pooliehi all00:00
pooliehi grid00:05
Noldorinhi jelmer00:20
michaelh1HI. http://doc.bazaar.canonical.com/plugins/en/ is a bit broken at the moment - shows a directory index.01:29
michaelh1I ended up there by Googling bzr colo which lead me to the missing http://doc.bazaar.canonical.com/plugins/en/colo-plugin.html01:30
pooliehi michaelh102:24
pooliei'll have a look at it02:25
poolieactually, i have to go out right now, but i'll have a look at it soon, thanks for pointing it out02:27
=== robbyoconnor is now known as hamsterdance
=== hamsterdance is now known as robbyoconnor
=== michaelh1 is now known as michaelh1|away
=== elmo_ is now known as Guest35863
mgzmorning all08:03
pooliehi all08:28
mgzhey poolie.08:28
pooliejust the man08:29
pooliecan you remind me - would python's default stream encoding change when it's run from cron?08:29
poolieit seems like it fails there but not when run from a terminal08:29
poolieeven with LANG set08:29
mgzI'll check, but eg sys.stdout.encoding possibly depends on having a terminal, it can sometimes be None certainly08:31
mgzyup, in pythonrun.c08:38
mgzfor some reason, the setting of the encoding attribute depends on isatty08:39
mgzif PYTHONIOENCODING is set to something (probably Python 2.7 only), that is used.08:39
pooliei thought so08:40
poolieand i'm lazy08:40
mgzit's not terribly clear why this logic is used.08:41
pooliekanbans have been failing from cron (and spamming me) because there is a unicode character somewhere in the output08:42
vilahey guys08:44
pooliei think i will just manually create an encoder08:45
mgzshortest fix is probably replace sys.stdout with a codecs module wrapper that encodes to locale.getpreferredencoding08:45
pooliehi vila08:45
vilajust recorded a new signature for lp being down, the package importer made tea happily and should do so for the next rollouts08:45
mgzas long as you're consistent with trying to output unicode, and not mixing in non-ascii bytestrings08:45
poolieit's actually writing to a file not stdout08:45
pooliebut yes08:46
vilaI'm still looking at the 6 failures which may also be recorded as identifying lp being down08:46
poolieuntil we deploy a new restfulclient it will continue to mishandle 503 in reading a collection, perahps08:46
poolie... or actually i think i put in a new workaround?08:47
vilahehe, one involves preload_root_objects :)08:47
vilapoolie: right08:47
vilain the mean time 'requeue --auto <pkg>' is good enough08:47
viladone, so 5 new signatures added,08:52
vilathe first one was *during* the downtime and as such represents the first access to lp which is the most important one08:53
poolieok running now, hopefully better, off to squash08:54
vilathe others were easy to analyze which is good as it means we can progressively update them to cover all cases as they appear08:54
vilapoolie: enjoy :)08:54
pooliehappy udd hacking08:54
mgzhappy squashing08:57
vilaexcellent, imports aborted during lp down time have already been re-processed with success (with some 'requeue --priority' help ;)09:00
jnlhi i've ran into a problem where im now missing a couple of commits: i had a colo-workspace with two branches "trunk" and "foo". i wanted to switch betweem them but forgot (or actually thought i didnt need it) to add the "colo:"-prefix and ran "bzr switch trunk" while on the foo branch, and then switched back with "bzr switch foo"09:56
jnlbut.. the foo-branch im on now is not the colo-branch "foo"09:56
jnlin fact:09:56
jnlbzr colo-branches09:56
jnlbzr: ERROR: No colocated workspace in .09:56
jnlis there any way to get my colo-workspace/branches back?09:56
jnli had a couple of commits in colo:foo that are gone now...09:58
jamjnl: find . -path '*/.bzr/branch' ?10:09
jnljam: ah! ./.bzr/branches/foo/.bzr/branch10:27
jamjnl: right, that is how colocated stores the real branches10:28
jamI'm not sure how to get your working space colo-ified again10:28
jambut the history is still available10:28
jnljam: bzr log in that dir looks like my missing branch!10:28
jnlsweet, thanks!10:28
redwyrexjelmer: I managed to get everything into svn, though I had to limit it to 100 changes at a time because the propery hook would fail for some reason after doing many11:46
jelmerredwyrex: oh, that's odd11:49
jelmerredwyrex: what's the error?11:49
redwyrexumm something about the tip revision or something11:50
redwyrexI assume it's on the svn side11:50
redwyrexanyway, thanks for your help11:55
lamontjelmer: still around?12:21
jelmerlamont: yes, hi\12:21
lamontsee chinstrap:~lamont/bzr-logs/*12:22
lamontthe .1 on the version is me messing up the upload and repackaging the source with the right dsc format version :(12:23
jelmerHmm, I thought ~9 already had the source format changed back to "1.0" ?12:25
jelmerlamont: The problem does appear to be that the patches aren't applied12:29
jelmerlamont: is this ~9 from the cat-proposed archive, or yours?12:30
lamontjelmer: it had source/format=1, but the .dsc was still new or such12:32
lamontArchitecture: any all12:32
lamontso I rebuilt it.12:32
lamontlet me see how a .2 fares12:33
=== tchan1 is now known as tchan
jammgz: I poked at https://answers.launchpad.net/bzr/+question/17512713:32
