/srv/irclogs.ubuntu.com/2009/08/21/#launchpad-dev.txt

=== Ursinha is now known as Ursinha-afk
sinzuiLaunchpad UI 3.0 Question and answer meeting in #launchpad meeting in 6 minutes01:54
thumpersinzui: how long will it go for?01:55
sinzuiI will linger for 45 minutes. It will really on only if people have questions that I can answer01:59
thumpersinzui: I don't have any questions, and I have to head out to meet Rachel01:59
thumpersinzui: I'm not sure who else will be around, but good luck :)02:00
sinzuifair enough02:00
=== matsubara is now known as matsubara-afk
=== abentley1 is now known as abentley
wgrantOne month since the open sourcing!05:16
lifeless\o/ happy birthday lp:)05:22
lifelesshmm, sourceforge has useful timelines now ><07:08
wgrantlifeless: You mean the project feed?07:11
lifelesshttp://sourceforge.net/projects/japt-proxy/develop for instance07:12
wgrantYep.07:13
wgrantGoogle Code has had something like that for ages.07:13
wgrantNow somebody just needs to go and fix the karma system to do that.07:14
lifelessmmm07:14
lifeless'someone needs to go and do that' +107:15
lifelessusing karma for it.... different discussion07:15
wgrantI suppose you could also just collate lots of collections of other activity records.07:15
lifelesswe have rss feeds for projects07:16
wgrantBut, well, there aren't such records.07:16
wgrantThere are RSS feeds of new revisions and new bugs.07:16
wgrantAnd announcements.07:16
wgrantStuff I miss from Google Code is changes to and comments on bugs.07:17
stubAnyone looking into what revision broke xx-productset.txt ?08:08
noodles775stub: yes...08:15
adeuringgood morning08:22
wgrantHrmph. stable is ancient.09:03
mrevellMorning compadres09:19
bigjoolsmorgen09:23
gmbGot to love it when ec2test just sort of stops and shouts "No! Lalala, I'm not going to test your branch."09:35
wgrantIs PQM happy again?09:36
wgrantIt has been a bit unhappy today.09:37
allenapwgrant: I looked into the BugTaskEditView stuff, and basically concluded that we needed to change all the other call-sites to use transitionToMilestone(). However...09:43
wgrantallenap: Hm? Why? No direct callsites should care about the schema.09:43
allenapwgrant: It may actually make more sense to change mutator_for to allow use on a read-write attribute, perhaps by passing a allow_read_write parameter, just so it's not the default.09:44
elmo** bouncing apache on vanadium for emergency update09:45
allenapwgrant: My logic may have been flawed :) But the milestone widget was behaving weirdly in the form, and simply changing it to be handled like status and importance seemed like the simplest fix.09:45
allenapwgrant: Well, not simplest, but an already tried and working approach.09:45
wgrantallenap: status/importance are readonly=True. I was unable to work out what was magical about them to make them writable.09:46
wgranti'm not sure that callsites outside the form need to be updated.09:46
allenapwgrant: I think it's the bit around line 1235 in browser/bugtask.py that makes sure they can be edited, then there's some juggling in updateContextFromData to apply the changes.09:49
mrevelldanilo-afk: http://blog.launchpad.net/translations/screencast-sharing-translations-between-releases-of-the-same-project09:49
wgrantallenap: Right, updateContextFromData will probably need changing. But my main concern at this point is getting the right widgets to show up.09:50
allenapwgrant: Okay, I'll come back to you in a few minutes.09:50
wgrantallenap: 1235 seems to replace the writable status/importance widgets with read-only ones if the user lacks privileges.09:50
wgrantallenap: Which suggests that the writable widget is forced in place before that.09:50
allenapwgrant: In my copy it does the opposite :)09:51
gmbDid the session DB get dumped again?09:51
allenapwgrant: But I'll go and experiment some more.09:52
wgrantallenap: Thanks. I have a couple of ideas which I'll try...09:52
gmbOoh, timeout filing a bug.09:53
