[00:15] bac: do you need to move the test the LaunchpadpadFunctionalLayer because the file need the librarian? [00:15] * sinzui looks [00:16] sinzui: no, b/c i'm not really hitting the librarian. the form validation machinery should kick in and flag the missing input [00:16] wgrant: is that a permission problem? [00:16] didn't you say "Adding a file"...That is librarian [00:16] bac: Yes. [00:16] sinzui: but the form should never get there. [00:19] bac: It was accessible a couple of days ago, but wasn't last night and still isn't now. [00:19] wgrant: is there an open bug? [00:19] bac: No. Is it one? [00:20] * bac looks at permission settings [00:23] wgrant: that make no sense [00:23] bac: No private branches ATM? [00:24] And no easy way to track down the OOPS? [00:24] no [00:30] Ah, it's only a problem on edge. [00:30] bac: I do not think the cgi.escape is needed. base-layout escapes the page title using standard tales [00:31] sinzui: ok. i was following barry's example [00:32] bac: there is no structural call...and if there was a problem., all the labels in all the forms would be bad. [00:32] I think the form label and the page_title should be identical... what does one say file and the other say download [00:33] sinzui: i was being all mechanical. i didn't give it much thought. i'll change that. [00:33] i don't understand your "no structural call" comment. [00:33] I don't think about it either. I set the label to the page title, the page title is the label [00:33] page_title = label [00:33] wgrant: the page should be public. i don't see anything referenced on the page that is private. stumped. [00:33] launchpadFormView make it easy [00:34] sinzui: i think the wiki says some should append "in launchpad" in the page title [00:34] bac: There was an 1800 line revision making large changes to that page yesterday. I'm going to see if I can reproduce it locally. [00:34] bac: wgrant: This may related to the changes by thumper [00:34] bac: that is new then. Sorry [00:35] sinzui: Right, that's the merge I'm looking at. [00:35] sinzui: no, it needs to be sorted out. [00:35] bac: actually, if that is required, we should not change any pages...that must be done by base-layout [00:35] sinzui: so you have no ideas on the test failure? [00:35] * sinzui is not change 300 pages again [00:35] no, I looked at the request and see nothing happened. I do not thine the action is firing [00:36] sinzui: sorry, i'm being called to dinner. let me know if you have any thoughts. [00:36] sinzui: it was firing, because i got a code error earlier when i omitted the contenttype [00:37] it must be there, we assume it is, and so i got a keyerror [00:37] * bac off to eat. bbiab. [00:39] now? [00:39] * jml apologises [00:39] kfogel, earlier as in, now? === thumper-afk is now known as thumper [00:41] bac: you did not include all the information requires to get the error. You passed in some field, but none are in a bad state. Make the description an empty string to get an error [00:42] thumper: Would the 403ing of launchpad's +activereviews to unprivileged users (eg. me, or the unauthenticated user) be due to your super-branch yesterday? [00:43] wgrant: almost certainly [00:43] bac:remember that the web form is passing more fields, which is an opportunity for bad data. Consider playing with 'filecontent' if you really want a disaster. I think you can skip this though since the form is being tested else where (a unittest and story I think) [00:43] wgrant: file a bug [00:44] thumper: Thanks. [00:48] hello [00:48] are we out of testfix yet? [00:49] jml: should be [00:50] mwhudson, thanks. [00:50] jml: is there a bug "it's impossible to find out if we're in testfix" yet? [00:52] mwhudson, I'll file one [00:53] and link from the Glue page [00:53] thanks [00:56] done [01:16] sinzui: i expected to get an error because the file is not being passed [01:16] sinzui: but i will happily forge on [01:16] but the form clearly does not require it [01:16] please do. [01:17] sinzui: in the schema it is marked as required [01:17] in the web ui it is required [01:17] and what does the widget do.. or make it [01:17] s/or// [01:22] jml: woo for landing that branch! [01:25] sinzui: are we still using pagetitles.py in the 3.0 