/srv/irclogs.ubuntu.com/2010/04/13/#launchpad-dev.txt

* maxb grimaces at tests which only fail in a full test-run00:19
maxblp.services.scripts.tests.test_all_scripts.ScriptsTestCase is the testcase of doooooom00:27
maxbI worry that it's going to melt my laptop00:27
spmmaxb: it won't melt; but it might get a bit soft and wobbly/deform....00:38
maxbIt's when the laptop starts to smell a little like a soldering iron that I worry :-)00:39
spmmaxb: so long as the smoke doesn't escape? everything is gravy!01:04
=== matsubara is now known as matsubara-afk
mwhudsonspm: can i get you to apply a patch to strawberry pls?01:39
spmsure01:39
mwhudsonspm: http://pastebin.ubuntu.com/413405/01:39
mwhudsonspm: *should* apply to the tree on there, not completely sure though01:40
mwhudsonif it doesn't i'll just get you to replace the entire file i guess01:40
spmkk01:41
mwhudsonspm: in news that may please a sysadmin, this patch is essentially telling bzr to stop being clever and be very stupid instead01:42
spm:-) applied01:43
mwhudsonspm: can you kill the running import on strawberry ?01:43
spmmwhudson: with great and a deep sense of satisfaction and pleasure; done.01:44
spmmwhudson: fyi. afk for a sec. delivery arrival...01:46
mwhudsonok01:52
spmback01:52
mwhudsonstupid > smart01:53
mwhudsonin this case01:53
spm:-)01:53
mwhudsonnoop import before patch: 1hr 10 minutes01:53
mwhudsonnoop import after patch: three minutes01:53
jelmermwhudson: the create_tree_if_local argument to sprout() doesn't work as expected?01:54
spm\o/01:54
mwhudsonjelmer: it doesn't?01:54
mwhudsonjelmer: the problem seems to be mainly that 2a fetch is terribly slow01:54
jelmermwhudson: I'm not sure, just trying to work my head around what your patch does differently01:54
mwhudsonjelmer: my patch uses copy_tree_to_transport rather than sprout01:55
mwhudsonjelmer: i think that counts as fairly different01:56
wgrantOh, nice!01:56
jelmermwhudson: heh01:57
wgrantDo I spy a bzr bug coming soon?01:57
mwhudson(credit to abentley for the idea)01:57
jelmermwhudson: I missed that, surprised bzr doesn't do that itself..01:57
jelmermwhudson: (in those cases where it's possible_01:57
jelmer)01:57
jelmermwhudson: should the branch scanner work ?01:57
wgrantThe puller times out :(01:58
mwhudsonnext on the hit list: having the puller use a similar hack for import branches01:58
wgrantHm. So it takes 37 seconds to copy directly, and slightly under 70 minutes for bzr to do it?02:00
mwhudsonsomething like that yes :(02:02
mwhudsonaragh02:06
mwhudsonTransport operation not possible: Transport <bzrlib.transport.http._urllib.HttpTransport_urllib url=http://localhost:10899/0000004d/> has not implemented list_dir (but must claim to be listable to trigger this error).02:06
wgrantIt pulls over HTTP!?02:06
mwhudsonyes02:06
mwhudsonactually not on staging, so this patch will probably work there....02:07
wgrantHm, but that's probably actually faster than using the smartserver here.02:07
mwhudsonyes, smartserver is /mostly/ a squeeze on latency02:07
mwhudsonand bandwidth in some cases i guess, but we have lots of one and very little of the other so...02:07
mwhudsonthe error message seems to be lying fwiw02:09
* mwhudson hacks, mightily02:26
mwhudsonspm: hi, could you apply this patch to the tree on tellurium please? http://pastebin.ubuntu.com/413423/02:40
mwhudsonwarning: do no attempt to read said patch02:40
mwhudson*not02:40
spm:-)02:41
spmaaaaaaa. I am blinded and cannonit see02:44
spmmwhudson: applied02:44
spmneed codehost restarted? looks like not....02:44
mwhudsonspm: no02:45
wgrantIt's taking a while..02:50
mwhudsonspm: is there a mirror-branch.py process running on tellurium?02:51
spm000      3944 97.1 18.5 653364 382120 ?       R    02:48   3:25      \_ /usr/bin/python2.5 /srv/bazaar.staging.launchpad.net/staging/launchpad/scripts/mirror-branch.py /home/supermirror/importd-push-branches/0004ea4b lp-mirrored:///~mwhudson/linux/trunk 322123 ~mwhudson/linux/trunk IMPORTED 102:52
