[12:04] ...the launchpad test suite seems deliberately designed to make it near-impossible to test on production systems [12:11] Merge to rocketfuel@canonical.com/launchpad--devel--0: new improved fascist. r=SteveA (for parts spiv wrote), [trivial] for the rest. (patch-2043: steve.alexander@canonical.com, andrew.bennetts@canonical.com) [12:12] jamesh: ping? === asgeirf [~asgeirf@203-206-229-169.dyn.iinet.net.au] has joined #launchpad [12:36] Merge to rocketfuel@canonical.com/launchpad--devel--0: [trivial] make the title input autofocus on the product filing page. (patch-2044: brad.bollenbach@canonical.com) [12:38] jblack: pong [12:39] ok. [12:39] I'm in a conversation atm (thats making a lot of sense to me). I'm almost done though [12:40] New Malone bug 1263 filed on product Malone by Brad Bollenbach: Bug ID links don't change colour after they've been visited [12:40] https://launchpad.ubuntu.com/malone/bugs/1263 [12:43] lifeless: ok. =) [12:48] New Malone bug 1264 filed on product Malone by Brad Bollenbach: Bug keyword search is overly-literal [12:48] https://launchpad.ubuntu.com/malone/bugs/1264 [12:50] New Malone bug 1265 filed on product Malone by Brad Bollenbach: Bug search seems to ignore matching on assignee [12:50] https://launchpad.ubuntu.com/malone/bugs/1265 [12:51] lifeless: your patches are in prod [12:51] and there has been a daily sync since the rollout [12:51] so now is good time to sanity check the result [12:51] lot of red... apparently cvs.gnome went down [01:13] ddaa: patch-370 cleans up tempdirs [01:14] I'm used up for today. Care to send me a mail? [01:14] sure [01:14] I learned the hard way not to make production admin when tired. [01:14] have you documented the process yet so I can do it ? [01:14] hu no... [01:17] it would be a bit lengthy to document all... my rollout process is a mix of personal best practises and the use of few scripts and the ssh/rsync setup [01:17] dude [01:34] lifeless: morning [01:41] Keybuk: hey there [01:41] there are a bunch of bugs on hct in lp - are they still valid? [01:42] yeah, they ones I filed to start off with [01:43] what do you think of matts apt-get wrapper suggestion ? [01:44] sounds reasonable [02:02] so I had a cunning plan [02:02] mdz has a list of packages that turn him on [02:02] Keybuk: what should I do for FTP data here : http://dl.openafs.org/dl/openafs/ [02:02] lets get mdz laid. [02:03] I think that will work [02:03] so I wrote a script to download them, stuff the appropriate soyuz data in for them, and am now running sourcerer over them on my laptop to figure out why the fuck it's broken [02:03] and I'll fix the problems [02:03] sweet [02:04] and then he'll have a list of things he knows will work [02:08] ...ok [02:08] that's quite fascinating [02:08] the list of packages comes out of launchpad backwards [02:12] Merge to rocketfuel@canonical.com/launchpad--devel--0: add better distrorelease translation pages, r=salgado (patch-2045: mark.shuttleworth@canonical.com) [02:13] Keybuk: can you eyeball that url for me ? [02:13] Keybuk: and suggest what I should do w.r.t. ftp series :| [02:16] no, because I'm debugging, go away [02:16] ;) [02:16] lifeless: Doesn't dyson do recursive? [02:17] jblack: I don't know. Bugalogs heres does, which is why I'm askin' him. [02:17] it does [02:17] just stick that http URL as the root [02:17] ok [02:17] Thats what ddaa tells me, anyways, so I'd go with http://dl.openafs.org/dl/openafs/ openafs-*src.tar.gz [02:18] Today I got about 30 setup. Then, ddaa and I talked for a couple hours, and I finally undersatnd this. :) [02:18] fantastic [02:18] yeah. if you merged, I'm the one that bombed ya. [02:18] holy christ openafs is big [02:18] WARNING: revisions left: 90486 [02:19] [02:19] its one of the bigger repos. [02:19] what's bugalogs? [02:19] typo [02:19] bugalugs is what I meant. You. [02:19] what's bugalugs? [02:20] Google tells me it's a "BUM THIEF" [02:20] I suspect Google doesn't know [02:20] jblack: in that case you probably want something like openafs-1.3.*-src.tar.gz because there are overlapping series [02:23] ddaa: they are all main I think [02:23] hmm [02:23] no they aren't [02:24] no releases from MAIN at all [02:24] they have branches for everything [02:24] yuck. [02:25] I sort of understood that when series have overlapping time spans we want to make them distinct productseries regardless because otherwise sourcerer may get new releases with the wrong release ancestry [02:25] ddaa: I don't quite understand how to do series that aren't directly related to main. [02:25] in that case, if lifeless is correct, that's like automake [02:25] thus wanting to use the "these look mainish to me" [02:25] (hammer) [02:26] Keybuk: bugalugs == that guy == hey you - at least how I intended it, humurous 3rd person reference [02:27] ah [02:27] I think I see now [02:27] I thought you had some mystic software [02:27] I am full of mystic software === jblack mumbles something about lp being vaguely archish. [02:29] nah, it's like an elephant [02:29] Heh. I'm referring to the thou-shalt-not-delete feel. [02:29] hehehe... I guess that's a missing feature :) [02:29] if god didn't intend you to delete records from databases, he wouldn't have given you DELETE FROM [02:30] god(stub) hasn't given you DELETE FROM [02:30] ddaa: What's a good example of an overlapping series you've done? I want to pick it apart while I'm thinking about it. [02:31] Never mind. I just looked at cyrus-imapd [02:31] when I say overlapping it's to mean that you can see it from the ftp, the best one is cyrus-imapd, but apparently that one you have is just like automake, so the overlapping bit is really second to the fact they are all living in branches. [02:31] Merge to thelove@canonical.com/dists--bazaar--1.5: new build (patch-32) [02:32] that's really a bit where you have to psyschic and get in the shoes of the packager [02:32] Merge to thelove@canonical.com/bazaar--devo--1.5: added --switch to build-config (patch-25: Matthieu.Moy@imag.fr) [02:33] So help me, if he introduced more tabs, then we're having the next conference in france just so that I can kick him in the balls. [02:33] ddaa: I'm gnomic, not psychic. :) [02:34] gnomes have big feet, so you can feel the packager shoes better :P === jblack checks his shoe size... [02:35] Ok. [02:37] mpt: around? [02:38] you know what they say about men with big feet? [02:40] I used to know somthing like that about women [02:40] But maybe they say the same for men on the other side. === stub [~stub@203-217-37-199.dyn.iinet.net.au] has joined #launchpad [02:51] heh [02:52] I appear to have severed sourcerer's dependency on launchpad [02:57] stub: hey dude [02:57] Morning [02:57] your name is mud [02:57] try "make -C database/schema" as a user called "launchpad" with super-user privegles [02:58] That would fail. You have to run that as a PostgreSQL superuser. [02:58] stub: how much effort is involved in fixing a bug like: https://launchpad.ubuntu.com/malone/bugs/1264 ? [02:58] yes, we made launchpad a postgresql superuser [02:58] then security.py took it away again [02:59] launchpad should not be a PostgreSQL superuser or your tests will be dodgy and not fail when they should. [02:59] right [02:59] so we relented and just ran it as postgres [02:59] bradb: It would involve dropping the full text search stuff and never using indexes for searches, which would be bad. [03:00] stub: surely there's gotta be some solution that doesn't suck here? [03:00] i can't be the first person to have ever considered that users make typoes when they search [03:01] (and to note the other benefits of partial matching, of course) [03:01] bradb: We would have to dump the text to some sort of external search engine like agrep or write stuff for PostgreSQL. [03:01] bradb: Spelling can be done by improving the stemmer I think. [03:02] bradb: You can integrate ispell into the mix, but I didn't have much success setting it up. [03:02] does a stemmer take a misspelled word and try to link it to one or more of the correct spellings? [03:03] bradb: It could. Checkout textsearching.txt for what the stemming currently does (mainly plural forms and some suffixes). [03:04] bradb: I don't know if the ispell integration does that though out of the box - I would need to get out of the box functioning fist (since the docs are sketchy) [03:04] jamesh: ping? [03:04] stub: https://launchpad.ubuntu.com/products/malone/+bugs?field.searchtext=listings&search=Search&advanced=&status=10&status=20&assignee=all [03:04] are the results being weighted properly here? [03:04] why is there a bug with a title that doesn't have "listings" in it above a bug with a title that does? [03:04] bradb: Nope. I only just got the ranking into the indexes. No code is using them yet. [03:05] ah [03:05] bradb: It complicates the queries substantially unfortunately [03:05] Keybuk: yeah? [03:05] but it will happen :) [03:05] stub: i'm willing to assume the complexity so that the user's life becomes easier [03:05] jamesh: so, I'm confused [03:06] you say that hct-enable got merged into rocketfuel [03:06] stub: can you give me an idea of when you plan to look into the typo fu? [03:06] except that none of the patch logs are in it [03:06] but when I try and merge, I get 82,000 conflicts [03:06] Keybuk: Mark had half of your branch merged into his launchpad--debbugs--0 branch [03:06] but none of the patch logs? [03:07] Keybuk: it seems that the other half of the revisions in your branch were keeping it up to date [03:07] Keybuk: base-0 -- patch-4 are in rocketfuel [03:08] oh, silly me [03:08] I think I just merged it into the production config [03:08] yes 5 thru 8 are all ketchup [03:08] Keybuk: I took a quick look through the conflicts, and they look like areas I got Mark to fix when I reviewed debbugs [03:10] right [03:10] so we can just forget about that [03:10] and remove it from PendingReviews [03:12] bradb: The basics is " select title from bug where fti @@ ftq('listings') order by rank_cd(fti, ftq('listings')); ". I havn't through through how to do it though when the indexes are in multiple tables. Probably something like 'order by rank_cd(bug.fti | bugmessage.fti, ftq('listings') === stub buggers off for 40 mins [03:13] friend has an emergency, I'll be popping out for a while [03:15] stub: sorry, i was referring to when you might have a chance to look at the ispell (or whatever) integration [03:34] bradb: yo [03:35] mpt: hey: what do you think of these suggested titles? https://launchpad.ubuntu.com/malone/bugs/1266/ [03:38] bradb: bug 936 is a bug in Malone, not a bug in Firefox [03:39] So if we chop out the "Malone: " from your suggested titles, the only difference between them and what we have currently is "Bug 1234 (edit)" vs. "Editing bug 1234" [03:40] or rather, "Bug 1234 in Mozilla Firefox (edit)" vs. "Editing bug 1234 in Mozilla Firefox" [03:40] mpt: the other difference is that the edit and view pages have that same "Editing..." title currently :) [03:40] mpt: do you see why that's a fairly important difference? [03:41] well, that's just a bug [03:41] So we have three distinct things here [03:42] in order of definiteness [03:42] (1) the view mode shouldn't say "Editing" [03:42] (2) the edit mode should be "... (edit)" rather than "Editing ..." [03:42] (3) all the titles should start with "Malone: " [03:43] correct? [03:43] (4) the bug and bug task titles should be consistent with one another [03:43] but yeah, that's basically it [03:44] (4) is something I did months ago and had to revert [03:44] in the way described by the bug report? [03:45] yes [03:45] you could have used that sentence as your first reponse when i asked what you thought of the titles suggested in that bug report :) [03:46] IOW, if i baz get that branch, i'll find a bugfix for precisely the bug i'm showing you? [03:46] If I had, you would have said "What's (4)?" :-) [03:47] If you baz get, you'll get a smelly fix for (4) [03:47] you won't get a fix for (1), (2), or (3) [03:47] and you won't be able to use the fix for (4) anyway, because that was the pre-LaunchpadTitles days. [03:48] you've lost me [03:49] We now have pagetitles.py for specifying titles of pages in Launchpad [03:49] how can it be that you fixed the titles in the way described by the bug report without addressing 1, 2 and 3? [03:49] I said, "(4) is something I did months ago and had to revert" [03:49] I didn't say "(1), (2), and (3) are things I did months ago and had to revert" [03:50] ok, i'm just going to make this fix :) [03:50] well, this is interesting [03:50] def bugtask_display(context, view): [03:50] return 'Bug #%s in %s: %s' % (context.bug.id, context.contextname, context.bug.title) [03:50] def bugtask_editform(context, view): [03:50] return 'Editing bug #%s in %s: %s' % (context.bug.id, context.contextname, context.bug.title) [03:51] bradb: So (1) shouldn't be happening. [03:51] it should [03:51] that only worked when they were two different pages, which they are becoming again now [03:52] Are you using bugtask_editform even for the view mode? [03:52] i've already submitted the merge to break it into two again [03:52] then i'm going to make another small change to hide the Save Changes button [03:52] (on the view form) [03:52] bradb: I don't think I'll have a chance to improve the text searching any more than it currently is before 1.0. It was on my list, but I don't see myself finding time for it. [03:52] then fix the titles [03:52] stub: ok, no worries. maybe we'll have a chance to discuss it in brazil. [03:53] well, if you're splitting it into two using the same template names as before, you won't need to touch the titles [03:53] they'll start Just Working [03:53] i called them bugtask-view.pt and bugtask-edit.pt [03:53] bradb: Yup. Be good to have some brains on it too. This is a whole big area of research that I've only dipped my toe into. [03:54] bradb: So all you'll need to do is fix the first and third of the four lines I pasted above [03:54] bradb: But it seems wrong to me that there should be completely different templates depending on whether you're editing something or looking at it [03:54] stub: i can imagine. it would be a real usability milestone for all of launchpad to improve the searches in that way. [03:54] that just encourages gratuitous inconsistency in layout [03:55] can lp titles even work if it's only one template? [03:55] erp - make that rank(fti, ftq('foo'), not rank_cd. Need to experiment with the two, but I expect the former will give us better results. [03:55] (also, the task view Just Looks Different that the view view) [03:56] bradb: I only put "Editing " in the title of the edit page *because* they were separate pages. I'd much rather they were just a single "Bug 123 in Mozilla Firefox" page with a single title. [03:56] New Malone bug 1267 filed on source package gqview by Vince: Doesn't remember window dimensions/location etc... [03:56] https://launchpad.ubuntu.com/malone/bugs/1267 [03:57] and that, in the future, you can click a little Edit icon next to an item to turn it into a text field or whatever [03:57] without having to go to a separate edit page [03:57] mpt: enter sabdfl, who has said he doesn't like having one superform that handles all possible scenarios, etc. [03:58] mpt: as well, like i say, they'll be different pages. [03:58] yeah [03:58] *sigh* [03:58] Launchpad has too many pages. [03:58] for example, kiko has proposed that maybe the watch