[00:14] boo to not being able to restart the buildmaster ever :( [00:16] huh? [00:17] what changed? [00:17] well, there's always a build running [00:17] if i bounced it now, we'd lose the progress so far on the tests currently running [00:17] maybe it's worth it though [00:17] (it would be nice if the bzr builder ran today) [00:19] thumper: opinion? [00:19] hmm [00:20] * thumper wants a stop.patch [00:20] mwhudson: will we need to manually kick the builds to get them going again? [00:20] mwhudson: if we won't, do it now [00:20] thumper: i don't think so [00:21] thumper: let's find out! [00:21] * mwhudson restarts the wrong bit, oops [00:22] What's the bzr builder? [00:23] builds seem to be going now [00:23] wgrant: it runs the launchpad tests against bzr.dev [00:23] mwhudson: Ah. [00:25] mwhudson: we have a new image for bzr? [00:25] thumper: no, it was master only changes [00:25] (a new image wouldn't require a master restart) [00:36] mwhudson: ah crap [00:37] mwhudson: my tests run fine locally [00:37] mwhudson: but there are changes on devel that cause it to fail [00:37] thumper: wheeee! [00:37] mwhudson: but I really want this cherry pickable...# [00:37] mwhudson: if I fix it to land, it won't cherry pick [00:38] thumper: land it on devel first, then prepare a branch off production containing the fix? [00:38] the branch I have is prepared off production [00:38] oh right [00:38] mwhudson: I think I'll push my current branch to a new branch on lp [00:38] then fix to land on devel [00:38] right, two branches is the answer i think [00:47] ah fuck [00:48] * thumper walks away for coffee [00:48] I'll fix the problem later... [00:50] * mwhudson afk for a bit again [00:51] (well, for lunch) === matsubara is now known as matsubara-dinner === matsubara-dinner is now known as matsubara [02:03] * wgrant cringes a bit at IAnnouncement. [02:04] And related interfaces.. [02:04] wgrant: installed on jaunty now [02:05] meaton2veggies_: i386? [02:05] wgrant: do a bzr blame on that one [02:05] thumper: I did see some comments around which indicate what I think you might be suggesting. [02:05] * wgrant tries. [02:05] wgrant: yes, VM [02:05] Indeed. [02:06] wgrant: any idea how i fix self-signed cert with feeds.launchpad.dev:443? [02:06] meaton2veggies_: Tell your browser to trust it. [02:06] yep, doing that now [02:07] in firefox 3.5 it comes up on the front page launchpad.dev error alert [02:07] That's correct. [02:07] You need to tell Firefox to trust it. [02:08] do i have to create a proper cert for the local site [02:08] You could, but there's little point. [02:08] ok [02:09] thumper: How aggressively should method name conventions be fixed? [02:09] wgrant: what are you wanting to fix? [02:09] thumper: IAnnouncement.set_published_date and IHasAnnouncements.announcements [02:10] (to setPublishedDate and getAnnouncements respectively) [02:10] wgrant: are they methods? [02:10] thumper: They are. [02:10] And they both want to be exported. [02:11] announcements is a method not an attribute? [02:11] thumper: Correct. [02:11] wgrant: I'd review and approve those changes [02:11] thumper: It takes some arguments to customise the result. Maybe it used to be an attribute. [02:13] thumper: I shall start another branch and do the renames. Thanks. [02:13] Seems much nicer than using exported_as. [02:14] wgrant: agreed [02:14] And they're not used much, so it's easy. === cody-somerville_ is now known as cody-somerville [02:39] * mwhudson reappears [02:39] wb [02:40] :) [02:56] mars: "- Node.queryAll()'s signature changed. It now returns an empty sentinel object with size() == 0 if no results were found, instead of returning 'null'. " [02:56] hoo-bleeping-ray!! [02:57] that was a particularly stupid change in pr2 [03:05] hehe [03:05] well, it makes the test clearer, and universal [03:05] test for length: done! [03:06] I don't know what Y.all() returns though [03:13] oh right [03:14] in loggerhead i had to change a few Y.query(selector).each(...) to nodes = Y.query(selector); if (nodes) nodes.each(...) [03:14] s/query/all/ i guess [03:14] which just made me a bit unhappy [03:21] The sparklines