or a

[15:17] sinzui: it doesn't have to be UL, but the only situation that I can think of that would benefits from DIVs would be two rows of action links, which seems unlikely. [15:18] EdwinGrubbs: We should not design CSS for the specific. The reason the old CSS is giant is that we had to add duplicate rules because we could not reuse rules. [15:19] ok [15:19] EdwinGrubbs: We do not need to define an element until we know there is a conflict [15:19] EdwinGrubbs: I think this is a good rule. [15:20] sinzui: I'll make the change. [15:29] sinzui: There is one other problem that I keep ignoring. If the table doesn't exist because there are no entries, then I either have to remove the table-actions class or add style="float: right" on the element. Either way, it seems to break the generalized nature of the css class. It seems like I just have to choose one ugly solution or another. [15:31] EdwinGrubbs: That is a common problem without a solution in our design [15:31] EdwinGrubbs: witness the issue of a side bar without portlets [15:32] EdwinGrubbs: I think you need to look ahead to learn if there will be content. [15:33] EdwinGrubbs: I looked for a CSS3 selector that allowed me to ask if the element has children, I did not find one. maybe beuno knows [15:33] sinzui, I don't recall anything like that in CSS3 === salgado is now known as salgado-lunch [16:37] sinzui, beuno: here is a working css solution that seems a lot better. http://pastebin.ubuntu.com/297595/ [16:38] sinzui, beuno: I use text-align instead of float so that the UL will not end up outside the border of its container. This works since the LI is already set to display:inline. === matsubara is now known as matsubara-lunch [16:46] EdwinGrubbs: : I award you a gold ☆ for cleaverness [16:52] cleaverness? [16:53] I'd make a joke, but I'm not sharp enough. uh. heh? [16:54] malapropism. You can take a cleaver to a solution fit the problem. It can also be clever at the sametime [16:54] gary_poster, :D [16:55] EdwinGrubbs, rockin solution [17:00] sidnei: your lazr-js branch has a huge diff, i think it builds on another unlanded branch? [17:00] flacoste: correct. here's the diff: https://pastebin.canonical.com/23563/ [17:03] sidnei: ah ok. So that's Bjorn's branch, I assume? Has his been reviewed? If so, is it ready to land? Would waiting for his branch to land before a review make sense? [17:03] gary_poster: you reveiwed bjorn's branch. it's marked as approved, supposedly ready to land. i can wait for his branch to land yes. [17:05] sidnei: it's your call, and the reviewer's call I guess. [17:07] sidnei: add that pastebin link to the m+p, and I'd suggest also explaining what widgets.conf and src-js/yui-path.js are for [17:08] flacoste: ok === deryck is now known as deryck[lunch] === salgado-lunch is now known as salgado === matsubara-lunch is now known as matsubara === beuno is now known as beuno-lunch [18:44] * rockstar lunches [19:07] g'night folks. === EdwinGrubbs is now known as Edwin-lunch [19:26] allenap, deryck, yesterday you reminded me of a branch that was landed to lazr.restful and that launchpad should have by now. the branch fixed a unicode-related bug. do you remember the bug number for that? [19:27] leonardr, bug #331990 [19:27] Bug #331990: The inline editor widget reports a JSON error when saving non-ASCII characters [19:28] deryck, thanks [19:28] leonardr, np [19:38] thumper, ping [19:59] deryck, ping [19:59] gar === beuno-lunch is now known as beuno [20:00] deryck, intellectronica, jml, how controversial do you think moving bugtask tables beneath the description is? [20:01] why am I experimenting with that, you say? [20:01] well [20:01] it reads better [20:01] it's more consistent with the rest of the site [20:01] makes people read descriptions [20:01] beuno: the problem is, the height will always vary [20:01] sure, if you're familiar with the bug, it sucks a bit [20:01] hmmmm [20:02] yes, it's less predictable, although we probably need to have a max-height, and let people expand [20:02] beuno, have you talked with mpt at all about the bug page design? [20:02] jml, I have not [20:02] beuno: even if you're not familiar with the bug, but familiar with LP, under the current scheme, the controls for changing the bug task are always in the same location. if the description pushes them to a different location every time it will be harder to hit them [20:02] hm [20:02] unless we make the description fixed height [20:03] beuno, there's a lot of history. You shouldn't be bound by it, but it might be good to know which bits of it you are ignoring :) [20:03] beuno, I think this will be quite controversial, just because it's such a big change. and heavy users are soooo used to looking top for the needed info. [20:03] yes, some people will be pissed [20:03] but lets think about that after we figure out what the best thing is [20:03] but will *all* people be pissed, that is the question. ;) [20:03] * deryck kids [20:03] then we can think about how