layout? [01:27] mwhudson: no [01:27] mwhudson: only as a fallback [01:27] thumper: hooray [01:27] mwhudson: prefer page_title in the view [01:27] thumper: when can we kill it? [01:27] mwhudson: we are slowly killing it [01:27] yay [01:27] mwhudson: people should be removing stuff as they migrate pages [01:27] at least I am [01:28] there is a bug that basically says "pagetitles.py is full of obsolete cruft" [01:28] probably not [01:28] my preferred option for fixing it is deleting it [01:28] but I added a comment to the top of the file [01:28] thumper: an oooold bug [01:28] that says DON'T ADD SHIT HERE [01:28] :) [01:28] * thumper heads for lunch an dcoffee === thumper is now known as thumper-fud [01:29] https://bugs.edge.launchpad.net/launchpad-foundations/+bug/1584 fwiw [01:29] Bug #1584: Tests don't check for titles of non-existent pages [01:30] mwhudson: Not officially. we want to remove it next cycle. base-layout preferred the view.page_title. once we have converted all pages, we need to check that pagetitles.py is empty [01:31] sinzui: cool, sounds like a plan [01:36] OOPS-1342ED479 [01:36] fine, i'll make my own url then [01:39] * mwhudson lunching [01:39] biab [02:26] hi mwhudson [02:26] flacoste: hello [02:27] mwhudson: any questions regarding the JS builder? [02:27] flacoste: the js builder seems simple enough, i just need to torture the OSAs into making suitable images === thumper-fud is now known as thumper [02:28] ok, cool [02:29] flacoste: we can't use mars generate script directly, root ssh login to the slaves is turned off [02:30] what is the workaround? [02:30] flacoste: the osas have accounts in sudoers [02:30] flacoste: so i'll just make them do it by hand [02:30] can't we update the script to work with sudo? [02:31] so that the OSA simply have to run the script? [02:31] flacoste: it seems all the script really does is apt-get -y install xvfb firefox xfonts-base [02:31] mwhudson, yep, that's about all that is needed [02:31] mars: oh hello [02:31] hi [02:31] mars: shouldn't you be changing nappies or something? [02:32] just passing by [02:32] :) [02:32] hmm, ok, then it should be fine [02:32] I just changed two [02:33] flacoste, mars: should we run the jscheck builder after every landing or nightly? [02:33] mwhudson: nightly for now [02:34] mwhudson, be sure to have the slave monitored for memory leaks, too [02:34] part of the suite crashes in bad ways [02:34] mars: it'll run on ec2 [02:34] mars: i'm not sure what you mean? [02:35] by crashes? or 'our JS test suite may have memory leaks'? [02:36] mars: "be sure to have the slave monitored for memory leaks" [02:37] mwhudson, when I was running 'make jscheck' on my local system, part of the suite appeared to be segfaulting, or at least prematurely terminating Firefox and Windmill [02:38] mars: well, that will certainly cause build failures [02:38] heh [02:38] * thumper is fiddling with loggers for the base mailer [02:39] mwhudson, basically, someone needs to keep an eye on the build slave, to see that it doesn't die mysteriously. I would hate to see someone's time wasted tracking down issues, when a bit of forewarning would have saved them. [02:40] mars: ok [02:40] mars: otoh, i'm worrying about making it run [02:40] mars: making it pass gets the SEP treatment [02:41] SEP? [02:42] mwhudson, that will be great, I'm looking forward to it :) [02:43] mwhudson: any more questions, i'm about to head off afk again? [02:43] hey mars! [02:43] mars: Someone Else's Problem [02:43] flacoste: nope, go have an evening [02:43] thanks, talk to you next week probably [02:43] hi sidnei [02:43] mars: have you not read the hitch-hiker's guide to the galay books? [02:43] +x [02:43] mwhudson, I have === sidnei is now known as sidnei-away [02:44] oh look, there's mars [02:44] heya deryck [02:44] mars, hey hey [02:44] mars: the spaceship powered by bistromaths in life, the universe and everything is hidden using a SEP field [02:44] ah! [02:45] * deryck is out for the night. Until tomorrow y'all... [02:54] hi mars! [02:54] hi bac [02:54] how's it going? [02:54] mars: when are you back? [02:54] pretty good. Had a moment, so decided to drop by [02:54] bac, not for a month yet :) [02:54] :( [02:54] good for you, though [02:55] I'm using it to my