[00:05] <m_tadeu> hi everyone....is there a way to revert a push?
[00:22] <jelmer> m_tadeu: you can push the old tip and use --overwrite
[00:25] <jelmer> sigh
[00:27] <bob2_> hah
[10:19] <Marqin> hi, i've found some old bzr repo.. how can i check from where is it pulling without pulling??
[10:50] <jelmer> Marqin: bzr info
[10:55] <Marqin> thx
[15:13] <junrrein> Hi people, I got a question. How can I configure the default text editor (under Linux)? I tried "bzr config editor=gedit" but when running bzr commit, nano is still used instead.
[15:20] <junrrein> I found out, thanks
[15:21] <mgz> setting EDITOR works, as should that config command
[18:58] <kolbe> is there some way to branch/checkout only a single file (or directory, or something) from a tree? so that i can keep that file in sync with the file in the tree without having to branch the entire thing?
[19:02] <rozzin> kolbe: "bzr export"?
[19:02] <kolbe> hmmm i don't see how that would help?
[19:03] <rozzin> I think it lets you export a subdiretory.
[19:03] <kolbe> i just want to grab http://bazaar.launchpad.net/~maria-captains/maria/5.5-noga-hf/files/head:/plugin/server_audit/
[19:03] <kolbe> well in order to export a subdirectory i would have to first branch the entire repository, which is exactly what i am trying to avoid
[19:04] <kolbe> i have a feeling this is not possible with bzr
[19:04] <rozzin> kolbe: I don't think that's true.
[19:04] <kolbe> you don't think what is true?
[19:05] <rozzin> kolbe: I think you can export directly from a remote repository.
[19:06] <kolbe> rozzin: hm ok i'll try
[19:06] <rozzin> kolbe: You can also do lightweight checkouts.
[19:06] <kolbe> *ideally* i'd like to be able to see version history for just things in this directory and be able to bzr pull to update files in it
[19:06] <rozzin> kolbe: ... which avoids downloading all of the history, and just downloads the current revision.
[19:08] <kolbe> rozzin: ok neat, export did work, in that it allowed me to grab just that directory but of course it doesn't have any revision info and i can't bzr pull inside of it to update things
[19:09] <rozzin> Yeah, that's just not how it works--you can't separate out `just the history for this subdirectory'. History isn't tracked like that; there's not a history-per-file like CVS or ClearCase.
[19:09] <kolbe> yeah
[19:10] <rozzin> Changesets that can span multiple files are `inside' revisions, not the other way around.
[19:10] <kolbe> i think bzr doesn't match up very well with how my brain wants to do things, i feel like everything i do is an epic struggle that just ends with me kind of tired and frustrated
[19:10] <rozzin> What's your frame of reference?
[19:10] <rozzin> i.e.: what system are you coming from?
[19:11] <kolbe> no frame of reference, just trying to do certain things that are apparently impossible with bzr
[19:11] <rozzin> Like?
[19:11] <kolbe> two major frustrations so far are how absurdly slow it is to branch a large tree
[19:11] <kolbe> and that if my internet connection is interrupted in the middle of that it's impossible to resume
[19:11] <rozzin> kolbe: You've never used another version control system before?
[19:11] <kolbe> rozzin: no, not really
[19:12] <rozzin> On the up side, once you go through that huge download, you never have to do it again.
[19:12] <kolbe> right
[19:13] <kolbe> until i go set up some ec2 instance or some other VM that i want to use for building things and i end up having to branch it again :)
[19:13] <rozzin> You can avoid the big initial hump, if you want, at the expensive of repeatedly paying it in small increments.
[19:14] <rozzin> kolbe: No.
[19:14] <rozzin> kolbe: I'm serious: once you get the data into your shared repository, you never need to download it again.
[19:15] <kolbe> rozzin: if i want to set up an environment on some other machine, i either have to manually copy my own shared repository there or i have to branch again
[19:15] <rozzin> Well, yeah. If you want to copy the data to another machine, then you have to copy the data to the other machine. :)
[19:16] <kolbe> i don't want to copy it, i want to be able to set up a new branch without it taking an hour and without me having to copy anything around on my own ;)
[19:16] <rozzin> Then do a lightweight checkout.
[19:16] <kolbe> but then i don't get revision history?
[19:17] <kolbe> i'm just whining now, don't mind me :)
[19:17] <rozzin> Dude--you can't have it both ways.
[19:17] <rozzin> Either you want to copy the data, or you don't.
[19:18] <rozzin> lightweight checkout `gives you access to the revision history', it just doesn't give you /local/ access to it. It all says remote, which is why the intial action is so much quicker.
[19:18] <rozzin> it all _stays_ remote, rather.
[19:19] <rozzin> You can also use a stacked branch.
[19:19] <rozzin> Thay might be more like what you're looking for.
[19:20] <rozzin> All of the original revision-history data stays remote, but additional revisions that you add onto that are local.
[19:20] <kolbe> rozzin: i think it's mostly frustrating that the branching operation requires so much processing instead of being able to just grab a snapshot of the historical information and download it in a big chunk
[19:21] <rozzin> kolbe: You can always start with a lightweight checkout and then `reconfigure' it into a full branch in the background.
[19:21] <kolbe> yeah? how do i do that?
[19:22] <rozzin> I think it's something like: bzr checkout --lightweigth foo bar; bzr reconfigure --tree bar &
[19:23] <kolbe> neato, maybe i'll try that sometime
[19:23] <kolbe> thanks!
[19:23] <rozzin> I mean, you'd really want to run the `reconfigure' in another xterm or screen session or something; you wouldn't actually want to background it, because it'll keep printing stuff out....
[19:23] <kolbe> yeah
[19:24] <rozzin> Also, I didn't think there was really that much processing overhead involved in downloading the revision history--isn't it really mostly just the cost of downloading the data?
[19:25] <rozzin> e.g.: if you have 50 MB of data that you want to download, you have to download 50 MB of data.
[19:25] <rozzin> ?
[19:26] <rozzin> I know you were hoping that you could say `just download the 10 MB of history relevant to this subtree'....
[19:26] <kolbe> it seems to be much more than that
[19:28] <rozzin> Have you tried just transferring the repository with scp or `ssh ... tar -c ...', and comparing times?