[00:20] lunchtime [00:21] mwhudson, I'm going to request a review of you in just a bit. You can get to it after lunch. [00:22] rockstar: okay [00:29] mwhudson: Cool. [02:06] barry: hey, hows the 2.6 going? [02:07] thumper: sort of good ;) i'm very confident we'll be on py2.5. we have an intermediate branch that is zope toolkit + 2.5 and if that goes well, we'll concentrate on ztk + 2.6. [02:07] cool [02:08] what's the zope toolkit stuff focusing on? [02:08] as in [02:08] what is different [02:08] ? [02:08] it gets us to a blessed combination of zope.* packages [02:08] blessed by zope corp? [02:08] or the zope community [02:08] what does it give us in reality? [02:08] anything useful I would use? [02:09] i guess it means things probably don't break in exciting and unexpected ways [02:09] i think that's right. also, we had to do it if there was any hope of getting to py2.6 [02:09] mwhudson: I'm having a weird lazr-js issue [02:09] barry: makes sense then I guess [02:10] barry: how much work is the zope toolkit stuff? [02:10] thumper: talking to anyone else is likely to be more use than talking to me :-) [02:10] thumper: a surprising amount. we're making good progress, but api's have changed [02:10] mwhudson: heh [02:10] mwhudson: I was going to file bugs, and then chase flacoste [02:10] thumper: good plan [02:40] How much ex-Z3 stuff does LP use from outside ZTK? [02:59] fuckity fuck fuck [02:59] mwhudson: I created a new branch and al [02:59] mwhudson: but the remote side is packs [02:59] mwhudson: and mine is 2a [03:00] mwhudson: lazr-js that is [03:00] thumper: doom, i think? [03:00] gah [03:00] * thumper makes a local packs repo [03:00] well, maybe a diff and apply it to a local copy [03:01] ah ffs [03:02] I can't branch it into local.... [03:02] nope [03:02] can I merge uncommitted from a different repo? [03:03] I could uncommit [03:03] then grab the changes [03:03] worth a try [03:04] bzr merge --uncommitted? [03:04] yep [03:04] works [03:04] Although that might care about richrootiness. [03:04] I don't remember. [03:04] ugly as sin [03:05] mwhudson: I'd file a bug saying lazr-js is still in packs format, but it'd just be me fixing it [03:06] thumper: most likely [04:09] mwhudson, thanks for the review. How long are you going to be about? [04:09] rockstar: probably quite a while yet [04:10] (when i stop working i really need to write my kiwipycon pypy talk so i'll be by the computer...) [04:10] mwhudson, great. I'll respond to your review in a bit. [04:11] * mwhudson boggles [04:11] does anyone recognise this error? http://pastebin.ubuntu.com/298762/ [04:12] oh nm [04:13] (i still don't know what that's about, but i don't think it's my fault) [04:41] mwhudson, so, switching to get_transport(self.branch.getPullURL()) gives me: NoSuchFile: No such file: u'/tmp/tmpSIhIst/lp-hosted/~person-name9/product-name4/branch11': [Errno 2] No such file or directory: '/tmp/tmpSIhIst/lp-hosted/~person-name9/product-name4/branch11' [04:41] rockstar: you need to pass hosted=True to create_branch_and_tree i think [04:42] mwhudson, ah. So my test didn't ever really work. :/ [04:42] well it did [04:43] just the understanding about where the branch is was a bit off [04:56] mwhudson, so, when you lock_write, and then move the .bzr branch, you get: Exception exceptions.UserWarning: in > ignored [04:57] rockstar: yeah, i wondered if you might [04:57] rockstar: i guess we should talk to a bzr developer about this [04:57] mwhudson, okay. lifeless? [04:58] * thumper EODs [05:19] ? [05:19] rockstar: you'll need to unlock the branch object [05:19] rockstar: and yes, there will be a race condition [05:19] lifeless, well, yeah, but we want to lock it so we can replace it. [05:20] lifeless, actually, I could probably get away with unlocking it right before rename('.bzr', 'backup.bzr') [05:21] yes [05:21] thats what I mean ;) [05:21] lifeless, gotcha. [05:21] it is a race condtion [05:21] filesystems need to grow transactions already [05:21] but its tolerable small [05:21] upgrade has a similar one [05:22] you can't lock stock and barrel replace the locking mechanism [05:22] and have it keep locking [05:22] the bootstraps are not large enough [05:22] lifeless, yeah, it seems like the time between unlock