[00:04] <StevenK> wgrant: Even worse block of SQL is strange. Running it by hand returns 3 columns, but only two turn up in list(results)
[00:13] <wgrant> StevenK: What does the Python look like?
[00:15] <StevenK> wgrant: Ugh, the diff is horrible
[00:16] <StevenK> wgrant: I can pastebin the diff, but it's nigh on unreadable. Or I can pastebin the garbo's __call__
[00:17] <wgrant> The latter would be nice
[00:17] <StevenK> wgrant: http://pastebin.ubuntu.com/1424454/
[00:18] <wgrant> StevenK: It looks like that SQL is all one big string
[00:18] <wgrant> Which is not going to work very well
[00:19] <StevenK> wgrant: I need to split it up? Postgres deals ...
[00:19] <wgrant> StevenK: But Storm has no idea that there's going to be an extra column.
[00:19] <wgrant> It needs to be two separate items in the findspec
[00:19] <StevenK> wgrant: So (PackageUpload.id, SQL(...), SQL(...)) ?
[00:19] <wgrant> Yes
[00:20] <StevenK> Hmm, that's a little bit of a trap
[00:20] <StevenK> But it makes sense
[00:26] <StevenK> wgrant: I currently have seperate tests for searchable_names and searchable_versions, shall I pull those together too?
[00:26] <wgrant> Might as well
[00:38]  * wgrant plays buildbot roulette again
[00:38] <StevenK> I wonder if you'll lose as hard today :-)
[00:38] <wgrant> Hopefully not
[00:38] <wgrant> These are more isolated changes
[00:59] <wgrant> 1 failure so far...
[04:06] <StevenK> wgrant: When you're finishing sharpening knives for buildbot, testrespository and/or subunit, https://code.launchpad.net/~stevenk/launchpad/populate-searchables-for-pu/+merge/138906 is ready for a second close-up.
[04:09] <wgrant> Just fixing 18 doctests
[04:09] <wgrant> will be a few minutes
[04:19] <StevenK> wgrant: Where a few is ten or more? :-)
[04:19] <wgrant> shh
[04:36] <StevenK> Or more like 30 ...
[04:41] <wgrant> The first 16 were easy :)
[04:41] <wgrant> The last two are some odd double-escaping...
[04:41] <StevenK> Haha
[04:44] <wgrant> Ah
[04:48] <wgrant> Hmm
[04:48] <wgrant> The URL linkification regexp is the problem here
[04:48] <wgrant> It's run over the HTML, not the text
[04:48] <wgrant> But it pretends it's running over text
[04:51] <StevenK> Hmmmmmm
[04:52] <wgrant> It has historically works because & and ; are unreserved characters
[04:52] <wgrant> But ' is now escaped as &#x27;, and # is not unreserved
[04:52] <StevenK> I think the set() magic used by add{Source,Build,Custom} doesn't happen in the garbo job
[04:52] <wgrant> But we don't care about the specific parsing of the structure of the URL
[04:52] <wgrant> So I'm just going to pretend that # is unreserved too
[04:52] <wgrant> StevenK: What do you mean?
[04:52] <StevenK> wgrant: Consider ' 1780947 | libfile-finder-perl libfile-finder-perl               | {0.53-1}
[04:53] <wgrant> Ah
[04:53] <wgrant> Ah
[04:53] <wgrant> The source and binary have the same name, of course
[04:53] <StevenK> Right
[04:53] <StevenK> And our magic query doesn't do that
[04:55] <wgrant> Might as well throw in a 'SELECT DISTINCT name FROM unnest([... the array expression ...]) ORDER BY name' or similar, I guess
[04:55] <StevenK> In which bit?
[04:56] <StevenK> Oh, around the array_to_string()
[04:56] <StevenK> No, wait
[04:56] <wgrant> Inside the array_to_string
[04:56] <wgrant> unnest() is rather special
[04:57] <wgrant> It converts an array into a set of rows
[04:57] <wgrant> So you can perform operations on it as if it was a table
[04:57] <StevenK> We already have array_to_string( (...) || (...) || (...) || (...))
[04:58] <wgrant> Right, but before you convert it to a string you want to make it unique and ordered.
