lifelessman, where did the day go00:57
mwhudsonlifeless: good question00:58
pooliehello lifeless, mwhudson00:58
mwhudsonpoolie: i've gotten a bit confused, sorry to say -- do you want me to land your mbp-trivial branch?00:58
pooliei'd like someone to do it01:01
pooliei don't think i have the right to land myself atm01:01
* thumper -> modaks for lunch01:01
lifelessare you in ~launchpad? if so you should modulo someone putting your key in pqm01:01
lifelesswhich is a simple losa request :)01:02
mwhudsonpoolie: i'll land it then01:02
* poolie looks01:02
pooliealso, landing is kind of a many-roundtrip thing01:02
poolieand it seems to mess up my "i'll spend one day on launchpad" plan01:02
lifelessyeah, I can get that01:03
mwhudsonwell, if you get in the position that you can run ec2 land yourself it's a lot easier01:03
mwhudsonbut yes01:03
lifelesshappy to do for you in general, just noting that policy wise, you should be allowed01:03
pooliei realize it may be partly my fault but the mean latency is somewhere over 24h01:03
poolies//well over01:03
pooliejust got a "thankyou" mail for bzr, specifically for mgz01:04
poolieand also for gnukeyring01:04
pooliewhat a nice way to start the week01:04
pooliebtw i'm sure this landing latency will get better01:04
mwhudsoni was saying to jml the other day that one thing i miss least from the days of officially being a launchpad hacker is the battle to land branches01:06
lifelessI've figured out why we get N running librarians01:06
mwhudsonpoolie: your branch is playing in pqm now01:06
lifelessmwhudson: you're now doing that what, 50% of the time ;P01:06
lifelesswe get N running librarians because:01:06
lifeless - we use atexit (which isn't what you might think it is)01:06
mwhudsonlifeless: only when cody-somerville is being slow at reviewing branches01:06
lifeless - the existing kill-stuff code looks in $root/librarian.pid01:07
lifeless - we rm -rt $root01:07
lifelessmwhudson: oh, so your soyuz-skin is going into a different tree?01:07
lifelessor you're working on this bash stuff?01:07
* lifeless twists the knife01:07
* mwhudson goes for a walk in the sun instead01:07
lifelessmwhudson: I am genuinely interested in the soyuz skin project, its very unclear to me where its at.01:08
mwhudsonlifeless: yes, a fair bit of bash hacking01:08
mwhudsonlifeless: i think "a bit stalled" would be the current summary01:08
mwhudsonlifeless: salgado would know more than me01:08
mwhudsonback in a bit01:08
poolieheh, how can anyone not be in ~launchpad-users? :)01:09
poolie(i realize it's a list)01:09
pooliei am in ~launchpad01:09
lifelesswe could add new accounts to it automatically01:09
pooliecan i send a mail or should you?01:09
pooliean RT, i mean01:09
lifelessthen they would get a prompt to subscribe straight away.01:09
lifelesspoolie: you should; usual thing.01:09
pooliek thanks01:10
lifelessoh yeah, I know where the day went. daylight savings.01:17
lifelessnz is UTC+13 now01:17
lifelessand with my typical early start01:19
lifeless 5 files changed, 60 insertions(+), 85 deletions(-)02:05
lifelessmwhudson: and here is your monday present https://code.edge.launchpad.net/~lifeless/launchpad/test/+merge/3667402:16
=== _thumper_ is now known as thumper
michaelh1Afternoon.  Can access the blueprints on a project from the API?02:34
pooliehi michaelh1, i think you can02:34
beunoI thought blueprints didn't have an API at all?02:35
mwhudsoni don't think blueprints are _very_ exposed02:35
beuno:)  hi mwhudson!02:35
mwhudsonhi beuno02:35
michaelh1poolie: I see there's a 'specification' object in the API docs but no 'specifications_collection' on a project02:36
lifelessjames_w put them up on the API02:36
lifelessthey are probably still quite vestigial02:36
mwhudsoncertainly, there is a lot of permission-by-view in the blueprints code, so i really hope those bits aren't exposed02:36
lifelessprobably it is02:36
lifelessacid test on whether stallman was right about needing passwords02:36
mwhudsonseems like the only bit that's exposed is for linking/unlinking branches and specs, probably for ajax love to that feature?02:37
ajmitchmwhudson: I think that's what was blocking that branch from being merged, last I saw it was still a work in progress02:37
wgrantmwhudson: That was my vague recollection, yes.02:37
mwhudsonmichaelh1: you can ask the linaro infrastructure team to work on this if you like :-)02:38
lifelessmwhudson: but they might go blind :P02:38
michaelh1mwhudson: I might do.  I'm looking for some type of meta blueprint that has others attached to it.  The current Launchpad UI makes this hard to navigate02:38
ajmitchlifeless: I would say "it's not that bad", but then again it's why I didn't go any further with it :)02:39
lifelessajmitch: its primarily a) old and b) unmaintained.02:39
wgrantc) pointless02:39
ajmitchboth of which don't really lend themselves well to having someone jump in & do a few fixes02:39
lifelessits currently unique in LP02:39
wgrantlifeless: Howso?02:40
lifelessI have sympathy for the view that blueprints and bugs should be facets of one thing; but at the moment thats the way it is02:40
lifelesswgrant: ^02:40
wgrantlifeless: How is it unique?02:40
lifelesswgrant: dependencies primarily02:40
lifelessalso some of the reports02:40
wgrantIt also shells out to dot.02:41
wgrantI died inside when I saw that.02:41
ajmitchon page view?02:41
ajmitchoh my02:41
lifelessold skool02:41
ajmitchI bet that won't fall under the 99% rule very easily02:42
lifelessfor rendering?02:42
lifelessmmm separate resource I think; also deleted nw.02:42
lifeless(At least, i think it was deleted)02:42
wgrantI didn't think it was deleted.02:43