wgrantgmb: Um, people have been complaining about that for weeks...09:53
gmbwgrant: s/weeks/years (on and off). True, but I'd never experienced it filing a bug on Malone before.09:53
wgrantbigjools: Why is the CoC required?09:57
wgrantIt seems quite irrelevant.09:57
bigjoolswhy?09:57
wgrantbigjools: Hm, I guess there are a couple of relevant points in it.09:59
wgrantMy memory was foggy.09:59
* wgrant goes back to beating formlib into submission.09:59
bigjoolsit needs it10:00
allenapwgrant: Heh, sorry, you're right about line 1235.10:00
wgrantbigjools: Not really. BugTaskEditView is just purest evil.10:00
wgrantallenap: Phew.10:00
wgrantallenap: Concerning that it's so hard to understand, though :/10:01
bigjoolsthere's a lot of purest evil, most of it in old Soyuz code :(10:01
mrevellbigjools: doc fixed10:01
intellectronicawgrant, allenap: the magic is using a mutator and i agree that this is a good approach10:01
wgrantbigjools: But I can understand most of the Soyuz code.10:01
bigjoolsmrevell: fanks10:01
wgrantintellectronica: That's not the dark magic here.10:01
wgrantintellectronica: The real magic is making the widget editable on +editstatus.10:02
wgrantAnd BugTaskEditView's existing method of doing that is not transparent.10:02
bigjoolswgrant: care to let your eyes bleed over the perl in the buildd? :)10:02
intellectronicawgrant: wow, you understand soyuz but not the bugtask edit view? do you also wear your socks over the shoes? :D10:02
wgrantbigjools: I've hacked sbuild before for doing archive rebuilds... not pretty.10:02
wgrantbigjools: Soyuz is easier to setup than dak/wanna-build/buildd!10:03
bigjools:)10:03
noodles775+1 for soyuz ;)10:03
intellectronicawgrant: oh right10:03
* gmb genitally hates BugTaskEditView.10:03
gmbEr.10:04
gmb*genially10:04
gmbIs the word I was going for there.10:04
wgrantintellectronica: Hm?10:04
bigjoolshow Freudian10:04
wgrantNicely done.10:04
al-maisanintellectronica: after all these Soyuz reviews you should know it by heart now ;)10:04
intellectronica:)10:06
wgrantSo...10:07
wgrantIs there any good reason that it does things in this order: maybe overrides the status widget with something writable, maybe overrides the status and importance widgets with something read-only, then maybe overrides the importance with a writable widget?10:08
wgrantOr is that just there to confuse people?10:08
allenapwgrant: Just to confuse I think.10:08
allenapwgrant: But probably something will break if done in a different order :(10:08
* wgrant finds a similarly nefarious order in which to place the milestone code.10:09
allenapwgrant: Something I noticed. On the main bug page, milestone_source is not None. On the +editstatus page it's None, so the milestone field is not replaced (at around line 1221 in browser/bugtask.py)10:09
wgrantallenap: Ahaa.10:09
wgrantSo.10:09
wgrantWe need to dig a restrictive milestone vocab up from somewhere.10:09
wgrantBut formlib must magically do that already.10:09
wgrantAs IBugTask.milestone's vocab looks global.10:10
wgrantOr we convince formlib to generate a writable widget to start with. Argh.10:12
allenapwgrant: Have a look at http://pastebin.ubuntu.com/256793/ - it seems to do something approximating correct. I haven't tested it much (and I intend to use copy_field at some point).10:22
allenapwgrant: I'm not sure if lines 30 and 31 are needed either. I just threw them in.10:23
wgrantallenap: copy_field then overriding readonly was going to be my next attempt. Let's see...10:24
allenapwgrant: The tests pass at least!10:24
wgrantallenap: I think those two are needed.10:24
wgrantOh.10:24
wgrantSo you don't need to override readonly?10:24
wgrantAh, you're not copying yet, right.10:24
allenapwgrant: Yeah, the actual desired effect is obscured.10:25
allenapwgrant: Here's my attempt with copy_field: http://pastebin.ubuntu.com/256798/10:28
wgrantallenap: Mine is similar, although I didn't know that one could give it kwargs like that.10:31
