=== abentley1 is now known as abentley [00:58] lifeless, have you seen the recent T-I-P post? [00:59] lifeless, there's a python-ideas thread on introducing setUpClass to the stdlib [01:01] jml: noooooooooooooooooooooooooooooooooooooooooooooooooooooo [01:02] yeah, I know. [01:02] I have to follow python-ideas now :( [01:02] well [01:02] isn't it basically a write-only mailing list? [01:03] yeah, but Michael Foord kind of likes the idea, from scanning. [01:04] http://article.gmane.org/gmane.comp.python.ideas/6715 [01:04] I'm not sure how to actually reply. [01:05] since I'm not subscribed [01:05] oh dear [01:15] jml: maybe a small tactical nuclear weapon [01:16] * mwhudson stabs the interwub [01:25] jml: I've seen it, no time to fire appropriate responses. [01:25] jml: the argument I'd present is that class is the wrong scope. === jamalta is now known as jamalta-afk === jamalta-afk is now known as jamalta === jamalta is now known as jamalta-afk [02:04] ok. replied. [02:05] to python-ideas? [02:06] jml: Um, I'm getting that LocationProxy thing now. [02:07] mwhudson, yes [02:07] wgrant, change launchpad.Admin to zope.Public in the zcml (or merge my branch again) [02:07] wgrant, this is for adapters specifically [02:07] jml: Ew. OK. [02:09] jml: I suspect that this is symptomatic of a deeper issue, though... [02:09] wgrant, yes. [02:09] wgrant, BjornT understands [02:10] Ah, good. [02:10] So I'll fix that in my earlier branch, then. [02:12] Actually, I might as well just drop the permission entirely if it's in yours. [02:38] wgrant, yeah. it's probably worth merging mine in again. [02:41] does anyone know how I can use buildout to manage a library that includes Python stuff but is built with autotools? [02:54] jml: this might help: http://pypi.python.org/pypi/zc.recipe.cmmi [02:57] BjornT, thanks! there's also a bug filed on subunit https://bugs.edge.launchpad.net/subunit/+bug/499775 [02:57] Bug #499775: build easy_install/pip installable packages at release time [03:18] * mwhudson has an impossible failure in his bzr-2.1c1 integration branch [03:19] :( [03:22] Do people get spammed if I create an in-progress MP? [03:23] i think currently yes [03:23] wgrant: why don't you ask halfway through thumper's talk, coming up next? [03:23] :) [03:38] jml: Lots of conflicts merging from you :( [03:39] :( [03:52] jml: Want to review my hopefully final non-slave branch? [03:52] wgrant, sure. [04:12] wgrant, I've done the review (although not changed the status) [04:13] jml: OK, thanks. I don't have an email yet, oddly. [04:14] wgrant, let's see if we can get it approved before thumper's talk ends [04:14] ... where did you ask the questions? [04:14] I see no questions. [04:14] (I honestly can't believe that he chastised people for not having good cover letters!) [04:14] wgrant, via email [04:15] Ah, it's there now. [04:16] argh [04:23] lifeless, ? [04:25] p-i === abentley1 is now known as abentley [04:54] BjornT, that recipe thing hurts my brain. [08:57] stub, hi, you around? [08:57] like a donut [08:57] stub, :) [08:58] stub, I accidentally landed a branch on devel that I wanted to land on db-devel instead (don't want it rolling to edge). Can I have you rs to revert it on devel and re-land on db-devel? [08:58] Sure [08:58] stub, cheers. [09:08] Morning [09:22] Hi all! [09:23] I have updated the launchpad-dependencies package in the launchpad ppa for karmic. [09:23] https://dev.launchpad.net/LaunchpadPpa tells me that I should now copy it to the other series on +copy-packages. [09:24] Why is intltool a dependency now? [09:25] never mind [09:25] wgrant: for the automatic template generation code. [09:25] I suspected that. [09:25] But that's wrong, then. [09:26] intltool is run inside the chroot, on the slave. [09:26] wgrant: oh [09:26] Not outside the chroot, on the master. [09:26] wgrant: ah yes, but the code has tests in the test suite. [09:26] wgrant: I could not land my code so far because the test won't pass on ec2 [09:27] wgrant: what other option could I have? [09:27] Hmmm, I see. [09:28] I guess that's not too unreasonable, then. [09:28] wgrant: maybe we need to add "lp-slave-dependencies" and "lp-test-dependencies" ? [09:29] to get this more