best advantage [02:55] it still takes hours to get the day rolling [02:55] great. i'll talk to you later. was on my way out. [02:56] we have relatives over, playing with the boys, so I can actually sit at the computer for more than 5 minutes :) [02:58] good night! [03:00] mwhudson: I need some suggestions for testing my email problem [03:00] mwhudson: got a few minutes? [03:00] thumper: yes, but i'm in a cafe [03:01] thumper: i'll move to the library and skype you? [03:01] hmm... [03:01] ok [03:01] thumper: talk to you in a few minutes [03:01] * thumper nods [03:08] * mwhudson fights with skype a bit [03:10] ffs [03:10] thumper: have you used the new pulseaudio skype with a headset yet? [03:12] no [03:13] thumper, I has returned. Can we talk on the phone now? [03:13] oh ffs [03:13] mwhudson, abentley said you can't redirect the ringing the computer speakers and the actual phone call to the headset. [03:13] thumper, sorry. We don't have to have the call. [03:13] rockstar: just about to have a quick call with mwhudson [03:13] rockstar: that wasn't directed at you [03:13] rockstar: i'd be happy to have everything going through the headset [03:14] thumper, :) I figured it wasn't. Just wanted to be slightly comical [03:14] thumper: :/ [03:14] rockstar: doesn't translate too well over irc :) [03:15] thumper, yeah, but I was going for quotes page material, so it had to be taken entirely out of context. :) [03:15] * thumper stabs rockstar [03:21] thumper: i win! [03:21] thumper: still want to skype? [03:22] mwhudson: yes [03:23] thumper: sorry [03:33] thumper: Also, I think the faint vocals you were complaining about yesterday were because PulseAudio was routing from my laptop mic, not my headset mic. [03:37] thumper: it's gone quite [03:38] quiet even [03:38] :) [03:38] thumper: i could call your cell phone... [03:38] paramatise from above [03:38] mwhudson: I think I have enough now [03:38] thumper: ok [03:38] mwhudson: thanks [03:38] thumper: np [03:38] rockstar: call now? [03:39] thumper, yes [03:54] abentley: it sounded like that, actually [04:48] * jml is back [05:06] jml: ping [05:06] kfogel, pong [05:08] jml: hey, so re tomorrow's meeting (uh, trying to think if it's also tomorrow for you...) [05:08] kfogel, it would be tomorrow for me [05:08] jml: yeah, tomorrow morning [05:08] kfogel, Sat 8am. [05:08] jml: it's already 8am for you? [05:08] kfogel, no, that's the scheduled time for the meeting [05:09] jml: sorry, let's talk in absolute times. that was what I meant: "the meeting time we have set is already 8am for you?" [05:09] so, moving it earlier doesn't work unless we move it *much* earlier [05:09] kfogel, that's right. [05:10] jml: let me get my timeanddate.com on [05:11] jml: ok, you are 14 hours ahead of me [05:11] jml: it's 5:11am UTC right now, 12:11am for me, and 2:11pm (fri) for you. Right? [05:11] kfogel, yes. [05:12] * mwhudson should move to the cook islands or tahiti for extra confusion of this sort [05:12] jml: whew, wow. Okay, I guess we keep the original meeting time for this one. [05:12] mwhudson: don't you dare [05:12] mwhudson: but let's be glad we have no interplanetary developers [05:12] kfogel, yeah. thanks :) [05:12] oi, the lag :) [05:13] jml: you're thanking me for a Saturday 8am meeting? I don't even have to train you :-). [05:15] kfogel, heh [05:17] kfogel: imagine the lag on skype for someone in orbit let alone another planet :) [05:19] or someone on satellite internet, i guess [05:19] (which you would be on the cooks) [05:36] hmm [05:49] anyone around who does LP dev on karmic? [05:54] jml: me [05:55] wgrant, I'm now trying to get my laptop set up for karmic dev, rather than inside a chroot [05:55] wgrant, I hit a bug, but solved it by RTFMing :) [05:56] jml: What was the bug? launchpad-dependencies not installing due to a missing python-xml? [05:56] wgrant, I didn't have https://edge.launchpad.net/~launchpad/+archive/ppa in my sources.list [05:56] jml: Doesn't rocketfuel-setup do that for you? [05:57] wgrant, what makes you think I use rocketfuel-setup? :P [05:58] jml: Ah, I see. [05:59] gah, now postgres won't install [06:09] * jml reboots ... maybe that helps [06:13] ok that proved to be unnecessary [06:14] it was because I had postgresql on