mwhudsonspm: okay02:53
mwhudsonseems maybe my hacks aren't helping all that much02:53
mwhudsonspm: can you du -sh /home/supermirror/importd-push-branches/0004ea4b /srv/bazaar.staging.launchpad.net/mirrors/00/04/ea/4b ?02:53
spm620M    /home/supermirror/importd-push-branches/0004ea4b02:54
spm61M     /srv/bazaar.staging.launchpad.net/mirrors/00/04/ea/4b02:54
spmthats an impressive mismatch?02:54
mwhudsonbah02:54
mwhudsonspm: can you apply this patch to tellurium too? /home/supermirror/importd-push-branches/0004ea4b02:58
mwhudsonno02:58
mwhudsonspm: http://pastebin.ubuntu.com/413428/ <- this patch02:59
spmsure03:00
spmpatched03:01
mwhudsonspm: can you zap the mirror-branch process?03:02
* spm reaches under the desk for the larger hammer.....03:02
spmyo stub03:03
stubput the hammer down...03:03
spm<whine> but *stub* I was gunna make you plural!03:04
mwhudsonspm: ok, can you sync the puller logs from tellurium to devpad please?03:05
spmmwhudson: 1000      4772 99.5  8.0 436900 165584 ?       R    03:04   0:53      \_ /usr/bin/python2.5 /srv/bazaar.staging.launchpad.net/staging/launchpad/scripts/mirror-branch.py /home/supermirror/importd-push-branches/0004ea4b lp-mirrored:///~mwhudson/linux/trunk 322123 ~mwhudson/linux/trunk IMPORTED 103:05
spmmwhudson: synced03:07
=== poolie1 is now known as poolie
mwhudsonoh goody oops reporting is broken03:12
mwhudsoni guess i'll give up for now and have lunch03:14
adeuringgood morning07:58
lifelessthumper: mwhudson: if you're still around, I need a pointer to the twisted vfs stuff in the lp code base - the policy on permitted file names specifically.08:06
lifelessthumper: mwhudson: nevermind, found it, drive by patch submitted08:15
wgrantbigjools: How many years is it since buildd-sequencer has been used?09:19
bigjoolssince Jan 200909:20
wgrantOh.09:20
bigjoolscprov wrote it in the back of the van at UDS Mountain View :)09:21
wgrantI thought buildd-slave-scanner was used before buildd-manager, and buildd-sequencer came before slave-scanner.09:21
bigjoolswait09:22
bigjoolsit's *not* been used since 200909:22
bigjoolsI mean cprov wrote b-m in the van09:22
bigjoolsslave-scanner is the same thing as buildd-sequencer09:23
bigjoolswe like to use as many different confusing names as possible09:23
wgrantEr, is it?09:23
wgrantlib/canonical/buildd/sequencer.py seems reasonably unrelated to cronscripts/buildd-slave-scanner.py09:23
wgrantOr is buildd/sequencer not actually buildd-sequencer?09:25
=== wgrant changed the topic of #launchpad-dev to: Launchpad Development Channel | Week 1 of 10.04 | PQM is open | https://dev.launchpad.net/ | Get the code: https://dev.launchpad.net/Getting | On-call review in irc://irc.freenode.net/#launchpad-reviews | Use http://paste.ubuntu.com/ for pastes
bigjoolswgrant: I've no idea what the former is, since it's in the buildd dir then it's part of the slave09:35
wgrantAh:09:36
wgrantThe task sequencer is a simple twisted daemon which looks after making09:36
wgrantsure that the buildd tasks (E.g. slave scanner, queue builder etc) are09:36
wgrantonly run one at a time, and potentially more often than once per09:36
wgrantminute.09:36
wgrantSounds useless, is in the wrong place, and breaks slightly with Python 2.6.09:37
bigjoolsno kidding09:37
wgrantAnd hasn't been touched significantly since 2005.09:37
wgrantbigjools: So, will anybody miss it if I remove it and its doctest as a step towards Python 2.6 support?10:12
bigjoolswgrant: where's the test?10:18
wgrantbigjools: lib/lp/soyuz/doc/buildd-sequencer.txt10:18
bigjoolsgood grief10:19
wgrantHm?10:19
bigjoolsdelete it with prejudice10:19
* wgrant destroys.10:19
wgrantIt's impossible to remove config schema elements at the moment, right?10:40
bigjoolshuh?10:46
wgrantDidn't rollouts explode last week because a config key had been removed, so the production configs (which still had a value for it) were invalid?10:47
