[00:25] <EdwinGrubbs> wgrant: ping
[00:29] <wgrant> EdwinGrubbs: Hi.
[00:55] <EdwinGrubbs> wgrant: hi, I was wondering if you had any opinion about hiding the links at the top of the +related-software page if they are empty. There would always be a link on the person/team page pointing to +related-software, but the +uploaded-packages, +maintained-packages, +ppa-packages, and +related-projects links would be hidden if they were empty.
[01:21]  * mwhudson tries to pound the puller into breaking locally
[01:36] <spm> mwhudson: :-)
[01:41] <mwhudson> completely unsuccessfully so far :(
[01:42] <wgrant> EdwinGrubbs: That's how I imagined it was intended to be.
[01:42] <spm> mwhudson: "if it was easy, we wouldn't have given it to you" <== story of my life and one you could take to heart as well ;-)
[01:43] <wgrant> What is the normal slave replication lag, and what is the maximum before requests all go to the masteR?
[01:53] <kfogel> dobey: pong
[01:56] <spm> wgrant: bugger and all; don't recall offhand - but not that long. a few mins max.
[01:56] <wgrant> spm: A few minutes is enormous...
[01:56] <spm> tradeoffs. most folk aren't modifying.
[01:57] <wgrant> API requests will now use a slave where possible. This is going to cause a lot of clients to get very angry.
[01:57] <gary_poster> wgrant: I don't think that is the case.  What makes you think that?
[01:57] <wgrant> gary_poster: See devel r9703
[01:58] <gary_poster> I think I know what that is, but double-checking.
[01:58] <wgrant> Oh. session cookie only.
[01:58] <wgrant> I guess that's less unreasonable.
[02:00] <gary_poster> right.  this is for JS browser users.
[02:03] <mwhudson> wgrant: spm is wrong btw, lag is usually 5-10 seconds
[02:03] <gary_poster> mwhudson: hi.  I have a bzr plugin problem with eggs in the branch where I'm trying to apply voodoo to make the build happy on karmic.  I've "fixed" the horrible namespace package problem.
[02:03] <gary_poster> Now I only have two failures, both for the same reason. test_successful_start_then_stop and test_unexpected_error_logs_oops fail because they start up bzr, which tries to load plugins from all possible places including the system's site-packages, even though I'm not using the system 's bzr.
[02:04] <gary_poster> If you have plugins installed then bzr complains, because they cannot be imported.  This causes test failures in assertFinishedCleanly: return code is 0, yes; stdout is empty, yes; but stderr is not empty: in my case, it is the following string:
[02:04] <gary_poster> "No module named dbus.bus\nUnable to load plugin 'dbus' from '/usr/lib/python2.4/site-packages/bzrlib/plugins'\nNo module named dbus.bus\nUnable to load plugin 'avahi' from '/usr/lib/python2.4/site-packages/bzrlib/plugins'\n"
[02:04] <gary_poster> I have stepped through the bzr code and I do not see a way to tell the bzr commandline to please only use the plugin path I'm providing, and don't look anywhere else, especially not in the system site-packages.
[02:04] <gary_poster> Therefore, what I'm planning on doing is changing assertFinishCleanly to check that the return code is 0; check that the stdout is empty; and check that stderr is entirely comprised of 0 or more of those plugin error messages, using a regex.
[02:04] <gary_poster> Can you get behind that?  The only other option I can see is making another LP-specific bzr egg change, which I am loathe to do.
[02:04] <gary_poster> (done)
[02:04] <spm> mwhudson: heh. that's a variation of definition. 5-10 secs to me is bugger and all. :-)
[02:04] <mwhudson> gary_poster: where are these tests?
[02:05] <wgrant> mwhudson: It's sometimes a lot more, though.
[02:05] <mwhudson> spm: you said "a few minutes max" -- it doesn't get as high as a minute very often though does it?
[02:05] <gary_poster> mwhudson: lp.codehosting.tests.test_lpserve.TestLaunchpadServe
[02:05] <mwhudson> wgrant: generally when something breaks, i think
[02:05] <mwhudson> gary_poster: looking
[02:05] <gary_poster> thank you
[02:05] <spm> mwhudson: very rarely - is something critical that we alert on.
[02:07] <mwhudson> gary_poster: ararargh!
[02:07] <gary_poster> mwhudson: well, yes
[02:09] <mwhudson> gary_poster: i guess your solution is ok
[02:09] <mwhudson> gary_poster: so long as there's a big fat XXX pointing at a bzr bug i'll find in a sec
[02:09] <gary_poster> mwhudson: heh, that sort of XXX would make me feel better, yes.