wgrantMaybe the full project overview thing.02:43
* ajmitch woudl think it could take awhile to do if it was a large graph of blueprints, and that page hadn't been hit recently02:43
mwhudsonfortunately it's also totally unreadable when there are a lot of blueprints02:43
wgrantLaunchpad also doesn't know how to capitalise its own name, apparently.02:43
ajmitchwgrant: that's amusing, given how https://blueprints.edge.launchpad.net/launchpad-project links to it02:44
wgrantThe blueprint listing reminds me of the good old days.02:46
wgrantOf the really old 2005-ish theme.02:46
ajmitchthat was stylish02:46
ajmitchat least compared to the ubuntu bugzilla :)02:46
lifelessanyone know how to get buildout working outside of LP?02:47
mwhudsonlifeless: svn cat <some url> > bootstrap.py; python bootstrap.py02:53
james_wthe branch is blocked on making the model safe to expose03:02
james_wI have a partial branch to do that, I think the main thing before it can be reviewed is some permission checks which involve moving some code from security.py to the model03:03
lifelessmwhudson: I meant in zope.testing specifically03:06
lifelessmwhudson: this is what I was meaning:03:28
ajmitchoh, that problem - I ran into that in the weekend & didn't go any further03:29
mwhudsonlifeless: that looks odd03:29
lifelessmwhudson: this happens every time I attempt to use buildout.03:29
lifelessmwhudson: I've a pretty firmly entrenched opinion now.03:29
mwhudsonthere are some arguments you can pass to bootstrap i think...03:29
ajmitchlifeless: this is why I was complaining in #nzpug this morning about buildout :)03:30
lifelessit trying to write to /usr/local is spectactularly naughty.03:30
mwhudsonlifeless: bootstrap.py --eggs=eggs maybe03:30
lifelessmwhudson: this isn't in the LP tree, I don't see why that would help ?03:31
mwhudsonlifeless: i think this will make it install eggs in a directory called 'eggs'03:31
lifelessits finding my existing setuptools deb package, then going nana03:31
mwhudsonlifeless: rather than /usr/local03:31
lifelessmwhudson: ah, interesting.03:31
mwhudsonlifeless: i'm cargo culting, clearly, but try it03:31
lifelessI probably shouldn't be so negative about it03:32
lifelessexcept that it seems to be a net time consumer03:32
mwhudsoni have mixed feelings about buildbot03:33
mwhudsonwhen it works, it's quite ok03:33
mwhudsonwhen it doesn't, it's ****ing mysterious at times03:33
lifelessrobertc@lpdev:~/zope.testing$ python ./bootstrap.py --eggs=eggs03:33
lifelessError: Invalid option --03:33
mwhudsoni meant buildout there03:33
mwhudsonalthough both statements are accurate03:33
lifelessmwhudson: spectacular03:34
lifeless$ python ./bootstrap.py bootstrap --help03:35
lifelessGetting distribution for 'distribute'.03:35
lifelesssee under 'above failure'03:35
mwhudsonmaybe the bootstrap.py in the zope.testing tree is very old?03:35
=== nigelbabu is now known as nigelb
lifelessits not an unknown thing :)03:36
mwhudsonyeah i tried that too :)03:36
lifelessI think this is a sign that I should migrate my lappy to maverick too03:38
lifelessso, I shall.03:38
wgrantBeware the wallpaper.03:39
lifelesswgrant: Doom3 >> all other.03:39
lifelessI've not had a default wallpaper in many many many years03:42
lifelessmy desktop does, so I already know what the mav paper looks like.03:42
mwhudsoni see my desktop about once a month03:43
wgrantI occasionally empty a workspace, and then am blinded.03:44
lifelessthe moire on my 1600x1200 crt is quite spectacular03:46
=== almaisan-away is now known as al-maisan
lifelesswheee wow dpkg is slow these days.05:19
lifelessfsync for the -pain-05:19
lifeless[on ssd no less]05:19
=== al-maisan is now known as almaisan-away
pooliehttps://bugs.launchpad.net/bugs/648504 has a complaint about the dupefinder being too tight now06:31
_mup_Bug #648504: exceptions.NotImplementedError: should resend request to http://feeds.edge.launchpad.net/bazaar/, but this isn't implemented <Bazaar:New> <https://launchpad.net/bugs/648504>06:31
pooliei realize it's a tradeoff06:31
=== almaisan-away is now known as al-maisan
spivWhat would be magic would be some sort of "yes this traceback and version match (and this traceback isn't a symptom of multiple separate bugs)" dupe-matching.06:37
spivBut there's a fair bit of handwaving involved in describing exactly what would be awesome there :)06:37
wgrantspiv: apport does that.06:44
wgrantThe retracers compare stack traces and dupe bugs as appropriate.06:44
spivwgrant: not exactly06:44
spivRight, after I've either filed my bug or, as a reporter, decided one of the suggested bugs is probably the same even though I'm not a developer and hardly qualified to know.06:45
wgrantIt has been argued that the dupefinder should be disabled for crashes, since the retracers can match them much more accurately.06:46
spivI've been regularly asked by Launchpad recently to judge if random Xorg crash reports from other people with vaguely similar laptops are the same as mine, whic his a bit erid.06:46
spivAlmost as weird as that typo!06:46
spivYeah, that would make sense to me.06:46
=== al-maisan is now known as almaisan-away
=== almaisan-away is now known as al-maisan
=== poolie_ is now known as poolie
jtvmwhudson: did you see the oopses for the recipe branch build?  BuildFarmJob without specific_job.08:00
jtv(See the launchpad-dev list.)08:00
wgrantjtv: You sure it was a recipe build?08:01
jtvwgrant: BuildFarmJob.job_type == 308:01
jtvThat's a RECIPEBRANCHBUILD, according to my map.08:01
wgrantThat kind of BuildFarmJob.08:02
wgrantOh, wait, misread.08:04
wgrantIs there a packagebuild with build_farm_job=1977174?08:05