wgrantAh, yes, bug 557271.10:49
mupBug #557271: Unable to remove config entries from the schema <Launchpad Foundations:New> <https://launchpad.net/bugs/557271>10:49
bigjoolshmmm11:00
deryckMorning, all.11:01
bigjoolshowdy deryck11:05
sidneijml, where's that patch you wanted me to land again?11:06
jmlhttps://bugs.edge.launchpad.net/zope.testing/+bug/56025911:07
mupBug #560259: Subunit output formatter doesn't handle layer setup errors <zope.testing:In Progress by sidnei> <https://launchpad.net/bugs/560259>11:07
jmlsidnei: it's the branch linked to that, and the attached patch11:07
sidneijml, both? or either?11:08
jmleither11:08
jmlsame patch11:08
sidneioh, looks trivial.11:08
sidneijml, you need a new release too?11:23
jmlsidnei: yes please11:24
jmlsidnei: and yes, it's really trivial :)11:24
sidneijml, done11:31
jmlsidnei: woot! thanks!11:32
sidneinp!11:32
=== mrevell is now known as mrevell-lunch
=== mrevell-lunch is now known as mrevell
henningesidnei_: can you come over to #lp-review, please? ;)13:30
=== sidnei_ is now known as sidnei
Ursinha_leonardr, ping13:56
leonardrhi ursinha_13:57
Ursinha_leonardr, hi, I hit an error with one of my lpapi scripts, I wonder if you know what could it be13:57
Ursinha_leonardr, "SyntaxError: no element found: line 16363, column 186"13:57
Ursinha_this is the error, when I try to login_with13:58
leonardrthat sounds like there's a problem with the wadl13:58
Ursinha_leonardr, yeah, but I have no idea why, since it works with another user13:58
Ursinha_leonardr, invalid credentials?13:58
leonardryou would get a 400 error. is that the entire error?13:58
leonardrthere's no headers?13:59
Ursinha_leonardr, let me show you the traceback13:59
leonardrno, i guess there wouldn't be since it's not an http error13:59
Ursinha_leonardr, this is the traceback: https://pastebin.canonical.com/30486/14:00
leonardrursinha_: i suggest catching the exception in _browser.py#get_wadl_application and printing out 'content'14:01
Ursinha_leonardr, sure, a moment14:04
Ursinha_leonardr, this is weird, it seems that the end of that 'content' is missing14:50
Ursinha_leonardr, line 16363, column 186 is the last line of that content, and it's not complete14:54
deryckkfogel, ping15:02
kfogelderyck: on phone, bbiab15:04
deryckack15:04
leonardrUrsinha_ that content comes from a file on disk. lib/canonical/launchpad/apidoc/wadl-*-*.xml15:04
leonardrcheck that file and see if it's truncated15:04
Ursinha_aha15:04
leonardrif so, remove that directory and make again15:04
Ursinha_leonardr, hmm, thing is I'm running my script on devpad15:06
Ursinha_works with one user and doesn't work with another15:06
Ursinha_leonardr, and I can't reproduce the error locally15:08
leonardrUrsinha_: i don't use devpad so i don't know what that means. you don't have access to the apidoc/*.xml?15:09
Ursinha_leonardr, exactly15:10
leonardrUrsinha_: i suggest you escalate to someone who does have access15:10
wgrantAren't the API scripts running against prod/edge?15:11
wgrantSo it's probably a client-side cache.15:11
Ursinha_leonardr, but I'm running the script with edge lp api15:11
Ursinha_what wgrant said15:11
Ursinha_wgrant, cache huh15:11
leonardrursinha_: on phone, sorry15:11
Ursinha_leonardr, no problem, thanks so far15:12
wgrantUrsinha_: Try obliterating/moving ~/.launchpadlib/api.edge.launchpad.net/cache15:12
Ursinha_wgrant, yeah, that worked15:21
Ursinha_wgrant, thanks15:21
deryckgmb, bug 294223 is done, right?15:26
mupBug #294223: Bugs missing after import from SourceForge <Launchpad Bugs:In Progress by gmb> <https://launchpad.net/bugs/294223>15:26
gmbderyck, Yep.15:27
gmbForgot about that15:27
deryckgmb, can you update and assign to the current milestone please?15:27
gmbderyck, Sure15:31
gmbDone15:31
deryckgmb, thanks15:32