and move is pretty trivial. [05:36] mwhudson, updated mp - https://code.edge.launchpad.net/~rockstar/launchpad/branch-upgrade-out-of-place/+merge/13747 [05:37] I apparently also accidentally voted on my own proposal. [05:52] rockstar: the question then becomes - did you vote yay, or nay? [05:52] spm, obviously, I voted yay. I'm my own best friend. [05:52] just checking. might have been .... interesting otherwise. ;-) [06:23] is the puller/mirror/scanner failing at the moment ? [06:23] https://code.edge.launchpad.net/~lifeless/bzr-builder/blocking has been pushed too but claims nothing [06:30] lifeless: graph looks ok to me, spm could check more closely though [06:30] looking... [06:35] interesting. the puller* is working fine afaict; but I can't find any matches in it's log for 'bzr-builder/blocking' [06:35] well its not even showing 'updating this branch' [06:35] and only 1 match for 'bzr-builder' [06:35] :!bzr push lp:~lifeless/bzr-builder/blocking [06:35] No new revisions to push. [06:36] :!bzr info lp:~lifeless/bzr-builder/blocking [06:36] Repository branch (format: unnamed) [06:36] so there is data there [06:36] * lifeless tries hitting it harder [06:36] a bigger hammer usually works - vs just harder. ?? [06:37] no change [06:37] can you manually trigger a mirror? [06:38] not sure I understand your meaning? to manually trigger? [06:38] the scanner and puller [06:38] for this branch [06:38] mwhudson: any ideas about how to recover [short of rming the branch] [06:39] lifeless: something that will lock and unlock it [06:39] lifeless: (i don't think push with no revisions to push does that) [06:39] it does [06:39] we take a lock before finding out there is nothing to push [06:40] otherwise race conditions [06:40] still, let me do it manually [06:41] done [06:41] mwhudson: ^ [06:42] lifeless: no idea, sorry, have to plead nearly-19:00 [06:42] fair 'nuff [06:44] oh I see [06:44] tickled the 'sometimes we push a checkout' bug [06:49] Good morning === ara_ is now known as ara === mthaddon` is now known as mthaddon [08:02] what's the easiest/quickest way to get an empty storm result set? [08:04] found it: from storm.store import EmptyResultSet [08:19] good morning [08:20] Good morning adeuring [08:20] hi al-maisan! [09:15] Morning! [09:33] good morning === danilo-afk is now known as danilos [11:01] Morning, all. [11:05] deryck, good morning. [11:13] yo deryck [11:27] Is "OOPS" an acronym? [11:31] operational ongoing problem symptom [11:34] intellectronica, you're making that up. [11:35] yes [11:35] any better suggestions? [11:35] no, that one is pretty good. === deryck is now known as deryck[afk] [12:09] I wonder why we capitalise it, then === matsubara-afk is now known as matsubara [12:10] If anyone knows, matsubara will know if OOPS is an acronym :) [12:10] hehe [12:10] The rest of the characters are full-height, so it would look strange if it wasn't allcaps. [12:10] that's a good one intellectronica [12:11] i think it's capitalized because it's SERIOUS [12:12] :) [12:12] Perhaps it's Portugese [12:12] Obridgado Oh Problemo Superioro [12:16] mrevell, I think it's partly because 'OOPS' in the code itself is capitalized. [12:16] mrevell, so we talk about OOPS Reports, rather than Oops Reports. [12:17] mrevell, IIRC it's not capitalized in the code. [12:18] jml: I think you've contradicted yourself but maybe I've misunderstood. Either way, if it's not an acronym I don't need to explain it anywhere :) [12:18] mrevell, oh sorry. [12:19] mrevell, we have OOPS codes and we also have source code to handle OOPSes. [12:19] Oooh, right. [12:19] in the later, I think it's OopsHandler rather than OOPSHandler, for example. [12:19] I see [12:20] but then we also say getUrl [12:20] (blech) [12:23] mrevell, it's not an acronym [12:28] thanks matsubara === mrevell is now known as mrevell-lunch === salgado-afk is now known as salgado [13:03] i guess it's "OOPS, something went wrong" :) === mrevell-lunch is now known as mrevell [13:39] sinzui: what if the release notes on the download page was a collapsible field? [13:40] sinzui: i like that as it would give symetry with the changelog [13:40] and would address your concern [13:41] bac: maybe. How large is the bz/+downloads page now, and how large