[04:59] <StevenK> ... I think I have the brackets right
[05:00] <StevenK>             array_to_string(SELECT DISTINCT name FROM unnest(
[05:00] <StevenK>             (SELECT array_agg(
[05:01] <StevenK> ProgrammingError: syntax error at or near "SELECT"
[05:01] <StevenK> LINE 2:             array_to_string(SELECT DISTINCT name FROM unnest...
[05:02] <wgrant> At that point you might as well use string_agg directly, I guess. SELECT string_agg(DISTINCT name ORDER BY name) FROM unnest(...)
[05:05] <wgrant> Oh right
[05:05] <wgrant> Atom is stupid
[05:05] <wgrant> Feed content is *meant* to be double-escaped...
[05:05] <StevenK> reference = u'binarypackage-102763 unique-from-factory-py-line3339-102754'
[05:05] <StevenK> actual    = u'unique-from-factory-py-line3339-102754 binarypackage-102763'
[05:05] <StevenK> Bleh
[05:05] <StevenK> I guess we do not care about the order at all, just as long as they're both there
[05:06] <wgrant> Mmm
[05:06] <wgrant> Ideally it would be ordered
[05:06] <wgrant> And it's easy
[05:06] <StevenK> That's from add{Source,Build,Custom}
[05:06] <wgrant> Ah
[05:07] <wgrant> Still easy
[05:07] <wgrant> sorted()
[05:07] <wgrant> done
[05:07] <wgrant> :)
[05:07] <StevenK> >>> ' '.join(set(['b']) | set(['a']))
[05:07] <StevenK> 'a b'
[05:07] <StevenK> So I think it is sorted already
[05:07] <StevenK> Just the test assumes SPR first
[05:08] <wgrant> Oh, so the reference/actual are swapped?
[05:10] <StevenK> So it would seem, because:
[05:10] <StevenK>          names = '%s %s' % (
[05:10] <StevenK> -            upload.builds[0].build.source_package_release.name, bpr.name)
[05:10] <StevenK> +            bpr.name, upload.builds[0].build.source_package_release.name)
[05:17] <StevenK> Blah
[05:17] <StevenK> ./2012-12-06/OOPS-f4b0427a8f57ec8a0619d40b49fc93c8 exists on neem, but the string 'OOPS' does not
[05:26] <StevenK> Oh, it does, just that less doesn't admit to it, which is handy
[05:27] <StevenK> ProgrammingError: syntax error at or near "' '"
[05:27] <StevenK> LINE 50:         )), ' '),
[05:27]  * StevenK whimpers
[06:01] <StevenK> Hmmmm
[06:02] <StevenK> I *think* r15555 fixed bug 835645
[06:02] <_mup_> Bug #835645: DistroSeries:+queue timeout paginating <lp-soyuz> <queue-page> <timeout> <Launchpad itself:Triaged> < https://launchpad.net/bugs/835645 >
[06:53] <wgrant> StevenK: Unlikely, why?
[06:59] <StevenK> wgrant: Because r15555 changed get_all() to use a DRS
[07:01] <wgrant> Hmm
[13:43] <bac> gmb: did you get your kanban issue fixed?
[14:58] <jcsackett> rick_h_: mp for you, when you have a moment: https://code.launchpad.net/~jcsackett/launchpad/no-releases-for-projectmilestone/+merge/139221
[14:59] <rick_h_> jcsackett: will do
[15:07] <rick_h_> jcsackett: r=me, but man I hate when something doesn't have what the interface says it has and it implements the interface
[15:12] <jcsackett> rick_h_: does it make you feel better that product_release *isn't* on the interface?
[15:12] <jcsackett> :-P
[17:31] <czajkowski> sinzui: thoughts on https://bugs.launchpad.net/launchpad/+bug/1088959
[17:31] <_mup_> Bug #1088959: launchpad.net crashes webkitgtk browsers. <Launchpad itself:New> < https://launchpad.net/bugs/1088959 >
[17:33] <sinzui> I will look into it. I think it is bogus. Midori and Epiphany are very reliable and I have not experienced any issues.