fine-grained [09:29] btw, I managed to copy the ppa to all series now [09:30] It's not a slave dependency; the script needs to install it itself. [09:30] I think lp-test-dependencies' purpose is fulfilled by lp-dev-dependencies, but I'm not quite sure of that. [09:31] wgrant: you mean that the ec2 test code pulls in lp-dev-dependencies? [09:31] then I guess that would suffice [09:32] I believe so. [09:32] * henninge wonders where to find that out. [09:32] ec2test images do, but I'm not sure about buildbot. [10:31] BjornT: So I'm creating a stub in scripts/page-performance-report.py which bootstraps code in pageperformancereport.py. Where should pageperformancereport.py live. [10:32] jml: That ec2 test seems to have fallen into an abyss. This time it hasn't landed or emailed. [10:36] wgrant: hey [10:38] bigjools: Hi. [10:38] what's up with not landing aaron's branch? [10:38] bigjools: You suggested that it shouldn't land until it's all been well tested on dogfood. [10:38] Which, AFAICT, it has not. [10:39] And it doesn't actually matter much whether the slave is merged... [10:39] ok [10:39] well it doesn't matter if it is, either, because it's rolled out separately [10:40] Right. [10:40] actually do you know if the 58~1 image had these changes in? [10:41] because these are 58~0 [10:41] and I tested 58~1 on DF [10:41] 'these are'? [10:41] I think 58~1 contains the latest recipe changes. [10:41] these changes [10:42] so the MP has the changelog bumped to 58~0, so I am guessing I have tested that branch [10:42] So they are. I wonder if that last increment just didn't get committed. [10:43] Anyway, I need to sleep. [10:43] I don't see 58 in the committed code [10:43] it's up to 54 [10:43] so who has ~1 [10:47] https://code.edge.launchpad.net/~abentley/launchpad/build-recipe [10:48] * wgrant 's internet connection is about to die. [10:48] That doens't have the ~1 changelog entry, but it is the latest code. [10:48] hmm ok [10:48] so wtf did the ~1 come from! === abentley1 is now known as abentley === mrevell is now known as mrevell-lunch [12:58] good morning === mrevell-lunch is now known as mrevell [13:22] morning beuno [13:31] BjornT, around? Trying to investigate the windmill speed issue as well === joey is now known as Guest44127 [14:16] BjornT, around yet? [14:16] or salgado? [14:16] mars, ? [14:17] hi salgado. Just looking through the css_combine makefile target, to look at removing Mochikit from the JS rollup [14:18] salgado, for excluding an individual file, I'm guessing an exception in buildout-templates/bin/combine-css.in, correct? [14:19] mars, erm, you talking about css roll up or JS roll up? [14:19] salgado, nm, looking at the JS combination would help :) [14:20] mars, just remove Mochikit.js from the jsbuild target in the Makefile [14:21] salgado, ah, I see, Mochikit.js is explicitly included in the Makefile, in the jsbuild step. So excluding it there should kill it [14:21] ? [14:21] salgado, so I'll assume you agree :) [14:21] heh === Guest44127 is now known as joey === salgado is now known as salgado-lunch [15:36] Chex, gary_poster, rockstar, bigjools, sinzui, allenap: hi, are you able to join the prod. meeting that should happen in 30 minutes? [15:36] I am [15:37] Ursinha: I am not [15:37] gary_poster: do you have a replacement? :) [15:38] Ursinha: not yet :-) I'll try to arrange one [15:38] thanks gary_poster :) [15:40] Ursinha: yep === bigjools-afk is now known as bigjools [15:41] Ursinha: I'm sprinting, but I'll watch for pings, if that's okay? [15:45] allenap: sure, thanks :) [15:54] Chex, someone from foundations, jtv, rockstar, bigjools, sinzui, allenap: prod. meeting in 8 min. @ #launchpad-meeting [15:56] Has anyone done anything about the failed build on devel buildbot? === salgado-lunch is now known as salgado [16:01] Chex, someone from foundations, jtv, rockstar, bigjools, sinzui, allenap: prod. meeting now @ #launchpad-meeting [16:02] s/jtv/henninge/ :) [16:02] Ursinha: s/ indeed [16:02] matsubara-afk, come back, dude, you've got a meeting to run. ;) [16:29] Ursinha, is there a bug for https://lp-oops.canonical.com/oops.py/?oopsid=1479S1000 ? === herb__ is now known as herb [16:36] gary: i just discovered