[06:40] <nmb> Any devs awake who could take a look at bug 1019954 ?
[06:41] <nmb> It seems like the PathFilteringTransport that is used in the server jail is not letting the RemoteBzrDir access the paths at .bzr/branches, but I can't quite track down why.
[08:08] <mgz> morning!
[10:51] <vila> jelmer: the fix for weave I mentioned last week is up for review: https://code.launchpad.net/~vila/bzr/1020007-branchformat4-lock/+merge/113014
[10:53] <jelmer> vila: cool, I'll have a look
[11:05] <jelmer> vila: r=me
[11:06] <jelmer> now we just have to wait for python-launchpadlib to be uninstalled before it can land..
[11:07] <vila> urgh, you filed an RT for that ?
[11:08] <jelmer> vila: jam replied to the RT about installing it
[11:08] <jelmer> vila: 54023
[11:09] <vila> cool, thanks
[11:11] <vila> oooh, the pkg_resources warning popping its ugly head, good, hopefully this will help address this mess ;)
[11:12] <jelmer> vila: well, if we just uninstall python-launchpadlib that won't really address the issue
[11:12] <vila> indeed, but someone may get more incentive to fix the root issue (which has always be a total mystery for me)
[11:15] <jimis> hi, how can I checkout only specific files from my +junk repo into current dir?
[11:25] <jimis> For now I'm using "bzr checkout lp:~user/+junk/dir ."
[11:25] <jimis> but it brings all files so it's a bit of a mess
[11:26] <mgz> go back a step, what are you actually trying to do?
[11:27] <jimis> fetch a script of mine into a cgi-bin directory
[11:27] <jimis> but I don't need all the rest
[11:28] <jimis> hmmm I'll work around that by fetching elsewhere and hard-linking
[11:29] <jimis> if you think of a way to checkout a file in bzr please let me know, I'll be back in a while
[11:29] <mgz> that's a reasonable approach
[11:30] <jelmer> jimis: bzr cat can do that for you
[11:30] <jelmer> bzr cat -d lp:~user/+junk/dir name-of-file > name-of-file
[11:30] <mgz> then chmod, then try and remember not to edit the file
[11:30] <mgz> having an actual branch and a link is better
[11:31] <mgz> partial checkouts into an existing dir would have a worse problem,
[11:32] <jelmer> I guess views could in theory do this for you?
[11:32] <mgz> you'd have files both in the repo and the current dir that the vcs has no idea how to sync
[11:33] <mgz> views would work provided the local dir was kept clean
[11:36] <mgz> I'm not sure if they actually do anything useful with the tree though
[11:39] <mgz> right, they just affect what bzr sees, not what's on disk
[11:40] <mgz> jelmer: bug 1001169 is a dupe of bug 413430
[11:41] <jelmer> mgz: marked as such
[11:42] <mgz> ta
[12:22] <jimis> jelmer: with bzr cat I wouldn't be able to update
[12:23] <jimis> mgz: local dir not clean, but thanks anyway
[12:23] <jimis> I'll follow the link way :-)
[12:24] <jelmer> jimis: you can just rerun the command to update
[12:27] <jimis> jelmer: that wouldn't merge my local differences :-p
[12:27] <jelmer> jimis: true
[12:28] <jimis> (some values are set on top of the script)
[15:25] <jelmer> vila: could I tempt you to do a review of https://code.launchpad.net/~jelmer/bzr/branchfmt/+merge/100220 ?
[15:29] <vila> jelmer: hmm, yet another branch with bzrlib/tests/per_branch/test_bound_sftp.py.THIS where are you getting that from ?
[15:30] <vila> jelmer: note the final '.THIS'
[15:31] <vila> jelmer: and by 'another' I mean I already encounter that in one of those suspicious merges I mentioned last week
[15:34] <jelmer> vila: hmm, I'm not sure... it must have been from a long time ago
[15:35] <vila> the scary bit is that that kind of file comes from a conflict and resolving the conflict should delete it...
[15:37] <jelmer> I guess I had a local change to it at some point..
[15:42] <vila> introduced in lp:bzr at revno 6503 removed at 6519 (guess who committed them ;)
[15:43] <vila> so, no harm done on trunk (also 2.6b1 apparently was released with it), but it would be nice if you could kill for good ;)
[15:47] <vila> jelmer: reviewed as needs info, please enlighten me ;)
[16:10] <jelmer> vila: commented
[16:17] <vila> jelmer: So, this is about reducing branch.py size ? Will that really make a difference on startup ?