cody-somervilleIs anyone here familiar with how pip interprets package versions?08:06
wgrantI know just enough to not drown. What's the issue?08:06
cody-somervillewgrant, sent you pm since its offtopic08:08
adeuringgood morning08:40
pooliedoes the lp webapp have a default log file in production?09:03
bigjoolsgood morning09:07
pooliehi bigjools09:07
thumperbigjools: morning / evening09:07
pooliehi jml ?09:07
* bigjools has too much email09:08
wgrantMorning bigjools.09:10
bigjoolshello William09:10
bigjoolsmorning mrevell09:11
wgrantbigjools: Ubuntu wants ubuntu-release to be able to accept from UNAPPROVED in the Release pocket, and ubuntu-sru from UNAPPROVED in Proposed. Do you have any objections to extending queue admin perms to include optional pocket and status restrictions?09:12
bigjoolswgrant: provided I see a bug filed with a comment from Colin, that's fine09:13
wgrantbigjools: Great, thanks.09:13
bigjoolswgrant: however I don't want those values hard-coded09:14
bigjoolsI guess we can do it through ArchivePermission09:14
wgrantbigjools: Certainly not. I'll just extend ArchivePermission to have upload_status and pocket fields.09:14
wgrantLike we do with components already.09:14
bigjoolsyay for nullable FKs09:14
jmlhello everyone.09:48
bigjoolsmorning jml09:57
jmlbigjools, hi09:58
bigjoolsjml: did you do any more work on the tests over the weekend?09:59
jmlbigjools, fwiw, I didn't get around to buildd-slavescanner.txt. I got fed up with having to do that TwistedLaunchpadZopeless dance so I started work on Deferred support for testtools.09:59
jmlbigjools, and then my batteries ran out.09:59
bigjoolsyou need to run powertop09:59
jmlbigjools, I wanted to land a branch that killed buildd-slavescanner independently of the refactoring09:59
jmlbigjools, I do so from time to time09:59
bigjoolssomething's screwing you over09:59
jmlbigjools, I think it was Chrome, actually09:59
jmlno, I don't think so.10:00
wgrant:( A power problem not caused by Flash?10:01
bigjoolsjml: ok did you figure out those adaptation errors?10:01
jmlbigjools, no. just focused on buildd-slavescanner10:01
jmlbigjools, only had 90mins of hacking time :)10:01
bigjoolsso you didn't change any of the builderslave-resume branch?10:01
bigjools(which appears to  be the new integration branch!)10:02
jmlbigjools, I'll push up what I've got.10:04
jmlbigjools, I don't think I did.10:04
bigjoolsI will spend the next million hours reading email then get back to that branch10:04
lifelessyay upgrade glitches10:19
jmllifeless, yay indeed.10:21
* bigjools WTFs at the wtf bug tag10:23
bigjoolsStevenK: will i-f-p skip disabled arches?11:24
jmlstub, will -vvv on ./bin/test also give transaction debug messages?11:25
stubjml: Doubtful - I don't think test runners can use the python logging module as that is being used by the tests they are running.11:26
jmlstub, cool. checking since we run w/ -vvv on ec2.11:26
StevenKbigjools: Doubtful, since that didn't exist when I was working on it.11:26
lifelessstub: jml: You could sensible capture debug messages to a detail.11:27
* lifeless crashes completely.11:27
bigjoolsStevenK: ok can you add it to the list please, it's critical to get fixed before natty11:27
wgrantDo we have a rollout before natty?11:27
wgrantDoesn't the current i-f-p stuff live in db-devel?11:27
bigjoolswe don't have a db rollout11:28
stublifeless, jml: Yes. We could make bin/test add a handler to getLogger('txn') that emits messages where we want, although we might have to reset it after every test in case the test tore things up.11:28
lifelessstub: not bin/test - use a Fixture and register it with self.addDetail.11:28
=== jtv is now known as jtv-afk
StevenKwgrant: No11:29
StevenKbigjools: Right, I'll sort it out tomorrow11:29
wgrantStevenK: Ah, good.11:29
bigjoolsStevenK: thanks - I'll add a card now11:29
StevenKIt should be quite quick11:29
jmlstub, I think lifeless is talking about the facility in TestCase where you can add details to test output in response to certain events (generally exceptions)11:29
* wgrant pokes at apt-ftparchive.11:29
* bigjools gets caffeine11:30
stubRight. We could do it, but we haven't done it :)11:30
StevenKbigjools: It can be worked around with application of -a11:31
stubAs it emits a log message when the transaction starts, this might be useful for tracking down code that holds open long running transactions which can be sucky (such as the transaction being opened by Storm refreshing a property...)11:31
wgrantGnargh lucilleconfig kill kill.11:35
wgrantOh wtf.11:35
wgrantbigjools: We need to stop creation of new arch-indep publications before maverick, or we'll need to do some manual cleanup before maverick gets changed to CURRENT.11:39
wgrantSince we need p-d-r to remove all the files.11:40
bigjoolswgrant: gnar11:41
wgrantHm, except that p-d-r doesn't obviously respect the Release pocket freeze, so it might be OK to have it remove them afterwards.11:42
bigjoolsI'd rather it worked now11:43
wgrantAt least there appear to only be two places that create arch-indep publications.11:45
wgrantSo, fix those two locations, tweak lucilleconfig to exclude disabled DASes, Delete existing publications, and we may be able to get maverick into a sane state.11:47
wgrantI think.11:47
wgrantHas the on-disk archive ever been compared with what Soyuz thinks should be there?11:50
bigjoolsnot to me knowledge11:50
wgrantp-d-r looks buggy, so I suspect there's a lot of cruft in the pool.11:52
deryckMorning, all.12:06
=== jtv-afk is now known as jtv
jmlderyck, good morning12:16
bigjoolswgrant: fyi, I just filed bug 64871512:19