information doesn't make sense on the view only page if there's no watch info set [03:59] A separate page for subscribing, a separate page for marking as duplicate, a separate page for editing title/description ... it's painful. [03:59] also, the view page could have, for example, and actions portlet that has an action to go to the edit page (like the bug page has for editing bug details). [04:00] mpt: we agree fully on that horror :) [04:00] the uncensored version of my opinion on that model is that it's PURE SHIT [04:00] but changing things will require a consensus that we haven't yet reached, i don't think [04:01] oh, now that's just kooky [04:02] the success *and* failure functions just got called for everything [04:02] Heisenbug! [04:02] bradb: I disagree with (3), however [04:03] and iirc kiko does too [04:03] mpt: you want to keep "Malone: " in there? [04:03] i'm happy to go either way, really [04:03] No, I want to keep it out of there [04:03] It's currently out of there [04:04] i'm happy to not put it in there too [04:05] i don't mind if we don't put (edit) in there either, to be honest [04:05] bradb: Bug 936 was a bad example for you to use in bug 1266, because bug 936 is a bug about Malone-the-product :-) [04:05] that's why it has "Malone" in its titles. [04:06] i'm going to change the description to make things clearer [04:07] New Malone bug 1268 filed on source package gqview by Vince: File browser doesn't use Gnome's widgets/icons [04:07] https://launchpad.ubuntu.com/malone/bugs/1268 [04:10] there [04:11] mpt: btw, did you notice: https://launchpad.ubuntu.com/malone/bugs/1263 ? [04:11] Malone seems to be a bit sporadic in sending out notifications today; not yet sure what's going on [04:11] bradb: yes, I saw when dilys mentioned it [04:12] ok [04:12] that's going to be ... fun [04:13] mpt: can we perhaps create two kinds of links, one for menus, and one for "content" links? [04:13] It's not that hard [04:13] the "content" kind of links could be underlined; the menu ones not, perhaps [04:14] telling menu links not to change color is either 10 or 20 more characters in launchpad.css [04:14] it's easy [04:14] 20, I think [04:14] i was wondering about that: is it bad for them to change colour? [04:15] well, why else did you want two different kinds? [04:15] oh, the underlining [04:15] right [04:15] well, ironically, the links in the bug listing sidebar shouldn't have been underlined, to match the portlets elsewhere [04:16] exactly, i screwed that up [04:16] no, that was my fault [04:16] or did you use something other than
? [04:17] it was my fault dude: [04:17] https://launchpad.ubuntu.com/malone/bugs/1263 [04:17] er, bad paste [04:17] style="text-decoration: underline">untriaged (18) [04:17] oh man [04:17] i like the sometimes F11, sometimes Shift-Ctrl-V paste feature of Linux [04:17] ok, I admit it, it was your fault :-) [04:17] mpt: yeah, i went a bit underline happy there. i've since done some more reading. [04:18] but, but, here's the ironic bit [04:18] (particularly when i looked at it and went "ugh, that can't be good") [04:18] SteveA had trouble finding the link for reporting a bug [04:18] because it was the only link that wasn't underlined :-) [04:18] heh heh [04:18] i'm not sure how that got there not-underlined [04:19] i don't recall intentionally meaning to do it that way, so either i screwed that up, or somebody else intervened to screw it up [04:19] *but*, in any case, the *real* *true* screwup is the underlining of the stats links [04:19] once they're un-underlined, all should be good [04:20] so SQLobject doesn't appear to do just an INSERT [04:20] but an INSERT and then an UPDATE [04:20] *sigh* [04:24] of all the complaints i have about the software i use to do my job, none ranks so highly as the fact that pqm stands between me and rocketfuel [04:25] bradb++ [04:29] lifeless: on a more serious note, how much effort would it be to hack pqm to send a confirmation message when it's received a merge request from someone? [04:29] oh, right, he's out [04:30] are you in the same black hole as me right now [04:31] . o O { I know I sent the merge ... why haven't I heard anything? does pqm not love me? hello? is this thing on? } [04:31] yer damn right i am [04:31] @!*!@ [04:31] pqm hung a number of times yesterday [04:31] did you read your mail.log to find out whether the problem was at your end too? [04:32] root@oxygen:~/launchpad/lib/canonical/launchpad/templates # tail -n 100 /var/log/mail.log | grep pqm [04:32] the stuff got through eventually [04:32] Jul 6 21:46:27 localhost postfix/smtp[20441] : 399EC1FDC4: to=, relay=relais.videotron.ca[24.201.245.36] , delay=0, status=sent (250 2.5.0 Ok.) [04:32] heh [04:32] but pqm needed to be kicked manually [04:32] i.e. 45 minutes so far [04:33] Jul 7 02:20:36 localhost postfix/smtp[14828] : 1EEA440FCCF3: to=, relay=mailgate.netsplit.com[62.49.129.34] , delay=3, status=sent (250 ok 1120699236 qp 21409) [04:33] 1h10 ... I got you beat ;) [04:33] hehe [04:33] actually, it might be cool if the confirmation message said what number in line we were too [04:34] and had an embedded midi of greensleeves? [04:34] but who am i kidding? it would still suck. [04:34] [04:34] actually, that's probably a pretty UK-centric joke [04:34] there was a time when pretty much every PABX shipped with Williams/Fantasia on Greensleeves as its hold music [04:35] simply cause it was a generic card BT could fit cheaply [04:36] pqm muzak, woo! [04:36] "your branch is very important to us. Please hold til one of our automated quality assurance bots can service your request" [04:37] though with pqm, I suspect the theme should be I HATE YOU SO MUCH RIGHT NOW)(!(")$)!$ [04:37] we could get elmo to record it === bradb chuckles depressively [04:38] right, it's in god's hands now. later all. [05:03] stub: from caseyu [05:03] uh, caseyt [05:03] uh, casey [05:03] psql: FATAL: no pg_hba.conf entry for host "82.211.81.149", user "dyson", database "launchpad_dogfood", SSL off [05:04] Keybuk: Will sort that now [05:05] Keybuk: Let me know if you want to use the staging server for any of this btw. You get production data to test with, but it is blown away every day. [05:06] define "production data" ? [05:06] as in the data in the real server? [05:06] if so, yes please! [05:06] Yes. Restored from production backup each day. [05:06] And upgraded to the latest schema [05:06] cool [05:06] please [05:06] I assume staging has its own librarian? [05:07] that's more useful to me right now than dogfood, in fact [05:08] dogfood sorted [05:08] Keybuk: Who else has access to casey? [05:08] elmo [05:08] ok. secure enough. [05:08] and probably Mark, if he asked [05:08] ;) [05:09] Do you want just dyson, or sourcerer as well? [05:09] both [05:10] what time (UTC) does the database get wiped? [05:10] Roughly 01:00 UTC. Depends on how long the production backup takes. [05:11] You will know because your connections will be killed brutally ;) [05:11] ok, I'm pretty well-behaved connection-wise these days though [05:14] Keybuk: ok. access to dyson, sourcerer and ro PostgreSQL accounts on staging.ubuntu.com should now be open to all accounts on casey. Let me know if more people get access and I need to tighten that. [05:15] ok [05:15] thanks === stub quickly changes to ssl connections to keep elmo happy [05:16] ok... all done for real ;) [05:21] if I've made a configs/+casey/launchpad.conf [05:21] what do I put in LPCONFIG? [05:21] LPCONFIG=+casey [05:22] ok ... here goes nuffin [05:25] Looks like you have a permission problem - need access to LibraryFile tables [05:25] oh, yeah, I keep tripping various permission problems as I go [05:26] random question [05:26] I'll need to fix them on staging, or feed them back through rocketfuel and they will be applied the next day [05:26] can I create an "hctapi" group, and just stick dyson, sourcerer, etc. in that [05:26] actualy, no [05:26] ignore me [05:26] that's no good because sometimes we don't want write (trebuchet) [05:27] I have a bunch of config changes [05:28] I'll pop them in a branch, hang on [05:45] scott@canonical.com--2005/launchpad--security-cfg--0 [05:45] stub: ^ do those look ok to you? === stub merges [05:52] the main permission problems I get now are where SQLobject goes off and does unexpected things [05:52] so I want to look at just one record, but it has to go fetch records from other tables there and then, rather than on demand [05:52] (there's a SELECT on Person for that, somewhere) [05:53] and that's ignoring the entire mess around SourcePackagePublishingHistory and the umpteen views for it [05:57] stub: hmm [05:57] here's a new one on me [05:57] psycopg.ProgrammingError: ERROR: permission denied for sequence libraryfilecontent_id_seq [05:57] Yup. Understandable. [05:57] I thought the librarian was supposed to be self-contained [05:57] and you weren't supposed to give yourself magic secret access to it [05:58] The Librarian API changed because of the transaction logic. The inserts are now done by the client [05:58] Erm.. I should say the API didn't change, but who does what behind the curtain has. [05:58] oh [05:59] so what do I need? [05:59] SELECT, INSERT, UPDATE on libraryfilealias and content [05:59] and, seemingly, the sequence ? [05:59] LibraryFileAlias and LibraryFileContent. SELECT and INSERT should be enough. Sequences are handled magically with the table permissions. [05:59] right [05:59] SQLobject seems to behave oddly sometimes though, and UPDATE when you create a record? [05:59] you ever seen that before? [06:00] Keybuk: It shouldn't need to, but I wouldn't be surprised if it did. In general, no database modifications are done (except selecting the nextval from the tables primary key sequence) until commit time. === stub taps his foot impatiantly waiting for baz diff [06:01] ok, updated that branch to give SELECT, INSERT on those two tables [06:06] 'm gonna crash now [06:06] have stuck that in your pending queue ;) [06:06] if it's good, I'll merge it into rf in the morning [06:06] can you update staging with it so I can play? [06:10] Keybuk: That all looks file. Send to rocketfuel if you are still up. [06:10] ok, done -- pqm seems stuck, but it'll go through when lifeless kicks it :p [06:56] morning [06:56] stub: i think i got a deadlock running the test suite yesterday [06:57] so i removed locking from the sqlobject cache on my machine, and it didn't deadlock any more [06:57] i'm not sure if it was coincidence or not [06:59] PQM seems to be getting stuck a lot this week [07:21] SteveA: I guess we don't need locking since we don't share the cache or instances between threads [07:21] right [07:22] Although if you could trigger the deadlock, that indicates something was behaving badly. A place for a check? [07:27] yeah, i thought about storing some state in a dummy lock to detect deadlocks [07:27] but it's quite tricky to do it in a way that would actually detect a deadlock [07:27] between different locks [07:27] i can easily detect asymmetric use of a single lock === Fluffels [~Fluffels@tpr-165-243-114.telkomadsl.co.za] has joined #launchpad [08:03] SteveA: You don't have to detect a deadlock - you just have to detect another thread attempting to acquire it. [08:04] Oh... or do you mean *normal* operation triggers this? [08:05] i don't know. i'll think about it more. [08:06] are you okay with me removing the locking from our sqlobject, and replacing it with something that barfs when strange things happen? [08:08] SteveA: Sure. [08:08] warning or full barf? [08:08] I'm defacto SQLObject expert this week am I? :-) [08:08] Full barf [08:08] okay [08:08] BARFAMATIC [08:08] Warnings are for pussies. We have to many warnings we ignore already. [08:08] i also have sqlos changes that hopefully fix the "bind the object to this thread" [08:09] that one got tricky, due to creating cyclic garbage with __del__ methods [08:09] i solved it with a weakref [08:09] i'm going to rip out sqlos over the course of the next month or two [08:10] yay [08:10] i've started by ripping about 1/2 its code out [08:10] so, don't expect us to sync with upstream sqlos ever [08:10] SteveA: Replace it with something Launchpad specific, or perhaps we can stuff the remains into SQLObject somehow? [08:10] i've told upstream about the bugs i've found [08:11] replace it with launchpad specifics [08:11] there isn't much to sqlos [08:11] and what is there is obsured by lame attempts at cacheing connections [08:11] it will just get folded into transaction event subscribers and a slightly fatter launchpad database base-class [08:11] Indeed. That is why I was wondering if perhaps it could be replaced by sticking a few hooks in SQLObject (_connection descriptor etc.), and the Z3 driver would be pretty small. [08:12] we can do that on our sql base class for launchpad [08:13] This could allow us to drop Zopeless too... hmmm.... [08:13] it could? [08:13] okay, what's the roadmap: [08:13] - simplify sqlobject cache, progressively remove sqlos (underway) [08:14] - totally remove sqlos (next few weeks or so) [08:14] - add some hooks into sqlobject to make it easier to interface with stuff, swap out the cache etc. [08:15] (not now but after) the Zopeless stuff could be folded into SQLBase, and we replace the zopeless transaction manager with hooks to transaction.commit() etc. Database connections could be done in one script setup method that brings up the CA at the same time. [08:16] I guess it doesn't win much of anything though, except the transaction module integration which we could do anyway [08:17] i'm all for making the zopeless environment less "special" [08:42] BjornT: ping [08:45] SteveA: hi [08:45] there are a couple of fascist warnings about malone code. [08:45] There were 8 imports of names not appearing in the __all__. [08:45] You should not import BugSetBase from canonical.launchpad.database.bug: [08:45] canonical.launchpad.database.bugattachment [08:45] canonical.launchpad.database.bugwatch [08:46] You should not import newBugTracker from canonical.launchpad.browser.bugtracker: canonical.launchpad.browser.project [08:46] [08:46] can you take a look at those, and see whether the names should be added to the __all__ of those modules, or whether the code importing the names is wrong? [08:48] it should be pretty quick to do, and will improve the import fascist spew for everyone [08:48] (the 8 above is the total number. there are just three to do with malone in some way) [08:49] SteveA: sure [08:49] thanks [08:54] i'm going to make the fascist allow importing any names into tests [08:54] even when they don't appear in the __all__ [08:54] tests are identified as code in /test/ or /ftest/ directories. === bradb-away [~bradb@modemcable082.64-130-66.mc.videotron.ca] has joined #launchpad === elmo [~james@217.205.109.249] has joined #launchpad [09:52] morning [09:52] elmo, hi [09:52] elmo, around? [09:57] carlos: kind of [09:58] elmo, did you installed gettext on production? [09:58] elmo, It's a launchpad dependency to export .mo files [10:16] done [10:17] elmo, thanks [10:18] elmo, would you install it on the new servers where we are going to move production so we don't forget it? [10:18] already done [10:18] elmo, it's only needed if they will host launchpad website [10:18] ok, thanks === sivang [~sivang@box79162.elkhouse.de] has joined #launchpad === sivang [~sivang@box79162.elkhouse.de] has joined #launchpad [10:40] stub: i'm pretty sure all