=== yofel_ is now known as yofel
pooliehi riddell06:40
vilahey poolie and others !06:51
pooliehi there06:53
pooliei'm giving udd config options to talk to staging lp to make it easier to debug/test stuff06:53
vilapoolie: great !06:53
vilapoolie: don't worry too much about using the stacked-based configs, I intend to upgrade udd when 2.4.x is deployed on jubany anyway06:54
poolievila, so how about https://code.launchpad.net/~mbp/udd/configure/+merge/7182207:46
poolieoh also, while i'm here, can you see if you can get into the blog?07:46
* vila looking07:46
vilapoolie: this won't work until 2.4x is deployed I think07:47
vilapoolie: meh, I'm out-of-date about iconfig07:49
vilagee, I didn't remember introducing a forward-compatible layer there...07:50
poolieah, are you saying iconfig's present but can't be used?07:50
vilano no no, you're doing fine, my bad07:51
pooliethanks for thinking of it though07:53
pooliehi thar07:55
jammorning all07:56
pooliehi jam07:56
vilapoolie: reviewed08:02
vilapoolie: BB:tweak08:02
pooliei'll roll that out then08:31
poolieactually i might try to actually fix the httperrors first08:38
jimisI did a bzr switch, and an unexpected conflict happened08:54
jimisCan I undo and go back to previous branch to commit?08:54
pooliehi jimis08:55
poolieyou can go back but i think that will carry back the conflicts08:56
pooliethere ought to be a way to say 'i wish i never switched (or, merged, or whatever)' but there isn't yet08:56
jimisI was afraid of that :-(08:58
jimisanyway I understand that this workflow is not considered as mainstream for bzr08:58
* jelmer waves08:58
pooliehi jelmer08:58
jelmerg'evening poolie08:59
jimisimho it should ask before switching, saying that a conflict has occured08:59
vilajimis: out of curiosity, what kind of unexpected conflict did you get ?08:59
jimisvila: I had changed a file but forgot to commit09:00
jimisso when I did "bzr switch other-branch" a conflict happened09:01
jimisand now I have to resolve it09:01
vilaoh, so you wanted to commit it, not switch to the other branch carrying the change ?09:02
jimiswow, switching back created a double conflict :-o09:02
vilayeah, never a good idea to leave unresolved conflict09:02
jimisvila: no I wanted to switch, but I had forgotten to commit this particular file09:02
vilaright, that what I meant, you wanted to commit *before* switching09:03
jimisthat's what I should have done, yeah09:03
vilathere are cases when you want to switch but still want to carry the change over09:03
jimisbut bzr should have reminded me because I forgot09:03
jimisI intend to file a bug when I find some time, about missing functionality when working with checkouts09:04
vilayeah, kind of the --strict option on pull/push and... I forgot the other one send ?09:04
jimisI understand it is not a mainstream workflow, but it is the *only* possibility when working with branches sized as big as lp:gcc09:04
jimismultiple dirs is not an option then09:05
Riddellvila: where is baboun again?09:10
jelmerRiddell, under vila's desk :)09:10
* Riddell books a trip on the eurostar to test his branch09:10
jelmerjam: did you have more comments on my transport-segments branch?09:10
vilaRiddell: http://babune.ladeuil.net:2484209:10
vilajelmer: :)09:10
jelmerRiddell: :)09:11
jamjelmer: no, just the display thing, Riddell had already approved the branch09:11
Riddellvila: hmm, how do I build a branch again?09:14
Riddelloh, helps if I log in09:14
vilayeah, a lot ;)09:14
jelmerjam: thanks :)09:15
jamvila: do you have python-2.4 on babune?09:23
Riddellvila: I scheduled selftest-subset-natty but it says "pending - natty64.local is offline"09:25
vilaRiddell: yup, the slave needs to be started, just wait a bit09:25
vilajam: probably, depends on the slave09:25
jamwhen did FF go to 6.0?09:26
vilajam: why would you care, we don't support anymore ?09:26
jamdid they just decide to stop doing minor versions09:26
jamvila: I'm backporting something to bzr-2.1 for Lucide09:26
jamso we *do* care there09:26
vilathen it should be available on the lucid slave09:26
vilajam: note that installing a local lucid chroot will probably be more convenient for you09:27