something odd [16:37] i can use type() to define a class whose name contains invalid characters like periods [16:38] not surprised :-) leonardr. on call, though [16:38] gary: np, just thought i'd mention it === beuno is now known as beuno-lunch [16:40] Ursinha: gmb and I are creating a new job-based cronscript to calculate bug heat, and we wondered if we needed to give it a specific oops_prefix? It's set to none in schema-lazr.conf for now. [16:42] salgado: I don't think so [16:44] allenap: matsubara would be the best person to ask [16:44] allenap: he's out today but might return later, I can ask him [16:45] Ursinha: Thanks. [16:46] allenap: no problem, sorry not being more helpful [16:46] Ursinha: As far as I'm concerned, that's very helpful :) [16:46] allenap: :) I meant now :) [16:52] Ursinha, the typeerror OOPS is bug 403281 [16:52] Bug #403281: public xmlrpc requests broken during read only period [16:53] it happened because mthaddon was testing the new read-only switch on staging [16:54] salgado: oh, I see [16:55] thanks for investigating [16:55] np [17:09] rockstar, ping [17:09] re: upgrade icon [17:09] beuno-lunch, pong [17:09] beuno-lunch, I was gonna ping you about that, I swear. [17:10] rockstar, the icon I had made before, wasn't *that* for upgrading a branch? [17:11] or do I need to find better quality crack? [17:11] beuno-lunch, you told me it was for when a branch already had an upgrade in process. [17:11] * beuno-lunch looks [17:11] beuno-lunch, I thought one with a red exclamation point where the green animated arrow is would be good. [17:12] rockstar, I'm worried that something like that conveys that something is wrong [17:12] like a warning [17:12] beuno-lunch, on another note, do you have balsamiq mockups for current pages? Something I can build off of? [17:12] and not hthe positive action of "Upgrade it!" [17:13] beuno-lunch, well, according to kiko, something is wrong. The branch needs upgrading. [17:13] beuno-lunch, personally, I think a big red button would be the best. "This branch is out of date and should be upgraded!" [17:13] right [17:13] so we can use the standard warning icon [17:13] "If you don't upgrade, we're going to come to your house and give you two slap." [17:14] "This branch is using an old format. _Upgrade now_" [17:14] beuno-lunch, essactly. [17:15] so no special icon for that, and use the current icon for "This branch is currently upgrading" [17:15] rockstar, will branches be read-only while upgrading? [17:15] beuno-lunch, nope. [17:15] rockstar, shouldn't they be? [17:15] beuno-lunch, nope, not according the mark. [17:15] beuno-lunch, this is the reason it's taken so long. [17:16] ok, as long as people don't get into trouble for using it the the meantime [17:17] so, we can decide wether to use the icon for "Upgrade now" or "Upgrading" [17:17] beuno-lunch, basically, it upgrades on the side, pulls in all the revisions that have been committed, and then replaces the branch. [17:17] maybe upgrading is better, although it will be rarely seen [17:17] beuno-lunch, yes, and then, wherever we use the icon, we need another icon for the other. [17:17] rockstar, I think we don't [17:17] beuno-lunch: see new screenshot at end of https://bugs.edge.launchpad.net/malone/+bug/506018 when you get a chance [17:17] Bug #506018: Need a "+patches" view: report lists patches attached to bugs. [17:17] rockstar, in fact, we may not need any icon at all. We could use the standard warning and info icon [17:18] beuno-lunch, okay. If we don't do you have a proposal? I'm installing balsamiq right now to make a mockup. [17:18] rockstar, as for balsamiq, in the balsamiq branch, I committed all the mockups I have [17:18] Where's the balsamiq branch? [17:18] rockstar, I don't think I have mocked anything up [17:18] * beuno-lunch looks for the branch and looks at kfogel's screenshot [17:19] kfogel, I worry that it may be hard to click on the overlay [17:19] is that actual code? [17:19] beuno-lunch: yeah, that's in our branch right now [17:20] beuno-lunch: may be hard to click on the link in the box? [17:20] kfogel, it's an overlay, right? [17:20] rockstar, https://code.edge.launchpad.net/~launchpad/launchpad/ui-wireframes [17:20] beuno-lunch: hey, one sec, we'll give you a live page you can play with, would that help? [17:20] beuno-lunch, great, thanks. [17:20] kfogel, that would be awesome [17:21] rockstar, feel free to commit to it [17:21] beuno-lunch, okay, will do. [17:21] so we can all share :) [17:25] beuno-lunch: does this load for you? http://www.red-bean.com/kfogel/canonical/patches-view-popup.html [17:25] I'll check [17:25] yes [17:25] no CSS though [17:25] so it breaks a lot :) [17:26] the CSS links to launchpad.dev [17:26] beuno-lunch: oh, heh, yeah. is it useable? [17:26] well, it is, but I can't tell if the overlay will work or not :) [17:27] beuno-lunch: (start up a launchpad.dev instance? you don't need it to be our branch, just to get the CSS loading) [17:29] beuno-lunch, in branch listings, should we also show a warning icon for out of date branches? [17:29] kfogel, I'll try that. Did you add any CSS? [17:29] rockstar, yes, but only for people who have permissions to upgrade it :) [17:30] beuno-lunch: no, just some javascript in the page itself (which you should get via the page, of course) [17:30] starting up an instance... [17:31] beuno-lunch: so, enjoying that lunch? :-) [17:31] beuno-lunch, okay. [17:32] kfogel, it's looking good on the table [17:33] tasty, but I can't say that for certain yet :) [17:33] making schema is slow [17:33] beuno-lunch: (there's a throughout-Launchpad problem whereby in a vertical list, a popup on one item blocks you from mousing down to the next row's item. It happens on ~person/branches page too, for example. It would be nice if the popup could pop up somewhere else other than the mouse pointer -- that's exactly the place where the popup is most likely to get in the way :-) ). [17:34] kfogel, yeah, that's why I asked :) === beuno-lunch is now known as beuno [17:35] I'll stop pretending I'm going to eat lunch [17:35] * kfogel hopes we get to fix that bug in the default popup behavior [17:35] beuno: we can wait, go eat [17:35] beuno: it's not like we don't have other stuff to do. [17:37] kfogel, I'd rather unblock you [17:37] kfogel, it feels ok [17:37] just add some padding between the borders of the bos and the content [17:38] and, it comes up when hovering the row, not the link [17:38] making it a bit suprising [17:38] and it's not intuitive, so if we can think of something that points to there being information, cool. Otherwise, file a bug for the future ;) [17:50] beuno: thanks. what's a "bos", btw? [17:51] beuno: for pointing to there being information, I was thinking: let's put that patch band-aid icon to the right or left of the patch age, and do the popup only on the icon? [17:51] s/bos/box [17:52] beuno: got it. we've fixed that already by using class="listing" [17:52] kfogel, I think that's a good idea. You can see that information on the page you click on anyway, so it's just a convinience thing to have the popup [17:52] beuno: based on Bryce's reactions, I think the popup may be a bit more than a convenience optimization. The users are expected to have a very much scan-and-only-tentatively-dive-in kind of workflow. [17:53] super [17:53] this is great work kfogel [17:53] beuno: thank abel, sitting next to me :-) [17:53] * beuno waves at abel [18:02] gary: the goal is in sight, i have one more zope problem i'd like to sort out with you when you have time [18:02] leonardr: cool. maybe on call soon? [18:02] whenever you're ready [18:03] it's code-intensive so i think starting out in irc would be better [18:03] ok [18:03] night! [18:08] it's code-intensive so i think starting out in irc would be better [18:09] is it me, or is launchpad having some issues right now? [18:28] mars, ping [18:28] hi rockstar [18:29] mars, where are we with the js issue? [18:29] rockstar, I am slicing the Mochikit code out of launchpad.js, with the intent on pulling the code in only on pages that need it [18:30] mars, is that going to land today? I have a rather urgent branch that is being held up by this issue. [18:32] rockstar, need to figure out context, but I can work with you to figure this out [18:32] rockstar, so you can't land your branch because windmill will go nuts and block everything if you