[00:00] <wallyworld> maybe you have an old yui lying around?
[00:00] <wallyworld> perhaps make clean will make it show up
[00:00] <mwhudson> well, i just set up the devel environment just now...
[00:01] <wallyworld> hmmm.
[00:01] <wallyworld> when i get a spare moment later i'll try and see what we're missing from the jsbuild target to cause the issue
[00:02] <mwhudson> ah
[00:02] <mwhudson> i need to load widget-position-align
[00:02] <wallyworld> makes sense
[00:04] <mwhudson> it has to be said, yui3 makes for very nicely greasemonkeyable pages
[00:04] <wallyworld> yeah, from what i've seen yui3 rocks compared to earlier incarnations
[00:05] <wallyworld> too bad yui isn't more popular. seems like jquery has all the mindshare?
[00:06] <mwhudson> yay i can use the lazr widget with my lp changes
[00:10] <wallyworld> mwhudson: when you say lazr, you mean from the lazr-js tree? for lp, we've stopped using lazr-js separately and pulled all the source in the lp tree and are mainatining it there
[00:11] <mwhudson> wallyworld: no, i mean, not my hack that's in my greasemonkey script
[00:11] <wallyworld> ah ok
[00:12] <mwhudson> wallyworld: ok, i'd like to merge this branch i guess
[00:12] <mwhudson> wallyworld: if i propose it will you review it?
[00:12] <mwhudson> i'll run the yui tests if you tell me how :)
[00:12] <wallyworld> sure. but i'm only a provisional reviewer. i'm being mentored atm
[00:13] <wallyworld> mwhudson:  bin/test --layer=YUITest
[00:14] <wallyworld> mwhudson: curtis has done an awesome job packaging the yui tests to be able to be run from the command line without  requiring a browser
[00:15] <wallyworld> you can also run individual tests by loading the html page corresponding to the test js file in a browser
[00:15] <mwhudson> yay
[00:15] <mwhudson>   Ran 44 tests with 0 failures and 0 errors in 1 minutes 8.693 seconds.
[00:15] <wallyworld> whoot
[00:17] <mwhudson> hm, seems i don't need the editicon change any more
[00:17] <mwhudson> think it's still dead code though
[00:18]  * mwhudson repurposes wallyworld's existing mp
[00:20] <wallyworld> mwhudson: i did that mp to quickly see the diff :-) i was just about to mention that the change now seems redundant.
[00:34] <G> (potentially silly question alert) when running 'testr run', 'make run' should be killed right?
[00:35] <wgrant> G: Shouldn't need to be. But it's possible some dodgy tests might require it.
[00:36] <G> wgrant: okay thanks
[01:19] <LPCIBot> Project devel build #1,019: STILL FAILING in 3 hr 59 min: https://lpci.wedontsleep.org/job/devel/1019/
[03:17] <thumper> hi hackers
[03:17] <thumper> how is the long-poll stuff going?
[03:18] <thumper> I've been eagerly awaiting the auto-diff updating
[03:18] <thumper> but it isn't happening :(
[03:24] <StevenK> Waiting for red squad to finish derivation, and then they switch to long-poll
[03:29] <spm> StevenK: isn't the correct repsonse "patches accepted"?
[03:30] <StevenK> Damn it. I go on holidays for a week and I forget everything.
[03:30]  * StevenK books in "re-training" for Monday
[03:38] <wallyworld> thumper: how's dx?
[03:38] <lifeless> digitally extreme :P
[03:38] <wallyworld> lifeless: you had a kid yet?
[03:38] <thumper> wallyworld: interesting
[03:38] <lifeless> yes :)
[03:38]  * wgrant gives lifeless a few looks of disapproval.
[03:38] <thumper> lifeless: really?
[03:38] <wallyworld> lifeless: congrats. details please
[03:38] <wgrant> It was on G+ and FB hours ago!
[03:38] <thumper> did I miss the announcement?
[03:38] <wgrant> Social media, people :P
[03:38] <lifeless> g+ 2 minutes ago
[03:39] <lifeless> FB hours ago.
[03:39] <wallyworld> wgrant: i hate social media
[03:39] <lifeless> EBUSY.
[03:39] <thumper> lifeless: congrats
[03:39] <wallyworld> lifeless: so why ffs are you on irc?
[03:39] <lifeless> wallyworld: easier than ringing the 100 off folk that Want To Know
[03:39] <lifeless> s/off/odd/
[03:40] <wallyworld> understood
[03:40] <wgrant> Heh
[03:40] <wallyworld> lifeless:  so, for those of us who hate fb, g+ - boy, girl?
[03:41] <wallyworld> thumper: got my new coffee machine :-D Breville BES900
[03:41] <thumper> wallyworld: awesome
[03:41] <thumper> wallyworld: so now you never leave home?
[03:41] <wallyworld> thumper: nope :-)
[03:42] <wallyworld> thumper: so feature freeze on oneric. everything looking good?
[03:42] <wallyworld> is the minimise behaviour or global menu stuff fixed?
[03:42] <wallyworld> please say it is
[03:42] <wgrant> wallyworld: Heh, good is not how Ubuntu release cycles ever look until the last two weeks.
[03:43] <wallyworld> wgrant: understood. i was being optimistic :-)
[03:43] <thumper> wallyworld: no
[03:43] <wallyworld> thumper: you suck
[03:43] <wallyworld> :-P
[03:43] <thumper> wallyworld: no, you suck
[03:44] <wallyworld> sometimes
[03:45] <lifeless> wallyworld: answered privately :)
[03:45] <wallyworld> lifeless: ah, so it is. thanks
[03:46] <wallyworld> lifeless: love the slogan :-)
[04:04] <lifeless> indeed ;)
[04:21] <G> $ testr run - 3hrs 20min and counting, is this normal?
[04:21] <StevenK> Yes
[04:21] <StevenK> Sadly
[04:21] <G> on the plus side, I did find a failing test in my changes :)
[04:22] <StevenK> A full testsuite run will take approximately 4 hours and change
[04:23]  * spm quaintly recalls complaining about the 90minute test runs
[04:25] <G> (Don't take it for complaining, was just shocked/surprised)
[04:25] <jtv> StevenK, wgrant: I tried an experimental domination run (source only and without judging phase) on dogfood's Debian yesterday.  I don't think it did anything because it ran suspiciously fast… I'll have to check.
[04:25] <StevenK> G: I complain about it all the time.
[04:26] <G> the only real problem I see is the inability to have 'make run' and 'testr run' going at the same time
[04:26] <StevenK> That's why we use ec2 for test runs.
[04:26] <StevenK> But that can get expensive.
[04:26] <wgrant> G: What happens when you try? It should mostly work.
[04:27] <G> wgrant: address already in use
[04:27] <G> I'm guessing because testr has zope running on the same port that make run starts zope on
[04:27] <wgrant> That's meant to not happen any more :(
[04:28] <G> wgrant: seems to be happening on my VM that I set up
[04:29] <wgrant> Which layer?
[04:29] <wgrant> AppServerLayer just started up fine for me.
[04:29] <wgrant> and I have a dev appserver running.
[04:29] <G> http://pastebin.ubuntu.com/679377/
[04:29] <G> runlaunchpad.start_launchpad()
[04:45] <G> StevenK: by the way, thanks for the help last night
[04:45] <StevenK> G: I didn't think I helped much, but you're welcome. :-)
[04:46] <G> I ended up by solving it by turning the second template into a Macro in the end, but you got me thinking in the right direction
[05:01] <nigelb> Good Morning!
[05:04] <G> nigelb: good $localtime
[05:04] <nigelb> heh
[05:05] <G> testr run just finished
[05:05] <G> lots of bzr tests failed hmmm
[05:07] <wgrant> G: Which Ubuntu release are you using?
[05:07] <G> natty
[05:08] <G> I'm guessing that is why
[05:08] <wgrant> Possibly. The test suite is only regularly run on Lucid, but it used to work on Maverick and mostly works on Natty.
[05:08] <wgrant> And partly on Oneiric.
[05:08] <nigelb> Ursinha: Happy Birthday!
[05:11] <nigelb> StevenK: I DID IT.
[05:11] <nigelb> One branch with TDD :)
[05:12] <nigelb> (except now my test doesn't pass)
[05:12] <G> nigelb: tut tut tut ;)
[05:12] <nigelb> heh
[05:19] <LPCIBot> Project devel build #1,020: STILL FAILING in 3 hr 59 min: https://lpci.wedontsleep.org/job/devel/1020/
[06:03] <nigelb> what time zone is danilos?