vilaunless you just want a final test run that is09:28
maxb  09:30
jamvila: yeah, pretty much09:30
jamI think everything is ok09:30
jamI just want to make sure before I submit it, and it goes boom when we release09:30
vilabut.. doesn't python defaults to 2.5 on lucid already ?09:32
vila   python | 2.6.5-0ubuntu1 |         lucid | all09:33
vilayeah, 2.6 even09:33
jamvila: sure, but we are still backporting to a 2.1 series, which *is* 2.4 compatible09:33
vilaI'm not even sure 2.4 is supported on lucid (imbw)09:33
jameven backporting to bzr-2.3 we should be verifying python-2.4 compatibility09:34
vilaright, but I'm not sure you can use lucid to test 2.4 you may need to dig deeper09:34
jamI thought you had that set up on babune09:34
jam(since it at least are excuse for why it was ok to upgrade PQM)09:34
poolieok this time for sure09:35
jampoolie: ? (this time you're gone for sure?)09:38
pooliethis time i think i fixed bug 81991009:38
jamor this time httperror is fixed09:38
ubot5Launchpad bug 819910 in Ubuntu Distributed Development "many packages fail to import due to HTTPError talking to Launchpad (eg ubuntu:transmission)" [Critical,In progress] https://launchpad.net/bugs/81991009:38
poolieit took kind of a silly long time09:38
vilapoolie: \o/09:38
pooliewe'll see09:38
pooliei have something a bit more realistic working locally though09:38
pooliewhich is a good step09:38
poolienot my best day ever09:39
pooliequick review anyone?09:40
jampoolie: looking09:43
vilanot at lot to look at ;) can't you add a test for that, even against a live lp ?09:43
jampoolie: str() is worrying for me, though it may be correct. And it certainly looks like something that Launchpadlib is likely to break09:43
jamsince you're poking at private attributes09:44
pooliethe object itself is a URI class09:44
vilaand yeah, why the str() ?09:44
poolieand it defines a str as apparently the right way to get an actual string09:44
jampoolie: ok, that part works for me.09:44
poolieother launchpad client infrastructure won't accept the URI09:44
poolieas far as using the underscoe09:44
pooliethere's no public equivalent09:45
pooliemaybe i'll ask in #launchpad09:45
jampoolie: there's no way to ask for a branch's LP URL?09:45
jamor is it that you have to go via a round-trip that you are trying to avoid09:45
jamlike lp.branches['~user/project/branch'].url seems like it should work09:46
jambut that is a round trip you might be trying to avoid09:46
pooliedoes that work?09:46
poolieit would be cleaner09:46
pooliei don't think this will be a performance problem09:46
poolieit's not hit all that much compared to the amount of other work this program does09:46
jampoolie: https://launchpad.net/+apidoc/beta.html#branches09:47
jamI'm thinking09:47
vilayeah, make it work, make it right, make it fast... and it doesn't work yet ;)09:47
pooliethat doesn't work; it seems to expect an integer index09:48
pooliebut there might bbe something similar09:48
jampoolie: so probably lp.branches.getByUniqueName(short_form).self_link09:48
poolieor no09:48
lifelessjam: I wonder if you could have a look at https://code.launchpad.net/~lifeless/python-oops-wsgi/extraction/+merge/7183909:48
lifelessjam: its a replacement for the oops middleware we use in launchpadloggerhead09:49
jamlifeless: no diff yet, but I'll give it a minute09:49
lifelessyeah, no panic09:49
lifelessbedtime for me, 7am meeting tomorrow ;)09:49
jamlifeless: maybe you have a hint. Trying to get the LP API reference for a branch known by its short name09:49
poolieyep, that's it09:49
jamtrying to get something like api/1.0/...09:49
jampoolie: getByUniqueName is correct?09:50
lifelessjam: you want the devel api, not beta09:50
jamlifeless: this is for the udd package importer09:50
jamI think we found it, though.09:50
lifelessbeta is -deprecated-09:50
lifelessdead dead dead09:50
lifelessfor a live service under maintenance, I would definitely be using the devel version of the API09:51
pooliewhere is he using beta?09:51
poolieoh in the doc api09:51
pooliei don't think that's a problem unless we still have an lplib that defaults to beta09:52
