[00:20]  * amanica thinks I should make the log-lowerbound-exclusivity change at the highest level possble, so that I break less stuff?!
[00:33] <ronny> hmm, http://bazaar-vcs.org/BzrVsHg is full of lies and massively outdated crap
[00:39] <SamB> jelmer: hmm ... it looks like svn_ra_replay_range calls svn_ra_replay
[00:40] <jelmer> SamB, in the subversion code you mean?
[00:41] <SamB> jelmer: well, that's what I see here:
[00:41] <SamB> (gdb) frame
[00:41] <SamB> #0  0xb76dc926 in svn_ra_replay () from /usr/lib/libsvn_ra-1.so.1
[00:41] <SamB> (gdb) frame 1
[00:41] <SamB> #1  0xb76dd553 in svn_ra_replay_range () from /usr/lib/libsvn_ra-1.so.1
[00:41] <SamB> (gdb) frame 2
[00:41] <SamB> #2  0xb76ef673 in ra_replay_range (self=0x8a7f770, args=0x8a7fd4c)
[00:41] <SamB>     at subvertpy/_ra.c:1240
[00:41] <SamB> it might not always happen, haven't looked at the source ...
[00:42] <jelmer> SamB, That seems perfectly fine, I don't think all transports implement ra_replay_range so falling back to ra_replay in that case seems sensible
[00:42]  * SamB looks at code now ...
[01:11] <SamB> jelmer: hmm, yeah, that does seem to be a fallback
[01:54] <SamB> huh, how would I run a testcase but break into pdb when it fails ?
[07:57] <spiv> SamB: there's no command line option for that, we really ought to add one.
[07:58] <spiv> SamB: insert "import pdb; pdb.set_trace()" into your test case by hand :/
[10:34] <abeaumont> hmmm, i'm wondering, any chance that bzr-git + git-svn works better than bzr-svn for large repos?
[10:38] <AfC> abeaumont: I seriously doubt it
[10:39] <abeaumont> i see
[13:46] <AfC> I put "parent_branch = file:///home/andrew/src/quill/mainline/" into .bzr/branch/branch.conf, but `bzr info` isn't reporting that as the parent branch. Any idea what's up with that?
[13:51] <LarstiQ> AfC: parent_location iirc
[13:52] <AfC> LarstiQ: ah.
[13:52] <AfC> LarstiQ: yeah, that did it. Thanks.
[13:53] <AfC> (why, oh why)
[13:53] <LarstiQ> why it worked? ;)
[14:02] <AfC> LarstiQ: no... why on earth a file that we force users to edit would have an asymmetrical naming pattern for things.
[14:02] <AfC> LarstiQ: submit_branch:
[14:02] <AfC> LarstiQ: parent_branch: would therefore make sense
[14:02] <AfC> but no
[14:02] <LarstiQ> AfC: I agree there is some consistency to be had, but I don't agree with user editing of branch.conf
[14:03] <AfC> LarstiQ: is there a UI for setting locations now?
[14:03] <AfC> (if I've missed something new and shiny, then I do apologize)
[14:03] <LarstiQ> AfC: not as it should be, specifically parent though, `bzr pull --remember`
[17:05] <ElMonkey> hola
[17:06] <ElMonkey> anyone happen to know if i can get files from bzr repo A into bzr repo B without making B a branch of A, while preserving file history?
[17:06] <ElMonkey> eg, i just want to pull out a specific subset of files from a repo to use in another project
[17:08] <fullermd> No, files are in history, not the other way around.
[17:08] <fullermd> (and in bzr terms, you probably mean 'branch' rather than 'repo')
[17:18] <ElMonkey> hmm
[17:19] <ElMonkey> there's no fairly doable workaround, i presume?
[17:19] <fullermd> You can sweet-talk it into using the same file-id's, which would make merging work, but that doesn't sound anything like what you'd want.
[17:19] <LarstiQ> ElMonkey: can you explain more fully what you are trying to solve?
[17:20] <fullermd> There are various ways one can go about rewriting a branch to create new, unrelated history that looks similar to the existing history.
[17:20] <fullermd> But you can't turn existing revisions into different content; that goes against the idea of history.
[17:21] <ElMonkey> well, its like this: i have a mac app, and i am porting it to the iphone
[17:21] <ElMonkey> now the mac app is older, and there's already an iphone scaffold in place from another app
[17:22] <ElMonkey> and now i'd like to load the relevant pieces of the mac app onto the iphone template
[17:22] <LarstiQ> and you want to keep the history of the iphone template?
[17:23] <ElMonkey> i'd like to keep the history of both :)
[17:23] <LarstiQ> I'd value the mac app more than the template
[17:24] <LarstiQ> ElMonkey: you could `bzr merge -r 0..-1` them together
[17:26] <ElMonkey> hmm
[17:26] <ElMonkey> well, thanks for the input, i'll have to think some more about how to do this :)
[17:27] <LarstiQ> ElMonkey: tried the merge command?
[17:27] <LarstiQ> ElMonkey: keeping histories/files of two branches is way easier than selectively transplanting some files
[17:28] <ElMonkey> LarstiQ, no, i havent tried it yet
[17:28] <ElMonkey> i think i have to adjust the directory structures a bit before i can
[17:29] <LarstiQ> ElMonkey: for best effect, likely. You can try and revert to see what happens before that though :)
[17:49]  * SamB wonders why he thought there were no failures with the replay code enabled ...
[20:40] <SamB> hmm ... 2a seems slow via HTTP, at least when samba.org is involved ...
[20:40] <LarstiQ> SamB: if -Dhpss shows a million requests, that would be why
[20:45] <SamB> actually, it just seems to be doing far too much computing
[20:45] <LarstiQ> SamB: there are a couple of bugs about poor http 2a performance at least
[20:46] <SamB> or, wait, samba.org doesn't have a smart server anyway I think
[20:46] <SamB> at least not people.samba.org
[20:50] <SamB> but -Dhttp doesn't seem to indicate that it's the making of many HTTP requests that is the issue
[20:51] <SamB> since I'm seeing a lot of BIG gaps between bzr recieving one response and sending the next request ...
[20:51] <SamB> which agrees with my bandwidth meter
[20:53] <SamB> hmm ... there is always a chance that my actual problem is that it's doing a lousy job in reporting it's progress ...
[21:21] <pigmej2> hey
[23:37] <GungaDin> How do I tell bzr diff to show only the names of the files that have changed and not the contents?
[23:38] <LeoNerd> Try   bzr st
[23:39] <GungaDin> thx