a ramdisk [06:19] hey [06:19] how do I get rid of a hardy chroot without blowing away my home dir [06:23] wgrant, have you got errors about not being able to import from Crypto.Cipher. [06:38] how do you escape $ in makefiles? [06:38] $$ [06:38] * thumper thinks [06:38] ah [06:40] * thumper is done [07:11] * mwhudson vanishes in a puff of hating EC2 [07:13] jml: I don't get such errors, no. Do you have a 2.4-compatible python-crypto? [07:13] jml: Karmic's isn't sufficient. [07:42] wgrant, I do now! [07:42] wgrant, I wonder why lp-dev-dependencies doesn't depend on it. [07:59] jml: It looks to me like python-crypto in ~launchpad/ppa is sufficiently 2.4-enabled. Did lp-dev-deps not pull that one in? [08:00] wgrant, apparently not. [08:00] or did you not run an upgrade after activating the PPA, just installing the new package? [08:00] it might have been that. [08:00] I just installed lp-dev-deps [08:00] Ah. [08:00] If you'd upgraded, it would have pulled the right one in. [08:00] I'll amend the wiki page. [08:01] rf-setup does it all for you. [08:01] wgrant, yeah, but... === abentley1 is now known as abentley [08:14] * jml blogoblags [08:15] What aboot? [08:17] wgrant, activereviews [08:18] wgrant, I have a queue of other things I'd like to publish, but I think I'll spread them out :) [08:19] jml: Bad choice of day. [08:19] wgrant, why so? [08:20] jml: launchpad's activereviews is borked. [08:20] on edge? [08:20] really? [08:20] Yes. 403 for unprivileged users. [08:20] grunk [08:20] I'll pull that one, and post the cranky one about layers then. [08:20] Oh. Fixed now. [08:21] \o/ [08:21] jml: hiya. Do let me know where your python-xml came from if you figure it out :-) [08:21] maxb, I lack apt-fu -- how can I find out? [08:21] well, dpkg -l python-xml will tell you the version [08:21] (10 years of using debian-based systems and I'm still a noob) [08:23] maxb, 0.8.4-10.1ubun [08:23] oh, it has chopped it to fit the column [08:23] dpkg -l python-xml | cat [08:23] will bypass the auto chopping to your terminal width [08:25] 0.8.4-10.1ubuntu2 [08:25] well that's the jaunty package. [08:26] I've no idea how you convinced a karmic system to install it without specific extra steps [08:28] umm [08:28] unless you still had it installed from upgrading from jaunty, and didn't let update-manager remove obsolete packages [08:28] good morning [08:29] maxb, well, I did some crazy things when I first upgraded to karmic [08:29] Hi adeuring [08:29] hi noodles775! [08:30] jml: I didn't actually say you needed python-xml; that was just my memory of why rocketfuel-setup failed for me on karmic. [08:30] wgrant, oh, my bad. [08:43] Was one of the 3.0 design goals actually 'eliminate absolutely all padding'? [08:44] :\ [08:44] I hope not. [08:44] Looking at https://code.edge.launchpad.net/~bac/launchpad/bug-421983, I think it was. [08:44] Look at that block of fields near the top. [08:44] wgrant, I've noticed the trend [08:44] jml: Ah, good. [08:47] Does LP's copy of devel suck, or does bzr? [08:48] wgrant, why do you ask? [08:48] jml: A checkout downloads around 250MB, and writes around 250MB to disk. A bzr pack takes it down to 110MB. [08:51] There's a couple of bzr bugs that were fixed very recently that might contribute to that. [08:54] The packs on bazaar.lp.net aren't quite so obscenely oversized, but they're still almost twice as big as they need to be. [08:55] it'd be nice to have a 'pack' button on Launchpad [08:55] (as well as an 'upgrade' button) [08:56] And a 'seriously fully remirror' button [08:58] oh yeah, that too. [09:00] Does LP's mailman do something silly like not send me a copy through the list if I'm in the To? [09:02] it would seeme so [09:02] annoying, that [09:02] Yes. [09:02] * wgrant files a bug. [09:03] http://paste.ubuntu.com/264827/ [09:03] maxb, do you know anything about that particular error? [09:03] ahhhh [09:03] jml: Just grab the right version from ~launchpad's PPA. [09:03] jml: You'll have to manually downgrade. [09:03] ahh ok. [09:03] It wants 1.5c IIRC. [09:03] Then it's happy. [09:03] Until it tells me that I don't have an image available. [09:04] Also I didn't bother repacking things for karmic that I couldn't find what they were needed for - and being that I