
[03:31] <StevenK> Bloody rollback tags
[03:32] <StevenK> ARGH
[03:33] <StevenK> I fix r14489 and r14490 and now r14491 is bad
[03:33] <StevenK>  /wrists
[03:34] <StevenK> poolie: Ping
[03:34] <poolie> pong
[03:35] <StevenK> poolie: Do you know what is going on with bug 509016?
[03:35] <_mup_> Bug #509016: Please load bzr-git, bzr-svn and bzr-hg in loggerhead <bad-commit-14991> <lp-code> <qa-bad> <trivial> <Launchpad itself:Fix Committed by jelmer> < https://launchpad.net/bugs/509016 >
[03:35] <StevenK> Can I roll it back so we actually deploy?
[03:37] <StevenK> cjwatson: Thanks for your exhaustive QA notes, I've tagged your bug qa-ok.
[03:37] <poolie> StevenK, i don't know anything beyond what's in the bug etc
[03:37] <poolie> what is breaking?
[03:38] <StevenK> poolie: Can we mumble?
[03:38] <StevenK> It will be quicker
[03:39] <poolie> after counting futzing with mumble?
[03:39] <poolie> can i just call you?
[03:39] <StevenK> Haha
[03:39] <StevenK> Sure
[03:39] <StevenK> Landline in the directory
[03:40] <nigelb> lol
[03:40] <StevenK> mumble works perfectly for me, so I don't know why poolie has to fiddle with it
[03:42] <nigelb> I've seen people complain about all 3 calling things at some point.
[03:42] <nigelb> mumble/skype/google hangout
[03:46] <StevenK> jelmer: Are you still around, by chance?
[03:47] <wgrant> StevenK: I think we should revert it.
[03:47] <StevenK> I'm just preparing a revert
[03:51] <StevenK> And buildbot started like 90 minutes ago.
[04:05] <wallyworld_> StevenK: should we just restart bb?
[04:05] <StevenK> Probably not.
[04:05] <wallyworld_> :-(
[04:07] <StevenK> DAMN IT, lp-land, use the correct key
[04:09] <nigelb> lol
[04:09] <StevenK> nigelb: I'm moving to a 4096 bit RSA key
[04:09] <wgrant> StevenK: Set gpg_signing_key in bazaar.conf
[04:09] <StevenK> I did
[04:09] <StevenK> That fixed ec2land, but lp-land is still wanting my DSA key's passphrase.
[04:10] <StevenK> And bzrlib.gpg is nearly incomprehensible
[04:10] <StevenK> Thanks lifeless!
[04:11] <nigelb> heh
[04:15] <poolie> StevenK, oops, i got totally distracted
[04:16] <StevenK> Heh
[04:16] <StevenK> poolie: I have a rollback prepared, just having trouble landing it since lp-land insists on signing with my 1024bit DSA key.
[04:16] <poolie> wgrant, in other news, i think python-markdown is tolerably safe
[04:17] <StevenK> Even though gpg_signing_key is set in bazaar.conf
[04:18] <wgrant> poolie: Hm, that is good news.
[04:18] <wgrant> poolie: Is that what GitHub uses, or do they have their own thing?
[04:19] <poolie> i think they use ruby
[04:19] <poolie> so different implementation, same format
[04:19] <wgrant> Probably.
[04:19] <poolie> similar format
[04:19] <wgrant> Yeah, I knew it was markdownish, but I thought they used Python in places.
[04:19] <wgrant> Perhaps not.
[04:19] <poolie> so, obviously any software can have bugs however
[04:20] <poolie> the code is not obviously silly
[04:20] <wgrant> Yep
[04:20] <poolie> it does have options for safety against xss etc
[04:20] <nigelb> wgrant: They use github-flavored markdown
[04:20] <nigelb> Btut they use a ruby thing
[04:20] <poolie> and it's fairly widely used for this
[04:20] <nigelb> to do the conversion
[04:20] <wgrant> Ah, that's what I was thinking of. They use pygments.