* wgrant is testing that.10:31
* gmb -> restarting proxy, bbiam10:34
wgrantallenap: Now I know that copy_field kwarg trick, I've spelt it http://pastebin.ubuntu.com/256803/10:35
* wgrant runs the tests.10:35
wgrantI don't think you need that 26/27 change.10:35
allenapwgrant: Cool.10:36
wgrantSo it was actually easy.10:39
wgrantThe existing complexity was deceptive.10:39
wgrantallenap: Thanks. Pushing the change that I pastebinned -- care to ec2test?10:41
wgrantAll the failing tests pass.10:41
allenapwgrant: Cool, will do :)10:42
wgrantNow I just need to convince cprov that the failures in my other branch were because devel was broken, and that will be landing #10! Yay.10:45
allenapwgrant: Wow, that's good going.10:49
deryckMorning, all.10:51
=== danilo-afk is now known as danilos
* wgrant considers looking at PPA package download counters.11:49
bigjoolswgrant: that needs some more foundations work11:50
wgrantbigjools: Really? What does it need?11:50
bigjoolswgrant: I don't remember exactly, but the current log parser only looks at librarian downloads11:51
wgrantbigjools: It's easy enough to refactor.11:51
bigjoolsoh for sure11:51
bigjoolsjust sayin' like11:51
wgrantbigjools: So, assuming I refactor the log parser...12:12
bigjoolswgrant: talk to salgado first, he wrote the existing one, he'll give you tips12:13
wgrantbigjools: Thanks.12:14
wgrantbigjools: Model-wise, I'm thinking an ArchivePackageDownloadCount, like LFDC except referring to an archive and a source or binary package. Does that sound vaguely in the right direction?12:15
bigjoolswgrant: I think so, although we need to think about whether nullable foreign keys are wanted12:17
wgrantbigjools: Why wouldn't they be?12:18
bigjoolsAPDC, with columns [source|binary]packagerelease, archive, count12:18
wgrantThat's right.12:18
bigjoolsbecause you can have source and binary12:18
wgrant[sb]pr are nullable, with a constraint to ensure exactly one isn't.12:18
bigjoolsmight be better to have 2 tables12:19
wgrantPossibly.12:19
* bigjools goes back to figuring out how a change in das:+index broke das:+builds12:20
=== Ursinha-afk is now known as Ursinha
* gmb lunches12:48
=== matsubara-afk is now known as matsubara
=== mrevell is now known as mrevell-lunch
derycksinzui, ping14:11
sinzuihi deryck14:12
=== salgado-afk is now known as salgado
=== mrevell-lunch is now known as mrevell
beunonoodles775, hi14:39
noodles775hey beuno14:39
beunonoodles775, how are you today?14:40
noodles775beuno: why, well thankyou :), and how would you be?14:40
noodles775Just doing some packaging in between planning a ui change :)14:41
beunoI'm the only way I can posibly be during a 2 month ui cycle: fantastically well14:41
noodles775heh14:41
beunonoodles775, have you seen that salgado's breadcrumb changes landed?14:42
beunofor example, bug 16529314:42
mupBug #165293: collisions through uploading same-named .pack files not handled correctly <packs> <Bazaar:Confirmed> <https://launchpad.net/bugs/165293>14:42
noodles775beuno: nope - haven't seen it, but great!14:42
noodles775beuno: very nice!14:43
beunonoodles775, while he works out some of the remaining issues, how do you feel about jumping in and moving them beneath the titles?14:43
Knut-HBHi, it's me... again. We got some problems while uploading files. A team member wants to upload template.pot and receives this error: http://paste.ubuntu.com/256915/14:43
noodles775beuno: has the formatting been separated out from the hierarchy? That was the main issue last time I checked (the hierarchy was rendering itself manually)...14:44
* noodles775 checks14:44
beunoah, good question14:45
beunosalgado should know14:45
beunosalgado, btw:  wooooooooooooooooooooooooooo14:45
salgadonoodles775, no, the hierarchy still renders itself14:46
beunosalgado, is changing that in your plans?14:47
salgadobeuno, not until it gets in my way or someone convinces me it's a good idea. ;)14:48
