[02:47] <Demosthenes> i need the ability to store keywords inline in my files like RCS did (ie: $Id$), and i'm looking at the bzr-keywords addon... has there been an "official" solution i should be aware of?
[02:50] <Demosthenes> plus the repo's a year out of date ;]
[03:04] <mlh> Demosthenes: filters
[03:10] <Demosthenes> mlh: the reference i pulled on that isn't quite complete. know any working examples? i'm at 2.0.3
[03:10] <fullermd> keywords IS built on the filter layer.
[03:10] <mlh> oh, there you go
[03:10]  * mlh gets his coat
[03:12] <mlh> I was always advise people to very carefully consider whether they actually need them, anyway
[03:21] <Demosthenes> its a typical issue where i need version number or branch id in something visible to the end user.
[05:31] <BlindWolf8> Hello?
[07:05] <BlindWolf8> Hello?
[08:13] <BlindWolf8> Hello?
[10:48] <dcoles> Interesting. When trying to use the https transport with a mercurial Google Code repository bazaar gives up at the first 405
[10:50] <dcoles> Though with http it seems quite happy to slowly fall back until it finds the right VCS transport
[10:50] <dcoles> Should the hg plugin handle https?
[12:00] <maxb> Is there a way bzr-svn could have messed up branch metadata such that bzr makes some *stupid* choices when merging?
[12:01] <maxb> such as reverting changes on the merged-into branch for no good reason
[12:03] <maxb> copying the base/left/right versions of the affected file into new bzr branches and trying the merge there produces a better merge
[12:09] <jelmer> maxb: I doubt that's a bzr-svn issue.
[12:10] <jelmer> dcoles: yeah, it should handle https but there are some known issues with https not specific to the hg plugin
[12:12] <dcoles> jelmer: Ah. Thought that might be the case.
[12:12] <maxb> jelmer: even though the merge is different if I recreate the relative ancestry the files would have had if the project had been in bzr all along, and the merge is better then?
[12:12] <maxb> s/files/versions of the file/
[12:13] <dcoles> jelmer: I'm just trying to look at that Google Go hg import bug
[13:45] <jelmer> dcoles: one thing that might work is to uninstall pycurl
[14:03] <dcoles> jelmer: For the moment using http should be fine - not need to push yet
[14:04] <dcoles> Just trying to work out why bzr-hg generates an inventory delta that seems to be missing one of the parent directories
[14:05] <jelmer> dcoles: just a word of warning; bzr-hg is still very much alpha code
[14:08] <dcoles> Ah. Ok. I'll keep that in mind :)
[14:48] <maxb> Is there a command to dump an inventory for debugging?
[14:49] <maxb> like bzr inventory --show-ids but more so
[15:03] <jelmer> maxb: Not aware of one, please let me know if you find one :-)
[15:04] <jelmer> dcoles: patches welcome :-)
[15:06] <dcoles> Heh. I'm trying :D
[15:08] <dcoles> People just keep insisting on using things OTHER than bazaar.
[15:11] <dcoles> ... or checking in files to SVN with paths like "connectors/perl/temp/\"...
[15:12] <jelmer> yeah, that one seems to be becoming more and more of an issue
[15:12] <dcoles> I see that there's already a bug for having '\' in a path.
[15:12] <jelmer> there isn't anything in bzr that inherently prevents us from using it
[15:13] <dcoles> Mainly to be safe on Windows correct?
[15:13] <jelmer> we've just been on the side of caution since it's impossible to check out those paths on windows
[15:13] <jelmer> yeah
[15:13] <jelmer> we know how we'd like to fix this:
[15:13] <jelmer> allow backslashes but refuse to check out paths with backslashes on windows
[15:13] <jelmer> when adding files with backslashes we'll warn the user
[15:14] <dcoles> That seems like the right idea. Much like when people check in files with filenames that differ only by case
[15:15] <jelmer> dcoles: right, that's another good example where we should allow files to be added (we already do atm) but need to warn the user.
[15:16] <dcoles> If I'm looking at migrating a repository, would the best thing be to "fix" svn's history and then import? Or is it safe to remap the filename at import time?
[15:17] <jelmer> dcoles: remapping the filename at import time isn't really possible, unless you use a custom version of bzr-svn and are never going to use the original svn repo again after you've migrated.