jampoolie, lifeless: sure, it was just what firefox quick-linked for me when typing "apidoc"09:52
poolieoh actually if we're going to call we can just use getByUrls and simplify this a bit more09:52
jampoolie: or getByUrl, just depends what you actually have.09:53
lifelesspoolie: its a problem to the extent that there is stuff in 'beta' that no longer exists, and conversely not stuff in beta that now exists.09:53
lifelessbest to read the docs for the api being used.09:53
pooliei am looking at the 1.0 docs09:53
pooliei mean i don't think there's a practical problem for the importer09:53
lifelessI get that; I'm just trying to ensure you don't do what I've seen others do which is try a gone api and be confused when its not there but is on the web page.09:54
poolieoh it is actually gone?09:55
lifelessthat particular one isn't.09:55
poolieperhaps the beta api doc pages should be removed09:55
lifelessbut, in beta particularly, a great many have been deprecated and are now gone09:55
lifelessand there is a great deal in 1.0 that won't be listed on the beta page09:55
lifelessso you might miss out on just the right api for the job09:55
lifelessjam: that diff is up; I realise I haven't glued the url inclusion in the report in yet, I'll do that as a separate landing I think09:56
poolieso, jam needs to just update his bookmark or browser history09:56
lifelessmore or less ;)09:57
pooliejelmer: hooray10:06
jelmerpoolie: glad that finally seems to be taking shape :)10:07
jamvila: so, do you have a test runner on babune that can run older versions of bzr against python-2.4?10:08
vilajam: not sure, do you know which ubuntu release supported 2.4 in the end ?10:10
vilahardy is 2.5 by default iirc10:11
jamvila: probably default, but still supported10:11
jamvs you can't even install python-2.4 on Lucid10:11
vilaafter a quick check, babune always use the default python so there is no existing job to test 2.4 explicitly, but you can probably create one pretty easily10:14
vilathe hardy slave has python-2.4 installed10:14
pooliejam, ok, how about https://code.launchpad.net/~mbp/udd/819910-service-root/+merge/71832 then10:17
jampoolie: lp_call is really how you have to spell that? ugh10:18
jambut otherwise, approve10:18
jampoolie: you have a comment "Sripo of the scheme, and hostname" that should go10:18
poolielp_call is a wrapper that handles retrying the operation etc10:19
poolieit really ought to be in lplib10:19
pooliealso, for that matter, it would be nice if you could declare special forms in python for things like this10:19
poolieok, i'll merge it and try it10:22
jampoolie: care to chat about my patch?10:30
pooliei'm just trying to roll this onto jubany and it seems to be breaking10:30
poolieafter that sure10:30
pooliethis is your get_parent_map patch?10:33
pooliejam, ok that _seems_ to be working10:55
pooliebit hard because of bug 82793510:55
ubot5Launchpad bug 827935 in Launchpad itself "consistent timeout on user branch listing page" [Undecided,New] https://launchpad.net/bugs/82793510:55
pooliejam i might have some dinner now, i'll try to ping you later11:00
jamI should have lunch as well11:00
jelmerhmm, it's kindof worrying that PQM is now so slow that we can keep it busy overnight :-/11:11
vilajelmer: ... and still has a queue of 5 :-(11:21
Riddellso it this my fault or baboun's? http://babune.ladeuil.net:24842/view/debug/job/selftest-subset-natty/12/console11:32
jamRiddell: isn't everything your fault? You're the new guy, after all.11:33
Riddellit might be less my fault if I learnt to spell babune11:33
vilaFATAL: Unable to delete script file /tmp/hudson2041655501694922969.sh11:34
vilashudder, long time not seen11:34
vilaslave died for mysterious reason, try again, using a smaller subset may help11:35
vilamaking chroots real nodes and switching the old jobs to these new nodes will also address the issue but I'm not there yet...11:36
vilaI've made some progress on upgrading the gentoo slave (which I couldn't do for.. months), currently rebuilding 12/258 packages...11:37
vilamaxb: ping11:46
vilajam: are you planning to build th windows installers for 2.4.0 ?11:47