noodles775salgado: you don't think it would be a good idea for the hierarchy to use a template to render itself, rather than the html code in the view?14:49
* noodles775 hasn't looked closely - maybe there's a good reason for the way it's done.14:49
salgadonoodles775, oh, absolutely!14:49
salgadoI thought you were talking about moving the rendering into the breadcrumbs themselves14:50
noodles775salgado: no, just using templates to render them so it's much easier (1) to change and (2) to add the new formatting html for breadcrumbs on baselayout while letting the old template use the current breadcrumb formatting.14:51
salgadoright, that makes total sense14:52
salgadobut it'd be easy to fix, no?14:52
noodles775salgado: I'm not sure - it looks like quite a bit of custom html construction going on in that render method.14:53
beunosalgado, none of that HTML is needed for the 3.0 design14:54
beunoer14:54
beunonoodles775, ^14:54
salgadolet's throw it away, then14:54
beunoit's simple text and links, with separators and a special rendering for the last item14:55
salgadoand break all pages not yet converted to 3.014:55
noodles775beuno, salgado: hangon - but that means all pages that *haven't* been updated to the 3-0 base-layout will go awol.14:55
sinzuiUI 3.0 question and answer session in #launchpad-meeting in 6 minutes. Everyone is welcome to ask questions. (no questions about lemurs please)14:55
beunonoodles775, I think that it doesn't matter14:55
beunothis is why we're not rolling out to production14:56
maxbIt turns out that mdz had a bzr branch for launchpad-dependencies back in the Dapper era (the source packages contain the bzr branch). Does anyone know where it might have gone? Did it see continued use after launchpad-dependencies went completely canonical-internal? It seems to have disappeared by the time launchpad-dependencies went to a PPA14:56
beunoit will just have breadcrumbs rendered on top instead of underneath the title14:56
noodles775beuno: no, they'll remain where they are, the formatting will just be 3.0 I assume.14:57
noodles775beuno: sorry, ignore me... that's what you said.14:57
beunonoodles775, right. I'm fine with that.14:58
beunomore pressure to move the old pages  ;)14:58
noodles775:)14:58
bigjoolssinzui: what's a lemur?14:58
noodles775beuno: OK, Sounds like another fun job - just pm bigjools to talk with him.14:58
beunonoodles775, think about it this way, if we drop the current breadcrumb rendering, we drop 14 HTTP requests on all pages14:59
sinzuibigjools: A primate native Madigascar. The most famous one is King Julian played by Sasha Baron Cohen.15:00
noodles775beuno: really? how is that? (the images are all cached right?)15:00
bigjoolssinzui: I should not have asked15:00
beunonoodles775, they are cached, but they are in icing, so every night, the URL changes. And, for people who come to LP from a google link, well, it's their first time visit15:01
noodles775beuno: sheesh...15:01
matsubarawgrant, around?15:12
barrybeuno, sinzui can you think of any reason why /projectgroups should not be viewable by ordinary or anonymous users?15:49
sinzuiI want it visible. I think we get questions about it because users cannot find it15:50
barrysinzui: it actually gives you an Unauthorized right now15:50
barrysinzui: i think i'll open that up in my /projectsgroup branch15:50
sinzuiyes, but I am not sure we need to do that anymore15:51
beunobarry no reason that I can think of15:51
barrysinzui: agreed.  beuno cool15:51
barrybeuno: why the 0.1% granularity in font sizes in style-3.0.css?15:57
beunobarry, I don't know, I still need to work on the font sizes  :)15:58
danilosallenap: did you not want to submit the fix in PQM to db-devel instead?15:58
barrybeuno: adeuring has a good point in a review of my branch:15:59
barry>I find the 0.1% "granularity" a bit exaggerated ;) If we assume15:59
barry>that the base font size is 20 pixel on some monitor, 0.1% would15:59
barry>mean a precision of .02 pixel. I don't think that any anti-aliasing15:59