[06:04] <G> nigelb: LP says UTC+02
[06:05] <nigelb> ah, right.
[06:11] <nigelb> How do I compare strings in regular expression
[06:11] <nigelb> I'm surpised that I have to ask this at all.
[06:16] <nigelb> Someone has thoughts on this? http://pastebin.ubuntu.com/679424/
[06:17] <wgrant> nigelb: It doesn't really look like you want a regular expression match.
[06:17] <nigelb> wgrant: I don't,no.
[06:17] <nigelb> what do I use instead?
[06:18] <wgrant> Does assertIn not do what you wish?
[06:18] <nigelb> Oh.
[06:20] <nigelb> Hrm, not sure how I would do that.
[06:23] <G> oh no....
[06:23] <G> my make lint was clean, until I modified the pagetests
[06:24] <wgrant> That's your first mistake right there.
[06:24] <nigelb> heh
[06:24] <G> now it includes various 'source has bad indentation' / 'source exceeds 78 characters'
[06:24] <nigelb> G: lint checks modified files
[06:24] <G> nigelb: yep, get that, just disappointed
[06:24] <nigelb> heh
[06:25] <wgrant> G: If it's not too huge, you should probably clean the doctest up, so it's not left for the next person to find the same way.
[06:25] <wgrant> But if it's a few hundred warnings like they sometimes are, meh.
[06:25] <G> should I try and fix it as part of the branch or propose a second merge to fix
[06:25] <G> looks /ike 68 errors
[06:25]  * G takes a look
[06:25] <wgrant> Most of them probably just need you to s/^   /    / or so.
[06:30] <nigelb> who's reviewer today? Technically lifeless I believe?
[06:30]  * nigelb checks
[06:30] <wgrant> Don't give him any ideas.
[06:30] <wgrant> I can have a look if nobody else is around.
[06:31] <nigelb> aha
[06:31] <nigelb> well, not ready yet.
[06:31] <G> wgrant: looks like the lint errors are a bit 'yeah right' to me
[06:31] <wgrant> G: Oh?
[06:32] <nigelb> OMG.
[06:32] <G> for instance, I can't see any bad indentation on ./lib/canonical/launchpad/pagetests/basics/notfound-traversals.txt
[06:32] <G> (line 7 that is)
[06:32] <nigelb> I totally didn't recognize this launchpad feature till now.
[06:32] <wgrant> G: The code blocks are meant to be indented by 4 spaces.
[06:32] <wgrant> To distinguish them from narrative.
[06:32] <nigelb> I was about to say that.
[06:32] <wgrant>     >>> if foo:
[06:33] <wgrant>     ...     bar
[06:33] <G> wgrant: oh....
[06:33] <wgrant> nigelb: Which feature?
[06:33] <nigelb> So, if I have a bug number in my branch name
[06:33] <nigelb> like foo-1234
[06:33] <wgrant> nigelb: That feature is about a month old.
[06:33] <nigelb> HA.
[06:33] <nigelb> Its awesome.
[06:33] <G> in that case I'd have to re-indent etc the whole file
[06:33] <wgrant> G: Yes. It's rather unpleasant. I just ignore it in that case.
[06:34] <wgrant> I might do a mass fixup over Christmas or something, when it's less likely to conflict with people.
[06:35] <nigelb> ..
[06:36] <G> nigelb: oh does it automatically associate the branch w/ the bug?
[06:36] <G> I just do 'bzr commit --fixes lp:<bug>'
[06:37] <nigelb> G: It doesn't. When I click the link to associate the bug, it already has the bug number from the branch
[06:37] <wgrant> So it suggests it.
[06:37] <nigelb> (well, I often forget to do that :D)
[06:37] <wgrant> But doesn't do it automatically.
[06:37] <G> oh yeah, I noticed that today too
[06:37] <nigelb> which is nice!
[06:37] <nigelb> But yah, I agree with wgrant, it needs to be a picker.
[06:38] <nigelb> wgrant: how would I use assertThat in this scenario
[06:38] <nigelb> https://code.launchpad.net/~nigelbabu/launchpad/specification-validation-59301/+merge/73631
[06:38] <nigelb> L77
[06:40] <wgrant> nigelb: You should probably use assertRaises.
[06:41] <wgrant> nigelb: It's normally just used to check the class of the exception, but it also returns the exception object.
[06:41] <wgrant> So you can then assertEquals on its content afterwards.
[06:41] <G> ha, just got my first buildbot e-mail :) a couple of hours ago
[06:42] <nigelb> Wait, why aren't you on Contributors page yet.
[06:43] <nigelb> G: FYI - https://dev.launchpad.net/Contributions
[06:43] <wgrant> Possibly the wiki upgrade... let me see.
[06:43] <nigelb> we're all competing with wgrant.
[06:43] <wgrant> I am invincible, nyahaha.
[06:44] <G> I can wait :)
[06:45] <nigelb> I'm landing 3 branches a week or at least trying to.
[06:45] <nigelb> That means about a year until I beat wgrant :D
[06:45] <G> hopefully I'll have 2 this week at least
[06:47] <nigelb> hrm, I don't understand how to use assertRaises.
[06:47]  * nigelb greps more
[06:47] <G> pondering snatching up bug 410331 tbh
[06:47] <_mup_> Bug #410331: PPA: should default to sensible/good name (or give example) <easy> <lp-soyuz> <ppa> <trivial> <ui> <Launchpad itself:Triaged> < https://launchpad.net/bugs/410331 >
[06:52] <G> nigelb: are you having trouble w/ 59301 ?
[06:59] <nigelb> G: I fixed the bug.
[06:59] <nigelb> I'm fixing the test.
[06:59] <nigelb> I always seem to have some or the other trouble with tests :D
[07:00] <G> nigelb: whats the command to run the broken test?
[07:00]  * G wants to have a look too
[07:00] <nigelb> well, you just run the test and you'll see its not passing.
[07:01] <nigelb> wgrant: I don't see a good enough example of assertRaises to understand how it works
[07:03] <wgrant> nigelb: e = self.assertRaises(SomeException, some_function, some, arguments)
[07:03] <wgrant> It will call some_function(some, arguments)
[07:03] <wgrant> And assert that it raises a SomeException.
[07:04] <nigelb> mmm
[07:05] <mrevell> Hello
[07:05] <nigelb> Morning mrevell!
[07:05] <mrevell> Hey nigelb :)
[07:05] <wgrant> nigelb: Contributions page unbroken.
[07:05] <nigelb> wgrant: And I can use the 'e' to get the error message for an assertEqual?
[07:05] <nigelb> wgrant: Ha, thanks!
[07:05] <wgrant> nigelb: Yeah. Use str(e), perhaps.
[07:06] <nigelb> Excellent, thanks.
[07:08] <wgrant> Woah.
[07:08] <nigelb> Heh
[07:08] <nigelb> THIS NEEDS CELEBRATION.
[07:09] <nigelb> At least we need to give him a certificate - "I purposefully quit IRC for X days"
[07:09] <nigelb> spm: ^
[07:23] <jtv> wgrant: well, I found out why domination on dogfood's debian was so fast.  All those active SPPHs for Debian are PENDING, not PUBLISHED.  Shouldn't gina set them to PUBLISHED?
[07:25] <spm> nigelb: haha
[07:25] <spm> nigelb: tbh, I call lies.
[07:25] <nigelb> hehe
[07:26] <rvba> Morning!
[07:26] <nigelb> Morning rvba!
[07:27] <rvba> G'day nigelb.
[07:27] <nigelb> wgrant: Could you review https://code.launchpad.net/~nigelbabu/launchpad/specification-validation-59301/+merge/73631
[07:27] <nigelb> I'm sure you'll find at least 3 things to be picky about ^-^
[07:29] <wgrant> jtv: Probably, yes. Not sure why it doesn't.
[07:30] <wgrant> jtv: Probably a holdover from when it imported Ubuntu.
[07:30] <jtv> It did?  Owww
[07:30] <wgrant> jtv: They needed to be Pending so the publisher would publish them.
[07:30] <wgrant> That's how Ubuntu ended up in LP, yeah :)
[07:30] <rvba> Unless you're playing with it ATM, could one of you guys update DF?