these sqlobject / sqlos things i've found are the cause of both the errors and the hangs in PQM, and probably some hangs in production [10:41] the hangs in PQM that are to do with the test suite are caused when the GC interrupts a thread while that thread is holding a cache lock [10:41] I'm not aware of any hangs in production. The major PQM issue is with cscvs unfortunately :-( [10:42] so, the GC calls __del__ on some old transaction, which goes to acquire an already acquired lock because it tries to clear an inappropriate cache [10:42] I think lifeless has seen three distinct ones, so if we can kill one of them it will be good. [10:42] ok [10:43] i'll start by killing the locking on the cache. next, land the sqlos connection improvements, then i'll tighten up the locking errors. [10:43] Ideally, the __del__ methods should be a noop in our system. [10:44] If we are customizing this enough, could be worth bitching loudly if this is not the case (?) [10:45] yeah, agreed. but, later on. [10:45] i want to land this bit by bit. [10:46] because i want some instant improvements, but i also have many other things to do immediately. [10:46] daf, hi, around? [10:49] hi [10:51] daf, I just added you to the CC of a mail from rosetta-users [10:51] daf, it has a question about scripts and Indian languages [10:52] daf, just in case you know something about that [10:52] ok [10:55] SteveA: yes it is [10:56] Kinnison: i checked in some minor changes [10:56] stub: How did gina do? [10:56] but i was surprised by some things [10:56] oh? [10:56] the exceptions did not seem to derive from Exception [10:56] That's almost certainly an oversight on my part [10:56] i fixed an import of * with no __all__ in the module [10:56] thanks [10:56] i'm surprised the exceptions work [10:56] After beating her around a bit with a blunt object, she imported warty and I'm about to kick off a full import of warty*, hoary* and breezy [10:57] stub: cool [10:57] so, when you fix them, please check their use is tested [10:57] You can see the results of the warty import on staging. I have no idea what to look for so please have a poke ;) [10:57] SteveA: Right, okay [10:57] stub: I'll do that in a moment === mdke [~matt@mdke.user] has joined #launchpad === SteveA wonders what pqm is doing [11:04] SteveA: it seems you're not checking against a current rf. BugSetBase is already in an __all__. it probably should be removed completely, but it's not a trivial thing to do. [11:04] stub: well, it system-errors when I look at the warty page :-) [11:04] stub: which is impressive [11:04] stub: I'll go get my keys so I can ssh to mawson and look at the db (assuming I still have access) [11:09] stub: Well it seems to have created 5450 sourcepackagereleases in the right distrorelease [11:14] http://artistic-insanity.net/song/daleksong.html [11:14] Oh dear === palmTree [~chatzilla@213.210.246.183] has joined #launchpad === palmTree [~chatzilla@213.210.246.183] has left #launchpad [] [11:17] mdke, hi, around? [11:31] Is there a way to combine a 'select distinct foocolumn' and a 'select count(*)' to get the obvious result? === ddaa [~ddaa@marvin.xlii.org] has joined #launchpad === ogra [~ogra@p5089CBED.dip.t-dialin.net] has joined #launchpad [11:35] and again... some user response from ubuntu-users : http://lists.ubuntu.com/archives/ubuntu-users/2005-July/041371.html [11:35] :) === ogra [~ogra@p5089CBED.dip.t-dialin.net] has left #launchpad ["Verlassend"] [11:39] Kinnison, I think it was SELECT count(DISTINCT foocolumn) [11:40] carlos: Hmm, ta [11:41] carlos, yo [11:41] carlos: yay [11:45] mdke, It's solved, thank you. The RosettaFAQ link was not a link at wiki.ubuntu.com/Rosetta but I already found a way to fix that [11:45] Kinnison, ? [11:45] carlos, cool [11:45] Kinnison, does it works? [11:45] carlos: it does with pgsql, and since we're using pgsql, rock on! [11:46] ;-) === sabdfl [~mark@host217-37-231-22.in-addr.btopenworld.com] has joined #launchpad [11:49] lifeless: is pqm stuck? i've sent various requests, and received no response. [11:49] and that was ages ago [11:51] checking [11:51] yes [11:53] can you purge all my outstanding requests from the queue? [11:54] I've added soyuz to the production launchpad products list [11:54] should I be adding it to a given project now? [11:55] grep steve ~/arch/queue/patch.11207* | sed -e 's/:/ /' |awk ' {print $1}' | uniq | xargs rm [11:56] ta [11:56] morgs: Remind me, how do I put a product into a project? [11:56] Hey Kinnison, edit the product. [11:57] Add the project there IIRC. [11:57] I don't see where [11:57] sabdfl: a reply from you on the direction of malone would help: http://lists.ubuntu.com/archives/ubuntu-users/2005-July/041371.html === morgs checks [11:57] morgs: can you put /products/soyuz into /projects/launchpad ? [11:57] OK [11:58] thanks dude [11:58] Ah, it's +review on theproduct [11:59] Kinnison: done [12:00] morgs: will the next production update offer that to launchpad admins? [12:01] Kinnison: It's there already, the product actions portlet has a Review link [12:01] Aaah I see it [12:01] righty [12:01] ta [12:08] morgs: hmm, looks like traverseProject is unused [12:09] daf: thanks, I'll take a look at that [12:09] BjornT: I just merged from RF [12:10] BjornT: the issue is that BugSetBase is being imported from database.bug [12:10] whereas it is defined in database.bugset [12:10] this is exactly the sort of mis-import that the facist should be finding [12:14] SteveA: oh, right. missed that. i'll change the imports [12:14] ta [12:17] Merge to rocketfuel@canonical.com/launchpad--devel--0: [trivial] break the task view/edit page into two separate pages (patch-2046: brad.bollenbach@canonical.com) [12:17] does dilys not report bugs here any more? [12:18] Hmm, actually, I filed a bunch of bugs (well, two) and I've not seen the mails yet [12:18] it launchpad's mailer backed up? [12:19] Kinnison, are they 'private'? [12:19] nope [12:19] bugs 1274 and 1275 [12:19] then I suppose it's daf fault [12:19] :-) [12:20] filed against soyuz and malone respectively [12:20] But I've not seen 'em on the launchpad list either [12:20] Good point [12:20] and I'm getting other mails from lists on lists.ubuntu.com [12:20] morgs: I'm trying to make the browser/product.py not import DB code, and I'm wondering if there are any page tests for the ProjectBugTracker stuff [12:21] blame Canada! === SteveA goes to visit the optician [12:21] daf: It seems, not... [12:22] daf: an updated fascist will be landing soon, which allow import of any names into test-related modules [12:22] daf: you mean browser/project.py? [12:23] morgs: yes, I do [12:23] :-) [12:23] :) [12:23] SteveA: that makes sense [12:23] morgs: looks like there are some tests for adding products, though [12:24] yes [12:24] so hopefully I'll know if I break something there ;) [12:28] morgs: would it be difficult to add a pagetest for creating a PBT? [12:29] daf: shouldn't be, let me have a quick look... === Kinnison wonders what people think about https://launchpad.ubuntu.com/malone/bugs/1275 [12:29] daf, There something wrong with the download form [12:30] daf, https://launchpad.ubuntu.com/products/schooltool/0.10-rc1/+pots/schooltool/+export [12:30] daf, try to select just a .pot file [12:30] you will get a system error [12:32] Kinnison: you could abuse the severity and priority (who needs those anyway) - that gives you what, 20 possible ways to classify a bug? ;) [12:32] carlos: hmm -- I looked at that yesterday, but failed to reproduce it locally [12:32] morgs: and for my next trick, watch as I make this pineapple disappear up your fundamental orifice [12:33] "All low priority bugs are UI bugs"... [12:34] Kinnison: so a product can have components, like a project has products... hmm. [12:34] Surely this would be handy from a bug tracking POV... [12:35] morgs: Except we can't call them components because that name belongs to part of soyuz's packaging infrastructure [12:35] err [12:36] featu^W [12:36] morgs: Project // Product // Subsystem [12:36] ? === Kinnison goes to ask his pedantic partner in crime [12:36] one sec [12:36] morgs: aha -- looks like there is a test for /products/+new, but not for /projects/foo/+newproduct [12:37] daf: I could knock up some tests, but given pqm's recent performance, you'd have to get them straight from me... [12:37] that would be wonderful [12:38] I just want to make sure I'm not breaking anything [12:38] morgs: My Rob suggests "Modules" [12:38] aspects [12:38] constituents [12:38] Ah, Modules. That sounds good... although... would every product consider UI to be a "module"? [12:38] constituents is a bit long [12:38] element [12:39] bit [12:39] "part" [12:39] thingy [12:39] monad === Kinnison lazily evaluates daf to nil [12:39] integrant [12:39] block [12:40] subdivision [12:40] head [12:40] segment [12:40] daf, did you tried with staging? [12:40] daf, seems like any .pot download request fails [12:40] Could also apply to a product that is written in more than one language - people would appreciate a way to mark a bug as being relevant to one or the other code base... [12:40] daf, even when I'm 100% sure librarian has that file as I just upload it [12:40] morgs: can you mark this as a spec to be discussed in brazil? [12:41] Hmm, speaking of which I hve plenty of wikiwork to do [12:41] carlos: mm -- I looked at the error logs yesterday, but couldn't make sense of it [12:41] carlos: https://launchpad.ubuntu.com/errors/showEntry.html?id=1120732784.830.918095340471 [12:42] daf, If I select a full export I don't get the .pot file anyway [12:42] carlos: that implies that self.context is None in the view class, which is mad [12:42] Kinnison: OK, BTW does this sound more like a malone thing or something that products could use more generally? [12:43] I think it could just be a Malone thing [12:43] BugCategory? [12:43] or possibly even just a text field [12:43] I do not see how it would make sense for anything but malone [12:43] but that would not be very relational of us [12:43] tags! [12:43] yes, tags :) [12:44] I wasn't going to be the first one to say it [12:44] morgs: It needs to be both [12:44] morgs: I.E. in products we need to define them [12:44] morgs: so that in malone we can use them [12:44] actually, better tags than that [12:45] In my experience a big fraction of bugs are annoying to classify [12:45] "Track your bugs in del.icio.us with our new firefox bookmarklet!" === mdke [~matt@mdke.user] has left #launchpad [] [12:46] like "this is a bug in a stylesheet, that in combination with a bug in the typesetter results in a confusing user interface" [12:46] should that be category "stylesheet", "typesetter" or "user interface"? [12:47] user will say user interface, stylesheet author will stay typesetter, upstream will say stylesheet :) [12:48] It's a UI bug which depends on another bug filed against stylesheet and another filed against typesetter [12:48] :-) [12:48] So, allow multiple tags... [12:48] this is getting into Philosophy PhD territory [12:49] Multiple subsystems is okay but potentially confusing [12:49] tags are fine so long as you can make sure noone misspells them [12:51] gmail's quite nice there - drop down to select a tag, or "new tag" [12:51] Do we have a spec template for launchpad specs? [12:51] "meta-spec, oui, mais je la mets o?" [12:52] SpecTemplate please [12:52] LaunchpadProposalTemplate [12:52] Kinnison: that way a pun of dubious taste [12:52] morgs: aah gotcha [12:52] re [12:53] ddaa: it relies on me knowing what 'spec' manages to translate back to :-) [12:53] Kinnison: nothing special [12:53] elmo: ping [12:53] Kinnison: btw, did I tell you about contrepteries? [12:53] daf, ok, the problem is a bit difficult to debug. Did you asked Steve or spiv? [12:54] Kinnison: SteveA: I'm sure you two guys would love these. [12:54] SteveA: ? [12:54] carlos: I'll look at it again shortly -- I'm in the middle of something right now [12:55] ok [12:55] daf, I'm going to file a bug so we don't forget it [12:55] thanks [12:56] ddaa: No I don't think you have === ddaa vaguely remembers having already talked about that before... [12:56] ddaa: Was I dunk? [12:56] erm, drunk [12:57] Kinnison: it's a kind of play of words that involves permutation of sonorities within a word or across words in a sentence. [12:57] is pqm hosed again? [12:57] ddaa: Hmm [12:58] Generally, the original sentence looks innocent, and the permuted one says something dirty. [12:58] sounds like a spoonerism [12:58] Right. [12:59] For example, say some female staff tells you she need to talk with you about something. You can answer: [12:59] BjornT, bradb-away, mpt https://launchpad.ubuntu.com/products/rosetta/+bugs [12:59] "Je vous laisse le choix dans la date" [01:00] (which is _not_ an idiomatic phrasing) [01:00] BjornT, bradb-away, mpt, the table title is missing a column so the title is not related with the content... [01:00] daf, https://launchpad.ubuntu.com/malone/bugs/1276 [01:01] ddaa: oh, so you mean it is to talk like fabionne always does ? [01:01] I was unaware that anything fabbione said had sexual undertones... but then I should have expected... he's italian... [01:01] s/anything/everything/ [01:02] "Je vous laisse le _ch_oix dans la _d_ate" [01:02] lovely === Kinnison gets awful mental imagines of fingering cats [01:03] well, the last word of the permutated sentence does have the dual meaning of "pussy" in english. [01:05] i thought it was suggestive enough without the spoonerism [01:07] mmm date pudding [01:08] So are various of the lists down, or is launchpad just not mailing out, or what? [01:09] Just found one I can imagine Kinnison saying with his lovely accent :) [01:09] "Taisez vous en bas" [01:09] the only mal i have experienced right now is that pqm seems halted again [01:09] "en tas" means what? [01:09] a cup? [01:09] in the cup? [01:09] in a heap [01:09] same place [01:10] cup == tasse [01:10] same place? [01:10] that is hung [01:10] which code? [01:10] launchpad webapp ? [01:10] stubs testrunner work would fix that [01:10] cscvs? [01:10] pqm 8283 3.8 1.7 105208 65544 ? S 11:24 1:43 \_ /usr/bin/python2.4 test.py -vv --dir hct --dir sourcerer [01:10] pqm 9224 0.0 0.0 0 0 ? Z 11:26 0:00 \_ [rev] [01:10] nope, something that shells out to 'rev' [01:11] I'm guessing sourcerer [01:11] what's rev? [01:11] "Les laborieuses populations du Cap" (morgs?) [01:11] I've seen that happen with the hct/sourcerer tests [01:11] it reverses the lines in a file [01:11] like tac, but in the other dimension [01:11] sabdfl: I've converted those two topics you put on the wiki into skeleton specifications [01:12] can some people please go to http://launchpad.ubuntu.com and tell me what URL you get redirected to [01:13] https://launchpad.ubuntu.com/ [01:13] thanks [01:14] Yep, you get a 301 to the https url daf gave [01:14] X-Pad: avoid browser bug [01:14] that's a good header [01:15] yummy [01:16] darn === Kinnison looks at his bank account and cries [01:17] 45 minutes to meeting, n'est pas? [01:17] i always seem to try "edit bug details" to close a bug [01:17] daf: ouai [01:17] daf: and unless I'm mistaken, you meant "n'est-ce pas?" [01:17] murky buckets, mon sewer [01:17] dee reen === ddaa pulls his hair [01:18] ddaa: quoi? [01:18] what's this murky bucket thing about, btw? [01:19] ddaa: "merci beaucoup" --> "murky buckets" [01:19] ooooow... === Kinnison grins [01:19] SteveA: did your X-Test-Authorization thing get merged? [01:19] SteveA: can't remember the exact header name [01:20] SteveA: the one that was for replacing the Authorization header in page tests [01:20] daf: not yet [01:20] ddaa: You should hear me say "je comprends tout ce-que vous dites" (or whatever the exact french is) in a brummy accent [01:20] bonnejour monnesieur, where is the tchampslizz? [01:21] (no need for an hyphen in that sentence) [01:21] SteveA: ok [01:21] looks like nothing of mine is getting merged at the moment :-( [01:21] ddaa: I heard someone say, in an incredibly thick birmingham accent "gee com-prends towt suckwee vows dye-ts" (approx) [01:23] oh, yeah, now that you spell it that way, that sounds like a typical sentence :) [01:25] folks, /msg me items for the agenda [01:26] morgs: ok, I think I have a page test for +newproduct [01:26] daf: ok === cprov [~cprov@200-171-140-32.dsl.telesp.net.br] has joined #launchpad === Kinnison goes to find food before the meeting [01:30] hi there someone knows what is causing this failure ? [01:30] File "/home/cprov/canonical/launchpad/launchpad/lib/canonical/launchpad/ftests/../doc/message.txt", line 311, in message.txt [01:30] Failed example: [01:30] msg.datecreated [01:30] Differences (ndiff with -expected +actual): [01:30] - datetime.datetime(2005, 6, 17, 9, 45, 13, tzinfo=) [01:30] ? ^ [01:30] + datetime.datetime(2005, 6, 17, 13, 45, 13, tzinfo=) [01:30] ? ^^ [01:31] stub and jamesh are the timezone lords [01:31] Merge to thelove@canonical.com/dists--bazaar--1.5: new build (patch-33) [01:31] yay, pqm is back [01:32] With those two ^^ in there, it looks like a diff issue to me [01:32] Merge to thelove@canonical.com/bazaar--devo--1.5: [fixes #1221] "baz show-changeset --diffs" no longer repeats output with added files (patch-26: Matthieu.Moy@imag.fr) [01:32] stub: like, chunkydiff picking the wrong lines? [01:32] cprov: try running with chunkydiff turned off [01:32] or hiding some whitespace in a dodgy fashion. Just looks suspicious. [01:33] there's not much point in using chunkydiff in system doc tests mostly [01:33] SteveA: right, let's see [01:33] just in page tests [01:33] probably need to add some option to explicitly turn it on only for pagetests [01:36] SteveA: btw, gaim damage the output I pasted, de difference is between 9 <-> 13 as you can see. [01:38] SteveA: Probably easiest to hook into that funky '#doctest: +FOO' notation I've seen floating about [01:39] yeah [01:39] it is a standard part of doctest, btw === Keybuk [~scott@descent.netsplit.com] has joined #launchpad === mpt [mpt@210-55-36-156.dialup.xtra.co.nz] has joined #launchpad [01:47] cprov: oh. Try running your tests as 'env TZ=Europe/London python test.py the_failed_test' [01:48] stub: yeah I did ... works === debonzi [~debonzi@200-171-140-32.dsl.telesp.net.br] has joined #launchpad [01:49] cprov: Great. We have a test that only works in London. I'll try and merge my 'run all tests in Calcutta' patch. [01:49] yes [01:49] stub: hehe great [01:49] stub, rosetta-package-po-attach.py needs to the '-q' argument [01:49] stub: I can do that right now [01:49] s/to/too/ [01:50] stub: TZ=Calcutta in the PQM precommit line [01:50] lifeless: If you can do it in PQM, great. Except that it will stop everyone committing until the dodgy test is fixed ;) [01:50] hi all [01:51] stub: or just do it in Makefile [01:51] lifeless: It is TZ=Asia/Calcutta btw [01:51] Merge to thelove@canonical.com/dists--bazaar--1.5: new build (patch-34) [01:51] hi debonzi [01:51] actually, pqm is better then we won't get tests that only work in Calcutta [01:52] ;) [01:52] Merge to thelove@canonical.com/bazaar--devo--1.5: Explicit error message when unable to connect to an archive. (patch-27: Matthieu.Moy@imag.fr) [01:52] so - tell me when its safe to change ;0 and I'll do so [01:52] stub, that funking teenage (gina) gave up again! :( === salgado [~salgado@200-171-140-32.dsl.telesp.net.br] has joined #launchpad [01:52] debonzi: I beat her up a bit and she is behaving better [01:53] stub, thanks god she got a tutor.. :) [01:54] debonzi: I think she is in shape to run in staging regularly with you fixing any glitches that pop up (there are a few exceptions still, mainly from some paranoia checks I added that might not be needed, but Gina continues). A full warty run has been done on staging and the full set is being run now. [01:55] stub, could you show me the output of: "SELECT COUNT(*) FROM POTemplate WHERE rawimportstatus=2;" and "SELECT COUNT(*) FROM POFile WHERE rawimportstatus=2;" on production? [01:55] If https://staging.ubuntu.com/distros/ubuntu/warty didn't raise a system error, you might even be able to see the results :-/ [01:56] carlos: -q added [01:56] stub, thanks [01:56] daf: morgan.collett@canonical.com/launchpad--project-tests--0--patch-1 for the project bugtracker test [01:56] launchpad_prod=# SELECT COUNT(*) FROM POTemplate WHERE rawimportstatus=2; [01:56] count [01:56] ------- [01:56] 105 [01:57] launchpad_prod=# SELECT COUNT(*) FROM POFile WHERE rawimportstatus=2; [01:57] count [01:57] ------- [01:57] 9401 [01:57] morgs: awesome, thanks! [01:57] cprov: that datetime test failure looked the same as the one kiko mentioned on the launchpad list (I checked in a fix for it). Have you tried merging from rocketfuel? [01:57] that will take a while .... [01:57] stub, ok, thank you [01:58] jamesh: are all the datetime specific tests fixed as far as you know? [01:58] jamesh: not yet , some pending code to arrange, I will within a hour or so ... ok can be now ... [01:58] stub, I got system error :( .. but anyway, unless somebody has changed it, launchpad will not show the gina imported packages because they are published by gina as PENDING, and atm 'soyuz' only shows PUBLISHED packages [01:58] stub: I think so. [01:58] lifeless: Throw the switch! [01:59] stub: you've committed the fix ? [01:59] stub: or its in the queue ;0 [01:59] lifeless: it is already fixed. cprov isn't in sync with rocketfuel [02:00] wtf [02:00] stub: yes, I'm not ... [02:00] daf, https://launchpad.ubuntu.com/distros/ubuntu/breezy/+sources/aegis/+pots/aegis/es/+translate [02:00] MEETING TIME [02:00] daf, did you saw that? [02:00] let's go [02:00] people is adding stupid things to the comments! [02:00] here [02:00] who's present? === carlos here === stub drop his pants [02:00] present [02:00] here === jamesh is here [02:00] here [02:00] I'm here. and kiko just arrived [02:00] i'm here [02:00] ok [02:00] yo === debonzi here === bradb here [02:01] ddaa, bob2, Keybuk ? [02:01] jblack: ? [02:01] _o/ [02:01] ah, jblack you're here [02:01] i need a bot to help with this [02:01] [02:02] Thats twice in a row. ;) [02:02] who else isn't here? [02:02] ddaa: "the appearance of a dragon walking" ?! [02:02] kiko said his irc client is having problems to connect [02:02] I'm not here. I'm merely present. ;) [02:02] 'CJK UNIFIED IDEOGRAPH-9F98' === stub needs a bot to attend IRC meetings [02:02] Keybuk: just have to find an opportunity from time to time to use that one :) [02:03] what does it mean? [02:03] okay, let's move on. === ddaa will look for a unicode dingbat usable as "I'm here" or "pong". [02:03] == Agenda == [02:03] - roll call [02:03] - agenda [02:03] - next meeting [02:03] - activity reports [02:03] - menus (daf) [02:03] - brazil topics [02:03] - where dogfood is going (Kinnison) [02:03] - three sentences [02:03] [02:03] /msg me new agenda items === bob2___ [~rob@crumbs.ertius.org] has joined #launchpad [02:03] next meeting, same time next week? [02:03] sure [02:03] hi bob2 === kiko [~kiko@200-171-140-32.dsl.telesp.net.br] has joined #launchpad [02:03] yup [02:03] hi ho [02:03] sounds good [02:03] for some reason the freenode hub I was connecting to kept timing out [02:03] sorry [02:03] spiv is away on vacation, btw [02:03] ddaa: there are several checkmarks === Kinnison is here [02:04] okay, same time next week [02:04] - activity reports === carlos is up to date [02:04] who's the dude, and who's the little lebowski ? [02:04] I AM THE DUDE === Kinnison is up-to-date (just) [02:04] DUDE === SteveA is up to date, dude [02:04] And a runny nose.... [02:04] dude! [02:04] dude [02:04] DUDE === BjornT is not up to date [02:05] five days behind === mpt cowers === debonzi is up-to-date [02:05] I'm a lebowski, and a little one at that :( [02:05] gtimelog is up to date, and so am I [02:05] up to yesterday [02:05] I'm up [02:05] jamesh: can you send yesterday's now? === daf catches up quickly [02:05] d00D [02:05] mpt: do you have them ready to send? [02:06] SteveA: I sent in yesterdays (haven't done todays) [02:06] jblack: how far behind are you? [02:06] jamesh: great. you're The Dude [02:06] SteveA: Yes, it's just that Ubuntu doesn't like dialup (I'm on IRC via Win98, don't hurt me) [02:06] SteveA: i'll send mine later today [02:06] I may not be present at the meeting next week due to Debconf [02:06] I'll send them when I get home [02:06] and I rock at activity reports [02:06] about a week. [02:06] I'll have them in this morning. === morgs keeps a stack of blank activity reports handy... [02:07] Merge to thelove@canonical.com/dists--bazaar--1.5: new build (patch-35) [02:07] cool [02:07] was that everyone? [02:07] bob2___: ? [02:07] - menus (daf) [02:07] I'm lebowskified [02:08] Merge to thelove@canonical.com/bazaar--devo--1.5: make import/tag/commit raise an error on missing summary (patch-28: daf@muse.19inch.net, Matthieu.Moy@imag.fr) [02:08] bob2___: sort out with lifeless how to get up to date [02:08] will do === daf dude [02:09] so, launchpad now has menus for lots of things [02:09] indeed [02:09] some of them are in production [02:09] and mpt has made them look nice in tabs and second-level tabs [02:09] but the staging server shows them off much better [02:09] at least, on staging [02:10] over the next week or so, we need to start looking at the other "actions" that are in portlets and elsewhere [02:10] and add them into the dubiously named "ExtraApplicationMenu" classes for contexts in launchpad [02:10] once we've seen how this menu gets used, i'll rename it [02:10] wow, staging looks hot [02:10] but for now, it's an "ExtraApplicationMenu" [02:11] https://wiki.launchpad.canonical.com/LaunchpadMenus explains about the different types of menu and how they work [02:12] although I'm responsible for maintaining the menus infrastructure, and fixing bugs, daf's been doing a great job of keeping the docs up to date, and implementing many menus in launchpad. [02:12] stevea: logs up to date. [02:12] I'm happy to try and help anybody who's trying to get menus working for their bit of Launchpad [02:12] haxoring daf! [02:12] there are still a couple of outstanding bugs / niggles in the infrastructure, that i will be addressing very soon [02:12] any questions about menus? [02:13] who's got menus and who's not? [02:14] rosetta: ? [02:14] pending review, yep [02:14] malone: ? [02:14] doap: ? [02:14] foaf: not [02:14] products do [02:14] salgado, suxorer [02:14] malone needs menu love [02:14] doap: not [02:15] guys [02:15] don't let this fall off [02:15] menus are how users will mostly find their way around [02:15] an important part of usability, and easy to implement [02:15] without menus launchpad looks like a page designed by a kindergardner [02:15] I need to do some for calendar [02:15] (gartner?) [02:16] look at rosetta in baz tip and the other apps [02:16] there is a serious shocking impact in the menuless apps [02:16] so get on it === bradb agrees [02:16] mpt, SteveA, what's the plan for the actions portlets? [02:17] jamesh: The most important thing to do is to implement a page saying "x does not currently have a calendar" for those products/distributions/people that don't have one [02:17] do we get rid of them now? [02:17] jamesh: then Calendar can get its own item next to Overview, Bugs, and Translations. [02:17] we could do a special kind of link that is disabled when there is no calendar, for that one [02:17] that is an option [02:18] kiko: App menus make some portlet items redundant, some in the action portlet, some elsewhere [02:18] products have a button for creating a calendar in a portlet [02:18] For example, the portlet on the left side of a person page is *almost* completely redundant now [02:18] having a disabled link sounds better -- saves people the frustration of clicking the link [02:18] and getting a "there is no foo" page [02:18] fair enough [02:19] jamesh: how do i tell if there's a calendar for a particular thing ? [02:19] that means menus infrastructure changes [02:19] no [02:19] just means one special link for calendars [02:19] SteveA: at the moment you can tell because there will be a portlet on the person, project or product page showing the month's calendar [02:19] pretty simple [02:19] mpt, I think you'll need to schedule time to discuss menu changes with /each/ responsible person [02:19] mpt, can you manage nagging and organizing that [02:19] jamesh: i'm thinking of something like ICalendar(object, None) [02:19] SteveA: if there is no calendar and you own the object, the portlet shows a "create calendar" button [02:19] guys, you need to schedule time with mpt between tomorrow and tuesday to talk about your app [02:20] ok [02:20] jamesh: i need to know how to tell from python code. let's talk about this after the meeting / tomorrow [02:20] SteveA: check ICalendarOwner(object).calendar [02:20] jamesh, oh, offtopic: mark wants you in Brazil on the first week. can you arrange? [02:20] jamesh: cool [02:20] mpt: what time will you start working again today (i.e. tomorrow)? [02:20] bradb: in about 13 hours [02:20] sleeper [02:21] mpt: ok, if i'm still around at that time, maybe we can discuss menus [02:21] we need to move on. anything else on menus? [02:21] SteveA, reiterate what I just said please [02:21] nicks in caps get more attention [02:21] WHAT HE SAID [02:21] kiko: I'd have to check if my ticket allows a change like that [02:21] (staging is updated in case anyone was wondering where it went) [02:21] jamesh, don't disappoint the sab [02:21] guys, you need to schedule time with mpt between tomorrow and tuesday to talk about your app [02:21] okay, moving on [02:22] - brazil topics [02:22] https://wiki.launchpad.canonical.com/BrazilTopics has appeared [02:22] you da man [02:22] heh... good luck. Flights to brazil from Australia will be tight. I suspect James will be Business class === Kinnison converted the topics sabdfl put for him into specification pages [02:23] branches look ugly, they were nice when they were bold [02:23] ddaa, where? [02:23] https://staging.ubuntu.com/products/automake [02:23] take note of what is written in BrazilTopics [02:23] especially under your own name [02:23] jamesh: try to work with cvd on it [02:23] I appear to have nothing but hardware stuff. Stuff I know little about and care even less about. [02:24] sabdfl: okay [02:24] sabdfl: the issue about the re-running failing tests will be sorted when we move to holger krekel's test runner. he's a core python guy who has written a very nice test runner over the last 18 months or so. [02:24] stub: maybe we could add db/ispell integration to your list? [02:24] SteveA, sabdfl when are we supposed to start the braindump for those specs? before .br or when we go there? [02:24] stub: those are just a few thoughts i've had over time. we definitely need to put some work into an optimised read-only db configuration [02:25] carlos: having basic stuff done now helps us all be on the same page