[04:21] <mwhudson> yes
[04:22] <mwhudson> you can tell, because it has the same performance bugs pygments has :-)
[04:22] <mwhudson> they cap the amount of cpu it can use though
[04:22] <StevenK> poolie: http://pastebin.ubuntu.com/770798/ explains the issue
[04:23] <nigelb> mwhudson: heh
[04:25] <poolie> StevenK, that might be bug 904550?
[04:25] <_mup_> Bug #904550: gpg signing key doesn't default to 'email' anymore <release-blocker> <Bazaar:Confirmed> < https://launchpad.net/bugs/904550 >
[04:25] <poolie> or just coincidental
[04:25] <poolie> um
[04:27] <poolie> StevenK, so try something like
[04:28] <poolie> echo merge bzr+ssh://bazaar.launchpad.net/~launchpad-pqm/launchpad/stable bzr+ssh://bazaar.launchpad.net/~launchpad-pqm/launchpad/db-devel | gpg --clearsign -u abcd1234 |mail -s 'commit message' launchpad@pqm.canonical.com
[04:28] <poolie> something like this
[04:28] <StevenK> Right, which just submits it manually
[04:28] <poolie> yes
[04:28] <poolie> if you're stuck, that's how i'd get unstuck
[04:33] <StevenK> PQMException: 'Failed to verify signature: gpgv exited with error code 1'
[04:33] <StevenK> :-(
[04:34] <poolie> ah pqm
[04:35]  * StevenK waits for logs to sync
[04:36] <poolie> ah
[04:36] <poolie> you need gpg -a too
[04:37] <StevenK> -a is the default if the output is a tty
[04:37] <StevenK> (Which it was in this case)
[04:38] <poolie> isn't the output a pipe to mail?
[04:38] <poolie> or you were pasting it
[04:38] <StevenK> Pasting it into Thunderbird and sending
[05:13] <StevenK> poolie: I fixed it with an enormous sledgehammer.
[05:36] <jtv-zzz> poolie: Hi!  Any luck with Q/A?  We've got a nasty qa-bad blocking some really urgent stuff, and I'm concerned about other pending Q/A holding up the rollback we need for the bad branch.
[05:38] <jtv> I thought openid was no longer supposed to send us to the wrong pages.  :(
[05:40] <jtv> poolie: superficially there's only 7 branches waiting for your Q/A and it's not the blocker yet, but if we need a rollback, that adds the 18 revisions that are blocked behind his.
[05:41] <jtv> Behind jelmer's, that is.
[05:41] <wallyworld_> jtv: StevenK is onto it
[05:41]  * jtv breathes out
[05:41] <jtv> thanks
[05:42] <wallyworld_> jtv: don't forget to breath in again
[05:42] <wallyworld_> and then out
[05:42] <wallyworld_> and then in
[05:42]  * jtv tries that
[05:42] <jtv> wallyworld_: thanks, that's really helping.  But we can't keep going this way all day, can we?
[05:42] <jtv> Maybe I should write a script.
[05:42] <StevenK> I did not do the QA.
[05:42] <wgrant> jtv: That rev is rolled back.
[05:42] <wgrant> 'twill be some days before you can deploy.
[05:42] <jtv> Which revision is rolled back?  jelmer's?
[05:43] <wgrant> Yes.
[05:43] <StevenK> Yes
[05:43] <jtv> I don't suppose we have a way of marking rollbacks as such after landing, just so the deployment report can be fixed?
[05:43] <StevenK> Only when it hits qas
[05:44] <jtv> no entiendo
[05:45] <jtv> StevenK, wgrant: isn't the buildfarm breakage something of a bad thing though?
[05:45] <wgrant> jtv: What's broken?
[05:45] <jtv> Oh, was my buildfarm fix rolled back?
[05:45] <wgrant> We rolled it back <5 minutes after it was discovered.
[05:45] <jtv> Ah!
[05:45] <wgrant> The world would have been on fire for days otherwise.
[05:45] <jtv> Didn't know that.
[05:46] <jtv> Right.
[05:46] <jtv> Wish I'd known that before working nights and holidays.
[05:46] <wgrant> Well.