[07:30] <wgrant> Feb 2006.
[07:31] <stub> jtv: I lost your review stamp when I resubitted a MP (added a new pipe so needed to change the dependency branch). Can you click some buttons, and optionally review the nonblocking_readline()? https://code.launchpad.net/~stub/launchpad/branch-rewrite/+merge/73563
[07:31] <jtv> rvba: I already Q/A'ed your ISD fix.
[07:32] <stub> jtv: nm. just hit refresh :)
[07:32] <rvba> jtv: I need it for another QA ;)
[07:32] <wgrant> nigelb: Could you convince someone else to review? I'm trying to disappear now.
[07:32] <rvba> jtv: but thanks for Q/Aing my fix for ISD.
[07:32] <nigelb> wgrant: Ok
[07:32] <jtv> stub: but I already approved that twice!
[07:33] <nigelb> jtv: Can I grab your attention for a quick review? :)
[07:33] <stub> yup. Needed to reload the page.
[07:33] <jtv> stub: ah, you found it.  Sorry, fell a little bit behind with 4 simultaneous conversations.  :)
[07:33] <jtv> nigelb: welcome, conversation #5!  :-)
[07:33] <nigelb> heh
[07:33] <jtv> Let me just catch up with the other 4.
[07:33] <nigelb> I'll grab lunch meanwhile.  Here's the MP - https://code.launchpad.net/~nigelbabu/launchpad/specification-validation-59301/+merge/73631
[07:34] <jtv> rvba: I tested your ISD fix because I thought we might be able to roll out some critical fixes.  But unfortunately it looks like abentley is the remaining blocker.
[07:34] <jtv> rvba: how fresh does dogfood need to be?  Because I updated a few hours back.  I can easily do it again if you need it though.
[07:35] <rvba> jtv: I tested it and I'm sure my changes are not on DF now.
[07:35] <jtv> wgrant: sorry, got a bit distracted with various other conversations (including phone call from a friend who it turns out is turning back from the brink of death :-)
[07:35] <rvba> It says my revisions where included in db-devel 2 hours ago.
[07:35] <jtv> rvba: ok I'll do that again
[07:35] <rvba> jtv: thank you.
[07:36] <jtv> rvba: it'll take a bit longer than usual (though not as long as when that db patch wouldn't apply!) because I'm currently testing a script.  Hence the question.  But right after that I'll update & notify.
[07:37] <jtv> wgrant: I'm currently running domination on debian, with the PENDING statuses updated to PUBLISHED.  Obviously it's going to be fairly intense.
[07:37] <rvba> jtv: There is no rush for me to QA that /right now/ please take your time. Please just ping me when it's updated.
[07:37] <jtv> rvba: btw is it appserver?  script?  both?
[07:37] <rvba> jtv: it's an api change so appserver.
[07:38] <jtv> OK
[07:38] <wgrant> rvba: Can't you do that on qastaging?
[07:38] <wgrant> jtv: Probably.
[07:39] <rvba> wgrant: Oh ... I'm so used to qa things on DF that I did not even think about that ... ;)
[07:39] <jtv> \o/
[07:39] <wgrant> jtv: How does gina currently handle the multiple versions case? I forget if it imports both or just the latest.
[07:39] <wgrant> jtv: If it imports all present versions, it's going to be amusing.
[07:39] <jtv> wgrant: urrr… I think it just imports everything it finds.
[07:39] <wgrant> jtv: Because each run will create all the old versions, then supersede them all.
[07:39] <jtv> I was shocked the other day to find out I may accidentally have made myself the go-to gina coder.
[07:40] <rvba> nigelb: If you're able to wait for 1 hour an official reviewer will be on duty.
[07:40] <wgrant> Then the next run will create them all and supersede them again.
[07:40] <jtv> wgrant: no I do think it skips everything that's already in the db; but if it finds multiple _previously unknown_ versions it imports them all.
[07:40] <wgrant> jtv: I know bits of it from 2 years ago, but try to suppress those memories. It's not very relevant to normal LP development, so it gets paged out quickly.
[07:40] <nigelb> rvba: Ha, just when I get back from lunch. Excellent.
[07:40] <wgrant> jtv: Ew, that would be even worse.
[07:41] <jtv> wgrant: it's like a clever woman I know who, when switching from microbiology to IT and clearly short of relevant experience, still left APL off her CV.  “Not getting the job scares me less than being made to work in APL.”
[07:41] <jtv> I told her to chance it.
[07:42] <wgrant> Heh.
[07:42] <wgrant> But the behaviour you suggest would be very inconsistent, confusing and wrong.
[07:42] <jtv> In what way?  I have a feeling we may be talking at cross purposes.
[07:42] <rvba> (QA ok on qastaging \o/)
[07:42] <jtv> OK
[07:43] <wgrant> jtv: The dominator and gina will fight, creating and superseding over and over again.
[07:43] <jtv> Why?
[07:43] <jtv> Ah, I see the misunderstanding.
[07:43] <jtv> wgrant: cancel
[07:43] <wgrant> If gina imports all versions referenced in Sources, it will import stuff that the dominator has marked as superseded.
[07:43] <wgrant> Heh.
[07:44] <jtv> wgrant: When I said “them all,” I was referring to the last bunch that I mentioned—when gina finds multiple previous unknown versions, AIUI it imports _all the unknown versions_.
[07:44] <wgrant> jtv: What's "unknown"?
[07:44] <jtv> Because it skips everything that's already in the db.
[07:45] <wgrant> I'm not clear on the definition of "in the DB."
[07:45] <jtv> I think the criterion was “not having an SPPH.”
[07:45] <wgrant> An active SPPH?
[07:45] <jtv> No, I don't think so.
[07:45] <jtv> But guessing.
[07:46] <wgrant>         # Create the Publishing entry, with status PENDING so that we
[07:46] <wgrant>         # can republish this later into a Soyuz archive.
[07:46] <wgrant> Heh.
[07:46] <wgrant> As I suspected.
[07:46] <wgrant> 5 years ago, maybe...
[07:47] <jtv> So… proper behaviour now is to make that PUBLISHED?
[07:48] <jtv> And maybe assert that the distro is debian for good measure, just so we find this spot back if we should ever change that?
[07:48] <wgrant> I guess.
[07:49] <jtv> And we'll also need a transitional measure of course.
[07:49] <wgrant> It looks like it will create new SPPHs for anything that doesn't already have a matching active SPPH.
[07:49] <wgrant> Probably including multiple versions of the same source.
[07:49] <jtv> That would be bad.
[07:49] <wgrant> But that needs testing.
[07:49] <wgrant> It would be probably correct, but makes use of the dominator difficult.
[07:49] <jtv> Well, it would be bad for my current purpoes.
[07:51] <jtv> The only scenario that springs to mind where it helps to check only for active records is when an old version of a package is made the most recent again.  But I'm not sure that should ever happen.
[07:52] <wgrant> Well, to check also for inactive records violates everything.
[07:53] <jtv> "Everything" is a little broad… could you narrow it down a little?  Let's take it as read that Albanian traffic law is not, as a practical matter, violated whatever gina does here.
[07:54] <wgrant> If it's published in the source archive, to say it can't be published in the target archive just because it has been published at some point in the past is a fairly hideous position to take. It causes gina to lose generality, and prevents us from partitioning SPPH.
[07:55] <wgrant> And causes gina to give incorrect results for Debian imports.
[07:56] <jtv> So you are saying that we must support Debian re-publishing older package versions.
[07:57] <wgrant> I'm saying that to not support that would be limiting gina's usefulness, placing new requirements on our data model... and it's not a large amount of effort to support that.
[07:57] <jtv> So you are saying that we must support Debian re-publishing older package versions.
[07:58] <wgrant> As a result of the other constraints.
[07:59] <G> wgrant: btw, thanks for my new-found fame on the launchpad wiki :)
[08:00] <wgrant> G: You should have been there a while ago, but the upgrade to moin 1.9 broke stuff :(
[08:00] <G> wgrant: it only passed QA this afternoon, which I don't class as a while ago :)
[08:01] <wgrant> It landed like 12 hours ago.
[08:02] <G> guess our definitions of a while ago differ ;)
[08:02] <jtv> wgrant: So apart from the known, pre-existing problem that only one release of a package is considered published per series/archive/component, it sounds like there would be just 1 problem: “if Debian withdraws publication of the most recent version of a package, the previous version is not automatically reactivated.”  How serious would that bug be?