jamif you could follow up, that would be great, as I spent a bit too much time on it already13:32
mgzjam: I've just been reading that as I ate my lunch and was wondering how to send you a teddy bear13:33
mgzwill follow up as needed.13:34
jammgz: so we could do something like fix up leo-editor, which may require a losa to run a script we write (branch into a clean repo, copy that back over the original).13:36
jamI've seen complaints about leo-editor in bug reports before13:37
jamI'm pretty sure he didn't intend to include a file with 800MB of '\' characters13:37
mgzyeah, and the good news is bzr does now have a check for that kind of accident13:40
lamontjelmer: I suspect that I know what happened here, verifying that13:40
mgzgetting rid of that revision from the leo-editor repo sounds like a good plan13:41
lamontjelmer: how much longer will you be aorund?13:41
jammgz: sort of. If you add a file in revision 1, and then in rev2 you bloat it to 800MB, bzr won't warn you13:41
jamI don't think13:41
jam(the check was put in 'add')13:41
jamwhich is actually what happened here.13:42
jamthe file has a parent revision.13:42
jamwhich is a wee bit smaller :)13:42
mgzjam: I noticed when subscribing to things last night I'm down as the approver for foundatons-p-bzr-workflows13:50
jammgz: I think that was mostly that I wanted you to be present at the meeting, and by making you an 'Important Person' it will try to schedule around that.13:51
mgzI'm now an essential subscriber, so I think the schedual should now know that I need to be there?13:52
jamI think it is actually supposed to be in the linaro-training track, which was added after I was setting up the blueprint13:52
jamI'll see if I can do something about it13:52
mgzif there's anything else you need or would like me to do for UDS, poke13:53
jelmerlamont: I'll be around for another 3 hours13:53
jammgz: it looks like the actually session is going to be: https://blueprints.launchpad.net/linaro/+spec/linaro-training-bzr13:53
jamAndy set up a different session, so we'll just switch to that one13:54
jamwow, I wanted to try and mark it superseded by the other session.13:55
jamBut if you try to go there13:55
jamit requires that you select the superseding blueprint from a drop-down list13:55
jamof *all* ubuntu sessions13:55
jamit is about 1900 pixels wide13:55
jamand probably 10,000 entries in it?13:56
jamon the plus side, I can view source to actually get a reasonable view of it, and linaro-training-bzr isn't in there ...13:57
mgzyep, that was pretty funny14:01
mgzdid you delete it then, or did my hack go horribly wrong?14:01
jammgz: I retargetted it to linaro, and then superseded it with the above spec14:02
mgzah, good good.14:02
=== zyga is now known as zyga-afk
=== Guest35863 is now known as elmo
jelmerjam: wasn't a name change reason for updating a file's last modified data?15:37
jelmernevermind, there's something else going wrong15:50
=== deryck is now known as deryck[lunch]
=== zyga-afk is now known as zyga
=== beuno is now known as beuno-lunch
=== deryck[lunch] is now known as deryck
=== beuno-lunch is now known as beuno
jamjelmer: yes, rename to another directory, and changing a filename are in the per-file graph. Note that renaming a directory isn't in the file's graph17:50
jelmerjam: Thanks17:55
jelmerjam: I just fixed the last failing bzr-svn test, which was related to that.17:56
* jelmer prepares a victory email for the mailing list17:56
jamjelmer: congrats18:05
lamontjelmer: your test suite runs forever on some architectures.  just sayin19:27
=== yofel_ is now known as yofel
jelmerlamont: what's your definition of forever ? :-) More than an hour?20:25
lamont8.3 hrs20:26
lamontyou know, forever20:26
jelmerlamont: ah, worse than Launchpad's testsuite in its dark days. That is forever.20:30
lamontit's also a bbg320:55
Noldorinhi jelmer22:51
jelmerhi Noldorin \23:00
lamontjelmer: ticket updated, still waiting on some of the bzr builds to finish23:06
jelmerlamont: thanks23:06
jelmerlamont: That's indeed very long. These were the arm/sparc builders?23:07
lamontand I'll be updating a wiki page somewhere with lessons learned from the packaging fun23:07
jelmerlamont: thanks, that'd be useful23:10
lamontjelmer: don't delete your final versions from your ppa - I'll want to go poke at them a little in doing that update23:11
jelmerlamont: sure, I'll leave them around23:12
Noldorinjelmer, how's things going?23:14
jelmerlamont: do you know when hardy is going to disappear from the builders? after 12.04?23:17
jelmerNoldorin: alright, how are you?23:22
jelmermanaged to get the full bzr testsuite to pass against bzr-svn earlier today, which I'm quite pleased with23:22
spivjelmer: \o/23:23
jelmerg'morning spiv :)23:23
jelmerhow are things at the big G?23:23
thumperjelmer: congratulations23:23
jelmerthumper: thanks23:25
jelmerI'll be really pleased when we manage to do this for git and hg, but that's some time away23:25
Noldorinjelmer, pretty decent, just as busy as ever heh.23:30
Noldorinjelmer, same with you? bzr 2.5 keeping you occupied?23:30
=== r0bby_ is now known as robbyoconnor

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