_mup_Bug #648715: Binary publications should not be created for disabled architectures <soyuz-publish> <Soyuz:Triaged> <https://launchpad.net/bugs/648715>12:19
gmbOkay, I know that I know the answer to this, but I can't remember what it is for the life of me: "psycopg2.InterfaceError: connection already closed" <-- what is that and why is it happening? Any ideas?12:33
wgrantgmb: Your PostgreSQL may be in recovery mode.12:35
wgrantCheck its logs.12:35
wgrantHm, no, that's a different thing.12:37
gmbwgrant, It is a pg problem though; make schema is blowing up. I'll dig into it further. Thanks for the pointer anyway.12:48
jmldanilos, hi13:07
danilosjml, hi13:07
jmldanilos, marjo has asked for a feature along these lines: "Enable translators working in Launchpad to seamlessly submit those translations to the upstream project to which they apply. "13:08
jmldanilos, I thought we did that already.13:08
danilosjml, well, "seamlessly" is very hard... that's next on our list after we've got the "seamless" imports going13:09
danilosjml, ("next on our list" if we are to continue on our bridging-the-gap theme for translations)13:09
danilosjml, what we've got know are partial tools that enable people to do lot of that work manually13:09
daniloss/got know/got now/13:09
* bigjools notices the EINTR bug that poolie fixed and doesn't know whether to laugh or cry13:10
jmldanilos, you guys are working on smoother imports now?13:10
danilosjml, "seamless" would be to have a button saying something like "submit these translations upstream" where they'd get into the upstream workflow for handling translations13:11
danilosjml, well, it's more like "we are still working on that"13:11
danilosjml, it's been going pretty slow13:11
danilosjml, "smoother imports" is getting translations directly from upstream into ubuntu13:11
jmldanilos, with what degree of interaction from users / admins?13:12
danilosjml, two differing steps: 1. get upstream translations into LP upstreams (requires code imports and admin interaction for now, but already possible); 2. set up packaging links between upstream series in LP and Ubuntu sourcepackages13:14
danilosjml, for 1, we want to improve the process further once we have the groundwork laid out to support actual sharing of translations between ubuntu and upstreams13:14
jmldanilos, I think Registry has done a lot of work on making code import, project registration & linking easier. Not 100% sure about series branches specifically.13:15
danilosjml, 1. already works for a wide range of intltool- and gettext-based upstreams, but requires admin set-up which is not nice13:15
jmldanilos, admin set up per package?13:16
danilosjml, I know, they are doing a good job with splitting the "official_*" tags as well13:16
danilosjml, in general, for translations it's pretty simple to set up so we just need to follow the same pattern and come up with good privilege models13:16
danilosjml, basically, set-up for translations should only include "this has external translations, IMPORT THEM" and "link the series with sourcepackage"13:17
danilosjml, "link the series" step is pretty nicely handled by registry folks already13:17
danilosjml, though, it'd be nice to integrate into one step if someone's interested in doing it all in one code (i.e. integrate set-up code branches, turn the translation imports on, link to source packages steps)13:18
daniloss/in one code/in one go/13:19
danilostyping ahead of me :)13:19
jmldanilos, I guess there are two potential issues with having them split up: multiple POSTs, and users not necessarily knowing what the next step is13:19
danilosjml, well, there's always ajax, and there's always code-reuse :) if we make those bits of pages nice reusable ajaxy "portlets" (not necessarily in zope sense, but just bits that can be plugged into other pages), it would make a nicer experience without increasing our workload too much13:20
danilosjml, fwiw, some of these things might already be easily reusable, I haven't looked13:21
jmldanilos, you said there's some ground work that needs to be done. I had thought you'd already done it.13:21
* jml nods13:21
danilosjml, well, we've got a huge feature branch in progress that I said is going slower than expected: that's to actually get upstream translations shared with Ubuntu directly13:21
danilosjml, we are nearing completion of that bit, but likely not this cycle after all13:22
jmldanilos, I understand :) huge feature branches being what they are.13:22
jmldanilos, with the other side, sending translations upstream seamlessly -- is there any reason to *not* do it after you've done the inbound direction?13:24
danilosjml, not really, it's just a lot of work because many upstreams will have differing needs... but we should attack it in a similar way, go for big "easy" upstreams first (gnome, kde, translationproject), then for big "hard" upstreams (mozilla, openoffice) and so forth13:26
jmldanilos, cool13:28
jmldanilos, it's rather nice that our broader plans happen to mesh with what Ubuntu says they need :)13:28
danilosjml, yeah, I am very happy with that as well :)13:29
jmldanilos, thanks for your time. I'll let Marjo know our current plans.13:35
danilosjml, cheers13:37
=== mrevell is now known as mrevell-lunch
flacostehi jml!14:17
jmlflacoste: welcome back :)14:17
flacostehello fellow launchpadders!14:17
noodles775Hey! Welcome back flacoste14:17
flacostehi noodles775! are still a launchpadder ;-)14:18
flacosteor did you move on already?14:18
flacostenot that I wouldn't greet you if that was the case :-)14:18
noodles775heh, yep - will be switching sometime soon, but just working on some soyuz UI work beforehand.14:18
noodles775(And I'll still be here on #launchpad-dev anyway)14:19
