[00:00] <jelmer> mwhudson: no - what address did you send it to?
[00:00] <mwhudson> firstname.lastname@canonical.com
[00:00] <mwhudson> i think
[00:00] <mwhudson> maybe i typoed
[00:02] <mwhudson> mm can't find it now
[00:04] <jelmer> typos are not uncommon with my surname :)
[00:06] <thumper> mwhudson: I'm sure jelmer@c.c would get to jelmer as well :)
[00:06] <mwhudson> Aug 11 14:47:37 grond postfix/local[5352]: 4973011603A: to=<jelmer.vernooij@canonical.com>, relay=local, delay=0.53, delays=0.21/0.26/0/0.07, dsn=5.1.1, status=bounced (unknown user: "jelmer.vernooij")
[00:06] <mwhudson> jelmer: that's right though isn't it?
[00:07] <mwhudson> maybe this was when the sender was wrong
[00:07] <jelmer> mwhudson, yeah
[00:08] <mwhudson> oh heh, the bounce got rejected with the same message
[00:08] <jelmer> mwhudson: if it's the sender that is the problem the error message is whacky
[00:08] <mwhudson> yeah
[00:09] <mwhudson> jelmer: anyway
[00:09] <mwhudson> jelmer: what does remain to be done to import non-master branches from git on launchpad?
[00:10] <jelmer> mwhudson: Basically, everything is there. The one thing that is missing is the UI in Bazaar to parse branch names in URLs
[00:10] <lifeless> which you just need to rework - push down to transport ?
[00:11] <lifeless> which will make a small patch too :)
[00:11] <jelmer> lifeless: yes
[00:11] <mwhudson> jelmer: ok, that was roughtly what i thought
[00:11] <mwhudson> thanks
[00:11] <jelmer> lifeless: and then it needs to be used in the BzrDir as well
[03:05] <thumper> lifeless: ping
[03:10] <lifeless> hey
[03:10] <lifeless> poolie: your webapp patch is landed
[03:10] <thumper> lifeless: I have some bzr related questions if you have a few minutes
[03:10] <lifeless> poolie: you might like to send a quick howto to the list or something.
[03:10] <lifeless> thumper: shoot
[03:11] <thumper> skype would be easier for me, but just finishing something off
[03:11] <lifeless> kk
[03:11] <lifeless> I'm signed in
[03:11] <lifeless> just connect when you're ready
[03:15] <poolie> hi lifeless, thanks  for that
[03:15] <poolie> lifeless: is it live on edge or staging yet?
[03:15] <poolie> good idea
[03:15] <lifeless> poolie: nope, just hit devel
[03:15] <poolie> apparently not live
[03:16] <lifeless> in ~ 6 hours it will get through buildbot and be deployed to edge at the next edge update
[03:16] <lifeless> and ~ 14 hours should see it on staging
[03:32] <lifeless> I mean
[03:32] <lifeless> from bzrlib import ui
[03:32] <lifeless> foo(ui)
[03:45] <thumper> I'm running everything in the launchpadzopelesslayer
[03:46]  * spiv misread "running" as "ruining"
[03:47] <lifeless_> thumper: so
[03:47] <lifeless_> internets fail
[03:47] <thumper> :)
[03:48] <thumper> spiv: I'd like to ruin most of them as they are probably doctests
[03:48] <lifeless_> error-report | subunit-filter -s --no-passthrough | subunit-ls > all-tests.klist
[03:48] <lifeless_> where error-report is the log file from ec2 (you might need a gunzip -c in there)
[03:49] <lifeless> then
[03:49] <lifeless> copy that file
[03:49] <lifeless> remove all the stuff *after* the failing test
[03:49] <lifeless> and remove 1/2 of the stuff *before* the failing test
[03:49] <lifeless> and then
[03:49] <lifeless> bin/test --load-list the-edited-file
[03:58] <spiv> thumper: yeah, I thought "he's ruining launchpadzopelesslayer?  fair enough..." ;)
[08:29] <adeuring> good morning
[09:17] <mrevell> Hi
[09:20] <lifeless> hiya
[09:20] <lifeless> have we have any feedback on the release?