deryckadeuring, can you update bug 261254 with linked branch, status, and assign to the current milestone?  I believe this is done, yes?15:43
mupBug #261254: Launchpad couldn't connect to ALSA Bug Tracker. <bugwatch> <oops> <story-reliable-bug-syncing> <ubuntu-qa> <Launchpad Bugs:In Progress by adeuring> <https://launchpad.net/bugs/261254>15:43
adeuringderyck: sure15:43
deryckadeuring, thanks!15:43
marssinzui, ping, do you have some time today to chat about changing some more /@@/ icons into sprites?  I have a question or two about it.15:49
marssinzui, if you have time, whenever you have time15:49
sinzuimars, I will after 3:00+. I do not know much about how icons become sprites. EdwinGrubbs wrote the tools that generate the sprite and css15:50
SlonUAmaxb: how is going ? =)15:51
maxbhello :-)15:51
marssinzui, ok.  EdwinGrubbs, do you have time to chat about sprites some time today?15:51
SlonUAmaxb: do u know how to enable karma .. just to see may digging =)15:51
EdwinGrubbsmars: sure15:51
marsEdwinGrubbs, what time for you?15:52
maxbI know nothing more than it involves a cronscript somewhere15:52
EdwinGrubbsmars: I can do it now, but it might be a little noisy since I'm at a coffee shop.15:52
marsheh15:52
marsEdwinGrubbs, that should be fine15:52
marsEdwinGrubbs, skype or mumble?15:52
SlonUAmaxb: i see 'karma has expired.' all time15:52
EdwinGrubbsmars: skype15:54
=== Ursinha_ is now known as Ursinha-lunch
=== deryck is now known as deryck[lunch]
marsintellectronica, online?16:16
intellectronicamars: yes16:16
intellectronicahow can i help you?16:16
marsTom, I'm looking at turning some of the /@@/edit icons in the bug status table into sprites16:17
marsintellectronica, can you see an issue with me doing so?16:17
marsthis is in the name of improved page performance, btw :)16:18
intellectronicamars: nothing i can think of off the top of my head. looks like a net win to me.16:18
marsintellectronica, cool.  I'll ask someone on the bugs team for a review when I'm done, just to make sure your team thinks my changes are sane.16:19
intellectronicacool16:20
marsthanks!16:20
=== barry is now known as flufl
=== flufl is now known as barry
=== deryck[lunch] is now known as deryck
jmlI'm forcing a rebuild of the recently failed 'lp' buildslave17:06
jml"no space left on device" seems to be the originating error17:07
=== beuno is now known as beuno-lunch
* jml discovers CachingIterator17:32
jmlit's just like lists in Haskell!17:32
jelmerooh, there is something like that?17:36
* jelmer has reimplemented it at least twice :-)17:36
jelmerjml: where does it live?17:37
jmljelmer, lp.services.utils17:37
jmljelmer, which is the best known home for such a thing17:37
jelmeroh, in Launchpad itself. It'd be a great thing to have in itertools...17:37
jelmerjml: Thanks anyway, I'm sure this will come in useful sometime :-)17:38
jmlyeah17:38
=== Ursinha-lunch is now known as Ursinha
* bigjools reboots from karmic into lucid17:38
jmljelmer, I should also add synchronize and dichotomy to itertools, I guess17:38
jmljelmer, I didn't write it, fwiw. I just found it.17:38
jmlgiven the docstring, I'll wager abentley wrote it17:38
jelmerjml: you can tell that just from the docstring?17:39
jmljelmer, "Some generators and iterators are expensive to calculate, like calculating the merge sorted revision graph for a bazaar branch"17:40
jelmerah, the reference to Bazaar17:40
jmljelmer, nope, thumper.17:41
=== leonardr is now known as leonardr-afk
=== beuno-lunch is now known as beuno
kfogelderyck: pong18:18
kfogelderyck: can't remember if your ping was internal or external channel :-)18:18
deryckdid I ping? :-)18:19
deryckoh, this morning... right18:20
deryckkfogel, are you working on stats for patch project?  Do I recall that correctly?18:20
kfogelderyck: yup18:21
kfogelderyck: https://code.edge.launchpad.net/~kfogel/tuolumne-lp-configs/patches-time-to-closing18:22
kfogelderyck: queries there; not much code yet18:22