really encourage me to read the project branch listing very quickly, before the horrible virus slowly infects all of the rows. [03:22] wgrant: thumper is landing a fix, or you could put one up :) [03:22] lifeless: They should be gone in a few hours, I think. [03:23] I'm surprised they missed the edge update last night. [03:23] wgrant: heh [03:24] wgrant: missed the cut off by about 2.5 h [03:24] thumper: thats ok, have a beer for commiseration. Or something. [03:24] * wgrant defeats the stupid uni firewall, and successfully pushes that IAnnouncement branch. [03:25] thumper: Shall I request a review from you? [03:25] wgrant: sure [03:25] Hmmm. [03:26] I attempted to search for 'devel' as the branch to propose the merge to, and the picker gives me one option: "undefined" [03:26] Oh. "Too many matches" is the error, "undefined" is the option. === matsubara is now known as matsubara-afk [03:35] ew [03:35] wgrant: try "lp:launchpad/devel" [03:36] wgrant: once you've done it once, it'll remember [03:36] and offer it next time# [03:37] thumper: Ah, useful. [03:37] thumper: Thanks. I'll file a bug. [03:37] ok [03:40] * thumper takes a break as 4 ec2's do their thing [03:54] So, a friend of mine just totally failed to realise that status and importance are AJAXy, because they have no icons. [03:55] wgrant: I filed a bug on that [03:55] search for mystery meat [03:55] bad web design since 1995 [03:55] I thought I remembered seeing them. [03:55] s/them/one/ [04:09] WTF? [04:09] got an invalid docstring :( [04:09] apidocs are processed from docstings [04:09] api doc compilation thingy [04:09] where can I find out what exactly is invalid? [04:10] thumper: i'd start with the diff against trunk :/ [04:10] woo bzr builder ran to completion today [04:10] its a new docstring for a new method [04:10] mwhudson: 8 failures [04:10] mwhudson: but yes, it ran :) [04:10] oh 3 failures, 10 errors [04:10] thumper: almost all uifactory failures [04:11] i think i have a branch that fixes that actually... [04:11] hm, no [04:11] i have a branch that was created to fix that :) [04:11] heh [04:11] mwhudson: add it to the todo list :) [04:11] https://bugs.edge.launchpad.net/bugs/405052 [04:12] Bug #405052: bzr 1.18dev breaks our uifactory [04:12] filed a while ago [04:12] i'm sure i fixed this, must have lost the changes somehow [04:12] :( [04:13] it's minor [04:13] off to get the car now [04:13] bbs [04:40] * thumper is no $806 poorer [04:40] s/no/now [04:40] d'oh [04:42] thumper: I'd be more worried about the $806 vs the no/now [04:42] spm: well, the car needed brake pads and discs [04:43] what do you need brakes for? don't NZer's just run into the nearest sheep to stop? [04:43] no way, waste of a good sheep [04:44] spm: it is the lack of sheep at the town lights that cause problems [04:44] you'd think they'd do something about that. what do you pay your taxes for eh! [04:44] spm: to pay for the politicians Wellington houses apparently [04:45] ah. some things dno't ever change.... [04:49] spm: how goes disk space on codehosting? [04:49] spm: also, what is the RT for more disk space? [04:58] wgrant, was your friend using Safari? Be sure to note that in the bug report. Safari doesn't show the icons, IIRC. [04:59] * mars goes off to sleep for real now [04:59] mars: I have noticed that sometimes they don't show at all [04:59] mars: on FF2.0 on Windows [05:00] mars: Firefox 3.0. It's intended that they don't show up unless hovered. [05:00] That they don't show up even if hovered over is another issue. [05:00] wgrant: even on hover they weren't showing [05:02] mystery meat sucks [05:05] lazr.restful seems to think that my zope.schema.URI is a hosted file :( [05:10] lifeless: there was an Indian resturant near where I lived once in London that had "meat chilli chilli" as a menu item :) [05:10] I was never game to ask exactly what the meat was [05:10] it was a mystery :) [05:11] :) [05:11] http://en.wikipedia.org/wiki/Mystery_meat_navigation is what I meant though [05:18] lifeless: found the community review tag bug [05:19] cool [05:20] 1 line fix, 50 lines of tests to write [06:17] concurrency is hard :( [06:40] !! [06:40] emacs just died on me? [06:40] ah no [06:40] just hiding [06:41] I'm done for now [06:41] will finish off these tests later [06:41] night [06:42] night [06:42] * mwhudson very nearly finished too [06:42] thumper: what was the problem for community reviews? [06:43] branch.reviewer vs. branch.code_reviewer [06:43] reviewer can be None [06:43] code_reviewer == branch.reviewer or branch.owner [06:43] I'm changing the view logic and renaming a method and adding tests :) [06:44] argh [07:23] Morning ppl [08:58] morning [08:59] morning jml [08:59] hey jml, noodles775; and g'night [09:00] :) [09:01] spm, g'night [09:09] Morning! [10:01] mrevell: Morning :) [10:01] hey allenap [10:03] Does anyone know why PQM is throwing out my branches? It seems to be claiming we're in testfix mode for lp/devel, but the waterfall is green a long way back. [10:08] allenap: we get into testfix even if db-devel is red [10:09] allenap: it's "stop the line" approach [10:09] danilos: Ah ha, thank you. [10:10] allenap: we should get out of testfix mode soon, though. it depends on when the script that turn it off sees the testfix commit. [10:11] BjornT: Fantastic :) [10:30] hey [10:31] liwii sent an email to the launchpad-dev mailing list [10:34] * wgrant hasn't seen it come through. [10:55] and it doesn't seem to be in the archive, nor in the moderation queue [10:55] wgrant, sorry I got distracted for a while [10:57] jml: Not even in the moderation queue? Huh. [10:59] jtv: ping [10:59] mrevell: pong [11:00] mrevell: saw your message about the getting-started guide, though haven't looked at the guide yet [11:00] mrevell: just what we need though, so definitely let's set a time. [11:00] (that is what this is about, isn't it?) [11:00] jtv: No worries. This is on a slightly different subject. Should the help wiki present bzr imports and exports as the default method? [11:01] but I don't know how the moderation queue works. [11:01] do we have an admin around who can take a look? [11:02] Hi, all. [11:02] mrevell: for exports I'd present both options side by side depending on use case (and it should be clear that it's for maintainers, reviewers etc., not for end-users) [11:03] mrevell: for imports, I would like bzr to be seen as the default source. Also makes it less inviting to set up proprietary projects for translation without contacting us. [11:03] deryck, g'day [11:03] hi deryck, hi jml [11:04] jtv: Great. I'm uncertain what you mean by " makes it less inviting to set up proprietary projects for translation without contacting us." [11:04] yo dezza [11:04] jtv, hi [11:04] * jtv feels an O(n²) process [11:04] jtv, henninge_: let's have a call [11:04] danilos: hands off, I'm talking to jtv === henninge_ is now known as henninge [11:04] mrevell: no way, it's in the calendar ;) [11:04] heh [11:05] mrevell: with bzr imports, you share your source first and the translation imports are then easy. With manual uploads, it can seem to make sense not to share source. [11:05] danilos: call [11:05] So flattering to have people fight about you... [11:05] jtv: I see, although of course people could just have a bzr branch with only their templates. But I get the point. Cool, will update das wiki now. [11:05] henninge: die, der or das for wiki? [11:05] mrevell: cool, thanks. [11:05] mrevell: das [11:06] danilos: call [11:06] lucky guess on my part, thanks henninge :) [11:06] das wiki ist mein wiki, henning get online on skype [11:06] danilos: I am [11:06] henninge: just seen you, calling [11:07] "ring" [11:07] "ring" [11:26] Now I have Abba in my head. Thanks :-P [11:27] jtv, henninge: https://lpstats.canonical.com/graphs/TranslationsCommitToBranches/ [11:30] al-maisan, who has launchpad.Edit permissions on a package set? [11:31] al-maisan, I can't find any security adapter for it [11:31] jml: it should be "techboard".. [11:31] jml: just a minute .. [11:31] al-maisan, techboard is the person with IPackagesetSet.new permissions. [11:32] * wgrant hopes there is never a need to access groups of package sets. [11:36] BjornT, do you know who has launchpad.Edit permissions to an object in the absence