[11:57] <deryck> Morning, all.
[13:22] <jml> hey
[13:22] <jml> I get a stack trace in Launchpad when I mark a bug as a duplicate of itself
[13:23] <jml> is that because I'm special?
[13:24] <wgrant> lazr.restful notices that you're in ~launchpad, so it gives you one. lazr-js responds by rendering the whole several hundred lines of it in the widget.
[13:24] <wgrant> Isn't it great?
[14:44] <jml> really want my testr branches reviewed.
[14:46] <jelmer> hi jml
[14:46] <jml> jelmer, hi
[14:47] <jelmer> jml: Is this for testr itself?
[14:47] <jml> jelmer, yes.
[14:47] <jelmer> jml: I would review, but I'm not a committer...
[14:48] <jelmer> (yet..)
[14:48] <jml> jelmer, thanks.
[14:48] <jml> jelmer, I'm hoping lifeless will make me a committer soon.
[15:00] <jml> jelmer, actually, some of my branches conflict with each other :)
[15:04] <al-maisan> jtv: thanks for pointing out the forthcoming bytea array syntax changes in postgres9 :)
[15:04] <jml> I can't believe we have to run our tests under xvfb
[15:05] <mars> well, only one suite needs that
[15:05] <jml> and yet all pay the price for it.
[15:05] <mars> the price being?
[15:06] <mars> I didn't think it took that much to start Xvfb
[15:06] <jml> I didn't say it was a high price :P
[15:15] <daker> hi
[15:15] <jml> daker, hi
[15:15] <daker> i have a proposal for the LP homepage
[15:15] <daker> http://ubuntuone.com/p/Bg6/
[15:17] <jml> daker, cool.
[15:18] <jml> daker, would the big images under "Get more contributors" etc. be links?
[15:18] <jml> daker, if so, where to?
[15:18] <daker> yes links
[15:18] <daker> and the rounded box should icons
[15:18] <daker> be
[15:19] <barry> EdwinGrubbs: ping
[15:19] <jml> on a completely different topic, the importfascist and the warninghandler work together so as to prevent one from getting a useful warning if something warns on import.
[15:19] <EdwinGrubbs> barry: hi, Barry. do you want to talk about mailman now?
[15:19] <jml> daker, where would they link to?
[15:20] <barry> EdwinGrubbs: sure! irc would work for right now.  i'm not set up for skype or mumble atm, but could be with a few minutes work
[15:20] <daker> jml, i don't konw may be a HELP pages, that shows the users how they can use LP
[15:21] <jml> daker, fair enough.
[15:21] <jml> daker, what would be different for people who are already logged in?
[15:22] <EdwinGrubbs> barry: ok, my first question is: can I look up the email that caused that oops so that I can understand how the decoding error was triggered?
[15:23] <barry> EdwinGrubbs: the losas will have to help.  the message should be in the shunt queue.  the trick is going to be finding it.  everything in the shunt queue is a python pickle of the message object and the metadata dictionary.  the file names do not encode the message-id
[15:23] <barry> and i think the shunt file name is not encoded in the oops
[15:24] <barry> EdwinGrubbs: i'm not even sure the message-id is available in the oops report
[15:24] <EdwinGrubbs> barry: is that too large for me to ask the losas for the whole thing?
[15:24] <mrevell> Thanks daker, that's very cool. The box in the centre at the top, I'm not sure what that is.
[15:24]  * jml is tempted to delete both the warning handler & the import fascist.
[15:24] <EdwinGrubbs> barry: does it have a date column I could limit it by?
[15:24] <barry> EdwinGrubbs: probably best would be to ask them to sync the shunted messages from the date of the oops
[15:24] <barry> yep
[15:25] <EdwinGrubbs> ok
[15:25] <barry> EdwinGrubbs: in a built launchpad lib/mailman/bin/dumpdb <messagefile> can be used to dump out the shunt messages
[15:25] <barry> EdwinGrubbs: it shouldn't be too hard to find the offender from that
[15:26] <jelmer> jml: it would be nice to have the import fascist as part of 'make lint' but not as something that breaks everything.