[05:47] <wgrant> Distro hadn't nuked you from orbit yet, so it should have been reasonably obvious :)
[05:47] <jtv> Explicit is better than implicit.
[06:08] <jtv> wgrant: what's the procedure with that rollback?  Am I responsible for undoing it once the fix is ready to deploy?
[06:09] <wgrant> jtv: Someone (likely Julian, since AIUI both you and I are away for some time) needs to request a deployment.
[06:09] <jtv> Are you talking about the obvious deployment (of the fix), or some extra deployment?
[06:09] <wgrant> The obvious deployment.
[06:10] <wgrant> There's nothing special about this sort of rollback -- it's just flipping the symlink back to an older tree and restarting services.
[06:10] <wgrant> So a normal deployment will clobber it.
[06:11] <jtv> OK
[06:11] <jtv> Well I was going to request the rollback anyway if nobody else beat me to it — which is why I'm chasing Q/A blockers in my free time.
[06:42] <huwshimi> A CSS/UI review for someone kind enough to do so: https://code.launchpad.net/~huwshimi/launchpad/inline-fields/+merge/85798
[08:54] <adeuring> good morning
[09:07] <mrevell> Hi
[09:12] <nigelb> Morning mrevell
[09:30] <bigjools> StevenK: do you know if there's a bug for adding package_name to SPPH/BPPH?
[10:05] <StevenK> bigjools: It's not package_name
[10:05] <bigjools> whatever then
[11:26] <gmb> Someone remind me: Are DB patches that are just constraints meant to go into devel or db-devel?
[11:29] <wgrant> gmb: db-devel
[11:29] <wgrant> They generally require exclusive table locks, which requires downtime.
[11:29] <gmb> wgrant, Thanks.
[11:46] <rick_h__> morning
[12:59] <bigjools> anyone know if there's a way of specifying that a parameter to @operation_parameters is for a particular api version?
[13:00] <bigjools> or do I need to repeat the whole lot using @operation_for_version?
[14:24]  * benji assumes jtv isn't around.
[15:16] <bac> hi abentley, frankban and i are using 'bzr lp-propose' but the template is not being populated.  i have the lpreview-body plugin installed as part of lp-dev-dependencies.  any idea what is missing?
[15:17] <abentley> bac: I expect it's not detecting that you're targetting a Launchpad branch.  Let me refresh myself....
[15:19] <bac> abentley: the contents of the editor are http://pastebin.ubuntu.com/771213/ -- so it does appear to know it is a LP branch
[15:19] <abentley> bac: the target branch needs to match this regex: lp:(~launchpad-pqm/)?launchpad(/(db-)?devel)?
[15:20] <abentley> bac: Sorry, it needs to match 'bzr\+ssh://bazaar.launchpad.net/~launchpad-pqm/launchpad/(db-)?devel'
[15:23] <bac> abentley: it looks like i have the full bzr+ssh URL in my locations.conf file but the shorthand is being used and thus not matching the regex
[15:24] <abentley> bac: If your submit branch is a local branch, you need to specify the bzr+ssh URL as its public location.
[15:25] <bac> abentley: here is 'bzr info' for my branch and my devel branch -- http://pastebin.ubuntu.com/771217/
[15:27] <abentley> bac: That looks like it would work.
[15:28] <abentley> bac: But also, it should be raising an exception if it's getting lp:launchpad as the target, and it's not doing that either.  So I'm not sure what's going on there.
[15:32] <bac> abentley: ok, i'll experiment on my dev machine at home and see if it still works and compare the differences
[15:33] <abentley> bac: The target branch is selected by lp-propose, so I'd stick a pdb somewhere in that code.
[15:33] <bac> ok
[15:54] <cr3> hi folks, you guys know there's a problem with bazaar.launchpad.net, right?
[15:54] <cr3> oops, not anymore. thanks for fixing it so quickly :)
[16:13] <adeuring> jcsackett: could you please review this MP https://code.launchpad.net/~adeuring/launchpad/bug-901016/+merge/85902 ? (100 lines diff)
[16:16] <rick_h__> abentley: have a sec? my changes for this bug run right into your refactor
[16:16] <abentley> rick_h__: OTP
[16:16] <rick_h__> abentley: ok
[16:30] <jcsackett> adeuring: sure thing.