of a security adapter? [11:38] my reading says no one has permissions [11:47] bigjools: Nice change on the Packaging/PPA help page, thanks. [11:48] mrevell: yeah, it kinda helps to know how to upload :) [12:10] jml: right. if there is no adapter, no one should have permission. although we have some general adapters, for example for IHasOwner, which might get used. [12:11] BjornT, hmm, thanks. Maybe that was being used in this case... [12:12] BjornT, are there any tools to find out which security adapter is being used? [12:12] currently I just use grep [12:21] jml: queryAdapter(ob, IAuthorization, 'launchpad.Edit') [12:21] BjornT, of course, thanks. [12:46] BjornT: does bugs.l.n offer full text search via the APIs? If not, are there plans to make it do so? (Someone from the Emacs project is asking.) [12:47] leonardr: ^^ (do you know answer to above?) [12:48] * kfogel goes to look [12:48] kfogel: i don't know the answer [12:48] kfogel: sort of. we offer the same search you get in the advanced bugs search [12:48] kfogel: yes, it does, through the searchTasks() method [12:49] kfogel: as intellectronica said, it's the same as the advanced search in the ui. [12:49] leonardr, intellectronica, BjornT: thanks. Yes, looking at advanced search, it seems to meet the description "full text search". I mean, it searches on the body as well as the description and summary of the bug, right? [12:49] anyone knows what might be the problem with PQM? my lazr-js branch is sitting there for quite some time with no progress [12:50] kfogel: yes [12:50] nm. it just got merged :) [12:51] BjornT: searchTasks() is not in launchpadlib, it is part of the ReST API, right? [12:51] kfogel: that's true for everything in the API. launchpadlib is just an automatically generated wrapper [12:52] kfogel: see https://launchpad.net/+apidoc/ for the complete reference [12:52] intellectronica: sure, but he could be saying a method name in launchpadlib [12:52] kfogel: the point is, the method will be there when you use launchpadlib, but you won't find it grepping the source of launchpadlib [12:53] intellectronica: oh, it dynamically generates the method names by querying the site? [12:53] at run time? [12:53] kfogel: exactly [12:53] intellectronica: inspired. I didn't know that. Thank you. === mrevell is now known as mrevell-lunch [12:54] kfogel: plus or minus a few modifications to make the interface a bit nicer [12:56] leonardr: we generate HTML id="foo" for each section name in the API docs; if we generated title="foo" attribute as well, then hovering over a given section would show the name, and one could write deep link URLs without having to View Source first. For example, view source around https://launchpad.net/+apidoc/#bug_target-searchTasks to see what I mean. [12:57] kfogel: that should be quite easy to modify. the documentation is transformed into html from wadl, which is an xml interface description language [12:57] kfogel: yeah, file a bug for it or jfdi by modifying the xslt file [12:58] that's done using xslt, which is heaps of fun, if you haven't tried it yet :) [12:58] leonardr: just wanted to make sure you thought it was a good idea; I'll have a look [12:58] jml: did you get my mail? [12:59] mwhudson, yes, I did. [12:59] mwhudson, I guess I should actually read it in detail. [12:59] jml: i would appreciate it [12:59] jml: no hurry though, i should be asleep :) [12:59] mwhudson, sure thing. [12:59] mwhudson, I've been puzzling through security & soyuz stuff [13:00] jml: so concurrency in the puller should be a nice relaxing change for your brain [13:00] jml: or something [13:01] mwhudson, yeah, for sure. [13:01] soyuz gets such a bad rap ... [13:01] bigjools, there's a fair bit of scope for cleaning it up [13:02] yeah, it's the oldest code in LP [13:02] and the problem domain itself is very complex [13:02] which compounds the problem [13:02] I personally never thought it was that complex, there's just a lot to know [13:03] bigjools, that's the sense I meant, I guess :) [13:03] bigjools, so, while I have you distracted... [13:03] I knew you'd say that :) [13:03] bigjools, why doesn't