do? [18:32] mars, apparently, yes. [18:32] ok [18:33] mars, if I run the tests that fail on ec2, they fail locally. However, if I run with -D, they all pass. [18:33] rockstar, two options: wait for me to finish excising Mochikit, or put your JS in a separate file that is not part of the rollup. I assume you are landing JS, correct? [18:33] eh? [18:33] rockstar, oh, a different issue from Bjorns then [18:34] rockstar, then lets work on your problem [18:34] mars, no, I think it's the exact same issue. [18:34] thumper and I were working on parallel branches, and we both saw the problem at the same time. [18:34] rockstar, can you successfully run the windmill tests in trunk? or a subset of them? [18:34] mars, yes. [18:34] interesting [18:35] rockstar, do a subset of the tests, such as a single test, fail? [18:35] mars, in my branch? Yes. [18:35] Unless I run with -D, in which case it passes swimmingly. [18:35] rockstar, and they fail locally? Tim and Bjorn can both run them fine locally. [18:35] afaik [18:35] I think it's all fishy. [18:35] lol [18:36] ok [18:37] mars, technically, the tests SHOULD pass, as they pass if I work through them interactively. The functionality is fine, windmill just pukes. [18:37] Nothing's technically broken. [18:37] ...'cept Windmill's ability to find out if it's broken. [18:39] why would *any* test you run on your system fail? [18:39] rockstar, do they fail because of... what? [18:39] please elaborate [18:40] mars, I can't. Windmill is just weird. [18:40] I don't know why Windmill thinks they fail. I am 99% sure this is the same issue that Bjorn and thumper are seeing. [18:40] rockstar, do the individual test assertions time out? [18:41] mars, I guess. It waits for elements that I can see show up, but it just decides they aren't there. [18:41] And it spikes my CPU. [18:41] ok [18:42] rockstar, if it is the same issue as Bjorn and Tim, then the workarounds I proposed still stand [18:42] mars, yes, but the workarounds aren't too much of an option, since I'm editing a file already in the rollup. [18:43] rockstar, what is the scope of the code? Page, App, or Core? [18:43] the scope of the code you added [18:43] Page. Branch merge proposal page specifically. [18:43] intellectronica, by any chance, is there any log of the people who say me too? dates they clicked? [18:44] bigjools: I presume that the ~1 changelog entry is an uncommitted change on abentley's laptop. [18:44] rockstar, then if you have to land it ASAP, you can probably split out the code you added into a new, non-rollup module. Or, you could include it inline on the page itself. [18:44] rockstar, the latter option works very well in a pinch [18:44] mars, so can I pull it out of the existing rollup, and how do I go about that? [18:45] rockstar, I assume you have a diff [18:45] since you changed something to trigger this bug :) [18:45] mars, yes, lemme find it. [18:45] wgrant: it could also be a version bump after we'd already installled ~0 as I remember them wanting to poke a tweak in [18:46] bigjools: Possibly. [18:46] Hmm, I think all the necessary master code is now in db-devel. [18:46] * wgrant tries a recipe build on that. [18:47] dogfood is pretty much up to date as well [18:47] I can try and poke something in there [18:47] mars, https://code.edge.launchpad.net/~rockstar/launchpad/update-review-table-on-comment/+merge/17439 [18:48] bigjools: Does it have a local codehosting setup? [18:48] I don't think it does. [18:48] anyone here use a Logitech Harmony remote and got the software running in wine? [18:48] rockstar, so comment.js is core: but this functionality in particular is only used on the MP page? [18:48] wgrant: why does it need that? [18:48] it can pull from staging [18:49] bigjools: Well, it references branches in the DB. [18:49] oh. cock. [18:49] mars, well, the one line in comment.js needs to be there. The other stuff should be moved to a lp-code javascript space. [18:49] bb in 10m [18:51] rockstar, is anything in comment.js used anywhere else? [18:51] ah ha! lp.CodeReviewComment! So I assume "no"? :) [18:52] mars, I think the bug comments use it as well. [18:53] rockstar, ok [18:53] Hm, what did I break in db_lp? [18:53] rockstar, this is a