before brazil. so spend a little time filling in what you know. [02:25] carlos: right away, any time you have a thought about something else you want to cover in brazil, put a note in the wiki [02:25] no matter how small [02:25] our wiki rocks [02:25] except when it logs us out [02:25] have we unpassworded it yet? [02:25] nope [02:25] yes [02:26] we have?! [02:26] stub: gathering some stats from staging will help. [02:26] I was told we had [02:26] yep, works without a password here === kiko tests [02:26] mpt: for editing? [02:26] wow === mpt tries [02:27] Yes, I just created TestAccount [02:27] anyway [02:27] SteveA, ok, it's just that I want to work on bugs and partial implementations before starting with new stuff and I thought next three weeks is a good time to do that, but will mix that with braindumps then... [02:27] sabdfl, ok, will do [02:28] in due course we'll merge the LP wiki and the main wiki [02:28] we just need a way to categorise pages and get recentchanges etc by category [02:28] anything more on brazil right now? [02:29] we should be moving on [02:29] notfromme [02:29] - where dogfood is going (Kinnison) [02:29] Right [02:29] so now we have staging [02:29] and so dogfood is becoming less useful for what it was originally intended [02:30] The intention is thus that the soyuz team will mostly be in charge of dogfood [02:30] We are preparing a buildd network associated with mawson (elmo has provided us with everything we need for that I think) [02:30] and we're essentially going to be shadowing breezy on dogfood [02:31] This does not mean that other people are not allowed to use dogfood, but it's going to be a little more confusing for a short while, during which time the soyuz team will be getting going [02:31] Does anyone other than soyuz have a vested interest in dogfood rather than stagin? [02:31] hct [02:32] though a regularly updated breezy is advantageous, provided you don't drop the database [02:32] Will that remain true as production ramps up with regular breezy imports? [02:32] debonzi, stub, Kinnison: you guys need to get on top of gina, or else we're going to flunk [02:32] I can't see us even starting breezy imports on production for a while yet [02:32] right [02:32] kiko: Last night, stub managed a hoary run on staging and he said he would be doing a proper production run tonight [02:32] Kinnison, the mail I got said otherwise.. [02:32] erm, sorry, a warty run on staging [02:32] SteveA: i'd like to discuss 1.0 announcements, ping me when it's a good time to do that [02:32] sabdfl: ok [02:33] thanks [02:33] Kinnison, I think Rosetta team would be interested to use dogfood from time to time to do some testing with a bigger database [02:33] Erm... full run on *staging* is going on now. I'm not running on Production until someone can confirm it is doing good and not harm. [02:33] stub, kiko, debonzi: Can we discuss gina after the meeting? [02:33] Kinnison, sure [02:33] carlos: that's not so urgent now we can access staging, right? [02:33] k [02:33] Okay, so hct has an interest in dogfood, and rosetta may [02:34] have all the relevant bugs been transferred from dogfood to production malone? [02:34] daf, staging helps to debug looking at data there, but dogfood could be used to change the data [02:34] daf, as staging is read only, we cannot debug all bugs [02:34] well, we run code on mawson, even if we're using the staging db [02:34] Kinnison, AFAIK a long time ago [02:34] carlos: true [02:34] We can duplicate the dogfood database at anytime - it doesn't have to be shared [02:34] we need to move on. can we have the rest of this discussion in a smaller meeting later? [02:34] Kinnison: not intentionally, but probably most of them have be refiled in prod otherwise. [02:34] daf, I'm just noting that we could be interested at some point to play with dogfood but that will not be somethig we will do all time [02:35] So if soyus and hct don't want to play together it is an easy fix. [02:35] Okay thanks [02:35] I just wanted to make sure people knew about what was happening with dogfood [02:35] we should have a wiki page that explains what staging, dogfood and production are used for [02:36] and the expectations of the databases there -- like when they get nuked etc. [02:36] That would be useful [02:36] Kinnison / stub: can you make sure we have such a page? [02:36] SteveA: Okay [02:36] stub: if you do prod/staging I'll fill in dogfood when you're done [02:36] sabdfl: 1.0 announcements [02:36] StagingServer exists [02:38] sabdfl: ping [02:38] ok, we want to have rolling 1.0 announcements [02:38] over the next couple of weeks === SteveA tried a phone ping, but the london mobile phone network is having problems :( [02:38] SteveA: it's shut down in places, deliberately i think [02:38] i'd like to see at least one announcement of some sort per week [02:39] also, i think we should get better about communicating landed features to the list [02:39] for each spec, could we make sure the summary is clear and can be copied into a weekly mail to the list of "things that just landed" [02:39] so, back to the 1.0 announcements [02:39] daf, carlos: could you announce that Rosetta 1.0 will happen next week [02:39] kiko: (weekly planned rollouts) [02:39] (btw, did you like your present this morning)? [02:40] heh [02:40] at least, as long as you are confident you can hit the goal next week [02:40] i think rosetta is looking excellent [02:40] jordi will shortly start working with you to facilitate bringing upstreams into rosetta [02:40] well [02:40] so we will need to spend some time putting together wiki and web pages that tell how to use it most effectively [02:41] sabdfl, I was not able to see your recent changes, sorry [02:41] sorry, lost network for a second there [02:41] so, the announcements come in two phases [02:41] I take the hit for not producing the report for this month earlier, sorry [02:41] kiko: let's make it weekly, just a summary of all the landings and their features [02:41] it should be a wiki page people can cut and past to when then land a spec-branch [02:41] sabdfl, about goals, I think the only big features not implemented are Karma and GNOME integration. [02:41] the first phase is a "LP XXX 1.0 will happen in one/two weeks" === SteveA notes that this meeting will be overrunning its allotted time. [02:42] i';d like that to happen for rosetta today or tomorrow [02:42] in the email, list all the exciting new stuff that's happened since the last announcement [02:42] Karma should be ready this week to be reviewed, so it will reach production in two weeks..... [02:42] go ahead without karma [02:42] I'm going to produce a report today [02:42] then, on the day, we'll put out the formal announcement everywhere [02:42] I need to anyway because Jane needs it [02:43] malone i think will miss by about a month [02:43] soyuz, i don't know [02:43] but I've had it in the cooker for a while [02:43] then it's a matter of wait for daf's work with menus is merged, daf? [02:43] soyuz still has major stuff in code review [02:43] we will also do announcements for the calendar system (yay james!) and the bounty system (yay me) [02:43] and gina hasn't run yet (AFAICT) [02:43] but i want you guys to think in terms of the two-phase announcements, starting with rosetta today [02:44] ok [02:44] that's all from me [02:44] any questions ? [02:44] daf, carlos: ping me with your draft announcement when its on the wiki [02:44] sounds good to me [02:44] sabdfl: sure [02:44] daf: you've got a knack for good writing [02:44] ok [02:44] use it :-) [02:44] sabdfl: sounds good, thanks [02:44] carlos: the menus stuff should be merged this week [02:44] mpt: the menus are not looking the way i imagined them to be, could we chat briefly after this meeting please? [02:44] daf, will it be moved into production on Tuesday? [02:45] we'll be overrunning by just 5 minutes. let's do the three sentences. [02:45] stub: are we branching for next production later today? [02:45] drum roll [02:45] write your sentences onto the channel now. [02:45] DONE: Gina [02:45] carlos: if it gets reviewed and merged bfore the production branch point, sure [02:45] WORKED ON: malone cleanups, linkification, meta-reviews, catching up [02:45] with application status, linting, sprint prep, etc [02:45] TODO: LP report (today, in particular because Jane needs it), [02:45] BrazilTopics, finalize sprint prep [02:45] BLOCKED ON: SteveA for linkification [02:45] DONE: BasicVoting, code review, sqlobject/launchpad random fixes. [02:45] TODO: More code review than I did last week, land the first round of BasicVoting and start the second one(which will hopefully be the last). [02:45] BLOCKED: No [02:45] DONE: Rosetta menus, get multiple format exports in production, pyflakes harness, import and other cleanups, write tests [02:45] DONE: More buildd stuff now that elmo has given us the dogfood buildd machines. Refinements to the buildd packaging. Uploader work. Assisting cprov. Assisted stub in getting production ready for gina. [02:45] DONE: RDF bugfixing [02:45] TODO: DoapSchemaNG, menus [02:45] BLOCKED: no [02:45] TODO: clean up traveral code, fix template export bug [02:45] DONE: Imports. talked lots about doap ftp details with keybuk. Taught productseries and ftpglob zen to jblack. [02:45] TODO: More imports. [02:45] BLOCKER: No time for anything else non-trivial. [02:45] TODO: Uploader, get dogfood buildds ready to roll and rolling. Get publisher going on production appserver to ensure distros stay clean once gina is running regularly. [02:45] BLOCKED: no [02:45] DONE: Launchpad general bits; more gina related work. [02:45] TODO: Some launchpad cleanup and bug fixes. [02:45] BLOCKED: None. [02:45] DONE: decruftification, BasicVoting design, lots of bug reporting [02:45] TODO: specific menus, spec backlog, HierarchyNavigation [02:45] BLOCKED: no [02:45] BLOCKED: Cprov's branch being merged (hopefully should happen real-soon-now), Gina runs on production completing. [02:45] DONE: land CalendarMerge branch, code reviews, some datetime bug fixing, follow-on calendar work [02:45] TODO: code reviews, follow-on calendar work [02:45] BLOCKED: no [02:45] DONE: sqlobject/sqlos spelunking, fascism, menus, reviews [02:45] DONE: finished up and submitted BasicBugAttachments for review. some [02:45] reviews. didn't feel entirely well this week. [02:45] TODO: general malone work, clean up, go through bug reports. improve [02:45] the email interface. [02:45] TODO: menus bugfixes, sqlobject/sqlos fixes, reviews [02:45] BLOCKED: no [02:45] BLOCKED: spec review from kiko [02:46] DONE: landed FormattingBugNotifications. lots of baz/pqm pain. landed some other bugfixes and cleanups. (e.g. dissolved BugMessageFactory, split the task page in two, fixed header bug a few days ago that carlos mentioned earlier, etc.) [02:46] DONE: dyson installed on casey [02:46] oh: TODO: Debconf [02:46] TODO: sourcerer debugging, dyson debugging, hct bug fixes [02:46] BLOCKERS: only one of me [02:46] DONE: entry [02:46] TODO: menus. fixing any bugs in FBN. MaloneFrontPages work. double-check that we've done all that's needed for 1.0 karma. any other random bits and bobs. [02:46] DONE: applyed GPG-NG review and improves assisted 2nd round of Buildd review, both assisted by jamesh [02:46] TODO: entry [02:46] TODO: Concentrate myself on buildd (I know I'm blosking dsilvers) [02:46] BLOCKED: stub review on DB patch for GPG and buildd review and test suite (myself) [02:46] BLOCKED: nothing. [02:46] DONE: bug #1036, POTemplateAdmin spec implementation changes, several bugs fixes, languagepacks [02:46] DONE: make-changeset is sane, TODO: apply-changeset saneification, BLOCKED: days too short [02:46] BLOCKED: entry [02:46] TODO: Karma implementation for Rosetta, more languagepacks fine tunning, GNOME imports [02:46] BLOCKED: GNOME imports are blocked until l10n-status.gnome.org new version is ready (I hope it will be ready this weekend) [02:46] TODO: Gina and bugfixes [02:46] BLOCKED: Nothing [02:46] DONE: yellow fever, imports. TODO: LOTS of imports [02:46] lifeless: please put sentences on separate lines. helps me grep [02:46] BjornT, that's BBA? [02:46] BLOCKED: None now [02:46] kiko: yes [02:46] sabdfl: I was going to tag from just now. You have bugfixes to land or features? [02:46] SteveA: ok [02:46] sorry. [02:47] kiko: i'm working on the linkification [02:47] thanks [02:47] should land today [02:47] carlos: shall we have a Rosetta meeting now? [02:47] SteveA, I'm going to verify jamesh' claim [02:47] stub: RF is golden for me right now [02:47] claim? [02:47] staging doesn't yet have everything [02:47] he reported a potential problem, SteveA [02:47] with... [02:47] so i wanted to confirm you will tag later today from RF [02:47] with linking? [02:47] daf, we should finish this meeting first... [02:47] yeah, i saw [02:47] sabdfl: Staging was updated a few minutes ago from head (during the meeting) [02:47] i'll take care of it [02:47] daf, and have lunch, please... [02:48] SteveA, yes [02:48] carlos: ok [02:48] Kinnison: do you see any problems with gina runs on production? [02:48] SteveA: The staging run of warty seemed good [02:48] Kinnison, what about stub's message? [02:48] SteveA: There are soyuz UI issues preventing me from interrogating via the staging UI [02:48] I seem to have missed the boat here [02:48] stub: hmm... system error on https://staging.ubuntu.com/distros/ubuntu/hoary/+translations [02:48] cprov: does stub know about your database patch? [02:49] SteveA: yes, just talked 10 min ago [02:49] sabdfl: https://staging.ubuntu.com/errors/showEntry.html?id=1120740503.640.503065040417 [02:49] ok [02:49] SteveA: somebody moved the arch branch back into the right column, that fucks up the page layout badly (as you can see on the staging automake product). Can you ask the appropriate person to get it fixed? [02:49] All my reviews are done. I'm not aware of blocking anyone atm. [02:49] ddaa: i need more context. a url, for example [02:49] https://staging.ubuntu.com/products/automake [02:49] ddaa, I was going to say the same, I saw your complaint earlier but had no idea what you were talking about [02:50] debonzi: can I see you on ##soyuz1.0 please to discuss the broken soyuz UI? [02:50] kiko: actually my complaint was about still something else [02:50] ddaa: for such things, filing a bug might be useful - it will get you to think abot what details are needed for folk looking at it ;0 [02:50] Kinnison, am on the way [02:50] daf: it's not the translations page, it's the general distro pages [02:50] debonzi: thank you [02:50] ddaa: i'm clueless. please file a bug. [02:50] it's the details-portlet [02:50] sabdfl: there is a problem with soyuz on staging [02:50] Any other BLOCKED issues that haven't been dealt with? [02:51] sabdfl: I'm about to go through it with debonzi and cprov to see if we can work out the fix [02:51] lifeless: sure, but the fact it has been right for a while, then broken again, suggest there might be an issue worth talking about in this post-meeting. [02:51] let's end the meeting [02:51] 4 [02:51] 3 [02:51] one other thing [02:51] ... [02:51] while lifeless is still here [02:51] bradb: is this important for the meeting? [02:51] for everyone? [02:51] SteveA, mine is an external dependency [02:51] 2 [02:52] ok, no [02:52] 1 [02:52] 0 [02:52] MEETING ENDS [02:52] yay [02:52] yay, I snuck in a ZERO [02:52] lifeless: would it be reasonable for pqm to send us a