[15:26] <barry> EdwinGrubbs: i'm around all day so please do ping me with any questions
[15:26] <jml> jelmer, yeah. landscape have an "import guardian" which is nicer in many ways
[15:26] <jml> jelmer, I once tried to split it out, make it a lazr project & re-use it
[15:26] <jml> jelmer, but got blocked on the "make a lazr project" part.
[15:27] <jml> maybe I should have another try now that I know a little more about buildout & friends.
[15:27] <daker> mrevell, the rounded boxes should take you to a some kind of help, and it should shows how to use the chosen functionality
[15:28] <mrevell> daker, what about the box at the top of the page that looks a little like an envelope?
[15:29] <EdwinGrubbs> barry: do you think that I should provide any different handling for a bad rfc822msgid right now, for example just bouncing the email back to the sender? If it should just raise an oops like it does now, I was thinking that the xmlrpc args in the oops should be improved to introspect the xmlrpc.Binary object so that we don't have to download the shunt messages in the future.
[15:29] <daker> mrevell, ah that's the logo of LP :D
[15:29] <mrevell> heh, ah right :)
[15:31] <mars> jml, I can probably do the lazr conversion for you pretty quickly.  I've lightened the lazr spec a bit, it just has not been formalized yet
[15:31] <barry> EdwinGrubbs: we could possibly be rejecting bad message-ids at the source, either in the mta, or earlier in mailman, but it's also possible this is a shallow bug.
[15:31] <jml> mars, that'd be great, thanks
[15:31] <mars> jml, I'm running the qa-tagger through a lazr conversion right now.  We'll see how long it takes.
[15:31] <jml> mars, although I'd also love to be able to do it quickly for myself.
[15:32] <jml> mars, I used lazr.newproject or something last time
[15:32] <jml> and all I got was incomprehensible errors :\
[15:32] <mars> jml, yep.  That is a bit heavier than what I am doing for new projects now.
[15:32] <jml> mars, should it be updated to be more like what you do now?
[15:33] <barry> EdwinGrubbs: i need to reacquaint myself with that code, but i don't have a launchpad build right now ;)  does lp build on maverick these days?  can i just do a rocketfuel-setup still?
[15:33] <mars> jml, it should, and I plan to.
[15:33] <jml> mars, cool. :)
[15:35] <EdwinGrubbs> barry: I think somebody is running launchpad on maverick, but I can't find the email to confirm that it works fine.
[15:35] <jelmer> EdwinGrubbs: Yes, I am.
[15:36] <jelmer> EdwinGrubbs: I am running with a backported version of python-psycopg2 though
[15:36] <jelmer> s/backported/older/
[15:36] <barry> i can bring up a lucid vm if necessary
[15:36] <EdwinGrubbs> barry: jelmer says it works for him.
[15:36] <barry> jelmer: cool.  i'll give it a shot
[15:37] <barry> EdwinGrubbs: cool.  i'll move upstairs in a bit and give it a try
[15:38] <jelmer> barry: you'll need the python-psycopg2 from lucid as 2.2 (which is in maverick) is a lot more pedantic and breaks lp
[15:38] <EdwinGrubbs> barry: will the losas know what to do if I just ask for "shunt messages" between specific dates, or do I need to point them at some docs?
[15:38] <barry> jelmer: k, thx
[15:38] <EdwinGrubbs> barry: or does that dumpdb thing take a date arg?
[15:39] <barry> EdwinGrubbs: queue/shunt should be enough i think
[15:39] <barry> EdwinGrubbs: no, it only takes a file name
[15:39] <barry> EdwinGrubbs: but the date is encoded in the file name (along with a sha1 hash of the file contents, but you don't care about that :)
[15:40] <daker> mrevell, i should be able to make it real
[15:43] <mrevell> daker, I'd love to see more. We're not going to change the home page right now but we plan to in the next few months. I'm about to start researching what LP users want from the home page, so that we have a basis for any changes we make.