bigjoolshey flacoste, flumper will be pleased to see you :)14:21
flacostebigjools: yeah, he told me that he was looking forward to give me my job back!14:21
flacosteyou guys gave him a hard time?14:21
bigjoolsdepends who "you guys" is I guess :)14:22
deryckhey wb flacoste!14:24
flacostehi deryck!14:25
wgrantbigjools: Disabling an arch is somewhat problematic. If you disable it (to prevent new publications) then mark its publications deleted, it never gets published to actually remove them.14:31
bigjools /o\14:31
wgrantBut I suppose we won't be doing this until powerpc dies in at least a couple of releases.14:31
wgrantbigjools: We shouldn't be disabling another DAS until powerpc dies. And that will be a while.14:33
wgrantAnd sparc/ia64 need manual cleanup anyway, so it's not too bad.14:33
=== mrevell-lunch is now known as mrevell
cr3buildout question: is there a way to leverage buildout someone to build against package versions available in a particular release to make sure the code will work on that release?15:00
jmlsinzui: you around?15:06
sinzuiI am15:06
jmlsinzui: there have been some requests about blueprints on the stakeholder list. I have this feeling that you're the right person to talk to to get a sense of difficulty of implementation15:07
sinzuijml :( Am I really the last remaining engineer to have hacked on blueprint?15:07
jmlsinzui: I don't know! You just have a blueish tinge to your aura.15:08
sinzuiI can help. I am pretty knowledgeable of the bug reports that often underly a blueprint feature request15:09
jmlsinzui: cool.15:09
jmlsinzui: what we want, probably, is easy ways to get a list of blueprints that are assigned to members of a given team15:09
sinzuiThat sounds a lot like my proposal to find the intersection of items assigned to a team for a milestone. I really would like a common way to say I want to see the items for the members of a team for any structure15:11
cr3jml: what that "huh" for me?15:11
jmlsinzui: yes, that would be ideal.15:12
jmlcr3: I guess when you say "release" you're talking about something other than a release of Launchpad.15:12
cr3jml: sorry, yes, I meant release of Ubuntu15:12
jmlcr3: the whole point of buildout is that you are insulated from the underlying OS.15:13
jmlcr3: well, maybe that's the whole point.15:13
cr3jml: that is indeed a point but I figured it would also be used to couple with a particular OS15:13
cr3jml: for example, I happen to used a dependencies branch from where buildout can only get dependencies, nowhere else. I use this as a snapshot of what is known to work, rather than relying on the latest code. I believe launchpad does the same under source-dependencies or somesuch15:14
cr3jml: however, if that list of packages were taken from the release of an OS, that would provide somekind of predictability that it should work on the OS. the problem with that is that: 1. the orig.tar.gz files on archive.u.c are named funny; 2. those files are also not quite what's on the release of the OS.15:16
cr3jml: so, perhaps another approach would be to bzr branch lp:ubuntu/release/dependency instead of using tarballs15:16
cr3jml: I was hoping this was already attempted by someone so that I could learn from their experience rather than wasting time myself finding out there's a real problem15:17
jmlcr3: perhaps. you might also want to ask on general zope channels.15:21
jmlabentley, jelmer: where are we at with being able to import non-master git branches?15:30
abentleyjml: it's not something I've been following.15:31
jmlabentley: np. just picking on you as a visible code team member :)15:32
abentleyjml: the issue is that bzr-git is tied into the standard Bazaar branch behaviour, and that doesn't support colocated branches.15:33
abentleyjml: so someone from the Bazaar team could update you on colocated branch support.15:33
jmlabentley: ta. I'll mosey on to #bzr.15:34
deryckI wish I had a notification system that would ping me -- "the builders are up and green.  go for it! now!"15:52
wgrantDon't worry, there's still gambling involved -- if you're lucky we'll be in testfix again by the time EC2 finishes :)15:54
deryckMy statement presumed I had already had a couple ec2 failures and was prepared to go to pqm directly.15:55
wgrantThus ensuring that the cycle continues.15:56
deryckmaybe.  In this case, I had a translations windmill test that failed in ec2, and I ran locally with no issues.  So submitted directly.15:56
=== deryck is now known as deryck[lunch]
jmlsinzui: hi16:02
sinzuihi jml16:02
danilosderyck[lunch], it might be a test that could be helped with a longer timeout somewhere for ec2 runs, it's useful if you file a bug so we can look at it and close it as 'invalid' because we can't reproduce it... I mean, perhaps increase a timeout for a page load or two if it's a clear-cut case like that16:08
=== Ursinha is now known as Ursinha-lunch
=== salgado is now known as salgado-lunch
jmlnoodles775: do I remember wrongly, or were you looking at generic comments code recently?16:33
noodles775jml: Yes I was.16:33
abentleyjml: I was, too.16:34
jmlwell, my question to both of you...16:34
jmlted was asking about attachments on MPs16:34
jmlI thought this was roughly analogous to sharing comments code between bugs & MPs.16:35
* noodles775 didn't realise the lp.services.comments code was used by bugs yet?16:36
jmlI don't know if it is16:36
jmlI'm asking you :)16:36
jmlI guess not then.16:37
noodles775AFAICS, the distroseriesdifference comments were the first re-use, but abentley might know more.16:37
abentleyjml: I'm pretty sure the display of comments is shared between bugs and merge proposals.16:38