[08:03] <wgrant> jtv: It also prevents us from reasonably archiving SPPH, which is problematic because it's a huge, slow table that we want to be able to archive.
[08:04] <wgrant> I continue to maintain that using the dominator here is an unnecessary complication and an incorrect solution.
[08:04] <wgrant> But I must concoct dinner.
[08:04] <wgrant> So I shall return later.
[08:05] <jtv> How does it prevent us from reasonably archiving SPPH?
[08:07] <adeuring> good morning
[08:07] <G> adeuring: good $localtime
[08:38] <wgrant> jtv: We would have to adjust the query to look for both, which is slow and silly.
[09:05] <nigelb> allenap: Hi, can I add https://code.launchpad.net/~nigelbabu/launchpad/specification-validation-59301/+merge/73631 to your queue? :)
[09:06] <nigelb> wgrant: Oh no. He's back.
[09:06] <wgrant> nigelb: I know :(
[09:16] <LPCIBot> Project devel build #1,021: STILL FAILING in 3 hr 57 min: https://lpci.wedontsleep.org/job/devel/1021/
[09:21] <lifeless> StevenK: your hudson cert has expired
[09:21] <nigelb> ...
[09:22] <nigelb> lifeless: We've been plotting ways to keep you away.
[09:22] <lifeless> why?
[09:22] <nigelb> Because you're "supposed" to be away :)
[09:22] <nigelb> Also, how's the baby?
[09:24] <lifeless> still in ICU
[09:25] <lifeless> I'm about to crash and sleep and then head back in in the morning
[09:25] <lifeless> so - gnight.
[09:25] <rvba> nn lifeless.
[09:25] <nigelb> g'nite lifeless
[09:31] <allenap> nigelb: Sure, I'll look at that.
[09:34] <nigelb> allenap: Thanks :)
[09:38] <nigelb> danilos: thanks for the review, updating.
[10:21] <StevenK> wgrant: No.
[10:24] <wgrant> StevenK: Well, they're certs, and they're free.
[10:25] <wgrant> Not sure if IE likes them, but meh.
[10:34] <nigelb> danilos: Hi, could you help me quickly with find_tags_by_class?
[10:34] <nigelb> I'm just wondering how I would grab the right one :)
[10:49] <G> danilos: can I put https://code.launchpad.net/~dev-nigelj/launchpad/bug61428/+merge/73632 in your queue (from our discussion last night)
[10:49] <nigelb> *faceplam*, the nigel in that gets me pinged.
[10:49] <nigelb> :D
[10:50] <G> nigelb: don't worry whenever I see your name in chat, I think it might be poorly addressed to me :)
[10:50] <G> nigelb: I think I'll just /nick Nigel ;)
[10:50] <nigelb> G: haha, how confusing :D
[10:51] <nigelb> if you /nick nigel, everyone who has to talk to both us are going to be surprised :D
[10:51] <nigelb> hmmm, you'll get pinged for my test failures.
[10:52] <G> oh... Happy Mailman Day :)
[10:52] <nigelb> heh, I tweeted that when I woke up.
[10:53] <danilos> nigelb, I can try, but generally, it should be simple to iterate over it and list all the returned items (the fact that you'll get only one should not hold you back from iterating over it)
[10:53] <nigelb> danilos: nevermind, trial and error saved the day.
[10:53] <danilos> G: in general, it is good to get more eyes to look at something, so I suggest you go over it with the current OCR, allenap :)
[10:54] <G> danilos: okay
[10:54] <nigelb> I printed the entire list and grabbed the index of the one I want. I hope that isn't evil.
[10:54] <G> allenap: could you please, when you have time, take a look at https://code.launchpad.net/~dev-nigelj/launchpad/bug61428/+merge/73632 for me?
[10:54] <danilos> nigelb, cool
[10:56] <G> nigelb: sorry for the ping ;)
[10:56] <nigelb> heh
[10:56] <nigelb> heh
[10:58] <nigelb> I wish bzr didn't throw a huge traceback when I Ctrl+C'd it.
[10:58] <G> nigelb: if you use irssi, you could maybe modify https://github.com/nigeljonez/misc-scripts/tree/master/irssi to prevent the highlights on 'nigelj' just a thought
[10:59] <nigelb> G: I could, but I don't want to :)
[11:00] <G> okay, so I was looking at https://bugs.launchpad.net/launchpad/+bug/810551 how does at the bottom: "Announced: <date>" sound?
[11:00] <_mup_> Bug #810551: +announcement/nnn lacks date information <anouncements> <trivial> <ui> <Launchpad itself:Triaged> < https://launchpad.net/bugs/810551 >
[11:01] <G> (like the Updated on: <date> that shows on some announcements)
[11:03] <G> that or: "Written for <project/distro/etc> by <person/team> on <date>. everywhere, and none of this: '<date>: Announcement title' thing on the +announcements list page
[11:07] <G> So like: http://dev.nigelj.com/announcementchange.png instead of say https://launchpad.net/mailman/+announcements
[11:20] <nigelb> Gah @ doctests.
[11:21] <jelmer> Gah indeed.
[11:21] <nigelb> woah, it passsed. *tries again to confirm*
[11:23] <nigelb> \o/
[11:25] <stub> allenap: https://code.launchpad.net/~stub/launchpad/pgbouncer-fixture-noca/+merge/73525 if you are free
[11:25] <allenap> stub: Sure.
[11:34] <rvba> allenap: wgrant I'm working on 837975 and perhaps one of you guys can give me a hand with something:
[11:34] <wgrant> Bug #837975
[11:34] <_mup_> Bug #837975: Distro +add and +edit page don't have settings for virtuality and restricted architectures <derivation> <Launchpad itself:Triaged> < https://launchpad.net/bugs/837975 >
[11:34] <wgrant> :(
[11:35] <allenap> rvba: I'm about to stop for lunch, but I'd be happy to help after that.
[11:35] <rvba> If I change the field "Enabled restricted families" (I unselect the only available option "ARM Processors" I get "CannotRestrictArchitectures: Main archives can not be restricted to certain architectures"
[11:35] <wgrant> rvba: Right, PRIMARY archives don't have architecture restrictions.
[11:35] <wgrant> rvba: I have argued that this is a mistake.
[11:35] <wgrant> But Julian has argued otherwise.
[11:35] <wgrant> It seems he may have relented.
[11:36] <rvba> wgrant: He told me you guys did not agree about something related to this bug ...
[11:36] <wgrant> rvba: That was a different thing.
[11:37] <wgrant> I argued there that we can just link to main_archive:+admin
[11:37] <rvba> wgrant: ok, so what do you think about my problem? I don't see the point of adding this field if you cannot do anything with it.
[11:38] <rvba> Am I missing something?
[11:38] <wgrant> rvba: I suspect that Julian has forgotten it doesn't do anything.
[11:38] <wgrant> It probably should do something, but that's more work.
[11:38] <rvba> wgrant: I guess I'll put that on hold until he returns then.
[11:39] <rvba> Unless you have a better idea?
[11:39] <wgrant> rvba: Not sure. Fixing this is probably mostly a matter of deleting code and adding a few rows to the production DB.
[11:39] <wgrant> It's unfortunate that he's off for a while :(
[11:40] <rvba> Ok, I'll work on others bugs for now ... thanks for your help wgrant.
[11:46] <nigelb> danilos: do you want to finish the review of https://code.launchpad.net/~nigelbabu/launchpad/patch-edit-684548/+merge/73514 or do you want me to ask OCR?
[11:47] <danilos> nigelb, looks good, thanks for all the improvements
[11:48] <danilos> nigelb, can you please set the commit message and I'll get it landed for you
[11:49] <nigelb> danilos: done! Thanks :)
[12:10] <StevenK> allenap: O hai -- how do you feel about reviewing a 1,300 line branch?
[12:30] <allenap> StevenK: Of deletions, fine. Of Soyuz or Translations, not so fine.
[12:32] <nigelb> haha
[12:32] <allenap> G: Was the "okay, so I was looking at https://bugs.launchpad.net/launchpad/+bug/810551 ..." bit meant for me?