confirmation message when it's received our merge request? [02:52] 08:52 < SteveA> MEETING ENDS [02:52] 08:52 < lifeless> 0 [02:52] i suck for not having written up last week's meeting [02:52] bradb: 22:52 < lifeless> 0 [02:52] 22:52 < SteveA> MEETING ENDS [02:52] bah [02:52] no bradb there === carlos -> lunch [02:53] bradb: I don't think so, AFAICT we've never ever had a dropped-receipt problem with pqm [02:53] see you later [02:53] bradb: if you want, I'll whip something up for just your email address, but it would bug the heck out of me. [02:53] BjornT, has the schema for BBA stabilized? I remember SteveA said something about mark considering linking attachments to messages... [02:53] lifeless: have anyone ever asked you if pqm has gotten their message? [02:53] s/have/has/ [02:53] stub: pqm's been very slow about merging stuff i submitted hours and hours ago. [02:54] SteveA: its hanging during the test runner [02:54] stub: so, maybe you can delay the branching for production until we get this sorted out? [02:54] SteveA: thats making it very slow as it waits for me to intervene [02:54] kiko: mark has edited the spec regarding the schema, so i consider it stable now [02:54] would anyone else want a confirmation message from pqm to let them know that their request has actually been received? [02:54] okay BjornT [02:54] SteveA: Sure. I'm not the one who was arguing for a fixed schedule ;) [02:54] will look at it [02:54] lifeless: it is stopping on a particular test? [02:54] SteveA: don't know [02:54] kiko: thanks [02:55] pqm 8283 3.8 1.7 105208 65544 ? S 11:24 1:43 \_ /usr/bin/python2.4 test.py -vv --dir hct --dir sourcerer [02:55] pqm 9224 0.0 0.0 0 0 ? Z 11:26 0:00 \_ [rev] [02:55] p [02:55] lifeless: i have a merge into sqlobject that will stop some potential deadlocks in launchpad. [02:55] SteveA: I'll choose a branch tomorrow (attempting to avoid new features if possible) [02:55] New Malone bug 1279 filed on product Registry by David Allouche: Bazaar branch appears on the side in staging [02:55] https://launchpad.ubuntu.com/malone/bugs/1279 [02:55] doesn't look like it though [02:55] stub: kiko very much wants to get linkification landed [02:55] something is running 'rev' - and there is no mention of it in the lp codebase [02:55] what is 'rev' ? [02:55] a unix tool [02:55] (man rev) [02:55] chmod og-rwx `which rev` [02:56] make check_on_merge [02:57] sabdfl: sorry, lost network again [02:57] ./lib/canonical/launchpad/tests/test_helpers.py: >>> print helpers.simple_popen2('rev', 'ooF\nraB\nzaB\n') [02:57] well there you go. Keybuk your grep skillz suck [02:57] sabdfl: the error report suggests it's something to do with the distrorelease details portlet [02:57] test_simple_popen2 [02:57] oh, bah, I didn't think to look for 'rev [02:58] I did look for "rev [02:58] daf: yes, i think it is likely to do with the lack of some data, that's in the sampledata [02:58] i don't think it's related to anything i've done [02:58] I think its likely that that test is giving us the grief [02:58] simple_popen2 isn't used [02:58] sabdfl: Expression: ? [02:58] we should just remove it [02:58] and its test [02:58] daf, carlos: this was your present: http://localhost:8086/distros/ubuntu/hoary/+translations [03:00] lifeless: are you removing simple_popen2 ? or should i do that? [03:01] SteveA: I'm going to bed if I have a choice [03:01] last night was way late. [03:01] sabdfl: just to confirm for when I go to the travel agent, you only want me to move my outgoing flight forward; not my return? [03:01] lifeless: can you jimmy in a commit to RF [03:01] that will make test_simple_popen2 a no-op [03:01] what would you like me to jimmy [03:01] def test_simple_popen2(): [03:01] r""" [03:01] >>> print helpers.simple_popen2('rev', 'ooF\nraB\nzaB\n') [03:01] Foo [03:01] Bar [03:01] Baz [03:01] [03:01] """ [03:02] change to [03:02] def test_simple_popen2(): [03:02] r""" [03:02] """ [03:02] that will allow pqm to continue functioning [03:02] so i can make a proper removal [03:03] it may still hang on cscvs, but that is nowhere near as common as this one is [03:03] file is lib/canonical/launchpad/tests/test_helpers.py [03:03] yup [03:05] why is cscvs hanging? [03:08] commit forced [03:09] race condition with the local run-up of a cvs server where sometimes it connects too early and then the server process stays in the process group [03:09] Merge to rocketfuel@canonical.com/launchpad--devel--0: disable simple_popen2 blame steve if this wedges anything (patch-2047) [03:10] sabdfl: oh, *nice*! [03:10] daf: thank you :-) [03:10] wanted to get that in before 1.0 [03:10] rocking [03:10] we need more language-oriented views [03:10] I saw you put it up for Brazil [03:10] yup [03:10] this is a great start [03:12] SteveA: so, I've gotten rid of all DB imports from the browser code, except for traversers.py [03:12] daf: waesome [03:12] um, awesome [03:12] traversers.py looks like the hardest of the lot [03:12] the malone folks need to clean up traversers.py [03:12] yes [03:12] as there's some crappy creapy stuff in there [03:12] plus there's some registry stuff in there as well [03:12] to do with mutable Sets / Subsets [03:12] daf, carlos, sabdfl: is the problem with broken/regressed translations the same as https://launchpad.ubuntu.com/malone/bugs/1155 [03:12] -- ? [03:12] which we could certainly do without [03:13] do you have any ideas for a better pattern? [03:13] it seems this sort of thing crops up fairly regularly [03:13] bradb-brb: talk to mark about how he wants to get rid of the mutable set / subset of things concepts, and replace them with using two-step URL traversal for dividing up the url space. [03:13] kiko: i don't think so [03:14] SteveA: the only issue i've found with that is the breadcrumbs still have the "+lang" stubs clickable === faassen [~faassen@a80-127-80-154.adsl.xs4all.nl] has joined #launchpad [03:15] sabdfl: i'll be fixing that [03:15] i think your approach is sound [03:16] hello faassen. can i interest you in localizing silva using rosetta ? [03:17] SteveA: i'm shifting all the Dummy's into the database class with which they are associated [03:17] hey faassen [03:17] how will they be imported into browser code? [03:17] sabdfl: just to confirm for when I go to the travel agent, you only want me to move my outgoing flight forward; not my return? [03:17] through an api on the xxxSet ? [03:17] sabdfl: ^^^ [03:17] hey. [03:17] jamesh: when are you due to travel back, currently? [03:18] SteveA: yes [03:18] SteveA: not at this time, but I'll check it out later. :) [03:18] sabdfl: okay. [03:18] sabdfl: my ticket is currently to fly out on the 24th and back on the 2nd [03:18] sabdfl: (i.e. the second week) [03:19] as soon as the malone guys fix up the imports in traversers.py, i'll turn on exceptions in the fascist, so that there will be no new imports of database code into browser code [03:19] i'll be moving all of the traversal around too, pretty soon. [03:19] i'd quite like the malone traversal fixed up before i do that === SteveA --> lunch [03:20] there's also ProductSeriesSet and ProductMilestoneSet [03:20] morgs: ^^^ [03:20] looks like PublishedPackageSet can just use getUtility [03:21] daf: sorry, more context please? I haven't been following... [03:22] morgs: sure [03:22] we're trying to get rid of all imports of database code into the browser code [03:22] the final occurrences of this are in traversers.py [03:23] traversers.py imports ProductSeriesSet and ProductMilestoneSet from the database [03:24] do you follow me so far? [03:24] OK [03:24] now these things are named *Set [03:25] but our policy is that *Set things have no state [03:25] and represent a set of all of something [03:25] e.g. ProductSet is all products [03:25] ProductSeriesSet, though, is instantiatied with a product [03:26] which violates this rule [03:26] and means it can't be used with getUtility [03:26] so we need to figure out how to refactor this [03:27] OK, other stuff uses interfaces here, like IBugTaskSubset(product) - is that what we want to move towards? [03:27] right, yes [03:28] IBugTaskSubset is an adapter === kiko [~kiko@200-171-140-32.dsl.telesp.net.br] has joined #launchpad [03:28] so we could have IProductSeriesSet(product) [03:28] hey bradb-brb [03:28] I think Steve believes there's an even better way of doing it which doesn't involve adapters [03:28] but I'm not sure what that is [03:29] we want to move away from SubSet's [03:29] the idea is traverse straight from Product to ProductSeries [03:29] the traverse says: [03:30] - i'm a product, [03:30] - ok, i see there is a '+series', let me look ahead.. [03:30] - ok, i see there is a 'main' series, and i have one, return that, and tell zope to jump traversal to that point [03:31] cool [03:31] browser/traverses.py line 96, for example [03:31] do we have an example for this? [03:31] aha [03:33] kiko, yes, I think it's related. That bug was because a .po file was imported and it failed so all translations imported before the break were stored instead of rollback the transaction [03:34] carlos: please could you and daf do an analysis of all rosetta scripts to make sure they properly use transactions? [03:35] sabdfl, sure [03:35] sabdfl: that's very neat [03:35] carlos: hmm, nothing should be stored in the database unless the transaction is being committed somewhere in the script [03:35] daf, I know, but it's happening so something is broken [03:36] sabdfl: how do index pages (e.g. /products/foo/+series) fit in with this? [03:37] daf: they should generally have a different name [03:37] for example: [03:37] /products/foo/+pot/potname => the potemplate [03:37] /products/foo/+potemplates => a page showing the po templates [03:37] right, that makes sense [03:38] that should let us rid of POTemplateSubset [03:38] sabdfl, isn't +potemplates our current +translations ? [03:39] carlos: it might be, but the point i was making for daf is that the +foo that is a traversal to foo's is different to the page which lists all the foo's [03:39] New Malone bug 1281 filed on product FOAF by Guilherme Salgado: Need some way to merge accounts without emailaddress or with only dead ones [03:39] https://launchpad.ubuntu.com/malone/bugs/1281 [03:39] ok [03:41] debonzi: around? [03:41] daf, here [03:41] sabdfl: it is possible to support both a +foo page, and +foo/thing traversal, but a bit more complicated. let me know if you want to have support for that kind of url scheme. [03:41] debonzi: https://chinstrap.ubuntu.com/~dsilvers/paste/fileF76lFd.html [03:42] SteveA: ok, thanks [03:42] debonzi: smells like untested code :) [03:42] daf, checking [03:42] daf, pyflakes ruling the world. ;) [03:43] sabdfl: i expect to do serious work on the traversal / breadcrumbs / urls simplification starting tomorrow afternoon. i would be easiest to add support for this at that time, although it can be added later too. === SteveA really goes to lunch [03:43] SteveA: i was going to do a quick-and-dirty list of non-clickable breadcrumbs in webapp/tales.py [03:43] i'll defer to a better idea from you :-) [03:44] i have no better idea for today's rollout. [03:45] i can do better for next week's === SteveA goes OUT to lunch ;-) [03:48] hey kiko, what's up? [03:48] stub: [03:48] psycopg.ProgrammingError: ERROR: permission denied for sequence libraryfilecontent_id_seq [03:48] didn't we fix that one yesterday? [03:49] Keybuk: I never applied your security.cfg updates. Did they get through pqm? [03:49] not yet [03:49] pqm hates everyone right now [03:49] it's being goth [03:49] I'll pull em from the branch [03:50] sabdfl: did you make changes to the productseries and release pages? They look nicer but there are no clickable links to the releases? === daf [daf@muse.19inch.net] has joined #launchpad === daf_ [daf@muse.19inch.net] has joined #launchpad [03:53] debonzi: sorry, fell off for a bit there [03:53] sabdfl, did you implemented the language selector for the suggestions? [03:53] stub, do you have one or two minutes to talk about the name change widget? [03:54] carlos: yes, it shows an alternative language suggestion [03:54] sabdfl, it's broken here... [03:54] sabdfl, I just selected Spanish with evolution in Hoary [03:54] and got a trace [03:54] sabdfl, http://www.usatoday.com/news/world/2005-07-07-london-claim_x.htm [03:54] grr [03:54] daf, its already removed on my tree... I just didn't merge yet [03:54] fucked xchat [03:55] sabdfl, https://chinstrap.ubuntu.com/~dsilvers/paste/fileANjzhf.html [03:55] debonzi: great [03:55] daf, thanks :) [03:55] sabdfl, also, I think we should add by default the language you are translating [03:55] s/add/select/ === bradb [~bradb@modemcable082.64-130-66.mc.videotron.ca] has joined #launchpad [03:56] carlos: i think that will confuse people [03:56] because they have already selected the "language" [03:56] if they see the selector with that language, they will think they can switch languages just by changing that selector [03:56] when in fact, that selector is just for more advanced use cases [03:56] the default is "no alternative language", not "the current language" [03:57] is pqm hanging? i sent a message where i accidentally left off the [trivial] , and resent again with the [trivial] , but have heard nothing back at all [03:57] sabdfl, I'm getting Spanish suggestions by default from: http://localhost:8086/distros/ubuntu/hoary/+sources/evolution/+pots/evolution-2.2/es/+translate [03:57] sabdfl, but the language selector is empty [03:57] I think that's confusing.... [03:57] its less confusing than making it look like the primary language selector [03:58] None is None, not "current" [03:58] which alt language is selected by default? None [03:58] i don't mind if you make it say "No alt language selected" [03:58] sabdfl, so you should not get suggestions at all, right? [03:58] sabdfl, my point is that I'm getting suggestions [03:58] carlos: you should suggestions normally [03:59] carlos: url? [03:59] http://localhost:8086/distros/ubuntu/hoary/+sources/evolution/+pots/evolution-2.2/es/+translate [03:59] I'm with latest rocketfuel version [03:59] sabdfl, I get suggestions, currently published elsewhere and from the global translation Wiki [04:00] with the default view [04:02] sabdfl, btw, shouldn't we add a $POFile.language/+admin link to the actions portlet? [04:02] that only appears if you have rights to use that page [04:02] carlos: yes [04:02] folks, there appears to be some confusion around the use of hidden links, based on permissions [04:02] New Malone bug 1283 filed on product The Launchpad by Guilherme Salgado: Need a better API for sending mail [04:02] https://launchpad.ubuntu.com/malone/bugs/1283 [04:02] sabdfl: did you make changes to the productseries and release pages? They look nicer but there are no clickable links to the releases? [04:02] please only hide launchpad.Admin links [04:03] morgs: yes, i did touch those pages a lot, might have messed up a link [04:04] sabdfl: the series releases portlet has no links [04:04] and the product page lists the releases in a table but again, no links [04:04] sabdfl, I hide too launchpad.Edit links because you will get an error page if you follow it [04:04] Keybuk: Permissions should be sorted (until the next update if pqm hasn't taken your patch) [04:04] carlos: i think you will just be asked to log in, right? [04:04] sabdfl, if you don't have the rights [04:05] you will see a page that says you don't have rights [04:05] malone