noodles775abentley: grep for "from lp.services.comments"16:39
noodles775In db-devel I see only code and (now) registry (for the distroseriesdifferences)16:39
stubDo we ever delete Revisions? Might we ever?16:39
abentleyrockstar: didn't we work on sharing comments code with bugs?16:40
noodles775jml: sorry, what was your question?16:40
jmlnoodles775: I still haven't formulated it yet :)16:41
rockstarabentley, I think thumper did something like that.16:41
jmlperhaps, a) how different are MP comments and bugs comments at the code level?16:42
jmland b) how easy would it be to extract the bugs attachment thing and make it re-usable16:42
noodles775jml: don't MP comments already support attachments?16:43
jmlnoodles775: I think there's no web UI for it16:43
wgrantThey're also implemented rather differently.16:43
wgrantMPs just look at extra MessageChunks.16:44
wgrantBugs have explicit BugAttachments created from MessageChunks.16:44
=== matsubara is now known as matsubara-lunch
rockstarabentley, jml, noodles, I'm pretty sure BjornT and thumper talked about how to unify them.16:51
noodles775jml: but fwiw, I don't see why the UI code couldn't be shared in lp.services.comments16:52
jmlrockstar: noodles775: thanks.16:52
cr3out of curiosity, where does the name "assets" come from in respect to static css files seem to be built?16:54
noodles775Night all.16:55
=== deryck[lunch] is now known as deryck
=== al-maisan is now known as almaisan-away
=== benji is now known as benji-lunch
=== Ursinha-lunch is now known as Ursinha
=== salgado-lunch is now known as salgado
=== beuno is now known as beuno-lunch
=== matsubara-lunch is now known as matsubara
maxb[Merge] lp:~mterry/gnome-control-center/ubuntu-2.32.0 intolp:~ubuntu-desktop/gnome-control-center/ubuntu17:59
* maxb is deeply confused why ~vcs-imports got attached as a reviewer to that17:59
bdmurrayIs there a convenient way to get <h1> from browser.contents in a test?18:03
deryckbdmurray, it's just beautiful soup underneath, so can't you do browser.contents.h1 or browser.h1 to get the first h1?18:09
=== benji-lunch is now known as benji
bdmurrayderyck: that didn' work out so well18:19
deryckbdmurray, have you seen lib/canonical/launchpad/doc/pagetest-helpers.txt ?  Nothing on h1 specifically, but might help you think about how to get at the content you want.18:22
jmlfindByTag('h1'), perhaps?18:22
deryckbdmurray, also the README and REFERENCE in lib/canonical/launchpad/pagetests might help.18:25
bdmurrayderyck: thanks18:25
=== beuno-lunch is now known as beuno
marsmorning lifeless18:47
lifelesshmm, no gary18:49
jmllifeless: good morning.18:49
jmllifeless: you're up bright and early18:49
lifelessjml: daylight savings18:49
marslifeless, gary won't be back until Wednesday18:49
lifelessmars: darn18:49
lifelessI have this buildout problem ...18:49
jmllifeless: oh. nice.18:50
marslifeless, what is the problem?18:51
marslifeless, I may be able to help, having set up buildout for a project or two recently18:51
lifelessrobertc@lpdev:~/zope.testing$ python ./bootstrap.py18:52
lifelessAfter install bootstrap.18:52
lifelessCreating /usr/local/lib/python2.6/dist-packages/setuptools-0.6c11-py2.6.egg-info18:52
lifelesserror: /usr/local/lib/python2.6/dist-packages/setuptools-0.6c11-py2.6.egg-info: Permission denied18:52
lifelessAn error occurred when trying to install distribute 0.6.14. Look above this message for any errors that were output by easy_install.18:52
lifeless  Bootstrapping.18:52
lifeless  Getting distribution for 'distribute'.18:53
lifelessError: Couldn't install: distribute 0.6.1418:53
marslifeless, weird.  What bootstrap are you using?  The one from the PyPi zc.buildout page/18:53
lifelessmars: the one in the source tree18:53
marshmm, that should work18:54
lifelessmars: bzr branch lp:zope.testing; python ./bootstrap.py18:54
lifelessmars: see what happens?18:54
marslifeless, ok.  That may be an old bootstrap.py18:55
benjilifeless: I /think/ this is cause by distribute trying to fake up a setuptools egg (part of the setuptools impersonation it does); it's somewhat icky, but you should be able to fix it by running your bootstrap as root one time18:55
lifelessbenji: hell no18:55
marsbenji, maybe just 'sudo aptitude install python-distribute'?18:55
lifelessbenji: if it wants to write to /usr/* it die a fiery death.18:56
benjimars: don't know; maybe18:56
marslifeless, yes, it should not be trying that18:56
lifelessstatik: hi; I think google calendar forgot that you're on the other side of the world: its moved the appointment to be the same NZT rather than UTC18:56
lifelessstatik: but, as it got me up for it, want to me now, or in an hour ?18:57
mars"Copyright (c) 2006 Zope Foundation and Contributors." - yeah, that's an old one18:57
mars2006 for bootstrap.py in zope.testing18:57
lifelessmars: this is in my lp vm - a lucid one; distribute isn't packaged there.18:57
lifelessmars: so, how does one fix this?18:57
james_wsetuptools in lucid /is/ distribute isn't it?18:58
lifelessjames_w: in maverick maybe18:58
lifelesspython-setuptools                                     0.6.10-4ubuntu1launchpad1                             Python Distutils Enhancements (setuptools compatibility)18:59
marsmine died on 'fetching distriubte', trying 'python bootstrap.py -vvvv'18:59
marsit worked that time?18:59
lifelessno, thats dpkg -l python-setuptools18:59
marslifeless, what happens when you run "python bootstrap.py -vvvv" ?18:59
lifelessso, copying the lp bootstrap into the tree worked19:00
marsrunning it again after it faults should not make it run.  That is some weird bug.19:00
james_wlifeless: http://packages.ubuntu.com/source/lucid/distribute19:00
