[02:24] <zenwryly> Since archival backup and DVCS have such similar semantics, I'm exploring using bzr for my backups.  The idea would be to collect changes locally, push them to a remote repo that holds all archival revisions, and then remove old history from the local revision.  Is there a way bzr might be able to accomodate this?
[02:29] <zenwryly> err, "remove old history from the local *repo*"
[03:19] <jelmer____> zenwryly, removing history is hard at the moment
[03:19] <jelmer____> zenwryly, it involves rewriting the repository from scratch
[07:06] <NEBAP|work> morning
[07:06] <NEBAP|work> how can I install a bazaar plugin right from launchpad?
[07:17] <NEBAP|work> k, I found
[07:17] <NEBAP|work> is it possible to get only the head revision of a branch?
[07:18] <NEBAP|work> would be usefull if I don´t have to download the complete history ..
[14:13] <jelmer> join #samba-technical
[14:13] <jelmer> argh
[14:38] <jelmer____> james_w, thanks :-)
[15:32] <Adys> http://dpaste.com/71605/ ochie
[18:11] <LarstiQ> jelmer____: sorry, I just had went to bed at that point :)
[18:13] <jelmer____> LarstiQ, :-)
[18:13]  * LarstiQ observes jelmer____ accumulating an ever longer tail
[18:14] <jelmer____> LarstiQ, Freenode doesn't have SSL support and I need authentication for jelmer_, jelmer__ and jelmer___ :-)
[18:14]  * LarstiQ blinks
[18:14] <LarstiQ> jelmer____: I see :)
[22:01] <ronny> anyone remembering the specific reason for .bzr/branch/tags being a set of netstrings?
[23:15] <DaffyDuck_> When I commit files, the files in the bzr repo get wrong permissions. This means that other users can't read/write the repository. Can these kind of things be solved with post-commit scripts?
[23:24] <lifeless> moin
[23:29] <lifeless> DaffyDuck_: they could, but why are they getting the wrong permissions?
[23:31] <DaffyDuck_> Because of users' umask, I guess.
[23:33] <DaffyDuck_> The problem doesn't occur when using the repositories remotely (bzr+ssh). But as soon as someone tries to work with the repository locally on the server, the permissions  get screwed up.
[23:41] <lifeless> DaffyDuck_: you could set their umask in bzr itself I guess
[23:41] <ronny> what api should i use if i want to do something like "bzr branch"
[23:41] <lifeless> or file a bug; I thought we used to have a facility to control this
[23:41] <lifeless> ronny: foo.bzrdir.sprout
[23:42] <DaffyDuck_> lifeless: Yes, in some cases it appears to be handled correctly, but in others it is not. I have filed a related bug, but no one has commented on it yet.
[23:46] <lifeless> ah, well I'm sure we'll get to it soon
[23:51] <DaffyDuck_> lifeless: Under normal circumstances we're working remotely, so this slight problem won't stop our conversion from subversion to bazaar. But sometimes when I'm anyways peforming work on the server, I'd like to be able to use the bazaar repositories without worrying about doing bad things to them in the process.
[23:53] <lifeless> DaffyDuck_: you could try setting your umask; if thats it then nothing bad will happen to them
[23:53] <lifeless> you could even replace bzr with
[23:53] <lifeless> #!/bin/sh
[23:53] <lifeless> umask <>
[23:53] <lifeless> bzr.orig $@
[23:56] <DaffyDuck_> lifeless: Yes, I've thought of that -- but it has its potential problems. I would prefer if bzr handles all the permissions as it knows best how to.  Looking at some of the change logs, it appears that bazaar does have some sort of infrastructure to manage permissions in repositories.
[23:57] <DaffyDuck_> I get the feeling that it's just that no one has come around to using those systems everywhere.
[23:57] <lifeless> setting the umask to allow group write shouldn't cause any problems for bzr
[23:58] <lifeless> if it does, they are definitely bugs