to make everyone as happy as we can [20:04] beuno, it's after hours for me and I have my happy Lean hat on [20:04] b.t.w i think fixed height for the description is definitely an option (assuming it's quite short) [20:04] sinzui, around? [20:04] on most visits to the bug page you don't read the description [20:04] beuno, 1. Figure out the constraints. [20:04] beuno, 2. Come up with two or three solutions that meet those constraints. [20:04] beuno, 3. Profit. [20:04] intellectronica, so leave white space for very short ones? [20:04] salgado: I am at this moment [20:04] intellectronica, beuno -- so if you don't read the description often, why top post it? uniformity? [20:05] jml, well, I am loosely doing that, but maybe I should do it more formally. If only I had time to play with this more [20:05] deryck, well, you don't interact with the same bug a lot very often either [20:05] so, for new bugs, you do interact with it [20:05] of course, if you use email, you will likely just go there to set statuses and such [20:06] which is probably why intellectronica has that feeling [20:06] beuno, the slightly more formal approach might save you time. [20:06] but if you live on the web ui, you read them most of the time [20:06] sinzui, we're getting http://paste.ubuntu.com/297719/ when trying to run LP on python2.6, but I don't think we have a new twisted release with that fixed, so we're hoping to silence the fascist for twisted imports. for now, at least [20:06] actually, deryck playing back my argument at me reminds me that as a user i'd hate that change. i almost never work with the description and i always want to quickly look at the bug tasks [20:06] sinzui, barry told me you know the fascist, so could help us figuring out how to silence it [20:06] jml, I have 2 hours left, but maybe you're right, I don't know [20:06] * sinzui thinks [20:07] salgado: I do, Let me try to remember this [20:07] beuno, yeah, as intellectronica said, I don't read descriptions first often. I look at the table first. and I do use the web UI mostly. but maybe that's function following form, too. [20:08] beuno: and an expanding description is kinda' against what we're trying to achieve with the ajaxafication. you will now need an extra click when you do need to read the description [20:08] beuno, if you do the constraint stuff, then it helps others to come up with designs. I'm torn between not wanting to derail you, and thinking that it's a bad idea to do a rush job of the single most important page on Launchpad :) [20:08] deryck, so what I'm going to do is, propose something crazy, write up the constraints I can think off, and hand it off to your team while I deal with ubuntu.com and the karmic release [20:09] jml, I'm not thinking I can solve the problem, bue maybe I can kick-start it [20:09] beuno, cool. [20:09] beuno, ok, but I'm a bit like jml and wonder what the value here is. maybe we should just table a redesign, and let us do simple bug fix improvements until we can think this through more. [20:09] jml, so I want to propose something more radical than polish [20:09] deryck, table means something different here :) [20:09] heh [20:10] beuno: also, not wanting to take the wind out of your sails, i must also add that there are many small things we need to fix on the bug page, and it's very likely that fixing them is much more bang for the buck than pretty much any new design we can possibly do right now [20:10] jml, postpone? [20:10] deryck, what's of most value to you for me to work on? [20:10] intellectronica, great information. What would help you guys the most then? [20:10] * deryck is thinking.... [20:11] beuno, how about a decision on open questions, rather than a mockup.... [20:11] beuno, I want you to do something more radical than polish too :) [20:11] beuno: i definitely need a bit of help deciding how to fix the top row of the bugs index page. that would really help me (that's a different matter, but since you asked) [20:11] deryck, if you point me at them, I will spend these 2 hours replying to that [20:11] beuno, mainly I'm think bugtask table stretch the top or not, and how to make the changing in statuses less easy? [20:11] beuno, let me get bug numbers. intellectronica, do you have bugs with open questios you can think of? [20:12] intellectronica, the search box, etc? [20:12] jml, want to schedule a call for that then? to start brainstorming? [20:12] deryck, http://www.economist.com/research/styleGuide/index.cfm?page=673901 (search for "table") [20:12] beuno, please do. [20:12] beuno: yes [20:12] beuno: https://bugs.edge.launchpad.net/malone/+bug/433854 [20:12] Bug #433854: Bugs home page: search box and project details don't fit on the same line [20:13] intellectronica, ok, that's on my list now, I will comment on the bugf [20:13] bug as well [20:13] jml, interesting. I was proposing postpone as an alternate to table? to suit the US usage. [20:13] beuno: awesome, thanks [20:14] deryck, so, as for the bugtask