lifelessmars: it didn't work at all19:00
lifelessmars: however many times it was tried; copying the lp bootstrap.py into the tree made it work.19:01
lifelessjames_w: thanks19:01
lifelessTraceback (most recent call last):19:01
lifeless  File "bin/test", line 17, in <module>19:01
lifeless    import zope.testing.testrunner19:01
lifelessImportError: No module named testrunner19:01
lifeless(this is zope.testing :/)19:02
marslifeless, yep, they split19:03
marslifeless, did you run bin/buildout?19:04
marslifeless, I'm checking buildout.cfg in the zope.testing source tree - it knows about zc.recipe.testrunner, so it should work correctly19:05
marslook at the built parts in that file19:05
marsbin/test is broken for me too19:06
lifelessmars: yes, I ran bin/buildout19:06
lifelessmars: you're serious? zope.testing.testrunner is not in the zope.testing package?19:06
marslifeless, yes, they split: zope.testing.testrunner became its own package, IIRC.  Jim Fulton coded it himself?19:07
lifelessI hope I won't seem shallow when I say that this is madness: I just want to fix layers for lp19:07
jmlz.testrunner 4.019:07
jmllifeless: the recent move has slowed me down from contributing upstream to z.testrunner19:07
lifelessperhaps we should just stop using it19:08
jmlyou won't see me arguing19:08
lifelessjml: I was going to run the test suite, so that the subunit fixing patch could be applied19:08
marslifeless, funny - zope.testing's setup.py does not reference zope.testrunner, and neither does the buildout.cfg19:08
marsI guess someone forgot to add the dependency19:08
lifelessI wonder how much money I could make, taking bets on how many years before this is packaged properly in Ubuntu19:09
marslifeless, add 'zope.testrunner' to the list of files in buildout.cfg [versions], then add 'zope.testrunner' to the [test] part eggs= list19:09
lifelessmars: zope.testrunner is what I need to be working on I suspect, or both at once.19:10
lifelessmars: how does one tell buildout to do that ?19:10
marslifeless, hack buildout.cfg19:11
marshere, just a sec19:11
=== lifeless changed the topic of #launchpad-dev to: Launchpad Development Channel | Performance Tuesday ! | Week 2 of 10.10 | PQM is open for business | firefighting: - | https:/​/​dev.launchpad.net/​ | Get the code: https:/​/​dev.launchpad.net/​Getting
benjiit appears that the comment in zope.testing's [versions] section applies; someone forgot to remove that version pin; if I remove it and rebuild the tests run and pass19:12
marslifeless, ^19:12
lifelessbenji: thats in buildout.cfg?19:12
marsbenji, cool.  I think sidnei has commit rights?  For all I know you do too :)19:12
lifeless  Ran 30 tests with 0 failures and 0 errors in 18.258 seconds.19:13
benjilifeless: yes19:13
benjimars: I do.19:13
lifelessI'm impressed at 0.5 sec each for a lightweight project :><19:13
lifelessso, I presume these projects needs to be changed in lockstep; whats the blessed way to do that?19:14
lifelessIf it was deb based I'd just set PYTHONPATH and be done with it.19:14
benjilifeless: >= version requirements in setup.py should be enough19:15
lifelessbenji: I'm not sure that I' asking the correct question.19:17
lifelessbenji: I need to edit code in *both* packages.19:17
lifelessbenji: and run one against the edited other. Not against pypi versions.19:17
lifelessbenji: how would you do this?19:17
james_wlifeless: develop = . ../other-project19:17
lifelessjames_w: in?19:17
james_wthe one that imports the other19:17
james_wbuildout.cfg in the main section19:18
lifelessjames_w: they import each other19:18
james_wthen in both?19:18
james_wit basically puts ../other-project on PYTHONPATH when you run ./bin/ in the branch you put that setting in19:18
james_wwith extra magic of cours19:19
benjilifeless: ah, I was talking about making the releases; you can use "develop = /path/to/the/other/thing" in the [buildout] section of each check out to make them see each other (possibly with changes to each [versions] section as well)19:19
lifelessis there a way to do that without editing a version controlled file?19:19
lifelessit will be frustrating to have an unclean tree all the time19:19
benjilifeless: you can use a different config (other than buildout.cfg) and run buildout with -c /my/hacked/up/config.cfg19:21
lifelessbenji: does that have to be a full copy?19:21
james_wbug 1300219:21
_mup_Bug #13002: Toshiba battery events flood <acpi-support (Ubuntu):Fix Released by thombot> <https://launchpad.net/bugs/13002>19:21
lifelesstheres no way to just override develop= ? e.g. via .buildout.cfg or something ?19:21
james_werr bug 61300219:21
_mup_Bug #613002: Please add a way to optionally extend <Buildout:New> <https://launchpad.net/bugs/613002>19:21
jmlg'night all.19:22
lifelessnight jml19:22
james_wlifeless: extends =, but you it's editing a version controlled file and you can't commit it19:22
benjilifeless: or you can give buildout the info on the command line, something like bin/buildout buildout:develop=...19:22
benjilifeless: yep19:22
benji(full copy)19:22
rockstarbenji, do you know much about buildout? (I'd ask gary, but he's not around)19:51
benjirockstar: a bit; what's up?19:52
rockstarbenji, a voice call would be great, if you can spare it.19:52
benjirockstar: sure; it'll be a few minutes though.  Do you want me to ping you when I'm available?19:53
rockstarbenji, that would be great.19:53
lifelessbenji: hi19:55
lifelessbenji: this is going to remain broken forever I think19:55
lifelessbenji: it took me 8 weeks to figure out how to run the test suite; I humbly suggest that practicality should win here.19:56