[16:29] <jelmer> vila: a bit - it also makes it easier to move the rest of this stuff out into a oldfmt plugin or something like that
[16:30] <jelmer> and it simplifies the overall amount of code in branch.py, which should make it easier to read
[16:34] <vila> hmm, right, would be nice to put your explanations in branchfmt/__init__.py then or newcomers will have a hard time understanding the in-progress layout
[16:44] <jelmer> vila: is that really necessary? We don't have something like that for repofmt either
[16:46] <vila> jelmer: repository.py and repofmt are more cleanly separated, you're not there yet with branchfmt, so my review is: please add comments. It will take you less time than trying to convince me to not put them :)
[16:46]  * vila off to dentist
[18:57] <mgz> heh
[18:58] <mgz> Jelme is so fast he lost his 'r'
[18:58] <mgz> "it's part of bzrtools" is an increasingly bad answer
[19:03] <jelmer> using the R hurts my throat too much :P
[19:03] <mgz> it's been rolled too much.
[19:04]  * mgz bounces bug 1019984 on again
[19:05] <mgz> maybe looks like bzr-git should override the dulwich SSHVendor with that from bzrlib?
[19:14] <jelmer> mgz: I guess that would be an option
[19:43] <Noldorin> heh
[19:43] <Noldorin> push issue again with bzr-git eh!
[19:43] <Noldorin> jelmer: does that tempt you to fix the other dpush issue while you're at it? ;)
[19:46] <jelmer> Noldorin: as berfore, patches welcome :)
[19:46] <Noldorin> jelmer: still can't persuade you damnit
[19:46] <Noldorin> lol
[19:46] <Noldorin> i only mention this again because someone is using my git mirror now, :P
[19:46] <Noldorin> and it's annoying
[19:47] <Noldorin> you know that a) i don't understand the codebase very well, b) i'm shit at python, c) i don't have the time...
[19:47] <Noldorin> jelmer: but you said you'd get on it if someone paid you ha
[19:47] <Noldorin> maybe i should find out how much now lol
[19:50] <jelmer> Noldorin: I don't really have the time either (or rather, I choose to spend my spare time on other things)
[19:50] <Noldorin> yes i know
[19:50] <Noldorin> you've said before...
[19:50] <Noldorin> but a) and b) are true still ;)
[19:50] <Noldorin> hence why i asked if someone paid you heh...
[19:50] <jelmer> Noldorin: I'm not the one bringing it up again..
[19:51] <jelmer> :)
[19:51] <Noldorin> hmm i think you misunderstand
[19:51] <Noldorin> i'm not criticising you here
[19:51] <Noldorin> language barrier perhaps. never mind :)
[19:54] <mgz> window not having an ssh binary and bzr push not updating working tree aren't really bzr-git related
[19:54] <Noldorin> not what i'm talking about though
[19:57] <mgz> the thing with pet bugs is only you can care for them properly.
[19:59] <jelmer> mgz: bug 1019984 is definitely not a dulwich bug though
[20:00] <jelmer> mgz: dulwich shouldn't know about bzr's SSHVendor
[20:00] <jelmer> I'll reassign it to bzr-git
[20:00] <mgz> well, arguably dulwich should have a less lame SSHVendor
[20:00] <mgz> but I agree bzr-git is probably the easiest place to fix it
[20:01] <mgz> if I ever get downstairs I'll give it a go.
[20:01] <jelmer> mgz: I would argue that bzr's SSHVendor is the lamer one to be honest
[20:02] <mgz> it works on windows though! mostly...
[20:03] <jelmer> dulwich support for paramiko would be ace too
[20:03] <jelmer> that's quite different from monkeypatching bzr's SSHVendor into Dulwich though
[20:06] <mgz> right.
[20:06] <mgz> the code in dulwich seems to encourage monkey patching
[20:06] <mgz> I'm not sure if that was the intention.
[20:10] <jelmer> sortof, it was meant to allow bzr to plug in its own ssh thingies
[20:10] <jelmer> and mercurial
[20:42] <Noldorin> hah
[20:42] <Noldorin> oh well
[20:42] <Noldorin> jelmer ignores me now
[20:42] <Noldorin> can't even pay him ;)
[20:42] <Noldorin> byew
[20:49] <jelmer> Noldorin: ?
[20:49] <jelmer> Noldorin: what did I ignore?
[20:50] <wgz> jelmer: what else apart from dpush depends on that SSHVendor in dulwich?
[20:50] <wgz> just doing info and branch seem to work here anyway
[20:50] <jelmer> wgz: fetching over ssh depends on it as well
[20:51] <jelmer> wgz: you probably have ssh.exe installed?
[20:51] <wgz> my terminal claims not...
[20:52] <wgz> and I use plink normally and have bzr configured to use that
[20:52] <jelmer> bzr-git might already be inserting the ssh vendor
[20:52] <jelmer> there is some code there
[20:53] <wgz> is there a -D I can pass that will tell me what bzr-git is up to with transports?
[20:54] <jelmer> wgz: it doesn't really use transports, except for a dummy GitTransport
[20:58] <wgz> okay, I can repo with branch git+ssh
[21:44] <Noldorin> jelmer: read above ;)
[22:18] <Noldorin> jelmer: though i guess we can't even pay you these days hehe