[15:43] <daker> brb
[15:46] <barry> EdwinGrubbs: brb
[15:47] <jtv> al-maisan: it sure gave me enough trouble for me to notice it…  I still run daily tests of e.g. postgres 9.0 using libpq 7.3
[15:47] <daker> mrevell, good
[15:48] <al-maisan> jtv: ahm
[15:48] <al-maisan> s/ahm/ah, I see :)/
[15:52] <EdwinGrubbs> barry: mbarnett doesn't know what the shunt messages are? Do you know what server or directory they are in?
[16:08] <deryck> gmb, hey.  So we don't currently have plans among the work we divided up between you and Abel for the "view all my subscriptions page," right?
[16:08] <deryck> gmb, I chatted with bdmurray about taking this on, which is why I ask.
[16:09] <gmb> deryck, Nope no plans (I responded to your ping last night but you'd already left)
[16:09] <gmb> deryck, It would be great if someone could tackle it this cycle.
[16:09] <deryck> gmb, yeah, no worries.  I loose track of the TZ sometimes when I ping. :-)
[16:09] <gmb> deryck, Maybe I should make my proxy respond with the local time when I've disconnected :)
[16:10] <deryck> gmb, bdmurray is definitely interested, so I suggested he start the first part of it -- hooking up a view, displaying what we can for now -- and then when he's blocked waiting on your work and Abel's, then he can do individual bugs again.
[16:10] <gmb> deryck, That sounds like an excellent idea.
[16:11] <deryck> cool
[17:30] <mrevell> Hey, I have a failing test that I don't know how to fix. The failure message doesn't look like something I've seen before. It is:
[17:30] <mrevell> FAILURE: canonical.launchpad.webapp.tests.test_login.TestOpenIDReplayAttack.test_replay_attacks_do_not_succeed (subunit.RemotedTestCase)
[17:30] <mrevell> Anyone able to lend a hand?
[17:32] <benji> mrevell: I can take a look at it in a few minutes.
[17:32] <mrevell> Thanks benji :)
[17:34] <bac> hi mrevell
[17:35] <mrevell> hey there bac
[17:36] <bac> mrevell:  where did you see that failure?  i just ran that test in isolation and it passes
[17:36] <mrevell> bac, On my branch: https://code.edge.launchpad.net/~matthew.revell/launchpad/top-right-name-bug134957
[17:37] <bac> mrevell:  on ec2?
[17:37] <mrevell> bac, yeah
[17:44] <bac> mrevell:  did you consider displaying both the name and displayname?
[17:44] <bac> e.g. 'Brad Crittenden (bac)'?
[17:45] <mrevell> bac, Yeah, but decided that it could get very long ... I'd Matthew Revell (matthew.revell) for example. Would you particularly miss seeing the displayname?
[17:45] <bac> dunno.
[17:46] <bac> mrevell:  i think that test failure is valid
[17:47] <bac> mrevell:  It includes self.assertIn('Sample Person', login_status)
[17:47] <benji> mrevell: ok I'm available now, still looking for an extra pair of eyes?
[17:47] <bac> mrevell:  so i think you need to update the test for your new expected output
[17:48] <bac> mrevell:  have you tried running it locally?
[17:48] <bac> bin/test -vvt TestOpenIDReplayAttack
[17:49] <mrevell> Ahhhh, I see, that's the name of the test? Right. Thanks bac, that's what I was missing. I didn't know which test was failing.
[17:49] <mrevell> Thanks benji but I think bac has helped me.
[17:49] <benji> cool
[17:53] <jml> is there any particular reason for ec2 test to go out of its way to nicely format the html that is served from instance?
[19:20] <jml> g'night all
[19:20] <jelmer> have a good weekend jml
[20:23] <EdwinGrubbs> rockstar: ping
[20:26] <EdwinGrubbs> matsubara-afk: ping
[20:52] <rockstar> EdwinGrubbs, pong
[20:54] <EdwinGrubbs> rockstar: I was wondering if you were familiar with the job system and whether we could have a pre-impl call on that.