allenapdanilos: Shite.15:59
barry>algorithm will produce different result for font sizes specified15:59
barry>as 24.6 and 24.62 pixel.15:59
barry 15:59
danilosallenap: it should be rejected (there are, I hope, DB changes in the db-devel you merged :), but you'll need to resubmit15:59
danilosand I can't trust your [testfix] to land, so I'll need to resubmit as well :)15:59
allenapdanilos: Phew.16:00
beunobarry, valid point16:00
danilosmthaddon: perhaps, you can clean up the PQM queue (the testfix should go to db-devel, not devel, and the other one will fail anyway)16:00
danilosbeuno: fwiw, we've got 'stuff to review' up on your translations page :)16:01
beunodanilos, where!?16:01
danilosbeuno: on edge, try translations.edge.launchpad.net/~beuno16:03
danilosbeuno: the rest of the page is still a mess, but there's one nice looking box with interesting stuff :)16:03
adeuringbarry, beuno: consider my comment about the font sizes as, let's say, a nitpick about another nitpick: It really doesn't matter if we specify the font size with that precision or not. Leaving the 0.1% is fine for me, even if I find it a bit odd ;)16:03
beunoadeuring, I agree with you though16:04
beunodanilos, nais!16:05
danilos:)16:05
beunodanilos, any reason why you picked a 2 column layout?16:05
danilosbeuno: it's not a 2 column layout, we've just used the max-width so it doesn't extend over the entire page... once we get to actual migration of that page, we can think it over :)16:06
adeuringbeuno, barry: But it might make sense to move the relative font size into its own CSS class, it if appears not only in barry's two branches but elsewhere too.16:06
beunodanilos, awesome. Happy to see things from our sprint leaking through16:06
danilosbeuno: yeah, totally16:07
=== Chex_ is now known as Chex
intellectronicaleonardr: around?16:18
leonardrintellectronica, yeah16:18
intellectronicaleonardr: do you remember, by any chance, if we have an easy way of getting the _web_ url of an api object16:19
intellectronicaleonardr: from javascript16:19
leonardrintellectronica: i believe bac asked that question a few days ago16:20
leonardrand ended up filing a bug for it16:20
leonardrbac, do you recall?16:20
intellectronicai think gmb also asked that question a few days ago16:20
gmbI did.16:21
intellectronicaif we don't have anything in place, maybe i should do that16:21
bacleonardr: you said there was a bug already filed.  since my need was not urgent i let it go at that.16:21
gmbAnd I got encouraged to fetch the object as XHTML, as I recall ;)16:21
gmb(Although it made sense in context)16:21
intellectronicaleonardr, bac: if the solution to the bug is not too difficult i'll pick it up now, since my need is a bit urgent16:22
sinzuibeuno: can you confirm that the top-level objects (IPrimaryContext) (pillars, people) will never have breadcrumbs?16:22
leonardrintellectronica: in fact, i believe BjornT was working on a solution, but not really making it a priority16:22
intellectronicaone thing i was thinking about is simply adding a web_url attribute to each object coming from the webservice16:22
leonardror at least he was working on a way to turn a web service request into a website request, which is a prerequisite16:22
intellectronicaleonardr: yes, that's done, but not sufficient16:23
leonardrintellectronica: yes, that's the consensus solution16:23
intellectronicaleonardr: oright, now if you find the bug feel free to assign it to me16:23
leonardrintellectronica: actually would you call it web_link instead of web_url?16:24
leonardrthat fits our naming convention16:24
leonardror web_self_link if you want to be verbose16:25
intellectronicamakes sense16:25
intellectronicamy_web_self_link16:25
beunosinzui, yes16:26
beunoI confirm16:26
sinzuiokay16:26
sinzuinoodles775: beuno: I think we have a rule to automatically suppress the context heading when the context is IPrimaryContext. This will drop the redundant <h2> headings and may mean no one will every need to pass to the heading-slot...the <h1> in the content with DTRT16:28