will it be. Are we making something too big to Au and China to load? [13:42] * bac looks === matsubara is now known as matsubara-afk [13:43] bac: I am off to my appointment. [13:43] sinzui: ok. stand up delayed until your return? [13:44] or i'm glad to host [13:45] bac: you may host. I have meetings at 10 and 11 today. [13:45] ok [13:45] bac: I landed my +ubuntupkg fixes. I am starting on bug 193469 [13:45] Bug #193469: Unlink (Ubuntu) package from series === matsubara-afk is now known as matsubara === gary_poster is now known as gary-sprint [14:04] I am trying to optimise a page and see quite a few of the following queries: [14:04] SELECT ValidPersonCache.id FROM ValidPersonCache WHERE ValidPersonCache.id = %s LIMIT 1', (2116775,) [14:04] Can somebody explain what these are? [14:05] Probably a question for the LP foundations team? gary-sprint maybe ^^ [14:10] al-maisan: I'm afraid I don't know. Stuart is back tomorrow. Asking salgado :-) [14:11] al-maisan, most likely from IPerson.is_valid_person [14:22] OK, thanks === deryck[afk] is now known as deryck [15:03] jml, sinzui: coming to the TL meeting? [15:03] beuno: ^^^ [15:03] flacoste, yes, thank you [15:04] flacoste, yes. [15:04] oh crap [15:27] barry (or anyone), random launchpad question [15:27] i used to be able to do this but have forgotten how [15:28] i'm trying to test a new version of launchpadlib. right now it can only be tested within the context of launchpad [15:28] actually, nm. i bet i have to modify buildout.cfg now [15:28] i was trying to change the symlink in sourcecode [15:33] oh! that reminds me. [15:34] leonardr, I'll review your branch. [15:34] jml, thanks [15:38] leonardr, do you think it would make sense to have an exception in lazr.restfulclient that signals "permission error" [15:39] leonardr, the launchpad coding standards warn against raising ValueError, IIRC. [15:39] jml: a "permission error" would logically be a 401 response from the server [15:40] in theory these two problems could have the same exception but i question whether it's worth it === MTecknology is now known as MTeck-ricer [15:41] * jml looks at the context [15:42] leonardr, a 401 response is raised as an HTTPError, right? [15:42] yes [15:42] ta [15:42] this error "you're trying to access something that's not a real url. but i happen to know what this non-url string is, and i can tell you what you're doing wrong" [15:42] *nod* [15:43] that probably comes under the aegis of ValueError then [15:43] I really don't like raising ValueError though. :) [15:43] i can make up an exception [15:43] that's ok. I can't enforce my aesthetics all the time :) [15:43] RedactedValueAccessError [15:44] Chex, bigjools, gary-sprint, danilos, rockstar, sinzui, allenap: LP production meeting in 15 min @ #launchpad-meeting [15:44] I hate Thursdays [15:44] matsubara: thanks [15:44] bigjools: why, my day consists of only 3 calls and one meeting :) [15:45] danilos: quite :) [15:45] leonardr, that'd be great, thanks. [15:45] leonardr, I've approved it with options for you to make that change. [15:45] leonardr, thanks again for doing this patch. [15:46] ok [15:46] jml: Question: If I use ec2 land on a merge-proposal targeted at lp:launchpad, how do I ensure it lands on devel and not db-stable (or whatever lp:launchpad points at)? [15:48] gmb, you can't, I don't think. lp:launchpad points at db-devel iirc. [15:48] gmb, it'd be fairly easy to add an option to override the MP though. [15:48] jml: Ah, so I'm better off just using ec2 test to submit the branch, then? [15:48] deryck, flacoste sent you an email a little while earlier about your