[16:30] <adeuring> jcsackett: thanks!
[16:39] <jcsackett> adeuring: looks good. r=me.
[16:39] <adeuring> jcsackett: thanks!
[16:58] <abentley> rick_h__: back.
[16:59] <rick_h__> abentley: I've got it. Your changes got merged so updated with devel and redoing the changes
[16:59] <rick_h__> bzr issues making it a bit slow, but think I'm set, thanks abentley
[16:59] <abentley> rick_h__: Cool.
[17:10] <abentley> deryck[lunch]: http://blog.launchpad.net/general/new-approaches-to-new-bug-listings
[17:32] <rick_h__> jcsackett: ping, got a sec to peek at: https://code.launchpad.net/~rharding/launchpad/heat_help_894740/+merge/85901 ?
[17:42] <jcsackett> rick_h_: sure, in just a moment. :-)
[17:55] <jcsackett> rick_h__: r=me.
[17:59] <rick_h__> jcsackett: ty much
[18:00] <mrevell> Night all
[18:28] <deryck> abentley, thanks!  looks good.
[18:28] <abentley> deryck: np.
[18:29] <rick_h__> deryck: did you have some 'next bugs' in your mind? It seemed during the stand up you had some targets
[18:30] <deryck> rick_h__, yeah, any of the high bugs with the tag bug-columns would be good. see: http://tinyurl.com/d9nevd9
[18:30] <rick_h__> deryck: ok cool. I'm going through them but wanted to make sure you didn't have specific next items in mind
[18:32] <deryck> rick_h__, nah.  just ask me if you have questions about specific bugs.
[18:32] <rick_h__> deryck: abentley looking at: https://bugs.launchpad.net/launchpad/+bug/900900
[18:32] <_mup_> Bug #900900: navigation links are missing from bug listing bottom if there is only a single batch. <bug-columns> <Launchpad itself:Triaged> < https://launchpad.net/bugs/900900 >
[18:32] <rick_h__> deryck: abentley but the batching code specically calls out that "Only render bottom navigation links if there are multiple batches."
[18:33] <rick_h__> so are we looking to change that site-wide?
[18:33] <abentley> rick_h__, deryck: I think we should change it site-wide.  It's confusing.
[18:34] <deryck> I agree.  it might makes sense for some defined small number, i.e. 1-5 results.  but anything over half a screen we certainly need bottom links, so....
[18:34] <rick_h__> abentley: I think the argument would be that in small nubmers, doubling the batch ui would clutter the interface
[18:34] <deryck> to really simplify, I say always have bottom links.
[18:34] <rick_h__> deryck: ok, will do then. Thanks
[18:46] <sinzui> Hi jcsackett. I have a few questions about your branch in review
[18:50] <jcsackett> sinzui: shoot.
[18:50] <sinzui> Line 41 of the diff adds an extra space to the text and the test verifies it.
[18:50] <sinzui> This message does not cover the case when a private team is subscribed to a bug or branch. Can we make the message more general?
[18:51] <jcsackett> sinzui: good catch on the typo, and yeah, we can generalize the message.
[18:52] <sinzui> "This action will reveal this team's name to the public. (Choose again) (Continue)
[18:52] <sinzui> jcsackett, we avoid yes/no button because they require a re-reading of the question.
[18:54] <flacoste> abentley: nice post!
[18:54] <sinzui> jcsackett, My concerns are just the text and labels. The implementation is great
[18:54] <abentley> flacoste: thanks.
[18:55] <jcsackett> sinzui: thanks; i'm pushing up the text changes now.
[18:55] <sinzui> fab!
[19:00] <sinzui> jcsackett, I also wanted to have a mid-implementation conversation about base-layout. I have rudimentary version working with tests. I think I want to change the macro used to prevent the main and side slots from rendering
[19:02] <jcsackett> sinzui: sure, one sec while i hop on mumble.