beunosinzui, it sounds like a good idea16:29
sinzuiWe need to keep the slot for the templates that are using it, or update every template to not use it16:29
noodles775+116:29
sinzuibeuno: I am traveling Saturday, but I think I can do this for a Monday landing16:29
beunosinzui, perfect16:30
sinzuibeuno: noodles775. I am hiding from a beach next week16:30
beunoooooh16:30
beunothat sounds nice16:30
beunoFlorida?16:30
sinzuinoodles775: thanks for helping me understand the issue16:30
noodles775sinzui: I hope you're not taking your laptop then? (if you're on holidays!)16:30
noodles775sinzui: np!16:30
barrysinzui: i thought you were alergic to the beach :)16:30
sinzuibeuno: A North Carolina barrier island.16:31
sinzuiI can find the vacation from hell warthogs email to remind everyone why I am hiding16:31
bigjoolssinzui: I look forward to another holiday report :D16:31
* bigjools still has that email16:31
sinzuiI did not report that my wife set a box of crustaceans on me last year16:32
beunosinzui, maybe you should jusst give up on holidays16:32
* beuno hides16:32
sinzuiI like mountains and hiking16:32
noodles775yeah, me to :)16:33
bigjoolss/h/b/16:34
gmbArgh. Can we implement bug dependencies for 4.0, please?16:40
* gmb wonders why he forgot to add that to his "top 3" list.16:41
=== salgado is now known as salgado-lunch
=== deryck is now known as deryck[lunch]
=== abentley1 is now known as abentley
=== matsubara is now known as matsubara-lunch
=== gary_poster is now known as gary-lunch
sinzuibigjools: when is the DSP page landing?17:09
bigjoolssinzui: when it's finished! :)  Seriously, I need a few days yet, prob a week17:12
sinzuiokay17:12
bigjoolsI need to codify that complicated table and re-write all the tests17:12
sinzuiI have been removing a lot of tests that were verify the markup structure.17:13
=== beuno is now known as beuno-lunch
rockstarALL:  I'll be ignoring IRC for the next few hours. If you need me, please SMS or call me.17:34
gmbIs anyone else finding ec2test to be massively, painfully slow to start up today?17:39
gmbIt's bugging the hell out of me.17:39
intellectronicarockstar: congrats on branch->bug linking. when i use it, the spinner appears above the action icon, instead of over it. is that intentional?17:50
=== deryck[lunch] is now known as deryck
mrevellOkay guys, see you Monday!18:03
=== matsubara-lunch is now known as matsubara
=== gary-lunch is now known as gary_poster
rockstarsinzui, what is the role of the facet property on menus?18:29
barrybeuno-lunch: ping -> #launchpad-reviews when you return18:29
sinzuifacet binds the menu to a specific application bugs, or answers for example18:30
sinzuirockstar: it is a dead branch in zope evolution. we should be using ILayer18:30
rockstarsinzui, so my facets should be 'answers' ?  I'm seeing things like 'overview'18:33
rockstarI'm just not sure what they do.18:36
=== beuno-lunch is now known as beuno
=== abentley1 is now known as abentley
=== abentley1 is now known as abentley
EdwinGrubbsrockstar, abentley: Can I assign one of you a question concerning pushing a branch to LP and the user gets 'different rich-root support' errors?19:05
abentleyEdwinGrubbs: Sure.19:05
EdwinGrubbsabentley: thanks19:06
=== abentley1 is now known as abentley
EdwinGrubbsdanilos: can I assign you a translations question?20:14
=== salgado is now known as salgado-brb
EdwinGrubbsabentley, rockstar: could one of you respond to the email message on launchpad-users with the subject "merge request resubmission"?21:42
rockstarEdwinGrubbs, I'll take care of it.21:42
EdwinGrubbsrockstar: thanks21:44
=== salgado-brb is now known as salgado
=== salgado is now known as salgado-afk
=== matsubara is now known as matsubara-afk
gary_posterhave a good weekend!22:35
wgrantmatsubara-afk: I am now.23:20
EdwinGrubbsbac: ping23:36

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!