jamvila: at some point, though I was still dragging my feet based on what Jelmer has stated about bzr-svn (critical bug still remaining)11:47
vilahuh ? Where ? When ? jelmer ?11:48
vilacrap, just found the mail11:48
vilajelmer: is there a bug # for that ?11:49
jamvila: https://bugs.launchpad.net/bzr-svn/+bugs?search=Search&field.importance=Critical&field.status=New&field.status=Incomplete&field.status=Confirmed&field.status=Triaged&field.status=In+Progress&field.status=Fix+Committed ?11:51
jam3 critical bugs11:51
vilajelmer: are all these bugs relevant for 2.4.0 or only for prior versions ?11:54
vilajelmer: also, are they regressions and do they affect the majority of users ?11:55
jamvila: I think you scared him away11:55
vila*I* am the one who is scared, I know nothing scares jelmer ;)11:56
jelmerI'm lunchin' :)12:01
jelmer826946 is the one that's worrying me12:02
jelmer704716 is fixed so no longer an issue12:02
jelmer485601 is fixed I think but doesn't have tests yet12:02
vilapfew, good to know (bon app├ętit by the way)12:03
vilajelmer: should bug #826946 block the 2.4.0 release (or even the 1.1.1 ? bzr-svn release) though ?12:05
ubot5Launchpad bug 826946 in Bazaar Subversion Plugin "Commit to svn repo completes, but repo is not updated" [Critical,Triaged] https://launchpad.net/bugs/82694612:05
vilaHMPF (no space left on device, gentoo I hate you)12:06
jelmervila: if it is as serious as it appears to be, it's a serious regression12:06
vilajelmer: let me know how it goes12:06
pooliejam, are you back?12:15
jampoolie: yep12:15
poolievila, is enospace connected or something else?12:15
pooliejam do you have an mp up already?12:16
jampoolie: yeah, I posted it while you were dining12:16
vilapoolie: I went to /etc/fstab and found the comment I left ages ago when the same problem happened: don't use tmpfs for /var/tmp if more space is needed for emerge12:17
jampoolie: if you want to read through the merge proposal, I think I've captured everything I had to say about it. The only other bit that would-have-been-nice is to have you run some tests from Sydney12:21
jambut they take a while, and I didn't work up a script to automate all of it12:21
pooliei might do that tomorrow12:21
poolievila, just add more swap12:25
vilapoolie: or more memory or just comment out the fstab line which I did. After a reboot, the emerge is going on now12:26
pooliejam, i'm reading it12:27
poolieor tell emerge to use a different tmpfs12:27
=== Mkaysi_ is now known as Mkaysi_|ZNC
pooliewhat a nice analysis12:34
poolieand what a long commit message :)12:34
=== Mkaysi_|ZNC is now known as Mkaysi
pooliejam, hi12:43
poolieso the reasoning and the code make sense to me12:43
pooliei can't see any actual bugs (other than a couple of tweaks)12:43
pooliei'm not super confident that i would catch any bugs that actually were there though, cause it's moderately large12:44
poolieso, perhaps i should just try it out a bit tomorrow12:44
jamI did manually test it against launchpad a fair amount12:44
jamNot that it exhaustively tests all possible edge cases.12:44
pooliei think landing into trunk first makes sense12:47
poolieit's really nice it doesn't need a server change12:47
poolieis there anything in particular we should talk about?12:48
jampoolie: I think I've worked through it. I mostly wanted to do the summary live-on-the-phone, but it has been done12:54
pooliethe flood of mp mail indicates the lpapi connection is unjammed12:55
pooliei wonder if people will feel spammed12:56
pooliei guess it's a good chance to consider how much they are useful12:56
jampoolie: when don't people feel spammed by lp?12:56
pooliei don't know12:56
pooliei certainly do12:56
pooliei'd like to do built-in non-email notifications some time12:56
pooliehttps://bugs.launchpad.net/launchpad/+bug/827935/comments/1 might amuse12:57
ubot5Ubuntu bug 827935 in Launchpad itself "consistent timeout on user branch listing page" [Critical,Triaged]12:57
pooliei should sleep12:57
jampoolie: real quick, what time would be reasonable to load the global config to check the value?12:59
jampoolie: OFFSET 313000 seems particularly telling13:00