parser [15:48] gmb, exactly [15:48] jml: I'll file a bug. [15:48] gmb, use --dry-run to assemble the commit message, perhaps. [15:48] i can help you do things the right way. let me know when you want to talk it through [15:48] jml: Ah, yes, good idea. Thanks. [15:48] gmb, np. [15:49] leonardr, yes, indeed. let's talk. chatting with flacoste about this now. then we can talk. [15:49] ok [15:56] danilos, I was talking to someone about translating documents the other day [15:56] danilos, and they told me that some people break docs up into strings and then use Launchpad to translate them. Is this true? [15:57] jml: cool, what kind of documents? [15:57] jml: of course they do, I've even written one tool to do that that is used by GNOME and ubuntu-docs (xml2po) [15:57] danilos, wow. [15:58] danilos, I hope this isn't a silly question, but why not provide a way to translate documents in Launchpad without doing that? [15:59] jml: heh, time and effort involved, that's the only reason :) [15:59] danilos, it's a good reason :) [16:00] jml: fwiw, I am spending some of my time to make the mess that is xml2po today (it is python, but with very lousy high level acceptance tests) a much more nicely modularised, so we should be able to reuse it in Launchpad at some point in the future :) [16:00] jml: but, I wouldn't hold my breath just now :) [16:00] danilos, cool :) [16:00] danilos, and don't worry, I won't. [16:02] jml: there's other interesting stuff there, like handling images and such... many people are interested, and Kyle (from OEM) has started a doctemplate project to make a template translatable documentation project ("doctemplate" in LP) and we'll be discussing it during UDS (we already did some bits during our translations sprint) [16:10] allenap: What's the deal with ec2 giving me import errors about PQM? (ISTR you mentioning this before; can't remember when or why). [16:13] Ah, it's to do with python2.4, isn't it. [16:13] Rats. === salgado is now known as salgado-lunch [16:34] bac: sinzui assigned me bug 458169, which is very similar to your bug 455812. I was wondering if you had any more information than what I commented on in my bug. [16:34] Bug #458169: Distroseries:+index page timing out [16:34] Bug #455812: distroseries milestone timeout [16:35] can somebody give me a clue how to run the inline tests in lp/testing/__init__.py , and only those test? === beuno is now known as beuno-lunch [16:35] EdwinGrubbs: Bac: I am very confused by the fact that this is a 3.0 problem given that we did not update the code for distroseries. Both pages are generally changes to the position of the portlet [16:36] EdwinGrubbs: i have not yet looked at it. we can talk after i return from lunch [16:36] EdwinGrubbs: Bac: I am sure that when we make one query for the distroseries, all will be better [16:36] bac: sure [16:36] adeuring1, what do you mean "inline tests"? [16:37] jml: the tests in docstrings, e.g., in time_counter [16:37] I want to add a little function there, but that one should be tested ;) [16:37] adeuring1, I don't think there's a way to run the tests in time_counter [16:38] adeuring1, personally, I think it would be a good idea to add lp/testing/tests/ and put in some unit tests [16:38] EdwinGrubbs: I was able to load older Ubuntu series and milestones. There may be some clue in that [16:38] * jml regrets not writing the factory with unit tests to begin with. [16:38] jml: you might be right... But I like the idea that you can at the same time doument and test [16:39] adeuring1, sure. in which case, I think you'll still need to add lp/testing/tests/ I think and then add a test_doc to that which scans lp/testing/__init__ and runs the tests there. Gimme a sec to confirm. [16:42] adeuring1, http://docs.python.org/library/doctest.html#unittest-api -- you need to have something like the code snippet there in a test_suite() method in a file with a name starting with 'test_' [16:42] adeuring1, you only want to build a suite object and return it. [16:42] jml: thanks! [16:51] adeuring1, my pleasure. [16:52] EdwinGrubbs: flocoste pointed out that distroseries.currentseries can cause the problem. That is not a good method--it is taking a guess. We can either fix it or use a cachedproperty in the view. I favour a lower fix because this is a reoccurring issue === henninge_ is now known as henninge === matsubara is now known as matsubara-lunch [17:11] abentley, rockstar ping [17:11] barry, pong [17:11] barry: pong [17:12] abentley, rockstar: hi. i'm working on our python 2.5 + zope toolkit branch and we're seeing a ton of failures in the codehosting smoke test. stepping through the code, i see this output in the bzr subprocess that gets called: [17:12] http://paste.ubuntu.com/299140/ [17:12] abentley, rockstar any clues would be greatly appreciated! :) [17:13] barry, abentley would probably know more about that than