user testing session at jbailey's place tomorrow, w00t! [04:05] carlos: that's perfectly reasonable [04:05] sabdfl, ok [04:06] https://staging.ubuntu.com/products/automake [04:06] ^ sweet, sexy releases [04:06] daf, The changes you did to implement the scroll bars for language selection and pofile downloads are broken (or firefox is not rendering them correctly from time to time) [04:06] daf, It's missing the vertical scroll bar [04:06] yeah, I've seen it too [04:06] daf, until I reload the page again [04:06] I think it's a Gecko bug [04:07] it doesn't trat max-height properly or something [04:07] I didn't know bob2 was such a hardcore gnu toolchain dude [04:07] elmo: he's the man [04:07] daf, I think the problem is that the size is calculated before you get the whole page [04:07] daf, so it thinks the scroll bar is not needed [04:07] Keybuk: remind me, is the ordering of releases significant? [04:07] New Malone bug 1284 filed on product Registry by Morgan Collett: [staging] No links to releases from product page [04:07] https://launchpad.ubuntu.com/malone/bugs/1284 [04:07] carlos: it's a bug [04:08] ddaa: yes, no, and thursday [04:08] daf, also, could we remove the horizontal scroll bar? [04:08] carlos: it's not supposed to be there :P [04:08] Keybuk: yes it, no we don't care, and thursday you'll play rami with your grandmother? [04:08] daf, but it is ;-) [04:09] elmo: or, if you prefer, I haven't a clue what to put in the owner field so just copy the field from the parent product ;) [04:09] maybe we can do something to get rid of it [04:09] ddaa: today is thursday [04:09] /unset keybuk riddlemode [04:09] New Malone bug 1285 filed on product Registry by Morgan Collett: [staging] No links to releases from productseries portlet [04:09] https://launchpad.ubuntu.com/malone/bugs/1285 [04:09] actually, those all happen to be in the right order [04:10] daf, the main problem is that it's in production already and people will get confused.... [04:10] 1.4 preleases? [04:10] daf, btw, meeting time? [04:10] 1.4-p2 was the second PATCH release of 1.4 [04:10] ha... [04:10] rather than a pre release [04:10] lucky :P [04:10] I think launchpad uses dpkg-style sorting [04:10] somebody apparently told the maintainer about the strange modern notion of "dot releases"... [04:11] it's a GNU project [04:11] the maintainer changed to someone who had a totally different way of doing things [04:11] did I tell you that story with my favourite gnu maintainer... he kept applying patches incorrectly, for months, until one day he asked me if I knew an automated tool to apply patches.... [04:12] "yes, there's this tool call patch" "oh, thank you!" === Keybuk reaches for the screen-cleaner === daf snorts [04:12] so you know that entire discussion we were having about dyson, and version sorting? [04:12] and how I didn't do it because I was lazy [04:12] well, turns out my instinct for apathy was correct [04:13] because it would have been pointless anyway, as launchpad sorts stuff itself [04:13] I vaguely remember it matters for how release branches (at the bazaar-sourcerer level) relate to one another... but if the iterator you are using does some smart sorting for sourcerer... [04:14] I'm not very clear about that bit though, that's why I ask you to remind me, maybe I misunderstood. [04:14] yeah, the iterator sourcerer uses (SourcePackageSomethingUtililtyCruetSet or something) has an in-built version sort [04:15] mmm Cruet [04:15] not sure how smart it is exactly... but I guess the answer is we do not care... you know how creative upstream can be with tarball names. [04:20] is it TeX that has a version number that approaches ? [04:21] 3 -> 3.1 -> 3.14 -> 3.141 etc. [04:21] I liked that one [04:22] daf, https://launchpad.ubuntu.com/distros/ubuntu/hoary/+sources/gnome-panel/+pots/gnome-panel-2.0/es/+translate?offset=42&count=1 [04:23] urgh [04:23] that invalidates our theory that uses an entry box if the msgid has only one line [04:23] looks like the batching is screwed [04:23] daf, ? [04:23] oh, no [04:23] sorry [04:23] daf, ;-) [04:23] yeah [04:24] we have a bug open about that [04:24] special handling of translator-credits [04:24] not really, the bug about translator-credits is related to autofill it [04:24] hmm [04:24] atm this bug prevents people to add their name to it... [04:25] it's related but different [04:25] Keybuk: I think tex approaches e, actually. [04:25] no [04:25] Right, tex approaches pi, metafont approaches e [04:25] This is TeX, Version 3.14159 [04:26] This is METAFONT, Version 2.7182 [04:26] daf, why? [04:26] shouldn't that technically be TeX version 3.14160 ? :p [04:26] and the next version be 3.141597 [04:26] carlos: yes [04:26] and then 3.1415965 [04:26] now _that_ would be hard to sort [04:26] daf, you are confusing me ;-) [04:27] apparently Don has some residual sanity. [04:28] carlos: I'm going to have food [04:28] carlos: let's have a meeting when I get back [04:28] daf, ok [04:28] ddaa: what screen resolution are you running? [04:28] 1600x1200 [04:29] ddaa: and you still find the bazaar branch portlet too big? ;-) [04:29] and 800x600 when I'm tired [04:29] stub, would you apply the patchset I sent you yesterday into production, please? [04:29] Aah... [04:29] Well, the problem is that the archive names get long - I'll look into that. [04:29] ddaa: or no sense of fun [04:30] Seems everything's moving into portlets these days... [04:30] besides, I like to have many things on screen, at least the browser, the chat client, and gtimelog [04:30] carlos: There was a conflict that will need resolving. I didn't look to see how much work would be involved but I can do that tomorrow. If you want, branch off launchpad--production--1.24 and resolve it and I'll cherry pick that instead. === morgs prefers to maximise apps, but hey! [04:31] Keybuk: I do not find it fun to have the evil of the arch namespace stab me repeatedly in the eye. I know it's bad, I do not need any reminder :( [04:31] stub, a conflict with my patchset? [04:31] dear pqm, please work, kthxbye [04:31] stub, so I branch from that production branch, merge my patchset, resolve the conflict, commit and ping you? [04:32] email me - I'm going to bed ;) [04:32] ok [04:32] stub, good night ;-) [04:34] Keybuk: symbolic names: [04:34] foomatic-3_0-branch: 3.1.0.2 [04:34] stub, just in case you didn't leave already... could you execute for me again the count queries I gave you before the meeting? [04:34] That's not a branch, isn't it? [04:34] yay, my no-empty-summaries baz patch got merged [04:35] It's just trying to confuse me? [04:35] lucky daf, for getting pqm to merge something [04:35] bradb: nah, Mattheu did it for me [04:35] stub, SELECT COUNT(*) FROM POTemplate WHERE rawimportstatus=2; and SELECT COUNT(*) FROM POFile WHERE rawimportstatus=2; [04:35] Matthieu, rather [04:35] ouch, /me turns speak volume down [04:36] launchpad_prod=# SELECT COUNT(*) FROM POTemplate WHERE rawimportstatus=2; [04:36] count [04:36] ------- [04:36] 105 [04:36] (1 row) [04:36] launchpad_prod=# SELECT COUNT(*) FROM POFile WHERE rawimportstatus=2; [04:36] count [04:36] ------- [04:36] 9401 [04:36] (1 row) [04:36] ddaa: yes, it's a branch [04:36] (it's a magic tag) [04:36] stub, thanks [04:36] 3.1.0.2 -> 3.1.2.x [04:36] I thought that branches had an even number of dots [04:37] how can I tell a branch from a tag, then? [04:37] branches have odd [04:37] revisions have even [04:37] _unless_ the last-but-one digit is zero [04:37] stub, could you kill the poimport script so I get some input about the import status? seems like it's stalled.... [04:37] in which case it's a branch pretending to be a revision === ddaa gargles, twitches, kicks [04:38] carlos: Which one? There are two (!) [04:38] jblack: you got that? === stub kills 'em both [04:38] stub, two? [04:38] stub, both, please [04:39] we should expand the lock file life a bit more when a new release is imported.... [04:39] carlos: One left lying around from the 6th... [04:39] stub, I suppose the lock file was removed after a day being executed [04:39] carlos: Might be worth trying that locking code from kiko, where the lock is cleared when the process dies. [04:40] Or just increase the timeout ;) [04:40] stub, so we don't remove it after a fixed amount of time? [04:40] stub, the timeout is only a problem when a new release is imported [04:40] ddaa: what? [04:40] carlos: if it works, yes. [04:41] stub, will it work with a kill -9 ? (don't think so) [04:41] carlos: aparently, yes. [04:41] I suppose it's a corner case we don't need to care too much [04:41] really? [04:41] kiko, ? [04:42] wow, those logs are huge... [04:42] carlos: if it is a kernal resource, it is released when the process dies. But I avoid system programming in general ;) [04:43] interesting... [04:45] Keybuk: you were meaning "odd/even" number of numbers, not number of dots, right? [04:48] is pqm hung? (no pun intended) [04:49] that is an oxymoron [04:51] Seriously, when has pqm been working today? [04:52] last merge I saw was 1h40m ago [04:52] morgs: *today*, he says, *today*...you're too kind, my friend. [04:54] pqm is busy doing groceries [04:54] it was going to get a haircut after that [04:54] would anybody else find it useful if pqm sent a confirmation message saying something like "Your merge request has been received. There are currently 2 other merge requests being processed before yours." (or something to that effect) [04:54] and then a sauna [04:54] heh heh...:/ [04:55] so call him in about 6h [04:55] You are merger number 23 in the queue. Please be patient, your branch is important to us. Your branch will be merged in Traceback (most recent call last): ZeroDivisionError: integer division or modulo by zero [04:55] yes, we were discussing pqm muzak last night [04:56] bradb, yes, I think such a message would be helpful, but it would be even more helpful if somebody profiled pqm and found out wtf it is busy doing while we are waiting [04:56] bradb: Knowing the queue would definitely help - or even a web page... [04:57] lifeless: it appears that i'm not the only pqm user who would find that confirmation message useful. === lamont [~lamont@15.238.5.49] has joined #launchpad [04:57] ddaa: yes [04:57] bradb: emphasize the "x in the queue" bit... [04:58] jakob nielsen says confirmation messages should be sent out quickly, to ensure that the user has some idea that something is actually being done with their request. [04:58] morgs: indeed [04:58] lifeless, pqm needs to do some aerobic workout, it's stamina is depressing [04:58] Aah, time to go... [04:59] as kiko noted, it would also be nice if there was a way of getting more detailed process info from it, but even the confirmation would be a step in the right direction [05:00] salgado: can i send you a one-screen long (approx.) diff for a quick review? it's almost a [trivial] , but i just wanted to be doubly-sure. [05:03] or BjornT, can you take a quick look at this? (or any other reviewer) [05:04] bradb, not *right* now, but I can do it after lunch. at most 2h from now [05:04] or kiko, could you review my title fix? [05:05] salgado: ok, i'll send it your way, thanks [05:12] https://launchpad.ubuntu.com/+login [05:12] is that where I should point people to register new accounts? [05:13] i think so. salgado: ^^? [05:13] yes, it definitely is === faassen [~faassen@a80-127-80-154.adsl.xs4all.nl] has left #launchpad [] [05:21] elmo: can you bounce pqm again? (sorry, i hate asking this all the time too) [05:21] as far as i can see, pqm is still blocked on rev [05:21] which is odd [05:21] because rev shouldn't be doing anything any more [05:22] SteveA: does the process move along if you send it SIGCHLD so that it cleans up the rev? [05:22] i'd need to be pqm or root to try that [05:22] true [05:23] hmm, simple_popen2 is actually used, in the poexport-template-tarball.txt test [05:23] but, that doesn't call re [05:23] but, that doesn't call rev [05:25] i'm going to grab a new rocketfuel launchpad tree to check that rev isn't there [05:26] New Malone bug 1286 filed on source package nn by Ralph Corderoy: nn suffers segv after empty NNTP LIST reply from sn. [05:26] https://launchpad.ubuntu.com/malone/bugs/1286 [05:28] so, 'rev' isn't in RF any more [05:28] daf, what about package-only dilys in #ubuntu-bugs :-P [05:28] bounced [05:28] i'm not sure why a merge would be using it [05:29] thanks elmo [05:30] Merge to rocketfuel@canonical.com/sqlobject--test--0.6: [rs=stub] remove locks from the sqlobject cache. step one of three improving steps. (patch-23: steve.alexander@canonical.com) [05:34] SteveA, you have no flight number listed in the sprint homepage. Why? [05:36] SteveA? [05:37] i only just received the flight # [05:37] i'll update it soon [05:37] can you tell it to me now? [05:37] I'm setting up pickups for you guys === kiko is on the phone [05:38] SteveA? [05:39] Flight - KLM (KL) - 797 Fri 15 Jul 2005 [05:40] Arrive: 8:50 PM [05:40] GRU Terminal 1 [05:40] thanks you the man [05:40] please update the wiki [05:40] kiko, I think you will like this one.... [05:41] kiko, https://wiki.launchpad.canonical.com/LaunchpadPoImportFeedback <-- Could you review that spec when you have sometime? [05:41] kiko, should I add it to the PendingReview page? [05:41] daf, around? [05:41] yeah [05:41] I forgot to have lunch [05:41] carlos, add it to my queue [05:41] X-) [05:41] daf, really? [05:41] yeah [05:41] kiko, it's already done [05:42] daf, dude... [05:42] daf: don't neglect to eat [05:42] it will burn you out [05:42] I know [05:42] daf, anyway, when you have some time, please, take a look to that spec too [05:42] I got distracted [05:42] daf, not before you have lunch [05:43] daf, I will leave in about 30 minutes, should we leave the meeting for tomorrow? [05:43] hmm [05:43] I think we have quite a bit to talk about, so perhaps we should leave it for today [05:43] leave it for today? [05:43] not have a meeting today [05:44] ok [05:44] I will ping you tomorrow as soon as I wake up [05:44] so we have time enough to talk about anything before you leave for holidays [05:46] how much of a pqm backlog is there right now? [05:49] SteveA: just to make sure i understand your suggestion: are you suggesting something like this? [05:49] class BugPageTitle: [05:49] def __call__(self, context, view): [05:49] return "Bug #%d - %s" % (context.id, context.title) [05:49] and then: [05:49] bug_index = BugPageTitle() [05:52] yeah [05:52] ok, thanks [05:53] or even just a simple function should work [05:53] def BugPageTitle(self, context, view): [05:53] i think [05:54] and then bug_index = BugPageTitle? [05:54] yeah, if that works [05:54] sorry -- working somewhere else [05:55] write a "helper", but it with the others, and always use 'context'. [05:55] do whatever seems clearest / most consistent [05:57] ok, thanks [06:07] New Malone bug 1288 filed on product Bazaar by Colin Watson: resolving conflicts in .arch-ids is far too painful [06:07] https://launchpad.ubuntu.com/malone/bugs/1288 [06:22] BjornT: I've replied to your review of DisplayingParagraphsOfText. [06:40] i guess pqm is hung again [06:41] bradb, I guess so too [06:41] I sent a merge request 2h ago [06:42] it's doing twisted and autobuild-snaps [06:42] same here (two, in fact, one missing [trivial] for which i haven't even received the failure email :/) [06:42] but not actually doing anything [06:43] SteveA: can i send you a diff of just what i changed for take