nascentupload check pocket permissions? [13:03] because they don't exist? :) [13:04] bigjools, what's IDistroSeries.canUploadToPocket all about then? [13:05] oh, it's in uploadpolicy [13:05] oh that's a state check [13:05] right [13:05] it's not a permission as such [13:05] what's the guiding principle for a check being in uploadpolicy rather than nascentupload? [13:05] I guess that makes your job interesting [13:06] hysterical raisins? [13:06] honestly, no idea, it doesn't seem that sane does it [13:06] bigjools, that's fine. I'm happy to clean things up as I go along. [13:06] just as long as I'm justified in doing so :) [13:06] * bigjools owes jml a lot of whisky [13:07] it's also nice doing this in the same room as the ubuntu foundations team :) [13:07] ha :) [13:07] * bigjools waves to cjwatson [13:07] bigjools, btw, this is what the verify_upload function looks like now: http://paste.ubuntu.com/247942/ [13:07] it's still missing component checking & pocket checking [13:09] actually, the tests might be a little more interesting: http://paste.ubuntu.com/247943/ [13:10] jml: is that a new file? [13:12] bigjools, yeah. the new file is at archiveuploader/permission.py -- http://paste.ubuntu.com/247946/ for the full thing [13:12] it seems to duplicate checks done elsewhere [13:12] bigjools, I'm happy to move it, but I figured that was as good a starting place as any [13:12] bigjools, it's extracting out verify_acl [13:12] right - so we can remove some other tests [13:13] I doubt they're as efficient as these [13:13] bigjools, yeah, that's probably a good idea. [13:13] since they're probably trying to upload a package [13:13] finding tests to delete is a lot harder than finding code to delete though. [13:13] heh heh. [13:13] lib/lp/soyuz/doc/nascentupload* [13:13] and [13:13] lib/lp/archiveuploader/tests/* [13:13] and I've been keeping the existing tests as guarantees that my refactoring maintains the existing behaviour. [13:14] good call [13:14] * bigjools is looking forward to seeing your changes [13:14] bigjools, do the interim versions look ok so far? [13:15] jml: I think so, I am still looking [13:16] cool. [13:16] well, I'm off to lunch now, but I'll be happy to read your comments in the backlog :) [13:16] ciao [13:16] jml: enjoy, I am eating now too [13:18] hmmm the new AJAX-y "request another review" for an MP doesn't go quite right [13:26] intellectronica, leonardr: http://lists.gnu.org/archive/html/emacs-devel/2009-08/msg00219.html [13:26] kfogel, col [13:27] kfogel: so what, emacs is moving or thinking of moving to LP? [13:28] intellectronica: Emacs is switching to Bazaar (but trunk not hosted on LP). I'm proposing that it switch to launchpad bugs, because right now only launchpad bugs and debbugs meet Emacs' criteria for a bug tracker: that it be free software, and that it be entirely operable by email. [13:28] very nice [13:29] leonardr: find . -name "*.xslt" -print ==> no results in Launchpad tree, same with "*.wadl". Give me hint as to what to search for? === mrevell-lunch is now known as mrevell [13:31] leonardr: ./lib/canonical/launchpad/apidoc/wadl-development.xml [13:31] ? [13:32] kfogel: no, that's a test file [13:32] the xslt file is in launchpadlib [13:32] now that launchpad is open source we could move it back into launchpad [13:32] but some people wanted to modify it [13:33] leonardr: cd ~/src/launchpadlib && find . -name "*xslt*" ==> no results [13:33] kfogel, debbugs could also do it [13:33] oh, doh [13:34] NCommander: :-) [13:34] kfogel, I thought bugzilla was adding support for that htough [13:34] leonardr: ah, found it I think [13:34] Does anyone know if the sync packages script can copy into a component [13:34] leonardr: wadl-to-refhtml.xsl, right? [13:36] y [13:39] leonardr: where is the process that drives the XSL conversion? I don't see any other file (like a Makefile or setup.py) that refers to wadl-to-refhtml.xsl. I.e., how will I test my changes? [13:40] kfogel: i believe there is something in the launchpad makefile that does the conversion and puts it in apidoc/ [13:40] leonardr: oh, launchpadlib is a dependency of launchpad? I didn't