salgadoEdwin-afk, we seem to have some spurious TeamParticipation entries which might be caused by those changes you did to the code which maintains that table: https://bugs.edge.launchpad.net/launchpad-foundations/+bug/597208 (last comment is the relevant one)19:58
_mup_Bug #597208: Run cronscripts/check-teamparticipation.py on production and make its output more visible <Launchpad Foundations:Triaged> <https://launchpad.net/bugs/597208>19:58
benjilifeless: I should be able to take a look in a few hours19:59
malariajtv: danilos wanted me to pop here about this bug: https://bugs.launchpad.net/rosetta/+bug/60863120:08
_mup_Bug #608631: Visual tag to represent narrow non-breaking spaces <diacritic:Invalid> <Launchpad Translations:Triaged> <https://launchpad.net/bugs/608631>20:08
malaria(anyway, hi everyone)20:09
lifelessmalaria: jtv will be around in about 7-8 hours20:11
malarialifeless: ok, it's the same every day?20:12
lifelessI don't know what you're asking20:12
lifelesshe sleeps like everyone else ;)20:12
malarialifeless: of, course, I mean it's because of it's timezone20:13
malariaso I have to schedule the best moment to go back here20:13
benjirockstar: skype?20:26
benjior mumble20:26
benjior smoke signals20:26
rockstarbenji, whoa, the people I needed both responded at the same time.20:26
rockstarbenji, let me chat with sidnei first.  It's possible I may not even need to bug you.20:26
rockstarbenji, okay, I do think I want to chat with you.20:35
rockstarbenji, what's your skype name?20:35
benjirockstar: benji_york20:36
=== matsubara is now known as matsubara-afk
flacostelifeless: morning21:11
lifelesshi flacoste21:12
lifelessare you back ?21:12
lifelessflacoste: I was just running out the door for a short errand21:12
flacostelifeless: i am!21:12
lifelessbut would love to catch up after that, if thats ok ?21:13
flacostelifeless: that was my goal, when will you be back?21:13
lifeless< 20 min21:13
thumperrockstar: please check QA on kanban21:18
rockstarthumper, ah!21:18
lifelessflacoste: ok21:26
rockstarthumper, hm, qa on one of my items is bad (it doesn't fix the bug it's supposed to, nfi).  Where should that card go?21:30
thumperrockstar: it stays in the qa lane and you fix it :)21:31
rockstarthumper, okay.21:31
flacostelifeless: https://chinstrap.canonical.com/~stub/splitit/22:06
Ursinhalifeless, hello22:12
rockstar"Get up, stand up. Stand up for your right"22:12
Ursinhalifeless, I believe r11579 landed the other half's fix of bug 506256, but the landing has no qa tags, so the bad-commit tag is still there22:16
_mup_Bug #506256: Remove Popen from buildstatus_OK <bad-commit-11566> <buildfarm> <qa-needstesting> <Soyuz:Fix Committed by jelmer> <https://launchpad.net/bugs/506256>22:16
Ursinhaand it's blocking further revisions22:16
lifelessUrsinha: hi22:20
lifelessbigjools was going to cp both revs22:21
lifelessI don't think he has yet, lets check prod-stable22:21
Ursinhalifeless, so I guess we could unblock other bugs by marking this one as qa-ok?22:28
Ursinhaassuming both parts have already been QAed, of course :)22:28
lifelessflacoste: https://devpad.canonical.com/~lpqateam/qa_reports/launchpad-stable-deployment.txt22:29
Ursinhalifeless, flacoste, prettier: https://devpad.canonical.com/~lpqateam/qa_reports/deployment-stable.html :)22:30
=== salgado is now known as salgado-afk
lifelessUrsinha: so, prod stable does not have both fixes22:44
lifelessUrsinha: and we have to deploy both - see the bad-commit-xxxx tag22:44
Ursinhalifeless, I know... let me try again22:45
Ursinhalifeless, are both on edge and QAed?22:45
Ursinhalifeless, that's what matters so we can unblock the bug, right?22:45
lifelessthis is the fixes=12345 stuff I was mentioning - same as rollback but rollforward22:46
UrsinhaI know22:46
Ursinhathat needs to be discussed and implemented22:46
UrsinhaI mean something else22:46
lifelessI'll be back in a minute, real-life interrupt22:46
lifelesssorry, back23:02
lifelesswgrant: around perchance?23:03
lifelessUrsinha: https://bugs.edge.launchpad.net/soyuz/+bug/12825923:05
_mup_Bug #128259: Buildmaster should not call "uploader" script for processing incoming binaries <buildd-manager> <qa-needstesting> <soyuz-build> <Soyuz:Fix Committed by jelmer> <https://launchpad.net/bugs/128259>23:05
Ursinhalifeless, hmm, what's your point? :)23:13
lifelessUrsinha: you were asking if the second half fixed it - its unknown ?23:13
Ursinhalifeless, I'm asking what's about that bug23:14
lifelessUrsinha: that bug is attached to the commit that is the other half of 1156623:14
UrsinhaI thought the other half of that bug was something related to that bug, just another branch ref. the same bug #23:17
Ursinhabut now I see23:17
lifelessUrsinha: yeah; the other half references two bugs.23:17
wgrantlifeless: Hi.23:38
lifelesswgrant: wondering if you have an opinion on the qa-ness of bug 12825923:38
_mup_Bug #128259: Buildmaster should not call "uploader" script for processing incoming binaries <buildd-manager> <qa-needstesting> <soyuz-build> <Soyuz:Fix Committed by jelmer> <https://launchpad.net/bugs/128259>23:38
lifelesssee backscroll23:39
wgrantI haven't heard.23:39
wgrantStevenK might know.23:39
lifelessalso, killing pids that happen to be in a file that was stale weeks ago is really pretty nasty.23:39
wgrantBut deploying that without thorough, thorough QA would be a seriously terrible idea.23:39
lifelessThis layer stuff really concerns me :)23:39
wgrantAre those revs in prod-stable yet?23:40
=== Ursinha is now known as Ursinha-afk
wgrantbigjools might have already CP'd them.23:40
=== matsubara-afk is now known as matsubara

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