two of the pagetitles.py changes? [06:43] yeah [06:43] thanks [06:46] i'm going to be changing the page titles stuff after brazil [06:46] to something that is actually included in the templates [06:46] but which retains the programmability we have now [06:50] that sounds like a good improvement to avoid the mysterious template name -> function-name-minus-dot-pt mapping magic [06:58] bradb: i don't see the email [06:58] i had to make one small whitespace change [06:59] just sent it now [07:01] bradb: looks good. i approve [07:01] thanks [07:02] SteveA: what do we do about pqm? i'll have a third merge request to make shortly after lunch [07:03] bradb, steve already approved that changes you wanted me to review? [07:04] salgado: yes, sorry, i should have told you that [07:04] bradb: i don't know what autobuild-snaps is [07:05] Kinnison: do you? [07:05] i think we'll just have to wait until lifeless gets up [07:05] bradb: pardon? [07:05] and get lifeless and stub to push back the production roll-out by a couple of days [07:05] Kinnison: do you know what autobuild-snaps is? [07:05] bradb: erm, never heard of it [07:05] bradb: in what context is it used? [07:06] ok. the word "build" make me think "Kinnison" [07:06] dunno [07:06] bradb: just keep a text file of the branches you want to merge, and then get on with other stuff [07:06] merge between your branches if you need to [07:06] we'll get this fixed tomorrow [07:06] right [07:06] bradb: worth checking with celso [07:06] i have that text file going already ;) [07:06] cool [07:06] bradb: he's taken on most of the autobuilder stuff from me [07:06] bradb: thank god === Kinnison waves [07:11] ciao guys [07:13] pqm is hung trying to generate a baz build [07:14] in fact, trying to build a snapshot [07:15] elmo, can you kill the autobuild-snapshot.sh? [07:16] done [07:17] uhuuu! thanks elmo. ;) [07:18] pqm vs. baz, a most appropriate duel [07:19] pqm has been fucked all day [07:19] it was pqm vs. coreutils earlier [07:20] it's pqm vs dead-concordia now [07:30] /bin/sh: TZ=Asia/Calcutta make check_merge: No such file or directory [07:30] WHAT THE FUCK?! [07:35] stub made the tests run under that TZ IIRC [07:39] I belive lifeless changed it in the end [07:39] in pqm, rather than in Launchpad [07:40] gotta go pays lotsa beels, bbl [07:46] hey ho === terrex [~terrex@84-122-69-8.onocable.ono.com] has joined #launchpad [07:54] Merge to rocketfuel@canonical.com/sourcerer--devel--0: [trivial] remove hctapi dependency, instead just take a url (patch-24: scott@canonical.com) [07:55] aha [07:55] LOOK MA! I got a merge through! [07:59] Keybuk, who are you bribing? [07:59] I've given up for the week [08:00] lol [08:00] it reject my merge NINE TIMES [08:00] "nine times?" [08:00] "nine times." [08:00] "I don't remember him being sick nine times." [08:01] "that's because he wasn't sick. he was skipping school." [08:01] a free brazilian dinner to the first person who guesses the movie [08:01] hmmm [08:03] kiko: dude [08:03] kiko: ferris bueller's day off, of course [08:03] mdz! [08:03] ffs [08:04] how are we going to ship a barzilian meal all the way to sfo [08:04] kiko: good timing; I'm starving [08:05] I think the third line is actually misquoted and she says [08:05] "I don't remember Ferris being sick nine times this year." [08:05] but that would have been a dead giveaway [08:09] kiko: my dinner would get awfully cold in sfo [08:09] considering that I'm in London, won't be back in LA for another week or so, and am not planning to visit san francisco for some time yet [08:11] you should visit sfo often, it's lovely [08:12] q. how can a box with 1gig run out of memory twice on the same day? [08:12] a. if rss is the problem, baz is the answer. === terrex [~terrex@84-122-69-8.onocable.ono.com] has joined #launchpad === JanC [~janc@dD5764BEC.access.telenet.be] has joined #launchpad [08:30] Merge to rocketfuel@canonical.com/sourcerer--devel--0: [trivial] don't call a rollback function if we don't get to the important bit (patch-25: scott@canonical.com) [08:31] bradb-bank, question [08:35] bradb-bank, why does bugtask-view duplicate large portions of bugtask-edit? [08:35] bradb-bank, why does bugtask-view still include a submit button [08:35] still include /code/ for a submit button, sorry [08:36] bradb-bank, and why does bugtask-view use the view's widgets still? [08:43] bradb-bank, rs=kiko to something that fixes any of that === SnakeBite [~SnakeBite@84.242.143.64] has joined #launchpad [08:55] off-irc to finish this report === terrex [~terrex@84-122-69-8.onocable.ono.com] has joined #launchpad === SnakeBite [~SnakeBite@84.242.143.64] has joined #launchpad === comadreja [~comadreja@80.224.108.186] has joined #launchpad [09:10] hmm launchpad doesn't send me my forgotten password, why could that be ? [09:11] New Malone bug 1292 filed on source package frozen-bubble by Loic Pefferkorn: Not installable because of broken package [09:11] https://launchpad.ubuntu.com/malone/bugs/1292 [09:17] Merge to rocketfuel@canonical.com/sourcerer--devel--0: [trivial] don't try to make a branch of a virtual orig, it just won't work (patch-26: scott@canonical.com) === terrex [~terrex@84-122-69-8.onocable.ono.com] has joined #launchpad [09:41] Merge to rocketfuel@canonical.com/sourcerer--devel--0: [trivial] create a Library automatically when I'm too lazy to pass one (patch-27: scott@canonical.com) [09:47] anyone with SWIG-fu available ? why can't use a swig-maid module by simply "import foo; bar = foo.child()", instead should I use "from foo import child; bar = children()" ? [09:49] right, that's that set of sourcerer bugs fixed === Keybuk sets it running again [09:49] let's see if we can get to 't' [09:50] need help, launchpad won't send me my password, I receive no mail [09:50] oops, sorry, received now === terrex [~terrex@84-122-69-8.onocable.ono.com] has joined #launchpad [10:12] Keybuk: did that calcutta thing end up getting fixed? it looks like it did, based on the merges above, but you never know... [10:12] I assume so [10:12] it must have been transient [10:13] Okay, I have written a small howto for importd rollout, per lifeless request. [10:13] Anyone knows where is the right place for this? [10:13] I mean, I could jus stash it on wiki.canonical.com [10:14] but I feel compelled to actually link it from somewhere [10:19] (hacking on the terrace)++ [10:21] SteveA: ping [10:29] aren't there too few bugs ? [10:35] New Malone bug 1293 filed on product The Launchpad by Brad Bollenbach: Customizing body attributes seems like a very secretive, if not almost impossible thing to do [10:35] https://launchpad.ubuntu.com/malone/bugs/1293 === kiko [~kiko@200-171-140-32.dsl.telesp.net.br] has joined #launchpad [10:52] Kinnison? [10:53] kiko: to answer your question from earlier, the reason Save Changes was still on that form was because of the one, the only, pqm [10:53] i fixed that bug yesterday [10:53] DIE PQM DIE [10:53] what about my other questions? [10:53] bradb, https://launchpad.ubuntu.com/errors/showEntry.html?id=1120767256.680.391418624412 === kiko whimpers at Kinnison's idle time === kiko gives up [10:54] not sure if you're already aware of that [10:54] i'll file a bug on the file private bug + Cc maintainer silliness [10:54] salgado: interesting, thanks for pointing that out [10:56] kiko: the only reason it's still using widgets is because that's not breaking anything. should we leave touching that until the next person who goes in there to do a formatting overhaul wants to change it? [10:56] (i.e. widgets are meant for display purposes too, etc.) [10:58] bradb, I want to be able to linkify assignee to his page. if you can do that, great. [10:58] k [10:58] thx [10:58] bye [10:58] :) [10:58] heh [10:58] how does the linkifying look/work? === dilys [daf@muse.19inch.net] has left #launchpad ["upgrade] === dilys [daf@muse.19inch.net] has joined #launchpad === dilys [daf@muse.19inch.net] has joined #launchpad === bradb notes kiko's Jakob Nielsen style emails to launchpad@ [11:06] kiko: I have a present for you [11:13] no way [11:13] really? [11:13] :) [11:14] and brad? [11:14] well, brad might appreciate it also [11:14] but it's kiko that's been hounding me [11:14] bradb, it looks and works marvelously -- apply the patch attached to the emails and you'll get the linkified vibe [11:14] and brad as in "and brad, what do you want at this late hour of the night?" [11:15] daf, you are DA MAN! [11:16] SteveA: this is not a "late hour of the night" question, to be honest. :) better wait till tomorrow. [11:16] lives are at stake here [11:16] kiko: do these regexes look good to you? https://chinstrap.ubuntu.com/~dsilvers/paste/fileXoJ9z1.html [11:17] jesus christ [11:17] what do I know about regexes [11:17] heh, ok [11:17] that's pretty cool though [11:17] well, did I miss any products out that we want? [11:17] looks pretty much correct [11:18] lemme 2-check [11:18] daf, does HCT not have a product? [11:18] maybe HCT bugs should go to #hct [11:19] you are a star daf [11:19] no worries :) [11:19] yeah, #hct is a good idea [11:19] hmm, I should make it so I can add/remove channels withot restarting her [11:21] bradb: go on, give me a hint [11:22] daf: you know what jwz says about regexes [11:22] daf: in theory, shouldn't there a check for the word /ubuntu/i for the sp regex? [11:22] "There's nothing worse then a gremmie out of control!" [11:22] no, that's not it [11:23] SteveA: i want to set focus to a field by default in a form, but there's two problems: 1. i don't appear to have anyway to override body attributes in main-template.pt and 2. this is not an add/edit form, so I don't think the browser:widget stuff is an option. [11:23] bradb: not all Ubuntu source packages have Ubuntu in the name though, right? [11:23] SteveA: so, i did this: [11:23] [11:23] (idea paraphrased from what kiko did on the malone front page, but modified in a way that works in bugs-for-context.pt). it seems less than optimal. [11:24] daf: they should [11:24] it's an interesting issue [11:24] it's not only optimal, it's the way it should work, bradb! [11:24] it doesn't make sense for more than one thing to have focus for a particular page [11:25] kiko: is it! [11:25] and what should have focus can only be decided, in the general case, when you're thinking about the whole page design [11:25] bradb: maybe in the title [11:25] I SAY SO THEREFORE IT IS [11:25] COWABUNGA [11:25] bradb: but the bug notifications use the displayname [11:25] bradb: (AFAICT) [11:25] SteveA: right [11:26] if the notifications have a distribution field in them, I can match on that [11:26] using javascript for it is a bit groty [11:27] but a reasonable solution since we don't have a clear way of selecting form fields once the PT is rendered [11:27] daf: [11:27] when i've rewritten page templates in woven, all this will be easy ;-) [11:27] Affects: ubuntu mozilla-firefox [11:27] Severity: Normal [11:27] Priority: Medium [11:27] Status: New [11:27] daf: that's the way FBN looks now [11:27] um, s/woven/nevow/ [11:27] beautiful [11:27] SteveA: 10 years from now! [11:27] :P [11:28] when is that going to land? [11:28] if .js for now is cool, then i'm cool [11:28] daf: it already has [11:28] *including* getting by pqm [11:28] not rolled out yet though, i don't think [11:28] ooh, so dilys gets broken next Tuesday while I'm away? [11:28] js for now is cool. But, this should go in the hackers faq [11:28] that's up to the good ole admins [11:29] SteveA: good point [11:29] like, the rationale for using javascript now [11:29] because we have no better way of doing it in general [11:29] yes [11:29] i'll give a little blurb in there right now [11:30] N + Jul 07 Arch Patch Queu ( 58) failure [11:30] mdk mdk mdk [11:30] ? [11:30] daf: i'm a mutt user [11:31] SteveA, is there a way of doing it without javascript, actually? [11:31] not a real, simple, here and now way [11:31] i mean, you can hard code it in your forms template code [11:31] you can add the "extras" stuff to autogenerated forms [11:32] i could add some post-processing pipelining to page template rendering [11:32] bradb: so am I, but I still don't get it :) [11:32] lots of ways we *could* do it [11:32] daf: mdk? murder death kill. [11:32] er, ok [11:33] that's my customized birthday card greeting to pqm [11:34] SteveA, bradb: uhm, how can we focus form fields without javascript. tabindex? Does that even work? [11:34] kiko: tabindex uses javascript as well [11:35] it does? [11:35] i thought tabindex was part of the html standards [11:35] or, more precisely, the .js uses tabindex to figure out what field to set focus to, if there was no error fields to set focus to (and perhaps it checks one other thing) [11:35] and it just works [11:35] this is an mpt question [11:35] yes, it might work without js too, but using tabindex is very non-trivial. [11:35] you can find the .js usage of it in launchpad.css [11:36] (it's from plone, i believe) [11:36] oh [11:36] okay [11:36] my JS hack works TODAY! [11:36] using tabindex will mean, in many cases, completely writing the form out by hand, which is a lot of effort if it was an autogen'd form, sadly :/ === kiko grins [11:36] autogend forms die die die [11:37] (though add/edit forms can use a custom widget and extra="" to do this, but anyway, back to the point, kiko .js hack is the best for now, i agree) [11:37] s/kiko/kiko's/ [11:51] bradb, are you using the malone email interface, or has it not been put into production yet? [11:52] kiko: no idea, i asked BjornT about it the other day, but i think he was on holidays that day [11:52] it hasn't yet worked for me [11:52] hmm, I'd really like to get my branches merged before I finish this week [11:52] I have two small ones and one big one [11:53] bradb, can you find out and reply to the report that I'm going to email out in 10 minutes? [11:53] thanks [11:53] it has an XXX on this topic exactly [11:53] i'll find out, yes [11:53] BjornT: around? :) [11:54] kiko: would you mind verifying the accuracy of this FAQ? https://wiki.launchpad.canonical.com/LaunchpadHackingFAQ?action=show#head-ee90f9fa4093c5ea120a9414069521234ecf94a1 [11:55] bradb, just leave the second JS format, I think, I prefer it [11:55] bradb, stupid question, the launchpad wiki just became non-password protected, no? [11:55] Burgundavia: AFAIK, yes [11:55] Burgundavia, right [11:55] 8-o [11:56] kiko: how does that format work for dotted fieldnames? [11:56] bradb, also, the JS snippet needs to appear after the form field is defined IIRC [11:56] this means that an OSS release comes soon? [11:56] bradb, is there such a thing? then I agree, the first one is a reasonable workaround [11:56] kiko: isn't that wiki full of information about our infrastructure and things like that? [11:56] Burgundavia, heh. we're getting there, but no hurry :-) [11:56] kiko: yes, it's what zope generates, and some fields manually place zope widgets in the forms, so they get field.foo names [11:56] ddaa, the sab has commanded it [11:56] aieee, bradb [11:56] s/some fields/some templates/ [11:57] I am OUTTA HERE === ddaa just put a ImportdRolloutHowto on that wiki... [11:57] bah... it's sabdfl's money after all... [11:58] We don't have nothing nothing to hide. [11:58] We're a pants-free company! [12:00] lol [12:00] lol [12:00] lol [12:03] kiko-afk: from a quick test, yes, it appears to need to be after the field is already defined. updated the faq. [12:03] thanks [12:03] great