[17:33] <czajkowski> nods
[17:34] <czajkowski> sinzui: how's you today? very quiet
[17:37] <sinzui> I have a bug just about fixed, but I got distracted a 403 when looking at code imports, bug 1089023
[17:37] <_mup_> Bug #1089023: Proprietary branches break code import listing <403> <branches> <regression> <sharing> <Launchpad itself:Triaged> < https://launchpad.net/bugs/1089023 >
[17:37] <czajkowski> ah nice one to fix
[17:43] <sinzui> rick_h_, do you have time to review https://code.launchpad.net/~sinzui/launchpad/allow-vcs-imports-to-rename-a-branch/+merge/139270
[17:44] <rick_h_> sure thing sinzui
[17:52] <rick_h_> sinzui: r=me thanks
[17:54] <sinzui> thank you rick_h_
[18:02] <timrc> hm one of our scripts caused LP to return a 500... no bueno
[18:02] <rick_h_> timrc: burn that script, it's faulty clearly :P
[18:02] <timrc> rick_h_, clearly :)
[18:03] <timrc> rick_h_, how do I look up an oops again? ID OOPS-b69df532ee375ff926e1605527d0b7d6
[18:04] <rick_h_> https://oops.canonical.com/
[18:04] <rick_h_> so https://oops.canonical.com/?oopsid=OOPS-b69df532ee375ff926e1605527d0b7d6
[18:04] <rick_h_> timrc: hmm, that change was reverted.
[18:04] <rick_h_> timrc: so that should be corrected now, but looks like you hit it while it was broken early this morning/overnight
[18:05] <czajkowski> that was fixed hours ago.
[18:05] <rick_h_> timrc: so maybe you can keep the script around a little longer since it was our fault
[18:05] <timrc> czajkowski, OK, the problems were reported to me while on vacation on the 10th
[18:06] <timrc> you guys should wait to introduce bugs when I'm not on vacation, jeez
[18:06] <rick_h_> shoot, but how will I meet my 'break timrc's stuff' quota for the year?
[18:07] <czajkowski> timrc: most odd but tis working now, it stopped last UK time yesterday and was fixed early UK time today
[18:14] <timrc> czajkowski, do you have an LP bug # handy by any chance?
[18:14] <czajkowski> timrc: we didnt log a bug this morning we just reverted the patch
[18:14] <czajkowski> what are you trying to do timrc ?
[18:15]  * czajkowski is also attempting to run out the door as past EOD but just so I can point you in the right direction
[18:15] <timrc> czajkowski, I'm trying to document what went wrong in our bug for the users affected by the regression
[18:16] <czajkowski> timrc: hmm there may be some confusion then, I assumed rick_h_ was referring to your oops which was like this mornings, where people could not create projects
[18:16] <czajkowski> what is your current issue
[18:17] <timrc> czajkowski, no, I think we're talking about the same thing... a user logged a bug saying as much and I wanted to respond with details of the problem... but if there was no bug filed against LP, that's fine
[18:17] <rick_h_> timrc: so it was a deploy that had a bug and was reverted.
[18:17] <rick_h_> there wasn't a bug file and the deploy is getting re-tested and worked on before being attempted again.
[18:17] <timrc> rick_h_, thanks
[22:29] <sinzui> wgrant, StevenK, can you one you look at the pastebin in the last comment: https://bugs.launchpad.net/launchpad/+bug/1077351
[22:29] <_mup_> Bug #1077351: SourcePackageRecipeBuild:+index LocationError build <oops> <recipe> <soyuz-build> <Launchpad itself:In Progress by sinzui> < https://launchpad.net/bugs/1077351 >
[22:58] <StevenK> wgrant: http://pastebin.ubuntu.com/1426332/
[22:59] <StevenK> wgrant, sinzui: http://www.youtube.com/watch?v=BgAlQuqzl8o
[23:09] <wgrant> StevenK: http://pastebin.ubuntu.com/1426362/
[23:09] <wgrant> I didn't fix the indentation, but the code is right
[23:54] <StevenK> wgrant: TypeError: Expected unicode, found <type 'str'>: '{'
[23:55] <StevenK> wgrant: I guess I need to cast back to text[], I'm just not sure where
[23:55] <StevenK> wgrant: SELECT array_agg(..)::text[] FROM (... at a guess?
[23:56] <mwhudson> and i thought i abused postgres as a hobby