table, everytime I think about, I get the feeling we need to do something different with it [20:14] deryck, oh yeah, that'd be fine. [20:14] deryck, which makes me need to change other parts [20:14] salgado: there I have two hacks in utilities/importfascist.py [20:14] deryck, I was just adding the link for interest. [20:14] which makes me ask questions like whether to move it below the description :) [20:14] beuno: i think that we need to align the tasks table, description and comments, and that when we do that it will improve things considerably [20:15] salgado: I think you wan to use warnings.filterwarnings from line 16 as your guide to ignore [20:15] DeprecationWarning: the md5 module is deprecated; use hashlib instead [20:15] beuno, ok, fair enough. we're stuck. :) but assuming we can't get a redesign this cycle, should the table stay where it is now, or move up above the two columns? [20:15] deryck, intellectronica, it's interesting, because if I look at random bugs, the thing that itches the most is comments [20:15] itches? [20:15] espcially activity [20:15] they use up too much horizontal space [20:16] they don't collapse when doing multiple actions at the same time [20:16] it's messy and hard to understand [20:16] right [20:16] yeah, that really sucks [20:16] so, if it was up to me, I'd work on that and tweak the bugtask slightly to address the problem with long names in the objects [20:17] to me, that's the biggest improvement I can see [20:17] but the ui solution is (i think) obvious. we just need to get them to collapse like bug mail [20:17] making conversations easier to understand and follow [20:17] yeah, I think that's the only way. even if I get the fascist out of the way we'd still get an error because our test runner tells python to convert warnings into errors [20:17] sinzui, ^ [20:17] yep [20:18] but there is a test that hacks around this. [20:18] sinzui, thanks for the help, I'll continue the chasing in #launchpad-sprint with the others [20:18] Or ther ewas a test [20:18] beuno, intellectronica -- that sounds a reasonable plan to me. [20:18] intellectronica, yes, get them to collapse, especially with the comments, and, if that's out of scope for this cycle, make them use up less space [20:18] but if you can make them collapse, I'm super happy [20:18] and I can propsoe a few small tweaks to the bugtask table so it has a bit more space [20:20] for example, the importance column does not need to be so wide [20:20] beuno, intellectronica -- I would rather finish our work already assigned this week and next, not focusing on the bug page too much, then use week 4 to have everyone on bugs team on the bug page, and land something nice week 1 of next cycle. [20:20] deryck, I'm cool with that, when would you need me then? [20:21] salgado: look in canonical/__init__.py. I puth something in there during the zope upgrades [20:22] hmm [20:22] beuno, how about we finish this dialog now or in email with your suggestions, then we do the work, and you take a look mid-to-late of week 4 at how we're doing? see if something else springs to mind. [20:23] intellectronica, does this type of plan seem reasonable to you as well? [20:23] yes, absolutely [20:23] ok, cool. [20:23] I'm on board as well [20:23] ok, cool. thanks guys! [20:24] deryck, I will write up that email now [20:24] beuno, awesome, thanks! === Edwin-lunch is now known as EdwinGrubbs [20:58] jml, I want a bit overboard, but I sent an email of what was floating around in my head [20:58] beuno, yay [20:58] left out the constraints because I felt people where going to derrail the email with their own crazy constraints [20:58] morning peoples [20:58] so I'd like that to be a separate thread when we have a better idea of when and what we'll be doing [20:58] beuno: got time for a talk? [20:58] hiya thumper [20:59] thumper, yes, I haven't started doing what I'm supposed to be doing yet [20:59] thumper, quick question [20:59] beuno: I have 15m before the standup [20:59] jml: shoot quickly [20:59] thumper, any objection to me adding IBranchLookup.getByUrls and exposing it over the API? [21:00] jml: none at all [21:00] thumper, yay. [21:00] thumper, I think I'm going to make a screencast of "how to patch Launchpad to expose an API" [21:00] thumper, and that would fit well, being simple and extremely useful to me personallye. [21:00] thumper, I'm bringing up skype [21:00] jml, I had been thinking about that for some time. [21:01] rockstar, getByUrls or the screencast? [21:02] jml, the screencast/howto about exposing things through the API. [21:02] rockstar, cool. it's a good way to get people started hacking on Launchpad, I think. [21:09] jml, yea, it's pretty easy. I think the testing infrastructure for the API could be a little better though. [21:12] rockstar, patches accepted :) [21:13] jml, really? :) [21:13] jml, as I recall, it was more a philosophical debate that prevented patches from being made. [21:13] rockstar, oh well, philosophy