simple refactoring then: extract lp.CodeReviewComment to a new module. [18:54] rockstar, that new module should not be part of the rollup by default. [18:54] mars, where's the rollup code? This has all seemed like magic to me. [18:54] rockstar, don't worry about minification or any of that junk if you want it to land. Just refactor, test, and go. [18:55] mars, so I create a js file and then include it in the bmp page itself, and not in the main template. Is that correct? [18:56] sinzui: ping [18:56] rockstar, ah, wait a minute, looking at the rollup code: Makefile:124 [18:56] Hi EdwinGrubbs [18:56] rockstar, confused: why is there an '-s' argument, and then a bunch of positional args... [18:57] mars, hellifiknow. I don't know how this stuff works. [18:57] rockstar, maybe '-s' is "grab everything", and the positional args are cherry-picked files? [18:57] rockstar, just thinking out loud. mumbling doesn't work so well in text-based mediums... [18:58] mars, no, I see what you're saying. [18:58] mars, I always just thought that I put a js file in the folder, it gets rolled up. Apparently, that was a correct assertion. [18:59] ok [18:59] so.. hm [18:59] rockstar, jsbuild takes a '-x' option.... [19:00] mars, yeah, I saw that. I should probably add an XXX comment. === elmo_ is now known as elmo === EdwinGrubbs is now known as Edwin-lunch [19:02] rockstar, so refactor to extract the module, then use the -x option on it. [19:02] rockstar, first test the -x option with something benign, see if it does what you need [19:05] rockstar, pick any JS file in the rollup that you don't need on your MP page, pass it through -x, see that it is excluded. Then run windmill on one previously failing test: the new, smaller rollup should pass, right? [19:08] mars, I'll go on an adventure. [19:09] rockstar, if it isn't too much bother, could you please try the fix first? :) [19:09] mars, that's the adventure. [19:10] wgrant: so since DF is a copy of production from just before Christmas, it will have branches in the DB [19:21] I love seeing the patch icon in bugs [19:21] it's so much clearer now [19:28] damn. This stupid last bread crumb bug is killing my test [19:28] bigjools: I wonder how the config is set up, though. [19:28] * wgrant checks. [19:30] Hm, it might just work. [19:30] It looks like it doesn't override the production codehosting URLs. [19:31] So lp: URLs in recipes should Just Work. [19:41] beuno: i'm not sure i understand exactly what you're looking for. we don't save the dates in the db but i suppose we could mine it from the web logs if we really want [19:43] intellectronica, dates on the db [19:43] there's a bug someone filed and I commented on [19:43] sinzui triaged it [19:44] I am sure it is on the first page of malone bugs [19:47] jml: the setUpClass discussion is making me sad [19:57] wgrant: when you tested copying private<->ppa copies did you select source only? [20:01] bigjools: I tried both, I think. [20:01] * wgrant tries agao. [20:01] s/o/in/ [20:01] wgrant: the original problem was when the guy selected rebuild [20:01] dunno if that's a coincidence === ajmitch_ is now known as ajmitch [20:06] bigjools: The target was ~ubuntuone/ppa? [20:06] * bigjools plays hunt the oops [20:06] ~ubuntuone/beta, sorry. [20:08] * bigjools fails [20:08] I think it was from memory [20:09] lifeless, :( [20:10] wgrant, your revision looks like 8903 [20:10] Although I guess you probably had to destroy the history to fix the publisher. [20:10] bigjools, hi [20:11] jml: Yeah, it landed 20 minutes after I asked. Very, very late. [20:11] wgrant, late how, I wonder. [20:12] ~5.5 hours after you kicked it off. [20:12] And it doesn't look like it was stuck in the queue, unless PQM was landing to something other than (db-)devel [20:12] rockstar, let me know if you have any luck with the refactoring [20:15] wgrant, that's about the same time as the other one. [20:15] beuno, sinzui: yeah, saw the bug. i have no opinion about this right now, and i'll defer to deryck to decide how to triage and schedule it, but no, we don't have that now [20:16] jml: hi - I'm not really here [20:16] maybe the test suite has become three hours longer :( [20:16] bigjools, hi [20:16]