[12:32] <_mup_> Bug #810551: +announcement/nnn lacks date information <anouncements> <trivial> <ui> <Launchpad itself:Triaged> < https://launchpad.net/bugs/810551 >
[12:34] <G> allenap: meant in general
[12:35] <StevenK> allenap: Based on the URL, you can guess. :-P https://code.launchpad.net/~stevenk/launchpad/no-more-staticdiff/+merge/72830
[12:35] <G> allenap: basically to anyone that had any thoughts on it
[12:36] <allenap> StevenK: I am happy to review that :)
[12:37] <allenap> StevenK: I assume a follow-on branch will DROP TABLE staticdiff?
[12:38] <StevenK> allenap: wgrant and I were not sure if slony would die horribly over that, but yes, that or something like it is the plan.
[12:39] <wgrant> StevenK: We can't drop tables at the moment. We drop FKs, move the table to the todrop schema, and then stub probably does something to it.
[12:39] <StevenK> So "something like it" it is. Fine.
[12:39] <stub> Nah, upgrade.py does stuff to it
[12:40] <wgrant> stub: Oh, I'd never seen that bit,.
[12:40] <stub> Its all automated. Just change their schema rather than DROP TABLE.
[12:40] <wgrant> But so it does.
[12:42] <allenap> StevenK: The chunk at line 730, does that make sense any more? Ah, I assume requestUpgrade() creates a job?
[12:42] <StevenK> So we land a patch to switch the tables namespace to 'todrop'?
[12:43] <StevenK> allenap: Yes, requestUpgrade creates a job.
[12:43] <StevenK> allenap: I've gone through and I'm reasonably sure that all of the tests pass and pyflakes/lint is happy
[12:43] <nigelb> allenap: ouch, ran into some disturbing trouble.
[12:43] <nigelb> allenap: its not actually escaping <script> blocks
[12:44] <allenap> nigelb: It will if you use structured() in the way I suggest, because I tried it here :)
[12:44] <StevenK> allenap: The diff at line 89{4,5} makes me sigh, once you hit it.
[12:44] <wgrant> nigelb, allenap: What are we up to?
[12:45] <nigelb> wgrant: https://code.launchpad.net/~nigelbabu/launchpad/specification-validation-59301/+merge/73631
[12:45] <nigelb> allenap: oh? I did make that change first. Let me re-check
[12:45] <allenap> StevenK: Is there no worth in modifying TestRevisionMailJob instead of ditching it? Erm, I guess I'm asking if removing it reducing our test coverage?
[12:45] <StevenK> RMJ still does stuff, sadly
[12:46] <wgrant> nigelb: Oh, you're not actually using a script tag, just using it to test injection. I am no longer terrified.
[12:46] <wgrant> allenap: Thanks for picking that up.
[12:46] <allenap> Hehe :)
[12:47] <nigelb> wgrant: haha.
[12:51] <nigelb> GAH.
[12:51] <nigelb> Forgot to save the file.
[12:54] <G> allenap: btw, if you do have views on my question, I'd be happy to hear them
[12:56] <allenap> G: Yeah, I'm finishing something off then I'm going to take a look :)
[12:56] <G> allenap: okay, thanks :)
[12:57] <StevenK> He's finishing of a review of mine that is making his eyes bleed
[12:58] <G> StevenK: does the code make his machine poke sticks of RAM in his eyes?
[12:58] <nigelb> I the "making eyes bleed" is understood with StevenK's branches.
[12:58] <nigelb> *I think
[13:07] <allenap> G: That looks good. I am not a Launchpad UI god, but a change like that is undoubtedly an improvement regardless of whether or not it completely satisfies the UI gods.
[13:08] <G> allenap: so basically that screenshot is okay?
[13:09] <allenap> G: Yeah, I like it.
[13:09] <G> (I was thinking of removing the trailing . though, because it looks odd)
[13:10] <allenap> G: Ah yes. Bug pages don't have the trailing full-stop so that's a good call.
[13:12] <StevenK> allenap: Thank you for the review.
[13:12] <G> now to work out what tests I have to run
[13:12] <G> don't exactly feel like waiting until 5am or later for tests to finish :)
[13:12] <StevenK> G: Well, what did you change?
[13:13] <allenap> StevenK: Welcome :)
[13:13] <G> StevenK: just lib/lp/registry/templates/announcement-macros.pt
[13:14] <nigelb> allenap: I fixed everything! Could you re-review https://code.launchpad.net/~nigelbabu/launchpad/specification-validation-59301/+merge/73631 ?
[13:15] <allenap> nigelb: Sure :)
[13:16] <StevenK> G: Right, so that seems a little far reaching. bin/test -vvt registry might be enough of a subset. Or you could try bin/test -vvt announcement and see what you hit
[13:16] <G> StevenK: thanks
[13:18] <StevenK> G: Or you can grep around and see which views use those macros and then see which tests use those views
[13:20] <G> StevenK: is -t <foo> basically  tests matching views with <foo> in their path?
[13:21] <StevenK> G: Right
[13:28] <G> well lib/lp/registry/browser/tests/announcement-views.txt in particular passed fine
[13:43] <benji> G: once you start running the same set of tests over and over again (doing TDD for example), you can use -m to tell the test runner to only look at a particular module; that will take 5-10 seconds off of the testrunner startup time
[13:44] <G> benji: ahhh okay
[13:44] <benji> (and -t and -m work well together too)
[13:45] <G> so I could do something like ./bin/test -m registry -t announcement to only do announcement tests related to registry?
[13:49] <StevenK> G: You can, yes
[13:50] <nigelb> Of course you'll break some doctest you have clu about sometimes.
[13:50] <G> thanks for that, learn something everyday :)
[13:51] <deryck> Hello, all.
[13:51] <deryck> adeuring, abentley, henninge -- sorry about missing the standup.
[13:51] <deryck> I have a patched tire now. :)
[13:51] <abentley> deryck: sorry to hear about the tire.
[13:52] <deryck> abentley, thanks.  not too big a deal.  I could pull of to a station easily.  Just slow waiting.
[13:55] <allenap> Oops, nigelb, I got distracted. r=me. Want me to land it for you?
[13:56] <nigelb> allenap: Yes, please :)
[13:56] <nigelb> ... \o/
[13:56] <nigelb> I just moved up to 4th spot :)
[13:57] <benji> G: to be exact, you could do "bin/test -m lp.registry -t announcement" (-m takes a module path)
[13:57] <nigelb> Enough Launchpad for 2 full working days :)
[13:57] <nigelb> Onto other projects
[13:57] <G> benji: ahhh right
[14:06] <Ursinha> nigelb: thanks :)
[14:07] <nigelb> Ursinha: \o/ :)
[14:07] <Ursinha> nigelb: :D
[14:09] <G> woohoo, all registry & announcement tests passed
[14:15] <G> allenap: for this template branch, would you be able to review it?
[14:15] <allenap> G: Yeah, sure.
[14:16] <G> allenap: any objection if I don't put a lot of content in the cover letter thing seems a bit hard to fit stuff in each section without repeating too much
[14:17] <allenap> G: As long as I understand it, it's fine. Don't treat it like a tax form :)
[14:18] <nigelb> That was...WOW. The best I've heard.
[14:18] <G> allenap: oh gosh, tax forms... I had to fill out an Australian one once, and nearly blew my top
[14:20] <mrevell>  /me wonders if we should rename the channel #nigel-dev
[14:21] <nigelb> haha
[14:21] <G> mrevell: haha
[14:21] <mrevell> :)
[14:21] <nigelb> I think G and I have led talking in this channel for the past few days
[14:21] <G> that reminds me, I should invite nigelb into the super secret social club ;)
[14:21] <G> (joking of course)
[14:21] <nigelb> And the fun bit? If G /nicks to nigel, this will be a confusing channel all around.
[14:25] <rvba> wgrant: About #827608, would you by any change have an idea where I could store announce_from_person=announce_from_person (passed to do_copy) to be able to use it in person._latestSeriesQuery (to populate ~xx/+related-software).
[14:25] <_mup_> Bug #827608: Sync requester isn't credited with upload <derivation> <Launchpad itself:Triaged> < https://launchpad.net/bugs/827608 >
[14:26] <G> allenap: MP: https://code.launchpad.net/~dev-nigelj/launchpad/bug-810551/+merge/73673
[14:26] <rvba> wgrant: I know it's late for you, feel free to ignore my question ;)
[14:26] <G> allenap: there is also https://code.launchpad.net/~dev-nigelj/launchpad/bug61428/+merge/73632 if you've got time
[14:27] <allenap> G: Yeah, I've got time for both.