accepted. [21:15] thumper, https://code.edge.launchpad.net/~mwhudson/launchpad/in-memory-launchpad-server/+merge/13613 [21:15] hello [21:15] that branch failed in ec2 test in the strangest way [21:16] but i guess you're talking about merge proposal pages in general :-) [21:17] mwhudson, hi [21:17] yes [21:17] the layout has something broken-ish [21:18] the "commit message: label and the actual message are jumbled together [21:18] abentley: skype? [21:18] beuno: yes, just one of the things I want to fix [21:18] thumper, was just about to ask you the same thing. [21:18] thumper, filing 2 bugs then [21:19] thumper: Do you see me? [21:20] thumper: I'll try hosting again. [21:20] I think thumper was the only one who couldn't hear any of us. [21:22] The benefit of Skype is that you don't need the T-Pain plugin. Everyone already sounds like him. [21:23] hello jml, I am considering tackling bug #401525, what do you think? [21:23] Bug #401525: Revision.getBranch is not source package aware

[21:24] al-maisan, I think you should ask thumper [21:24] ah, I see. [21:24] thumper, it is pretty easy to get counts of files and dirs in bzr [21:26] build https://lpbuildbot.canonical.com/builders/lp/builds/257 failed with: [21:26] OperationalError: could not write to file "base/289007/2691": No space left on device [21:27] I understand Gavin forced it once already .. how can I find out more about the status of that build? [21:28] al-maisan: That's a tricky one. The build slaves shut down shortly after a failure iirc, so it's difficult to know what's going on. [21:29] allenap: OK .. so what is the next step? [21:29] al-maisan: Something to check is if the db setup enables statement logging or not. I'll look into that in the morning. [21:30] al-maisan: If the instance can be caught soon after the failure, a losa can connect to it and see what's filling it up. I think. [21:30] allenap: I see. [21:30] al-maisan: If you see it happen, ask a losa immediately. [21:30] allenap: thanks -- will do. [21:31] +1 (+N, really) on jml's idea to make a screencast showing how to patch Launchpad to expose an API. [21:31] maybe I should say that to jml :-) [21:31] kfogel, glad to hear it. [21:32] kfogel, I was going to write an email to you & mrevell asking for help, but I got bored halfway through. :) [21:32] anyway, I'm going to rot my mind & teeth with video games and chocolate. [21:33] ok this is screwy [21:33] ./bin/test -vvct lp.code.model.tests.test_branchlookup.TestGetByLPPath fails [21:33] ./bin/test -vvc lp.code.model.tests.test_branchlookup TestGetByLPPath [21:33] passes [21:34] i guess i need to look at test_suite() functions [21:34] that is weird [21:34] jml: how would you feel about a db patch that adds file and directory counts to the branch table? [21:35] mwhudson, what does -c do? [21:35] rockstar: colorization [21:36] jml: I didn't think you were still here. Go do something healthy like play video games, but yeah, a walk-through of exposing APIs would be a huge enabling thing for people -- e.g., lfaraone could use it, thinking back to our conversation last night. [21:36] mwhudson, ah. I use -vvt to go all the way to a specific test by module all the time. I'm not sure why that would be failing. === salgado is now known as salgado-afk [21:38] rockstar: the difference is that if you say ./bin/test -vv module it doesn't load all the test modules [21:39] mwhudson, yeah, so you can specify the module, but not the test class. At least, that's what I've seen. Specifying -t means you can specify the class. [21:40] rockstar: you can also say -vv module class [21:41] mwhudson, ah, okay. So -t only loads SOME test modules then? [21:41] rockstar: no, other way around [21:41] -t loads all, then filters [21:41] without -t, it loads some, then filters [21:41] mwhudson, ah, okay. [21:46] * mwhudson has the feeling that he's going to be very annoyed when he gets to the bottom of this [22:05] circular imports :( === matsubara is now known as matsubara-afk [22:11] d'oh [22:13] it's really fucking weird [22:49] oh wow, found it [22:49] basically, my branch accidentally makes lp.code.model.branchlookup imported at test-loading time [22:50] and this module does adapter registration at import time [22:50] and test discovery time is too early for this to work [22:50] Adapter registration at import time? Ewww. [22:50] (in trunk i guess the module is only imported when the component architecture is set up and things work) [22:50] wgrant: i'm going to blame jml's hatred of zcml i think [22:51] Grok! [22:51] yes [22:51] Well, Martian. [23:08] wgrant: go on then, submit a patch to base launchpad on grok! [23:13] ME GROK SMASH THUMPER [23:26] mwhudson: has autocomplete on alt-/ stopped for you in emacs 23? [23:26] thumper: no [23:26] :( [23:26] thumper: C-h c M-/ says what? [23:27] M-/ runs the command pop-tag-mark [23:27] well that explains that i guess [23:27] it runs dabbrev-expand for me [23:27] gah [23:27] pop-tag-mark is on M-* for me [23:28] found it in my .emacs :(