kfogelderyck: abel is helping w/ queries btw18:22
deryckkfogel, ok, that's all I needed to know. I've been meaning to get back to stats for that project and wondered if your work included lpstats stuff.18:23
deryckkfogel, btw, you'll have to subscribe me to that branch to see it.18:23
kfogelderyck: oh, it's private?  one sec18:25
kfogelderyck: you want rev notifications too, or just branch attribute notifications?18:25
deryckkfogel, just attrs.18:25
kfogelderyck: done18:26
=== gary_poster is now known as gary-lunch
deryckkfogel, thanks.  and thanks for getting the stats together.18:28
kfogelderyck: you're welcome; but thank me when it's done -- it's hard to measure this thing!18:30
kfogelnot use, but benefit, I mean18:30
deryckyeah, that's why I hadn't got round to working on it yet.18:30
sinzuijpds, ping18:48
=== dpm is now known as dpm-afk
=== leonardr-afk is now known as leonardr
jpdssinzui: Hello.19:11
sinzuijpds, have you tested that you can control the official country mirror status using api?19:12
jpdssinzui: Yes.19:12
sinzuijpds, as you represent the primary user of the feature, you want to update this bug's tag to qa-oa: https://bugs.edge.launchpad.net/launchpad-registry/+bug/36165019:14
mupBug #361650: launchpad could know about official country mirror status <feature> <mirror> <qa-needstesting> <Launchpad Registry:Fix Committed by jpds> <https://launchpad.net/bugs/361650>19:14
jpdssinzui: Sure, done.19:14
sinzuithanks19:15
Ursinhahi gary-lunch, I filed a bug now, bug 562486, are you aware of this issue? and, is that really foundations?19:41
mupBug #562486: accessing pending_gpg_keys using the api fails to some users <api> <oops> <Launchpad Foundations:New> <https://launchpad.net/bugs/562486>19:41
=== gary-lunch is now known as gary_poster
gary_posterUrsinha: I was not aware of the issue.  It's definitely a foundations issue, and if it is a matter of how the code is exported rather than infrastructure, it is also a registry issue.  I'll ask leonardr for his opinion...once the oops tools are fixed :-)20:02
Ursinhagary_poster, :)20:03
Ursinhagary_poster, thanks20:03
gary_posterthank you20:03
sinzuigary_poster, Ursinha there are two issues that relate to this bug...20:05
sinzuigary_poster, 1, the seem to be available in one database and not another, leading to many oopses....20:06
sinzuibut there is a second debacle from last release where gpgkeys were renamed to gpg_keys, not not all the code was updated20:06
sinzuiUrsinha, EdwinGrubbs is landing a fix for the latter.20:07
gary_postersinzui: "seem to be available in one database and not another": which different databases?  Do you mean replication slaves and masters?20:08
sinzuigary_poster, yes, gpgkeys are manages via logintokens20:08
sinzuigary_poster, we also saw this happen in the CoC web UI20:09
gary_postersinzui: does that maybe mean that you need to always connect to the master for these queries, using one of stub's context manager things?20:10
sinzuigary_poster, I do not know20:10
* gary_poster doesn't really know context so is probably not helpful20:10
gary_posterit sounds like it20:10
gary_postersounds like a race condition20:10
gary_posterand if it is a race condition with the slaves, force using the master.20:11
gary_posterif you think that sounds reasonable, and you are not sure about stub's context manager, I can hunt up an example for you, sinzui20:11
EdwinGrubbssinzui: regarding the bug with +claim, does this mean that the Account.activate() method is also problematic since it sets a password that might be different from what was entered on the LoginService?20:29
leonardrjames_w, can you give me the specific failure you get when you trigger bug 561521?20:29
mupBug #561521: Success of PATCH request dependent on dict iteration order <lazr.restful:New> <https://launchpad.net/bugs/561521>20:29
james_wleonardr: it was in the mailing list post20:29
leonardrjames_w: ok, i'll find it and put it into the bug20:30
