[02:20] <thomi> I'm confused: Can I install a hook by placing it somewhere in the branch's .bzr folder?
[02:21] <thomi> in some places in the docs it says hooks need to be installed via ~/.bazaar/plugins, and in other places it seems to suggest tat you may be able to install the hook into the branch/repo directly
[03:00] <lifeless> thomi: you cannot install them in the branch/repo
[03:00] <thomi> :(
[03:00] <lifeless> thomi: it would be a huge security hole to run arbitrary code in something you clone; we permit /config/ but not /install/ there.
[03:00] <lifeless> thomi: if you want a plugin active for just one repo, the plugin should consult config in that repo.
[03:01] <lifeless> as in, branch.get_config() and drill in from there.
[03:03] <thomi> lifeless: I want to make peopel contributing to a particular project unable to commit code unless pep8 passes - it sounds like I can't do that without them having to install the plugin manually
[03:03] <thomi> I guess I can do it at package build time instead.
[03:03]  * thomi is feeling draconian today
[03:04] <lifeless> you could put the check in your test suite
[03:06] <thomi> lifeless: yup, but I'd still need to trust the developers to actually run the tests :-/
[03:13] <lifeless> thomi: ...
[03:13] <lifeless> thomi: your CI system should reject their merges before you see them, no?
[03:15] <thomi> Yeah, I'll implement the check at the CI level
[19:46] <dlpenguinlover> Hello, I'm having some trouble with a bazaar repo i'm trying to get working. I'm trying to install a launchpad instance on a virtual Ubuntu Server, however while i'm downloading a repo to install Launchpad, bazaar is erroring out saying the repo is corrupt. I do not want to redownload 400+ mbs of data, is there any way to "fix" the repo?
[19:59] <mathrick> dlpenguinlover: I'm not exactly sure what you ended up with, but if you have a bzr branch in an accessible state, you can bzr check it
[20:00] <mathrick> it will do so some sanity checks and tell you if/what issues there are
[20:04] <dlpenguinlover> "No working tree found at specified location."
[20:07] <mathrick> dlpenguinlover: bzr check --repo and possibly --branch
[20:10] <dlpenguinlover> mathrick: $ bzr check --repo --> Checking repository at 'file:///home/david/launchpad/lp-brakches/'. ----
[20:11] <mathrick> yes
[20:11] <mathrick> it will take time
[20:11] <dlpenguinlover> checked repository file:///home/david/launchpad/lp-branches/ format RepositoryFormat2a() -----> 0 revisions 0 file-ids
[20:11] <mathrick> uh
[20:12] <dlpenguinlover> $ bzr check --branch -------> No branch found at specified location.
[20:12] <mathrick> then it hasn't actually downloaded anything it seems
[20:12] <mathrick> dlpenguinlover: what does "du -hsc file:///home/david/launchpad/lp-branches/" say?
[20:12] <mathrick> err
[20:12] <mathrick> without file://
[20:13] <dlpenguinlover> 237M   /home/david/launchpad/lp-branches/       237M total
[20:14] <mathrick> well, not sure how to help you with that
[20:14] <mathrick> but it seems to be broken completely
[20:14] <mathrick> dlpenguinlover: how exactly did it error out in the first place, and is it consistent?
[20:16] <dlpenguinlover> mathrick: well, the launchpad installer (called rocketfuel-setup) was the one executing the bzr commands, but this is what it outputted. "bzr: error: No WorkingTree exists for "file:///home/david/launchpad/lp-branches/devel/.bzr/checkout"/ ERROR: Your trunk branch in /home/david/launchpad/lp-branches/devel is corrupt. Please delete /home/david/launchpad/lp-branches/devel and run rocketfuel-setup again."
[20:17] <mathrick> and did you do that?
[20:17] <dlpenguinlover> Yes, and it gives the same error message.
[20:20] <mathrick> dlpenguinlover: aight, so you have sub-branches under that dir
[20:20] <mathrick> what does bzr info file:///home/david/launchpad/lp-branches/devel give you?
[20:30] <jelmer> hi mathrick
[20:30] <jelmer> mathrick: there is also "co:" which behaves similarly to "colo:"
[20:43] <mathrick> jelmer: oh, I couldn't find that mentioned anywhere
[20:44] <jelmer> mathrick: it's undocumented - don't use development-colo, as it is unfinished
[20:44] <jelmer> (hence the "development-" bit :-)
[20:44] <mathrick> jelmer: I know, but it kinda beats having mysterious path errors when cloning colocated branches across OS
[20:47] <jelmer> mathrick: okay, up to you :) Just be wary of bugs
[20:47] <mathrick> I am!
[20:48] <mathrick> jelmer: in fact I ran into a couple where various things get horribly confused when I ask them to compare colocated branches
[20:48] <mathrick> diff says no changes, missing doesn't even understand what I'm asking it to do, etc.
[20:50] <jelmer> mathrick: please file bugs and tag them "colocated"
[20:50] <mathrick> OK