[20:54] <rockstar> EdwinGrubbs, sure, lemme fire up skype.
[20:57] <rockstar> EdwinGrubbs, I don't see you on skype.
[20:58] <EdwinGrubbs> rockstar: it doesn't seem to want to connect.
[20:58] <rockstar> EdwinGrubbs, :(
[20:59] <EdwinGrubbs> rockstar: I haven't used it in a while since I just use mumble.
[20:59] <rockstar> EdwinGrubbs, yeah, my headset that liked mumble has given up the ghost.
[21:05] <lifeless> hi EdwinGrubbs
[22:14] <EdwinGrubbs> salgado: ping
[22:14] <salgado> hi EdwinGrubbs
[22:16] <EdwinGrubbs> salgado: do you know if it is possible to include more information in oopses than just adding individual lines to the Request Variables section?
[22:19] <lifeless> EdwinGrubbs: it is
[22:19] <salgado> EdwinGrubbs, not really, but Ursinha or matsubara-afk should know
[22:19] <lifeless> EdwinGrubbs: we may need to tweak the api - but thats easy
[22:19] <Ursinha> EdwinGrubbs, matsubara-afk knows it best
[22:20] <EdwinGrubbs> oh, so you pass the buck to the guy who isn't around. I see out it works.
[22:20] <lifeless> EdwinGrubbs: what do you want to record?
[22:22] <EdwinGrubbs> lifeless: the email that mailman sends over via xmlrpc. Currently, the losas have to go look for a files in one directory based on the unix timestamp that is in the filename. The files themselves are pickles, so searching it isn't fun.
[22:23] <lifeless> I'm not familiar with the mailman integration
[22:23] <EdwinGrubbs> lifeless: I've found the part of the code where stuff gets added to the oops report, but I don't know if I'll blow up the oops.py.
[22:23] <lifeless> for my curiousity, could you give me a little extra context?
[22:23] <EdwinGrubbs> here's the bug with the oops: https://bugs.edge.launchpad.net/launchpad-registry/+bug/615655
[22:23] <_mup_> Bug #615655: UnicodeDecodeError in xmlrpc holdMessage <mailing-lists> <oops> <Launchpad Registry:In Progress by edwin-grubbs> <https://launchpad.net/bugs/615655>
[22:24] <EdwinGrubbs> lifeless: as you'll see in the oops, it just tells you that the xmlrpc args contains an <xmlrpclib.Binary> object. not very helpful.
[22:24] <lifeless> so, check oops-tools
[22:24] <lifeless> to see if it will go boom
[22:24] <lifeless> it shouldn't, but you'll need to tweak it anyhow to show new data
[22:25] <lifeless> unless you put it in req_vars or db-statements which would be a bit of a hack
[22:25] <lifeless> ok, so mailmain makes an xmlrpc call, giving us the email for <some reason> ?
[22:25] <EdwinGrubbs> I'll look at that.
[22:28] <lifeless> so, I'd like to have a free form 'extra detail' section
[22:28] <lifeless> oops-tools needs to change to show it in the lp-oops web service
[22:28] <lifeless> and we need an api to add stuff to it
[22:28] <lifeless> if we had that, would it be sufficient for your needs?
[22:29] <EdwinGrubbs> right, it looks like there is a bug open to make the oopsMessage() context method do that instead of sticking things in the req vars.
[22:30] <EdwinGrubbs> a free form extra detail area would definitely work for including email messages.
[22:34] <lifeless> \o/
[22:34] <lifeless> so why does mailman xmlrpc call us ?
[22:36] <EdwinGrubbs> lifeless: for moderating emails, which is done in a Launchpad page.
[22:39] <lifeless> thanks
[22:50] <lifeless> EdwinGrubbs: are you planning on doing the extension work on oops to make this happen ?
[23:01] <EdwinGrubbs> lifeless: well, assuming that matsubara-afk doesn't tell me any reason it shouldn't be done on Monday. I definitely want to do it, and it doesn't look like it could possibly take more than a day of work on my side.
[23:14] <lifeless> \o/