jam(sort all 500,000 entries, and give me the middle 48)13:01
poolieyeah that offset looks to me like it's either loading them all (i'm not sure if that's what the oops means) or slicing a big collection13:03
pooliei haven't dug into it13:03
pooliehow do you mean 'what time'?13:03
poolieas in, where in the code, or how long do i think it would take?13:03
jampoolie: where in the code13:03
poolieperhaps when the remotebranch or similar is constructed, make an instance and hold it?13:04
jamI don't think we want to check the config on every _get_parent_map_rpc call13:04
pooliethat should make sure it's read just once13:04
jamthis is in RemoteRepository13:04
poolieok, remoterepository then13:04
poolievila may have a more evolved idea, but putting it there should mean it's not loaded too many times in normal use, and we can move it later13:04
jampoolie: but you're thinking to cache the GlobalStack object, not just the config value13:05
jamvila: ^^ thoughts?13:05
poolieis even reading out of that expensive?13:05
poolieit may be13:05
jampoolie: I don't think it is expensive as in, will show up in real-user-time branching all of history13:05
jammaybe in shorter term. But for all of bzr it only does 600 or so RPCs13:06
jambecause of the preloading behavior13:06
vilajam,poolie: is there a global config on lp ?13:06
poolieok, good night then13:06
jamvila: a config for all branches on lp? no13:07
pooliejam, if it's nontrivial to change i wouldn't block on it13:07
jamthough you could do a locations.conf thing13:07
pooliewe can always add it if it's wanted13:07
vilaI suspect not (today)13:07
poolieanother option would be to go off a debug flag which we know should be cheap13:07
pooliegood night all13:08
vilajam: if you're talking about *local* config then the plan is that checking an option will be a bit more than checking an in-memory dict (or thrice) i.e. far lower than reading a file13:09
vilapoolie: g'night13:09
vilajam: or doing an rpc call13:10
jamsure, but "plan" and what will "land in 2.4" is pretty different13:11
jelmerg'night poolie13:12
vilajam: well, stop worrying about a bug that has been there since the beginning then13:13
vilajam: *all* options requires reading one or several config files so far13:14
jamvila: which is why I've generally *avoided* having things in config files so far13:45
vilajam: then keep doing so until we get a better config or join the effort to get there ;)13:49
* jelmer is generally also not too fond of adding configuration options13:55
jelmerit's often the easy answer, deferring to the user instead of thinking about what the right thing to do is13:56
vilayup, I know the trap13:56
bigjoolsas a KDE user, I say moar config!13:56
vilathat doesn't mean nothing should be configurable13:57
jelmerbigjools: ((-:13:58
vilajelmer: the other extreme example is OSX (or more generally Apple) where you *can't* configure what you want because *they* know better14:00
jelmervila: yeah14:04
* jelmer takes a quick break14:11
=== Mkaysi is now known as Mkaysi|ZNC
vilajelmer: back ?15:11
jelmervila: yep15:17
vilajelmer: you mentioned a fallout about my last patch around BZR_HOME, I can't find the thread anymore :-/15:18
vilajelmer: but from memory you mentioned bzr-svn tests failing because of it. I was wondering if these tests should be put into core instead15:20
jelmervila: basically, if BZR_HOME is set then bzrlib.config.config_dir() returns unicode strings15:20
vilayeah, I remembered that15:20
jelmervila: GMTA. lp:~jelmer/bzr/cachedir was an attempt to do just that :)15:20
vilaI don't clearly see the connection though :)15:23
vilaor was it just a first step ?15:24
jelmervila: bzr-svn falls back to storing its cache in ${config_dir()}/cache/svn if it can't find a cache directory15:24
vilaha, that's the connection, pfew, but... where does cache_dir calls config_dir ?15:26
jamjelmer: so it looks like PQM is still chewing through stuff you submitted yesterday. Not a good sign.15:31
jamI do wish we had some stats on time-to-run-the-test-suite15:31
jamanyway, I'm off for now. Have a good night y'all15:32
vilag'night jam, the issues you mentioned have been briefly discussed here ;)15:34