[02:11] <mwhudson> wth it's Fix Released
[02:12] <mwhudson> oh, but very recently
[02:12] <mwhudson> gary_poster: i think your patch is ok then, we should remove the hack when we upgrade bzr
[02:12] <mwhudson> gary_poster: https://bugs.edge.launchpad.net/bzr/+bug/316192 fwiw
[02:12] <mup> Bug #316192: bzrlib.plugin.load_plugins() loads system plugins even for non-system bzrlib <Bazaar:Fix Released by vila> <Launchpad Bazaar Integration:Fix Released by thumper> <https://launchpad.net/bugs/316192>
[02:16] <gary_poster> mwhudson: awesome that is fixed.  OK, thank you very much.
[02:17] <mwhudson> gary_poster: np
[02:57] <mwhudson> spm:  has u1 stopped exploding yet>
[02:57] <mwhudson> ?
[02:57] <wgrant> It's working enough for me to have discovered 5 XSS vulnerabilities in the past hour.
[02:58] <wgrant> (I am not amused)
[03:01] <mwhudson> :(
[03:08] <spm> mwhudson: sorta. in that the immediate stuff is now under control; now I move to the ZMOG 9999 priority stuff. I can do small simple things; else... busy :-/
[03:16] <mwhudson> spm: ok
[03:18] <mwhudson> spm: btw the thing i want to talk to you about is one of those things where priority -> \infty as we get closer to karmic being released :-)
[03:19] <spm> mwhudson: heh. figures.
[04:11]  * mwhudson updates his puller bashing to be less wasteful
[07:30] <spm> noodles775: 1. Heyo! 2. Bug 452730 for your attn :-)
[07:30] <mup> Bug #452730: lp_publish/ppa-generate-keys is broken atm <Soyuz:New> <https://launchpad.net/bugs/452730>
[07:30] <noodles775> looking...
[08:12] <noodles775> spm: do you know what the latency between the main and slave db servers is?
[08:13] <noodles775> (just chasing one possibility for the above bug)
[08:13] <wgrant> noodles775: Oh, I know exactly what that bug is.
[08:13] <wgrant> devel r9687
[08:13] <noodles775> wgrant: it's related to r9687.
[08:14] <noodles775> yep...
[08:14] <wgrant> There are now PPAs in production for users without a PPA named 'ppa'.
[08:14] <noodles775> wgrant: sure, but the new method that's been added doesn't assume the name is ppa - it just grabs the first one.
[08:14] <wgrant> noodles775: But that code is only on edge.
[08:14] <wgrant> noodles775: Not on whatever machine runs ppa-generate-keys
[08:14] <noodles775> wgrant: right.
[08:15] <wgrant> So Person.default_ppa still looks up a PPA named 'ppa', always.
[08:15] <noodles775> yep - I'll chat to bigjools about getting r9687 cp'd.
[08:15] <noodles775> Thanks wgrant
[08:15] <wgrant> That's what I suspected. Thanks.
[08:28] <wgrant> spm: U1 rollout not done yet, I hope?
[08:29] <spm> wgrant: partially.
[08:29] <adeuring> good morning
[08:30] <wgrant> spm: Great. I presumed it would be done after 3ish hours, but hoped it wasn't, given that it still seems broken.
[08:31] <spm> the app servers are done; but i'm having a mild pebkac issue with one of the back end services - unsure how related that is to the problems
[08:31] <wgrant> Uhoh.
[08:32] <spm> believed I didn't have the keys to push the new hotness up. turns out I did - in a different directory. le-sigh.
[08:37] <awilkins> jelmer: Why isn't Launchpad using subvertpy instead of python2.4-subversion  .... I'd guess it's an immense fiddle to have both subversion 1.6.5 and launchpad-dependencies because of the python bindings?
[08:40] <wgrant> awilkins: Launchpad doesn't use bzr-svn. It uses a cscvs, horrible horrid horror which uses some strange svn bindings and predates subvertpy by several years.
[08:42] <awilkins> wgrant: I'm presuming that means that LP-monitored SVN branches are essentially trapped and the only way to get patches back to the original project is to ... send patches?
[08:43] <wgrant> awilkins: Yes. There is an intention to move to bzr-svn soon.
[08:43] <awilkins> wgrant: Is there a roadmap the public can see? How about git integration? When will this pestering get too annoying? ;p
[08:44] <wgrant> awilkins: I am the public. I do not know of such a roadmap.
[08:44] <wgrant> 'git integration'?
[08:45] <awilkins> LP using git in the same capacity it uses bzr, where possible
[08:45] <wgrant> That is not on the roadmap even if it does exist.
[08:57] <wgrant> win 23
[09:04] <gmb> Morning padlaunchers.
[09:14] <mrevell> Morning!
[09:15]  * wgrant wonders whether to consider it safe to dive back into Canonical-produced code...
[09:15] <wgrant> maybe as long as it isn't webapp code.
[09:16] <awilkins> That on IP grounds, or just because it's scary? :-)
[09:18] <wgrant> Scary in bad ways.