I. [17:13] rockstar: cool === salgado-lunch is now known as salgado [17:13] barry: I think the warning on line 2 can be ignored. [17:14] abentley: k [17:14] barry: I'm not sure what the one on line 7 would be caused by, but it's in our VFS implementation. [17:16] abentley: any ideas where i should start to look? [17:16] barry: lib/lp/codehosting === EdwinGrubbs is now known as Edwin-lunch [17:17] abentley: does this call xmlrpc? [17:18] barry: Yes. [17:18] abentley: hmmm [17:21] barry: It looks like the message is due to raising InProcessTransport in lib/lp/codehosting/vfs/transport.py [17:23] barry: This may be an error in bzr's error reporting, rather than the underlying bug. [17:24] abentley: otp, thanks, will respond shortly === beuno-lunch is now known as beuno [17:32] abentley: okay. can you think of any issues related to python 2.5 or 2.6 here? [17:32] abentley: i guess the question is: if this doesn't fail in py2.4, what would make it fail in 2.5? and would it still fail in 2.6? [17:34] abentley: do you mean that there's some other failure going on that is being masked by this? and if so, can you think of a strategy to debug it? [17:36] g'night all [17:37] night jml [17:43] barry: I'm not aware of any issues with 2.5 or 2.6 on the bzr side, so I would expect it's a codehosting or twisted issue. [17:44] barry: You can debug it by changing the sites that raise InProcessError to return '' instead. [17:46] barry: There should also be logs / oopses for the twisted server, so it would be worth hunting those down. [17:46] barry: You can test it using "make run_all". [17:48] barry, quick question: can we let people unsubscribe from the ML archives? or is that getting too much into perl? [17:48] abentley: thanks for the tips. i'm going to look into this more after lunch [17:49] beuno: you mean an "unsubscribe" link in the archives? i think you're deep into perlworld [17:49] barry, I suspected as much, thanks [17:54] * sinzui wonders if the QA fairy will visit sinzui's landings === danilos is now known as danilo-afk [17:57] barry: you make need to make use of utilities/make-lp-user to get a working ssh key for bzr branch pushing [18:04] bigjools, ping [18:05] beuno: I am about to leave for the day, is it quick? [18:05] bigjools, it is [18:05] okidoki [18:05] with private PPAs [18:05] I see the deb lines ifI don't have permissions [18:05] ah yeah there's a bug about that [18:05] and need to go to another page to see the lines with the tokens [18:06] why don't we just hide them if you don't have permissions, and show you the debs with the tokens if you do? [18:06] why do we have a separate page? [18:06] it's a page that shows all your tokens for all subscriptions IIRC [18:06] so has a slightly different use [18:07] but yes we can put the specific one in the PPA page [18:07] perfect answer [18:07] last one [18:07] * bigjools knows all about beuno's "quick" things [18:07] in what situation can I access a private PPA, but not have permissions? [18:07] heh [18:08] beuno: there should be no situation like that [18:08] bigjools, can I have permissions but not have a token yet? [18:08] yes [18:08] I think [18:09] yes [18:09] bigjools, ok, good enough, thanks [18:09] barry, ping again [18:09] private mailing lists [18:09] are they exposed on the team's page, but the archive blocked for users without permissions [18:09] beuno: actually I am talking tripe [18:09] or completely hidden? [18:10] bigjools, I will assume you mean nonesense [18:10] beuno: you can't see the web page at all, ever, with a subscription, and you need to get a token to see the repo [18:10] however we want to allow users with tokens to see the new user-focused page [18:10] s/tokens/subscriptions/ [18:10] and now I really have to go [18:10] good night! [18:11] bigjools, so if I don't have permissions, i never even see the link to the PPA? [18:11] you don't [18:11] bigjools, thanks. [18:11] go before I think of something else === matsubara-lunch is now known as matsubara [18:18] night all [18:33] abentley, ping [18:33] rockstar: pong [18:36] abentley, so, I'm writing this script to run branch upgrades. How do I run the cronscript? I assume there's some fancy helper here. [18:37] rockstar: The current state of the art is