sinzuiEdwinGrubbs, I do not know. /me looks20:30
sinzuiEdwinGrubbs, isn't the account the SSO account, I do not think there is a Launchpad Account20:31
james_wleonardr: note that it's a followon error that is in the thread. If you want the exception thrown by the mutator then modify the test mentioned in the bug to print the response from the patch request.20:32
sinzuiEdwinGrubbs, the use case we are trying to avoid is a non-login user working with profiles and account. So if the user can access Account.activate() without being logged into the SSO server, there is a problem20:33
leonardrjames_w: i can't find the relevant post in my mailbox. what was the subject?20:35
james_w"Test failures on some platforms due to lazr.restful bug"20:35
EdwinGrubbssinzui: the reason I ask is that +claim gives you a login token that takes you to the ClaimProfileView in c/l/browser/logintoken.py and that calls emailaddress.account.activate(), so it would seem like the non-login user could access it.20:36
leonardrok, i got it20:36
sinzuiEdwinGrubbs, yes, that is taking the wrong path bad= person > email > account20:36
sinzuiEdwinGrubbs, good = account > email > person20:37
sinzuiEdwinGrubbs, so I think Account.activate is not a problem, but the callsite must know when/how to call it. claim does not20:38
EdwinGrubbssinzui: huh, does that just mean that we don't want to link a launchpad person that has the same email as the account but isn't already linked to that account directly?20:39
sinzuiEdwinGrubbs, we do want to link person to emails, we have to since emails are unique20:39
EdwinGrubbssinzui: I'm just trying to clarify why account>email>person is good.20:40
sinzuiEdwinGrubbs, I think this discussion is moot. there are two callsites, and you are removing one of them20:40
sinzuiEdwinGrubbs, So we can limit our discussion to Person.setPreferredEmail20:41
sinzuiEdwinGrubbs, Maybe we should change the guard in setPreferredEmail() and raise an exception instead of trying to fix the issue?20:43
EdwinGrubbssinzui: do you think it could be broken, or do you just want to get rid of the kludge?20:45
sinzuiEdwinGrubbs, We need to be careful in this method, I believe reset password switched the account from deactivated (by the user) back to activated. we may need to ensure the callsites (SSO!) has updated the account first.20:45
sinzuiEdwinGrubbs, ^ reset password switcheS THE ACCOUNT USING THIS METHOD20:46
sinzuiEdwinGrubbs, maybe I am over thinking this issue. setPreferredEmail() has no issues with active activated and deactivate accounts. suspend accounts cannot get here. so the problem is only when the account is not active.20:48
sinzuiEdwinGrubbs, I see scripts are calling setPreferredEmail. This could be very ugly.20:49
EdwinGrubbshmmm20:51
sinzuiEdwinGrubbs, I think we should remove the account hack in setPreferredEmail(), but I need to verify how we are resting password20:51
EdwinGrubbssinzui: well, I'll continue working on removing +claim for the time being.20:52
sinzuiyes, I suspect setPreferredEmail() is a separate bug20:53
sinzuiEdwinGrubbs, SSO reactivates the account itself. I think we can treat bug #248518 as a trivial deletion in a separate branch20:59
mupBug #248518: setPreferredEmail activated accounts <registry> <tech-debt> <Launchpad Registry:Triaged> <https://launchpad.net/bugs/248518>20:59
=== EdwinGrubbs is now known as Edwin-afk
mwhudsongood morning21:19
sinzuihi mwhudson21:20
* mwhudson tries to browse the code of the kernel import21:42
mwhudsonabentley, rockstar: https://code.staging.launchpad.net/~mwhudson/linux/trunk22:16
maxbtrunk != linux-2.6.31.y.git22:37
mwhudsonit's just the default in the form :)22:38
mwhudsonweee adding all the revisions in the scanner for that kernel import took 90 minutes22:55
=== matsubara is now known as matsubara-afk

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