[10:04] <jml> wgrant, that's simply not fair.
[10:05] <wgrant> jml: Hm?
[10:09] <wgrant> What have I done now?
[10:09] <jml> wgrant, Canonical-produced code isn't that scary
[10:09] <jml> and it's not a fair category.
[10:10] <wgrant> jml: It is scary in that I don't know how many more fucking obvious huge vulnerabilities I will discvoer.
[10:15] <jml> wgrant, oh ok.
[10:15] <jml> wgrant, yeah, that sucks. :)
[10:16] <wgrant> Although LP has less of them than U1, or they are less obvious. This is good.
[10:45] <jml> allenap, hi
[10:46] <jml> allenap, do you have any idea how to stop PQM from sending success messages to the Launchpad list every time we do an automatic merge from stable?
[10:46] <jml> (you don't have to, but I thought I'd ask someone)
[11:19] <jml> allenap, I've read your test parallelization email & lifeless's reply.
[11:19] <jml> allenap, I don't have much to add to what Rob said, except to say that I'm happy for you to call to talk about stuff if you think that'll help.
[11:39] <gmb> jml: Do you have any idea why I'd see http://pastebin.ubuntu.com/294573/ when running ec2 demo?
[11:39] <gmb> What do I need to do to ec2 to know who I am in LP?
[11:40] <jml> gmb, the unhelpful answer to your first question is that our ec2 tools lack tests.
[11:41] <gmb> Hah.
[11:41] <jml> gmb, give me a minute to finish what I'm doing now and I'll have a look & see what's going on.
[11:41] <gmb> jml: Ta muchley.
[11:49] <mrevell> Can anyone remind me where to find the API uses wiki page?
[11:50] <mrevell> Of course, it's https://help.launchpad.net/API/Uses
[11:50] <noodles775> lol
[11:55] <jml> gmb, ok. I'm all yours.
[11:56] <jml> gmb, can you paste me the full log of what you're doing, from command through to error?
[12:01] <jml> gmb, also, what do you get when you type 'bzr launchpad-login'
[12:01] <gmb> jml: So, full log: https://pastebin.canonical.com/23397/
[12:01] <gmb> jml: `bzr launchpad-login` gives me `gmb`
[12:02] <jml> gmb, thanks.
[12:02] <jml> gmb, I'm going to try reproducing the problem locally.
[12:03] <gmb> ok
[12:05] <jml> "I told him to go away and reproduce his problem, but not in those words exactly"
[12:05] <jml> (why I'm not Woody Allen)
[12:06] <jml> gmb, try applying this patch http://pastebin.ubuntu.com/294591/. You'll need to do it in a branch that's not the branch that you want to demo (I recommend stable)
[12:07] <gmb> jml: Okay, bear with me a sec...
[12:13] <gmb> jml: Works. Thanks!
[12:13] <gmb> jml: r=me on the patch if it needs landing ;)
[12:13] <jml> gmb, https://code.edge.launchpad.net/~jml/launchpad/fix-ec2-demo/+merge/13461
[12:14] <gmb> jml: Approved.
[12:14] <jml> gmb, thanks.
[12:16] <jml> gosh ec2 land is handy.
[12:17] <bigjools> :)
[12:17] <bigjools> is there a mode on it to send straight to PQM?
[12:17] <jml> bigjools, no, I thought of doing that.
[12:17] <bigjools> 'twould be useful for us heathens
[12:17] <jml> bigjools, but then mwh said "what's the point?" and "your branch is too big already"
[12:18] <bigjools> I can think of many points
[12:18] <jml> bigjools, I'll happily review a patch for adding a --no-test option or similar.
[12:18] <jml> alternatively
[12:18] <bigjools> ok when I have some time I'll ask you for a pre-imp on that
[12:18] <jml> I'll trade you that for a vfs backend on soyuz so people can rename their accounts if they have PPAs.
[12:19] <bigjools> heh
[12:19] <bigjools> I suspect that would push the renaming problem elsewhere, but I am not unsympathetic to it
[13:02] <bac> bigjools, jml: i though of adding a --no-test option, but it seemed a little perverse since the point of the script is to send things off to ec2
[13:02] <bac> bigjools: you can do 'ec2 land --print-commit' and combine the results with 'bzr pqm-submit'
[13:02] <bigjools> I think its use lies more in the automation of the perverse options we have to put in the commit string
[13:03] <bac> bigjools: i understand.  i've had the need myself but it still felt odd.
[13:04] <bac> bigjools: so an alias of:  bzr pqm-submit -m "`ec2 land --print-commit`" or some variation may work
[13:05] <bigjools> bac: that's so sick I love it :)
[13:05] <bac> :)
[13:05] <bac> bigjools: and it doesn't have to go through code review!