in update_preview_diffs.py [18:38] abentley, yeah, I used that as my template, since it seemed to be the most up to date. How does it actually get run though? [18:38] rockstar: That is, subclass JobCronScript, specifying config_name and source_interface. [18:38] rockstar: It is run by cron. [18:38] abentley, yeah, so how do I run it to test it? [18:39] rockstar: bin/py cronscripts/my_script_name.py [18:39] abentley, ah, okay, cool. [18:40] rockstar: You'll probably want to specify LPCONFIG=development as well. That will force it to use the development config files instead of the production ones. [18:41] abentley, okay, thanks. [18:41] rockstar: But of course, you then have to create development config files. [18:41] abentley, yup, chasing that now. [18:42] sinzui: thanks i wonder if that's it [18:42] beuno: correct. the archives are blocked by our openid plugin [18:43] barry: I found that I could not connect to the ssh server because changed the user name of our leader to ~mark [18:44] barry, but you can see the link to it? [18:46] beuno: well, no because if you're not a member of the team, you shouldn't even be able to get to the team page [18:47] barry, so you can't have a private mailing list in a public team? [18:47] if you can see the link, you can see the archive? [18:47] beuno: no, currently not though there's a bug open on that (a use case i agree with) [18:48] barry, cool, thanks. I just wanted to know if you could ever get in a situation where you clicked ona link that you could not access [18:48] beuno: i don't think so [18:48] beuno: well, modulo openid bugs of course :) [18:49] barry, that clears it up for me, thanks [18:49] cool [18:51] beuno: I am looking at the SP page for a packaging link that needs to be deleted: https://edge.launchpad.net/ubuntu/hoary/+source/jedit [18:53] sinzui, I'm having a hard time understanding what that is [18:53] or how you ended up there [18:54] beuno: 1) "Upstream release series" appears to be redundant with "Upstream associations", 2) I am surprised that in the https://edge.launchpad.net/ubuntu/+source/jedit, we do not ask the user to confirm the removal of the link. 3) I expect the unlink action to look like (-) Remove packaging link. [18:55] beuno: you can get the the SP page from the productseries index or +packages (https://edge.launchpad.net/jedit/+packages) [18:57] beuno: you may be familiar with this: https://edge.launchpad.net/bzr [18:57] ^ The extra edgy link is a data error. I want to allow users to remove the bad link to the SP. [18:57] sinzui, so how do upstream associations happen? [18:58] are they inferred from series linking? [18:58] https://edge.launchpad.net/bzr/trunk <- (+) Link to Ubuntu package and (+) Link package [18:59] wow, I'm confused [18:59] why do we have (+) Link to Ubuntu package *and* (+) Link package? [18:59] beuno: not inferred, explicitly set, though SP links cannot be deleted, only DSPs. Since there will never be a DSP because this is an error, it is not possible to correct the mistake. there are 81 broken SPs in production [19:00] sinzui, so it sounds like you need to delete the link when the DSP are deleted [19:01] beuno: that is point 4) Should re rename (+) Link to Ubuntu package to link the the current Ubuntu series? It is takes most of the thinking out of the task. The other one can link to any distro and you can specify primary or include package relationships [19:01] beuno: the DSP is a binary. there is not any code, so the DSP will not be made [19:02] sinzui, so one is for Ubuntu, and the other one is for any distro? wouldn't it make sense to have one page? [19:02] (or widget) [19:02] I tried and failed [19:02] sinzui, code-wise or UI-wise? [19:02] I can help you with one of thse [19:03] *those [19:03] beuno: I think I need a lot of work to reconciile the history operation that is not present in the any packaging operation. we do not know Fedora's history [19:04] sinzui, so one is to link 2.0 <> karmic [19:04] and the other one is 2.0 <> bzr2.0~ubuntu1? [19:06] beuno: one is for the Ubuntu who cares about what they are doing right now, the other is really for projects who care about supporting any project [19:07] sinzui, and what's the benefit (or use case even) of the latter? [19:08]