[14:27] <G> except, if youcan wait a second, I'd like to make one more change for the 61428 one
[14:27] <nigelb> mrevell: HAHAHA. Excellent tweet.
[14:28] <G> nigelb: got a link?
[14:28] <nigelb> https://twitter.com/#!/launchpad_net/status/109270580846534656
[14:28] <G> mrevell: haha thanks
[14:28]  * G follows @launchpad_net now :)
[14:29] <mrevell> I kinda feel we're going for a buy two, get one free deal.
[14:29] <mrevell> :)
[14:29] <nigelb> haha
[14:30] <G> nigelb: we could start a TV show... "The Two Nigels" :)
[14:32] <nigelb> We could start a podcast :P
[14:34] <jelmer> we used to have a Launchpad podcast, back in the day...
[14:35] <nigelb> jelmer: revive it!
[14:35] <jelmer> mrevell: this was one of the items I missed in your "my plans for Launchpad" email.
[14:36] <wgrant> rvba: There is nowhere.
[14:36] <rvba> wgrant: ok then maybe I need another plan.
[14:36] <wgrant> rvba: There's no way to track copies.
[14:36] <wgrant> rvba: I raised this a year ago, but it's been conveniently ignored for 12 months :)
[14:37] <mrevell> jelmer, I'm recruiting a new "communications" person. If they have time for a such thing, it might be fun to do it again. I'd want to think about what sort of benefit it was bringing Launchpad, though. Ad-hoc episodes now and then seems fine to me, though.
[14:37] <rvba> wgrant: and now it's on me ;)
[14:37] <jelmer> mrevell: I was mostly kidding, though I did enjoy the podcast.
[14:37] <wgrant> rvba: Fun. I think we need to discuss with bigjools next week. It requires a significant amount of thought.
[14:37] <wgrant> Launchpod was good.
[14:37] <mrevell> heh, I'm glad you liked it. Maybe we should do it again.
[14:38] <nigelb> One every meetup would be nice
[14:38] <nigelb> So, one from UDS, one from Thunderdome, etc
[14:38] <G> allenap: I just pushed the change to the bug61428 branch, just adds an extra test
[14:38] <rvba> wgrant: k
[14:39]  * rvba marks another card 'blocked'. Sigh.
[14:39] <mrevell> nigelb, That's definitely do-able.
[14:40] <nigelb> mrevell: \o/
[14:45] <jelmer> mrevell: that reminds, I wonder if it would be possible to do a blog post about the improvements for the code imports I've done recently
[14:45] <jelmer> mrevell: if that's ok, should I just send you an email with a suggested text?
[14:46] <mrevell> jelmer, Yes please!
[14:47] <nigelb> I need to blog to the planet on some of the stuff I've done.
[14:47] <nigelb> Some are nice!
[14:47] <nigelb> Let me wait for the last two to land.
[14:48] <mrevell> nigelb, If you have something for the Launchpad blog, let me know.
[14:48] <nigelb> hrm I could just give you stuff that should go on the Launchpad blog instead, yeah.
[14:49] <mrevell> Whichever you prefer.
[14:50] <G> haha just red the person picker post... people would have fun finding me by IRC nick there ;)
[14:50] <G> *read
[14:51] <G> I'd imagine everyone with a 'g' in their name, e-mail address etc would appear
[14:54] <allenap> G: I think three characters must be provided before it'll even attempt a search.
[14:54] <G> allenap: ahh yeah, I forgot about that rule
[15:42] <henninge> what's the reason for this failure again?
[15:42] <henninge> ComponentLookupError: (<InterfaceClass canonical.launchpad.webapp.interfaces.IPlacelessAuthUtility>, '')
[15:42] <henninge> in a test
[15:51] <allenap> henninge: Try a different layer perhaps? The utility is not registered.
[15:51] <henninge> allenap: yes, I am trying that currently. thanks
[15:51] <G> allenap: how did you see leading dates in the titles on +announcements?
[15:52] <G> allenap: the diff removed them
[15:52] <G> (lines 8/9/10)
[15:54] <stub> gary_poster: Do you happen to know where our current Storm is from? Trunk or a branch?
[15:54] <henninge> much better
[15:55] <stub> gary_poster: nm. annotate tells me it was StevenK. I'll sort it tomorrow.
[15:55] <gary_poster> stub, lp:~launchpad-committers/storm/with-without-datetime
[15:55] <gary_poster> ok stub
[15:57] <timrc> I'm creating PPA's on staging but they do not seem to be getting whacked every 24 hours, is there a way to do this? deleting ppas is only a partial deletion (it doesn't seem possible to delete a PPA and then create one with the same name)
[16:02] <allenap> G: Yes, of course, oops. Ignore me.
[16:02] <G> allenap: I like your tests, I should've thought of it, I'll add them, plus an extra test :)
[16:03] <allenap> G: Cool.
[16:09] <G> allenap: actually, the idea for the third test was a bit flawed
[16:10] <gary_poster> G, I am envious of your nick ;-)
[16:11] <G> gary_poster: haha, had it since late `05
[16:11] <cr3> if I have a question about login.launchpad.net, should I ask here or to the isd folks?
[16:11] <gary_poster> :-) cool
[16:11] <gary_poster> cr3, isd
[16:12] <cr3> gary_poster: cheers!
[16:12] <gary_poster> welcome :-)
[16:17] <G> oh whoops didn't mean to do that
[16:18] <G> I thought for some reason, selecting resubmit in the dropdown, would somehow mark the review as a resubmit request
[16:18] <G> allenap: I updated the merge proposal, but may have stuffed it up at the same time
[16:19] <allenap> G: Okay :) I'll take a look.
[16:21] <allenap> G: Btw, the Resubmit status is meant for use by the reviewer. It's confusing, and others have had made this mistake too (including me). I assume this is what you meant by stuffing it up?
[16:21] <henninge> adeuring: The script does not work.
[16:22] <adeuring> henninge: what's the problem?
[16:22] <henninge> "find" is not a method of ResultSet, it is a method of store.
[16:22] <henninge> Store, even
[16:22] <allenap> G: Want me to land it for you?
[16:22] <G> allenap: yeah, thats what I meant
[16:22] <G> allenap: yes please
[16:23] <cjwatson> allenap: is it meant to be "Needs resubmission" or something?
[16:23] <adeuring> henninge: ResultSet has such a method too. But it accepts only WHERE clauses
[16:23] <G> cjwatson: I think that'd be a better status
[16:23] <cjwatson> ah yes, https://help.launchpad.net/Code/Review
[16:23] <allenap> cjwatson: Yeah, that would be a better name.
[16:23] <henninge> adeuring: So I'll either need to extend getByStatus or put the Storm query in the script code.
[16:23] <cjwatson> "Needs reworking" based on that help page
[16:23] <henninge> adeuring: ForbiddenAttribute: ('find', <storm.store.ResultSet object at 0x104d09ec>)
[16:24] <henninge> adeuring: I'll have a look at the api doc to see what you might be thinking of.
[16:24] <adeuring> henninge: right... remove_security_proxy might work. but that's odd place to use it...
[16:25] <henninge> adeuring: no, "ForbiddenAttribute" is not about security.
[16:25] <henninge> adeuring: you'd see "AuthorizationError" in that case
[16:25] <G> I will take it as a sign that I need sleep, so have a good day all :)
[16:25] <henninge> ForbiddenAttribute is misleading ...
[16:25] <adeuring> henninge: rihgt. but let me check the class resultset
[16:25] <henninge> adeuring: I am doing that now
[16:26] <henninge> adeuring: hm, you are right, there is one.
[16:26] <henninge> in the api doc
[16:31] <henninge> adeuring: so, removeSecurityProxy helps
[16:31] <henninge> odd
[16:31] <adeuring> indeed
[16:31] <henninge> maybe I am wrong about ForbidenAttribute
[16:34] <henninge> adeuring: I won't be able to add the support for distroseries search.
[16:35] <adeuring> henninge: sure, no problem. Just try to finish what you have :)
[16:35] <henninge> adeuring: that is something that should be included in HWSubmissionSet
[16:35] <adeuring> right
[16:36] <henninge> adeuring: So, it's almost done now but using removeSecurityProxy.