jelmerjam: g'night15:36
jelmerjam: it does indeed look worrying :_/15:36
jelmervila: it doesn't, but the bzr-svn equivalent of it does15:36
vila....finally :)15:37
vilasooo, no need to search how to avoid return unicode right ? Better spend my time on reviewing your branch instead ? Or are you already addressing jam's review ?15:38
vilajelmer: ^15:43
jelmervila: Yeah, I'm addressing jam's concerns at the moment.15:45
=== beuno is now known as beuno-lunch
=== Mkaysi|ZNC is now known as Mkaysi
jelmerhave a good evening16:52
wilxHi, I am having problems installing/running bzr-fastimport: http://codepad.org/Uz43PkP417:19
wilxThis is the selfcheck fastimport as suggested by the README.txt output.17:19
wilxI run setup.py build and then I have copied the directory to ~/.bazaar/plugins/fastimport17:20
fullermdThat's happening in bzr-git, not fastimport...17:25
=== deryck is now known as deryck[lunch]
fullermdRiddell: ping   (if'n you're around still)17:38
=== beuno-lunch is now known as beuno
=== deryck[lunch] is now known as deryck
wilxfullermd: Oh, so it is not a problem with my installation steps of bzr-fastimport?19:00
wilxShall I report it as a bug as suggested?19:00
maxbPerhaps, but not on fastimport, because the problem is an version incompatibility between bzr-git and dulwich19:18
hexspriteI _think_ i'm using a checkout, why do I get this? ERROR: Cannot switch a branch, only a checkout.19:21
maxbwilx: Actually the problem is simply: If you install dulwich 0.8.0, you must also install bzr git >= 0.6.219:22
maxbhexsprite: Try "bzr info", see if it is as you expect. Pastebin it for more help.19:22
lifelessmgz: hi19:22
hexspriteMaybe that's not even the right command for what I want to do.  Which is to switch my working copy to be based on a new repository located in my homedir19:22
mgzhey lifeless19:24
hexspriteeg. i just merged all my changes from my feature checkout in ~/feature into ~/trunk repo. now i want to update ~/feature to be based on ~/trunk19:24
lifelessmgz: I can meet anytime, if now would be more convenient for you19:24
lifelessmgz: I just need 5 minutes to feed cats n stuff (I"ll be right back after that - let me know :P)19:25
maxbhexsprite: Please pastebin the output of "bzr info ~/feature" and "bzr info ~/trunk" - it will help explain the details of your situation19:25
hexspritebzr info for both: http://pastebin.com/eMF1iiVD  -- note ~/main a checkout of trunk and 1079_dojo is where I'm actively working on new features19:28
mgzlifeless: I'm ready when you're back19:28
hexspriteto get 1079_dojo up to date so should i be doing: bzr switch ~/main19:30
lifelessmgz: kk19:36
lifelessmgz: do you have skype ?19:36
mgzI don't have a decent mic to connect to the computer unfortunately19:37
lifelessringing in  a sec19:37
jxselfSays 2.4 was released August 8?19:41
jxselfhttp://bazaar.canonical.com/en/ disagrees19:41
memehi im trying to use bzr to push source code to a launchpad project but i have no idea what im doing i think im doing it right but i get bzr: ERROR: Not a branch: "/home/jadon/airscripts/.bzr/branch/": location is a repository.19:46
meme 19:46
memeanyone able to help?19:47
jxselfRelax. Be patient, young grasshopper. :)19:50
beunomeme, you're not in a branch, you're in a repostiry19:50
memeso i need to cd deeper tell im in a branch?19:51
hexspriteseems like bzr pull ~/main was what i needed19:55
memewell i did something that happend and now it does other stuff and it seems to work19:57
=== hexsprite_ is now known as hexsprite
=== Quintasan_ is now known as Quintasan
=== yofel_ is now known as yofel
Riddellfullermd: pong22:08
fullermdOhai.  Um.  Why was I pinging you?22:10
* fullermd scratches his elbow.22:10
fullermdOyah.  Are you doing much actively on qbzr?  I noticed you popping up in the commit logs somewhat lately...22:11
wilxmaxb: Thanks, that's probably it. I have only bzr-git 0.6.1 installed.22:56
controversiali've tried svn, git, and hg, and i'm using hg right now.  i'm curious about bzr though.  how does that compare against hg22:57
=== krow_ is now known as krow

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