realize that. Thanks. [13:40] (I thought we drove the ReST APIs directly... maybe we do, and the dep is only for the apidocs. dunno) [13:43] kfogel: i'll lay it down for you [13:43] leonardr: lay it on the line for me, brother [13:44] there are five pieces of software: lazr.restful, launchpad, wadllib, lazr.restfulclient, and launchpadlib [13:44] data flows out of the launchpad database from left to right [13:44] lazr.restful and launchpad are on the server. wadllib, lazr.restfulclient, and launchpadlib are on the client [13:45] lazr.restful has self-contained tests that use fake web services [13:45] wadllib has self-contained tests that use an old static copy of launchpad's wadl [13:45] lazr.restfulclient has tests where it acts as a client for lazr.restful's fake web service [13:46] launchpadlib contains launchpad-specific code and must be tested against the actual launchpad web service [13:48] we integrated launchpadlib tests into launchpad because changes to launchpad can break launchpadlib [13:48] but all the other pieces of software -- lazr.restful, wadllib, and lazr.restfulclient -- are tested standalone [13:48] leonardr: *nod* thank you. (please take a look at http://paste.ubuntu.com/247977/ to see if I'm on the right track, when you get a chance) [13:50] kfogel: that looks good to me, but ask sinzui, he's the xslt expert [13:50] * sinzui looks === carlos_ is now known as carlos [13:50] sinzui: thanks. Not tested yet (i.e., I haven't generated docs from it yet) [13:52] kfogel: what is the value of having a title in every heading? [13:53] sinzui: right now, if you're reading the API docs online (say, visit https://launchpad.net/+apidoc/#bug_target-searchTasks), you cannot know the name of the section you're in without doing View Source. With this change, hovering the mouse pointer over a section would show its name. That way people could more easily refer to specific sections within the documentation. [13:54] s/refer/refer other people to/ [13:54] heh, I mean s/refer to/refer other people to/ [13:55] In other words, I had to do View Source to come up with the "#bug_target-searchTasks" part of that URL. After this change, I could just hover and grab it. [13:55] okay. This change look good. [13:56] sinzui: do you know off the top of your had what command I should run to re-generate the HTML to test this? (If not, no worries, I'll poke around.) [13:56] make clean; make build [13:56] will do it [13:56] heh [13:56] thnks [13:56] thanks === matsubara-afk is now known as matsubara [14:01] morning everyone [14:01] flacoste: morning [14:02] hey flacoste, good morning. I'll let you settle in but I have a CP request when you have time to talk about it :) [14:04] sinzui: http://paste.ubuntu.com/247987/ Not sure how to interpret that error message -- there are 9 files named "_pythonpath.py" in the tree, and they are all under version control or a symlink to a file under version control, so just removing the file would be weird... [14:05] I wonder if this relates the the make distclean that was requested [14:06] kfogel: we want to get rid of _pythonpath, buildout allows it to a certain extent, but it's not complete yet [14:06] kfogel: I'm not sure what is wrong [14:09] sinzui: thanks [14:09] flacoste: ok. In the meantime, do you know what I should do to make my tree buildable? [14:09] sinzui: (yeah, I was wondering the same thing re distclean) [14:10] kfogel: i don't know what is the problem, and my tree is two weeks old here so... [14:12] flacoste: np [14:12] kfogel: top level _pythonpath [14:12] gary_poster: remove it? (it's versioned...) [14:12] it is? :-/ hm, checking further [14:13] uh, it shouldn't be. checking further. [14:15] kfogel: it is not versioned. it is in .bzrignore. buildout-templates/_pythonpath.py.in is versioned [14:16] gary_poster: sigh. So when you do 'bzr status -v' on an unversioned-but-ignored file, it does not say "unknown file", it just returns silently. [14:16] JamalFanaian, ping. can you set a commit message on https://code.edge.launchpad.net/~jamalfanaian/launchpad/bug-337582/+merge/9524 so that I can use it when landing your branch? [14:17]