[16:36] <adeuring> henninge: so, let's wait for a review :)
[16:37] <adeuring> anybody around with some ELementTree-fu? this method http://paste.ubuntu.com/679902/ is supposed to move two node from one parent to another. Or to add missing nodes to the new parent. (The exact content of the old parent does not matter). Running a RelaxNG validation later fails, but if I move the data around in the XML text, things work...
[16:38] <adeuring> flacoste: ^^^?
[16:42] <henninge> adeuring: also, the find does not even work. :(
[16:42] <adeuring> henninge: ouch.. which expression do you use?
[16:43] <henninge> adeuring: what do you mean?
[16:43] <adeuring> which parameter do you pass to find()? or... remember to use new_resultset = resultset.find(whatever)
[16:44] <henninge> adeuring: I did not change that code
[16:44] <adeuring> henninge: cyn you paste it?
[16:45] <henninge> adeuring: let me try something first
[16:45] <henninge> adeuring: there was no assignment
[16:46] <henninge> see, now it passes ;-)
[16:46] <adeuring> henninge: ahhh, sorry ofr this
[16:46] <henninge> yeah, I missed that, too
[16:59] <Ursinha> gary_poster: hi :) do you know if there's anything going on with launchpad? I can't view bugs, they're timing out
[16:59] <gary_poster> Ursinha, hi!  happy birthday, yeah?
[17:00] <Ursinha> thanks :)
[17:00] <mrevell> Night all
[17:00] <gary_poster> Ursinha, I'm about to go to a dr. appt.  deryck, benji, bac, are any of you all available to investigate the graphs and such?
[17:00] <Ursinha> thanks gary_poster :)
[17:00] <gary_poster> welcome :-)
[17:01] <benji> gary_poster: sure (once I figure out what you're talking about) ;)
[17:01] <henninge> adeuring, deryck: I have to run now, I am sorry. Got an appointment in an hour.
[17:01] <adeuring> henninge: np, have a nice evening!
[17:01] <henninge> I prepared the MP and made the branch owned to the squad.
[17:01] <Ursinha> benji: I'm facing timeouts since this morning
[17:01] <henninge> https://code.launchpad.net/~launchpad-orange-squad/launchpad/abel-broken-hwdb-reports/+merge/73697
[17:01] <Ursinha> now I just can
[17:01] <Ursinha> can't view bugs anymore
[17:01] <Ursinha> they're timing out
[17:02] <henninge> adeuring: I prepared the MP and made the branch owned to the squad.
[17:02] <adeuring> henninge: ok, thanks
[17:02] <Ursinha> benji: I'm trying to view this bug: https://bugs.launchpad.net/ubuntu/+source/gwibber/+bug/619654
[17:02] <_mup_> Bug #619654: gwibber-service crashed with GError in __getitem__() <amd64> <apport-crash> <maverick> <natty> <ubuntu-une> <Gwibber:New> <gwibber (Ubuntu):Confirmed> < https://launchpad.net/bugs/619654 >
[17:02] <gary_poster> benji, heh.  take a look at https://wiki.canonical.com/IncidentReports/2011-08-19-LP-confused-query-plan-degraded-performance .  It has links to a lot of tools I used to look at performance for a recent incident.
[17:03] <benji> k
[17:03] <gary_poster> benji in particular https://lpstats.canonical.com/graphs/OopsLpnetHourly/ ...
[17:03] <benji> Ursinha: if you log out do you see the same behavior?
[17:03] <Ursinha> haven't tried, a momen t
[17:03] <gary_poster> benji, https://lp-oops.canonical.com/reports/production/2011/08/19/ (adjust for today)
[17:04] <benji> gary_poster: thanks
[17:04] <Ursinha> benji: cool, logged out that works fast
[17:04] <gary_poster> benji, https://lpstats.canonical.com/graphs/OopsLpnetHourly/ looks fine
[17:05] <gary_poster> benji, so does https://lpstats.canonical.com/graphs/DBCpuLoadAppServers/
[17:05] <Ursinha> and logging in back times out again
[17:05] <benji> Ursinha: I wondered if that would be the case, I suspect there is /something/ different about you that it is spending too much time on.
[17:05] <deryck> henninge, ok, adeuring and I will make sure we get it landed.
[17:05] <gary_poster> Ursinha, could you give us an OOPS please?  In a perfect world, it would be old enough to actually be synced to devpad, like at least 20 min or 30 min
[17:05] <henninge> deryck: cool, thanks
[17:06] <henninge> deryck, adeuring: Good luck and see you on Monday.
[17:06] <Ursinha> sure
[17:06] <adeuring> henninge: thanks and a nice weekend
[17:06] <Ursinha> benji: I'm subscribed to tons of teams that are subteams of other teams and yada yada
[17:06]  * henninge hugs Ursinha
[17:06] <Ursinha> gary_poster: benji: OOPS-2070AW69
[17:06] <Ursinha> henninge!
[17:06] <henninge> Ursinha!
[17:06]  * Ursinha hugs henninge back
[17:06] <gary_poster> ack Ursinha, thanks, looking
[17:07] <henninge> Happy birthday! (if FB is not lying)
[17:07] <henninge> Ursinha: ^
[17:07] <henninge> :-D
[17:07] <benji> Ursinha: yeah, I guess that a recent deployment introduced a pessimization that hit you hard; we'll see if we can find it
[17:07] <Ursinha> it's not :) thanks!
[17:07] <Ursinha> thanks benji
[17:07] <Ursinha> should I file a bug about it?
[17:07]  * henninge has to run
[17:07] <benji> yeah, a bug sounds like a good idea
[17:08] <Ursinha> cool, I'll file it now
[17:09] <benji> Ursinha: reference that oops too (although I can't see it yet, hopefully it'll show up shortly)
[17:09] <gary_poster> benji, as I imagine you have discovered, OOPS is not there yet.  https://wiki.canonical.com/Launchpad/FreshLogs tells you how to request logs on devpad, and then you can look at them directly
[17:09] <benji> gary_poster: thanks
[17:10] <gary_poster> I'm requesting the oops summary page now...there it is
[17:10] <Ursinha> benji: do you want the traceback?
[17:10] <Ursinha> I can see it in the oops page
[17:10] <benji> Ursinha: that'd be great
[17:10] <gary_poster> Possible OOPS to investigate
[17:10] <gary_poster> https://lp-oops.canonical.com/oops?oopsid=OOPS-2070DT30
[17:10] <gary_poster> https://lp-oops.canonical.com/oops?oopsid=OOPS-2070DQ6
[17:11] <Ursinha> benji: http://paste.ubuntu.com/679935/
[17:11] <Ursinha> is the problem happening when trying to load the subscription portlet or I'm seeing it wrong?
[17:12] <gary_poster> benji, Ursinha, this is https://bugs.launchpad.net/launchpad/+bug/811447
[17:12] <_mup_> Bug #811447: BugTask:+index timeout - death by sql in PersonSubscriptions(user, bug) <timeout> <Launchpad itself:Triaged by gary> < https://launchpad.net/bugs/811447 >
[17:13] <gary_poster> I have a branch to fix it that I was working on just then :-P
[17:13] <Ursinha> gary_poster: cool! :)
[17:13] <benji> Ursinha: yeah, it's building the subscription portlet when it runs out of time; of course something else might have used up all the time before that, but it's a smoking gun
[17:13] <benji> gary_poster: very nice
[17:15] <benji> gary_poster: cool, I'll put it in my notes
[17:18] <Ursinha> thanks gary_poster and benji
[17:19] <gary_poster> Welcome Ursinha.  It won't be rolled out today as a birthday present, but hopefully early next week :-)  fwiw, the trigger should generally be bugs with many duplicates, to which one or more of your teams are subscribed.
[17:21] <Ursinha> I see.
[17:21] <Ursinha> thanks anyway! :)
[17:26] <gary_poster> jcsackett, I have three branches ready for review if you are willing to do it while I am absent for doctor's visit.  If not, no prob.
[17:26] <gary_poster> https://code.launchpad.net/~gary/launchpad/bug838869/+merge/73695
[17:26] <gary_poster> https://code.launchpad.net/~gary/launchpad/bug838878/+merge/73702
[17:26] <gary_poster> https://code.launchpad.net/~gary/launchpad/bug811447/+merge/73704
[17:26] <gary_poster> (The second two depend on the first.)
[17:26] <gary_poster> # 2 is 913 lines, but #3 is 181.