[19:06] <rick_h__> deryck: is this something I could probably just self-review? https://code.launchpad.net/~rharding/launchpad/batch_nav_900900/+merge/85935
[19:07] <deryck> rick_h__, you can't self review your own stuff yet, until you're a graduated reviewer yourself.
[19:07] <rick_h__> deryck: ah, makes sense.
[19:07] <deryck> rick_h__, lots of "self" there, but I hope you get the idea.
[19:07] <rick_h__> deryck: right, need to get rid of the "training" badge first. Understand
[19:08] <rick_h__> jcsackett: see you're doing some call, if you get a sec after that please? https://code.launchpad.net/~rharding/launchpad/batch_nav_900900/+merge/85935
[19:11] <sinzui> jcsackett: http://pastebin.ubuntu.com/771461/
[19:22] <jcsackett> rick_h__: looking now. :-)
[19:23] <deryck> I also have something for review jcsackett.  https://code.launchpad.net/~deryck/launchpad/avoid-extra-buglist-count-901124/+merge/85940
[19:23] <rick_h__> jcsackett: ty
[19:38] <jcsackett> rick_h__: i've left questions on your MP.
[20:01] <deryck> jcsackett, did you see my ping about a branch for review?
[20:02] <jcsackett> deryck: i did not, i will go grab your branch from the queue now.
[20:02] <deryck> jcsackett, cool, thanks man!
[20:09] <jcsackett> deryck: looks good; nice workaround the COUNT issue. :-)
[20:09] <deryck> jcsackett, thanks!
[22:42]  * huwshimi must. not. reload. mp.
[22:49] <benji> heh
[23:14] <lifeless> mwhudson: are your tags queries all copacetic?
[23:15] <mwhudson> lifeless: i don't know what that word means
[23:15] <lifeless> http://en.wiktionary.org/wiki/copacetic
[23:16] <mwhudson> well, i've written some queries that pass all my tests
[23:16] <lifeless> cool
[23:17] <lifeless> ftr what we do in lp is a little / moderately bloody
[23:17] <lifeless> I've been meaning to refactor
[23:17] <lifeless> e.g. to use an array equality
[23:18] <lifeless> the mathematical way is using intersections, of course
[23:18] <mwhudson> my approach is basically count(tags in a that are not in b) == 0
[23:19] <mwhudson> lifeless: where is the code in lp?  lp/bugs/model/bugtask?
[23:19] <lifeless> counting works, though wouldn't you just count(found tags) == len(wanted tags) ?
[23:19] <lifeless> mwhudson: yes
[23:20] <mwhudson> lifeless: well, not much difference really i guess
[23:21] <mwhudson> lifeless: ah so part of my fun is that the list of tags to search for is retrieved by another part of the query
[23:22] <lifeless> in one query ?
[23:22] <lifeless> can you share ?
[23:24]  * mwhudson rummages
[23:26] <mwhudson> lifeless: http://paste.ubuntu.com/771702/
[23:26] <lifeless> mwhudson: oh, did I tell you that there is django glue in oops-wsgi now ? courtesy jamesh
[23:35] <lifeless> mwhudson: so, uhm right
[23:35] <lifeless> how to stop unity raising the active window to the front.
[23:35] <mwhudson> lifeless: ah yes, will look at oops-wsgi
[23:38] <lifeless> mwhudson: so, slight diversion to re-re-tweak my de
[23:38] <lifeless> lava-scheduler-app-device-tags lists tag, device
[23:38] <lifeless> testjob tags lists tags for tests
[23:39] <lifeless> so you're grabbing the count of test tags that are not in the device tag list
[23:40] <lifeless> what you want to state is, I think, 'test jobs that have at least the devices tags'
[23:43] <lifeless> a first tweak would be 'has_tags': 'select count(*) = 0 from ...'
[23:45] <lifeless> meh, it works, and you're querying one device, shrug :)
[23:46] <lifeless> you could add a CTE if your tag table isn't well indexed.
[23:48] <mwhudson> it's django i assume there are no indexes :)
[23:48] <mwhudson> but easy to add though
[23:49] <mwhudson> the primary key on id is probably enough here though...