[12:28] <gary_poster> bac benji gmb, call in 2
[12:28] <bac> ok
[12:41] <gary_poster> bac benji gmb, if anyone wants to take Julian up on his offer/request on the internal mailing list to explain what's going on with process-death-row, please do, and let the rest of us know
[12:42] <benji> I have to admit I've ignored email thus far today.
[12:42] <gary_poster> :-)
[12:49] <gary_poster> btw, I have another visit to the garage in 30 min or so
[13:00] <bac> gary_poster: did you try upgrading to oneiric?
[13:01] <gary_poster> bac, I tried upgrading to natty first.  Horrible failure. :-/
[13:01] <gary_poster> but then my maverick was hosed it seemed
[13:01] <bac> gary_poster: oh, you *are* behind
[13:02] <gary_poster> bac, I have an clean oneiric waiting for me to set it up
[13:02] <bac> i thought grabbing beta 1 (like we're supposed to) was brave
[13:02] <gary_poster> I just keep on putting it off
[13:02] <gary_poster> it runs, I just don't have my dev environment
[13:27] <bac> gary_poster: would you run 'ec2 images' and tell me the newest one reported.  should be 521
[13:27] <gary_poster> bac, it is.
[13:27] <bac> thx
[13:39] <gary_poster> bac benji gmb, at wgrant's request, we need to get  r13891 deployed in about 4.5 hours after buildbot run, if possible.  he warns: "buildbot has been a little unreliable lately: if this test run fails, get a LOSA to check for and kill any orphaned processes on pigeonpea. bzr forking services and rabbitmqs have been leaking occasionally."
[13:39] <gary_poster> mthaddon adds: "I'd rather we figured out why these processes are hanging around - seems like we've had to do this kind of clean up quite a lot recently"
[13:39] <gary_poster> So, let's try to keep an eye on buildbot.  I have an alarm for when the suite is supposed to pass if nothing fails.
[13:39] <gary_poster> Meanwhile, I'm off to the garage, and will be back soon hopefully.
[13:40] <gmb> noted.
[13:40] <benji> k
[13:40] <bac> check
[14:38] <gary_poster> benji, our squad has two extra tasks today.  I kinda figure you and I maybe ought to take them, since bac and gmb had extra CHR fun on Monday and left us with such a nicely clean plate.
[14:38] <gary_poster> One is to try to understand what the heck bigjools is talking about with the email he sent.  He estimates 10 minutes, but honestly I'm concerned that this is insufficient for the time to be spent valuably, at least if I'm the one listening.  I tend to start nodding my head in desperation when I hear much about soyuz.
[14:38] <gary_poster> I suspect you know more than I about packaging.  I also suspect that this is the less onerous of the two tasks.  I'm happy for you to take this one, if you wish.
[14:38] <gary_poster> The second task is to be buildbot-daddy and deployment-daddy, and make sure that gets to completion, as I mentioned before.  AFAICT, tests are doing OK so far.  That said, if something fails, it may involve trying to figure out what is going on wrong with https://lpbuildbot.canonical.com/builders/lucid_lp/builds/1342/steps/shell_6/logs/summary
[14:38] <gary_poster> I'm willing to take that one.
[14:38] <gary_poster> To be clear, I think task one is easier, so I'd be happy to take that, if you want the other one. ;-)
[14:39] <benji> I'm glad to pretend to know more than you do about soyuz.
[14:40]  * benji finds the email to read.
[14:45] <gary_poster> thanks benji
[15:05] <gary_poster> CHR!
[15:10] <benji> What is it good for?!
[15:11] <gary_poster> absolutely nothing!
[15:15] <gary_poster> the song is good for working to though
[15:26] <benji> I saw someone reccomend working to dubstep, the first thing that came to mind was this video: http://www.youtube.com/watch?v=JEcj2xi9Ctc
[15:39] <gary_poster> And CHR is done!  ...and no, I don't think I'd work to that :-P
[15:46] <benji> I can't believe I understood what Julian was talking about. ;)
[15:46] <gmb> gary_poster: Remind me again: how are we supposed to do UI JS integration testing these days? ISTR that we killed windmill and yet lp/bugs/windmill is still staring at me like the creature from the black lagoon.
[15:47] <benji> IIRC, we are using windmill for what it's really meant for (very high level functional tests) and the YUI bits for other JS things
[15:48] <gary_poster> gmb, heh.  This was the project I was mentioning that Deryck hasn't been able to get back to, that I'm trying to get into a landable state just now.  It will be YUI tests that have simple Python fixtures for a backend.  It's all there (at least as far as it working) except hooking it into the test suite.
[15:48] <gary_poster> We intend to sump winmill entirely
[15:48] <gary_poster> dump
[15:48] <benji> (I actually like that song, the miming is hilarious though.  I rairly work to music though, too distracting.)
[15:49] <gary_poster> and Deryck will be getting rid of all windmill tests
[15:49] <gary_poster> we expect to add selenium for acceptance-test style things later
[15:49] <gary_poster> but that means more set-up of some sort that I've forgotten now
[15:49] <gmb> gary_poster: Right. I ask because Aaron has demanded tests - which is fair - but I don't want to write Windmill tests if I can help it. I'm tempted to file it under "tests will be added in a later branch once $foo_work is complete" but I don't know if that's acceptable.
[15:50] <gmb> (Well, since I'm saying it, I know it's not, acceptable, but I was hoping there was some non Windmill solution)
[15:52] <gary_poster> gmb, you could say you are blocked on me finishing this, then. I might be done today.  Then you will be using the barely tested new test functionality, of course, but somebody has to be first.... :-P
[15:52] <gary_poster> well, it has plenty of tests
[15:52] <gmb> I like a challenge :)
[15:52] <gary_poster> but no-one has actually tried to use it yet except to test itself
[15:52] <gmb> But yeah, it seems kind of pointless doing windmill stuff if we're just going to have to re-write it soonish.
[15:52] <gary_poster> heh, ok, gmb, I'll keep you apprised then, and try to get this done post-haste
[15:52] <gmb> Cool, thanks.
[15:54]  * benji woories for his connectivity, given this thunderstorm.
[17:40] <gary_poster> benji, I'm about to report on the list that your qa-bad r13889 (from bug 742662) is, because of an overlapping qa-bad that was introduced in r13862 that was just fixed *after* your revision, blocking any deployment.  :-/  Please roll that back asap :-)
[17:40] <_mup_> Bug #742662: Mixed new line markers causing OOPS while importing translations <bad-commit-13889> <oops> <qa-bad> <rosetta-imports> <Launchpad itself:Fix Committed by benji> < https://launchpad.net/bugs/742662 >
[17:41] <benji> gary_poster: will do (/me tries to remember exactly how to do that)
[17:42] <gary_poster> benji, can help if you want.  I've had to do it recently
[17:43] <benji> gary_poster: thanks; I assume it starts with a new branch to which I merge the reverse revisions and then either bin/ec2 land or pqm-submit
[17:44] <gary_poster> benji, yes, with a particular command-line argument there at the end.  Look in the --help output for it
[17:44] <benji> gary_poster: argument for which step?
[17:45] <gary_poster> benji, ec2 land or bzr lp-land (I almost never use pqm-submit directly these days)
[17:45] <benji> k, thanks
[17:45] <gary_poster> np
[20:01]  * bac chrs
[20:14]  * bac done!