can't ec2test, I didn't do anything with python-boto [09:06] wgrant, that works, thanks. [09:06] wgrant, have I filed a bug about getting a public image [09:07] jml: You haven't, I don't think. [09:07] you may want to copy the python-boto in the LP PPA to karmic series [09:07] maxb, I don't know how to do that? :) [09:07] jml: No, you haven't. [09:08] * jml does so [09:09] jml: Thanks. [09:10] maxb: That's not entirely useful, as it's an *older* version, so it would need to be explicitly installed. Better is to convince ec2test to work with 1.8. [09:11] +1 [09:11] how about I file a bug about that too [09:11] jml: A grand idea. [09:12] https://bugs.edge.launchpad.net/launchpad-foundations/+bug/424197 [09:12] Bug #424197: ec2test doesn't work with python-boto 1.8 [09:15] https://bugs.edge.launchpad.net/launchpad-foundations/+bug/424197 [09:15] Bug #424197: ec2test doesn't work with python-boto 1.8 [09:15] *ahem* [09:15] https://bugs.edge.launchpad.net/launchpad-foundations/+bug/424198 [09:15] Bug #424198: Public EC2 image for testing Launchpad [09:15] jml: Thanks. [09:16] 51 build-infrastructure bugs :( [09:16] mwhudson has quite a job. [09:17] we all do! [09:17] he's only BE for this month [09:17] True. [09:17] and personally, I don't see the role as an excuse to not fix the build system [09:19] argh. So remember I got a bunch of meta-lp-deps branches reassigned from ~maxb to ~launchpad? Well it looks like LP doesn't update the stacked-on URLs in the branches...... [09:19] fail [09:22] maxb, yeah, that's a known bug. [09:23] maxb, I can fix the broken branches for you, if you'd like. [09:23] (although my mana is running pretty low right now) [09:24] jml: Is ec2test happy now? It has probably not been run on Karmic before. [09:24] jml: lp:~launchpad/meta-lp-deps/{all but trunk} if you have time [09:24] wgrant, looks like it. [09:24] all of which are supposed to be stacked on trunk [09:24] wgrant, but I have a conflict [09:25] maxb, hmm. the easiest way might be to make a copy of trunk at the old location [09:25] there's only three though [09:25] Can anyone remind me of the runes for running a single story with bin/test [09:25] ? [09:25] <-- Person who has been doing lots of back-end work. [09:26] gmb: bin/test -vv -t the-test.txt [09:26] gmb, ./bin/test -cvvt foo.txt [09:27] in fact, I always use ./bin/test -1cvvt foo.txt [09:27] because the second failure never helps. [09:27] -1? -c? [09:27] wgrant, -c == color [09:27] * intellectronica doesn't even know what -c does [09:27] -1 I can now guess. [09:27] wgrant, -1 == show only the first failure for any given doctest. [09:27] Right. [09:27] (the other failures still exist, they are just hidden) [09:27] nice [09:30] I'm pretty sure it didn't work with the old zope.testing [09:30] jml, intellectronica Thanks. [09:30] That'll do nicely. [09:30] * gmb thought it was still different from running doctests [09:30] (e.g bin/test -vvf canonical $test [09:30] ) [09:31] gmb, I don't think you need -f ever. [09:31] Right. [09:34] I really want to see if I can use trial or nose or bzr selftest or something to run LP tests now that we've upgraded our zope.testing [09:36] or I should fix tribunal & double check that our subunit output works [09:36] that'd be fun [09:41] jml: would using those bring any advantages? [09:42] bigjools, yes. [09:42] bigjools, most of them are nicer, basically [09:42] bigjools, better error reporting or more fun or more extensible in some way [09:43] faster? :) [09:43] bigjools, unlikely [09:43] * bigjools thinks about porcine aviation [09:43] bigjools, maybe one of them is smarter about test discovery & searching... [09:43] bigjools, that'd save some time for local red-green-refactor cycles [09:44] indeed - the startup time for doctests gives me an earache [09:44] The world would be a better place if tests didn't depend on LibrarianLayer unnecessarily. [09:45] wgrant, earlier today I mentioned a cranky blog post about layers... [09:46] jml: So you did. [09:47] wgrant, basically, the bug is that our tests should declare what they actually need one by one, rather than name a layer that they hope contains everything they need. [09:47] jml: Right. [09:47] that'd be nice [09:47] Can the testing machinery do that? [09:48]