[17:26] <jcsackett> gary_poster, i saw. i'm finishing lunch now, then i'll start in on them. :-)
[17:26] <gary_poster> thanks jcsackett :-)
[17:43] <flacoste> adeuring: did you solve the issue?
[17:43] <adeuring> flacoste: no :( It's really odd. Let me paste some related stuff...
[17:44] <flacoste> adeuring: it's like if Relax-NG wasn't taken the new nodes into account?
[17:44] <adeuring> flacoste: exactly
[17:44] <adeuring> the odd thing is: If I write the modified tree to a file and read the file back into another tree, things are fine
[17:45] <adeuring> flacoste: http://paste.ubuntu.com/679965/
[17:45] <adeuring> ...like so
[17:45] <adeuring> flacoste: if I remove line 23, things break...
[17:46] <adeuring> flacoste: I am currently simply using regexes for the same job. ugly, but more promising :)
[17:49] <flacoste> adeuring: i don't understand line 23 is identical to line 9
[17:49] <flacoste> and it's not like if submission changed between the two?
[17:50] <adeuring> flacoste: that should have been etrre.parse(f,...)
[17:51] <adeuring> flacoste: let me check again...
[17:55] <flacoste> adeuring: the error is when validating which works with a string anyway, no?
[17:55] <adeuring> flacoste: arghhh sure!
[17:56]  * adeuring is an idiot
[17:56] <adeuring> so, regexes FTW
[17:56] <adeuring> thanks!
[17:57] <adeuring> so that's why writing to a file and reading (properly...) back worked
[17:58] <adeuring> flacoste, deryck: but: seems that the idea to move two nodes around works. Have tested so far just one HWDB report, so I am not _that_ sure, but anyway
[18:01] <deryck> cool
[18:01]  * deryck has fingers crossed
[18:03] <flacoste> adeuring: btw, getchildren is deprecated in 2.7, using list(node) would work in both 2.6/2.7
[18:03] <adeuring> flacoste: ah, thanks.
[18:03] <flacoste> adeuring: and you can node.findall('name') instead of the list comprehension you are currently doing
[18:04] <flacoste> there is also a findtext() not sure if that would be useful to you currently
[18:04] <adeuring> flacoste: ok, but I opted for a regex variant :)
[18:04] <adeuring> since we need the string representation anyway
[18:06] <flacoste> findall() takes a node name or a xpath subset: http://effbot.org/zone/element-xpath.htm
[18:11] <adeuring> flacoste, deryck: with the data modification intended by the broken method, 848 from 858 reports that we unusable before can now be processed :)
[18:11] <flacoste> adeuring: awesome!
[18:11] <deryck> adeuring, very nice!
[18:12] <adeuring> flacoste: what do you think: should I keep a regex based fix. or shouldd we goo back to tweaking the element trre, write it back to a string, and validate this string?
[18:13] <flacoste> adeuring: what does the regexp-based fix look like?
[18:13] <flacoste> it's doing substitution on the original document string?
[18:14] <adeuring> flacoste: http://paste.ubuntu.com/679992/ (7 is unrelated to the current discussion -- that's about ESC characters in lots of other reports)
[18:14] <adeuring> s7/line 7/
[18:15] <adeuring> flacoste: http://paste.ubuntu.com/679995/ (with regexes inlcuded)
[18:17] <flacoste> adeuring: it's not as gross as i thought it would be :-) probably faster also
[18:17] <adeuring> right, I suspect too taht it's faster
[18:17] <flacoste> since this is big kludge anyway... i'd be tempted to let it go
[18:18] <flacoste> i'd suggest making the _udev_node_exists and _dmi_node_exists non-greedy
[18:18] <flacoste> .*? instead of .*
[18:18] <adeuring> flacoste: of course, sure
[18:18] <adeuring> flacoste: they also need a DOTALL flag
[18:18] <flacoste> it will fail faster
[18:19] <flacoste> right, because of the newlines
[18:19] <adeuring> exactly
[18:20] <adeuring> I didn't test this branch against "working" reports yet...
[18:20] <adeuring> wanted to see first if the approach is usabel at all
[20:51] <adeuring> jcsackett: could you please review this mp: https://code.launchpad.net/~adeuring/launchpad/hwdb-test-natty-mess/+merge/73722 ?
[20:52] <jcsackett> adeuring: i'll do it as soon as i finish the one i'm looking at now. :-)
[20:52] <adeuring> jcsackett: ok, thanks!
[21:07] <flacoste> jcsackett, adeuring: already done :-)
[21:08] <jcsackett> flacoste: cool. :-)
[21:08] <adeuring> flacoste: thanks!
[21:10] <adeuring> flacoste: right, the test is really not very comprehensive -- but the later running parts of the processing script are quite paranoid, so I think we can be sure that we will not store nonsensical stuff
[21:10] <flacoste> adeuring: right, you might add a comment to the test saying that the call would raise an exception without the work-around
[21:11] <flacoste> or something like that
[21:11] <flacoste> so that it's obvious that the assert is really not the meaningful test
[21:11] <adeuring> flacoste: right
[21:11] <flacoste> it might even be superfluous
[21:18] <adeuring> night everybody
[21:25] <thumper> morning
[22:31] <jelmer> hey thumper, mwhudson
[22:41] <jelmer> hmm, the scanner doesn't seem to like ~vcs-imports/libreoffice/core
[23:00] <mwhudson_> jelmer: it oopses with 'job ran too long'
[23:01] <jelmer> oh, oops..
[23:01] <mwhudson_> jelmer: in getAncestryData i think, before even it starts inserting stuff into the db
[23:01] <jelmer> maybe 2.4 will help with that
[23:01] <jelmer> jams recent fixes for get_parent_map will definitely  help
[23:34] <wallyworld> thumper: you must be missing launchpad - you've been poking around in here a bit of late :-P
[23:34] <thumper> wallyworld: I've found some things, particularly in email, that I want changed
[23:35] <wallyworld> i can imagine :-)
[23:35] <wallyworld> have you raised bugs?
[23:36] <wallyworld> thumper: what do you want fixed?
[23:36] <thumper> bug email headers
[23:36] <thumper> imapfilter only checks the first occurrance of a particular header it finds, like X-Launchpad-Bug
[23:37] <wgrant> Really? That sounds like a pretty glaring flaw.
[23:37] <wallyworld> thumper: and we stuff multiple X-Launchpad-Bug header items into the message?
[23:39] <thumper> yep, one per task
[23:39] <thumper> wgrant: it is defined in the imap rfc
[23:39] <thumper> for remotely selecting messages
[23:40] <wgrant> Still sounds like a pretty glaring flaw.
[23:40] <wallyworld> thumper: so with the squad reorg, unless a bug is critical or part of feature work, it's unlikely to get done sinze we currently have so many criticals to burn down
[23:41] <wallyworld> unlikely to get done soon
[23:41] <wgrant> And even if it is escalated to critical, it's unlikely to be done within 6 months.
[23:42] <wallyworld> thumper: so in other words, patches welcome :-)
[23:42] <wallyworld> but changing mail headers would have a lot of downstream impact on people's existing mail filters etc, so I guess existing headers would need to be retained?
[23:43] <wallyworld> and new ones added?
[23:44] <wgrant> So we have to have three filtering methods.
[23:44] <wgrant> Yay.
[23:45] <wgrant> One for pathetic Google Mail, one for pathetic built-in IMAP filtering, and one for everything else.
[23:46] <wgrant> Maybe we should have an "I use a retarded mail server" option...
[23:47] <wallyworld> wgrant: why is gmail retarded?
[23:48] <wgrant> It can't filter on custom headers.
[23:48] <wgrant> At all.
[23:48] <wallyworld> wtf? serious?
[23:48] <wgrant> Yes.
[23:48] <wallyworld> why????
[23:48] <wgrant> That's why there's all this crap in the body of all emails.
[23:48] <wgrant> To appease Gmail users.
[23:49] <wallyworld> why would not being able to filter on custom headers appease gmail users?
[23:49] <wallyworld> surely it would do the opposite?
[23:50] <wgrant> Hm? We have to put all the filtering metadata in the body too, or Gmail users complain that they can't filter.
[23:51] <wallyworld> ah i think you misunderstood my question - why doesn't gmail support filtering on custom headers i wonder?
[23:51] <wgrant> NFI
[23:51] <wallyworld> seems like a pretty basic omission