[13:15] <allenap> jml: I don't know how to prevent PQM from sending success messages, but I'll have a look.
[13:15] <allenap> jml: Thank you for reading through my message. I think Rob's reply is very useful to help me approach the problem, but I'll ask you about implementation as I go along.
[13:17] <jml> allenap, cool. don't worry too much about the PQM thing, I've now just set up a rule so I don't get those emails any more :)
[13:17] <allenap> jml: Jolly good :)
[13:42] <bac> sinzui: https://edge.launchpad.net/gnome
[13:44] <sinzui> bac: are you hinting that I fix the licenses for all those projects?
[13:45] <sinzui> bac: This page looks better
[13:49] <sinzui> bac: Do you know why the projects state they are proprietary when they are have a subscription: https://edge.launchpad.net/launchpad-project
[13:51] <sinzui> bigjools: What is the license for https://edge.launchpad.net/launchpad-buildd
[14:29] <bac> sinzui: nothing so deep.  i was showing you that our tiny little fix allows the gnome page to load without OOPSing
[14:29]  * sinzui nods
[14:30] <bigjools> sinzui: good question
[14:44] <sinzui> https://edge.launchpad.net/launchpad-registry/+milestone/3.1.10
[15:43] <Ursinha> err, better here
[15:43] <Ursinha> salgado: ping :)
[15:44] <Ursinha> salgado: hey, I'm checking the oopses here and I see occurrences of UnicodeDecodeError in +filebug
[15:44] <Ursinha> salgado: OOPS-1384C2001
[15:44] <Ursinha> err
[15:45] <Ursinha> salgado: https://lp-oops.canonical.com/oops.py/?oopsid=OOPS-1384C2001
[15:45] <Ursinha> salgado: is this related to bug 61171
[15:45] <Ursinha> ?
[15:45] <mup> Bug #61171: +login page OOPSes if query string has accented chars encoded as ASCII <oops> <Apport:Invalid> <Launchpad Foundations:Fix Released by salgado> <https://launchpad.net/bugs/61171>
[15:47] <salgado> Ursinha, no, this is something else
[15:56] <dobey> hey guys
[15:56] <dobey> is it possible to change the comment string on a GPG key that was generated for a user/team for PPAs?
[16:15] <Ursinha> salgado: I've filed another bug, bug 453203
[16:15] <mup> Bug #453203: UnicodeDecodeError in +filebug: unexpected code byte <oops> <Launchpad Bugs:New> <https://launchpad.net/bugs/453203>
[16:15] <Ursinha> maybe a bugs' person could take a look on it
[16:19] <Ursinha> hi allenap
[16:19] <allenap> Ursinha: Hi!
[16:19] <Ursinha> :)
[16:19] <Ursinha> allenap: do you have some time to take a look at bug 453203, please?
[16:19] <mup> Bug #453203: UnicodeDecodeError in +filebug: unexpected code byte <oops> <Launchpad Bugs:New> <https://launchpad.net/bugs/453203>
[16:22] <allenap> Ursinha: Looking now. That's an odd one.
[16:26] <Ursinha> thanks allenap
[17:50] <jcinpv> Running Ubuntu 9.04 in Virtualbox on an Intel Mac; After a recent update, the Window Manager does not launch on start up or restart, but does launch after logging out and back in. Going to System->Preferences->Windows, get error message: 'Cannot start the preferences application for your window manager; Window manager "unknown" has not registered a configuration tool.'  I have deleted .gnome2, .gnome2_private, .gconf, and .gconfd to
[17:50] <jcinpv> no avail and I have reproduced the problem in a new user account. Where to look?
[17:56] <Ursinha> jcinpv: I think you may find more useful advices on #ubuntu
[18:01] <mrevell> Have a good weekend all.
[18:02]  * rockstar lunches
[18:52] <jcinpv> Ursinha: Been there. No help.
[18:52] <jcinpv> They suggested I submit a bug.
[18:53] <jcinpv> Where do I submit a bug?
[19:04] <Ursinha> jcinpv: https://launchpad.net/ubuntu/+filebug
[19:57] <jcinpv> Ursinha: Thanks. Bug submitted.
[19:58] <Ursinha> np jcinpv, sorry not being that helpful
[20:05] <jcinpv> Ursinha: You were very helpful in pointing me to the right place to file a bug. My first time.
[20:07] <Ursinha> jcinpv: in this case, glad that helped :)
[21:37] <rockstar> abentley, chat?
[21:37] <abentley> rockstar: 1 sec
[21:38] <rockstar> abentley, just call when ready.
[21:48] <barry> are there any bugs folks still around?  allenap, gmb, intellectronica, deryck
[21:49] <rockstar> abentley, that lag is terrible.
[21:49] <abentley> rockstar: and unpredictable...
[22:18] <gary_poster> night all