[15:20] <dcoles> jelmer: Hmm. I figured that might be the end result. Alright. Cleaning up history seems to be the go.
[16:04] <lool> Hi folks
[19:15] <dwt> Hi there, I'm trying to use bzr-svn with bzr 2.2b3 but it keeps failing with an error like this: Unable to load plugin 'svn'. It requested API version (2, 1, 0)....
[19:15] <dwt> is there a newer version of the bzr-svn plugin available somewhere?
[19:16] <dwt> I'm currently using http://people.samba.org/bzr/jelmer/bzr-svn/1.0/
[19:16] <dwt> (from launchpad that looked like the most recent branch)
[19:18] <dwt> jelmer: maybe you could give me a hint? (If it's daylight where you live… :)
[19:22] <dwt> Should I be doing 'bzr branch lp:bzr-svn' - and what difference would that make?
[19:30] <dwt> Oddly, enough that seems to have fixed my problem - even though I'm on the same revision now as before. ¿¿¿
[19:41] <dwt> well, thanks for the patience I guess.
[19:41] <dwt> adios!
[20:14] <marioxcc> hi all
[20:15] <marioxcc> ¿what benefits do Bazaar offers over GIT?
[21:46] <C-S> marioxcc: it has a very clean interface and above all: it has plugins
[22:00] <marioxcc> C-S: but what about the basic
[22:00] <marioxcc> branches, files, merges, and so
[22:00] <marioxcc> what are the differences?
[22:19] <C-S> marioxcc: to be honest, I can't tell you very much about the internal differences, however:
[22:20] <C-S> marioxcc: branching is a bit more traditional in bzr, i.e. you have a folder per branch
[22:23] <marioxcc> C-S: ok
[22:23] <marioxcc> thanks
[22:26] <C-S> marioxcc: I would just give it a chance. I personally switched from git to bzr a few month ago.
[22:26] <C-S> marioxcc: and I am really happy
[22:27] <C-S> marioxcc: but the differences are not that big in the end. These are all great VCSs
[22:28] <marioxcc> C-S: I like the model of GIT of commits as a directed graph and branches as references to commits
[22:28] <marioxcc> i can't figure out why did emacs now use bazaar
[22:29] <marioxcc> maybe BZR have a awesome feature than git don't
[22:29] <marioxcc> that is why i'm asking
[22:30] <C-S> marioxcc: well there are nice advantages. launchpad.net is great for example.
[22:30] <C-S> marioxcc: I think, its interface and documentation is just much more consistent.
[22:30] <C-S> marioxcc: but I am not sure why they switched. I did because git became buggy...
[22:31] <C-S> marioxcc: and git has a severe drawback in design with renaming.
[22:31] <C-S> marioxcc: it seems to guess the file names and there are quite a few examples where it guesses wrong.
[22:31] <marioxcc> actually some we consider the "renaming" thing a feature :)
[22:32] <C-S> marioxcc: in bzr you can also add empty directories. something not possible in git.
[22:32] <marioxcc> yeah
[22:32] <marioxcc> about launchpad, i don't like it
[22:32] <marioxcc> it promotes the void open source
[22:32] <marioxcc> instead of free softwre :(
[22:32] <marioxcc> i always use and contribute to GNU Savannah, when i can :)
[22:32] <C-S> marioxcc: I don't understand you.
[22:32] <marioxcc> oh :S
[22:32] <marioxcc> sorry
[22:33] <marioxcc> i mean
[22:33] <marioxcc> launchpad is about open source software
[22:33] <C-S> marioxcc: right
[22:33] <marioxcc> savannah is about free software
[22:33] <C-S> marioxcc: what's the difference?
[22:33] <marioxcc> free software talks about freedom and ehtical values
[22:33] <marioxcc> open source don't
[22:33] <C-S> marioxcc: is there free closed source software supported? I don't like that.
[22:34] <marioxcc> ¿?
[22:34] <marioxcc> the "free" in free software means freedom, not free beer
[22:34] <marioxcc> we the GNU supporters look for freedom, not zero price
[22:34] <C-S> marioxcc: I know, that's why I use a truly free software: BSD
[22:35] <C-S> marioxcc: I see
[22:35] <marioxcc> BSD-style licenses allow propietary derivates :(
[22:35] <C-S> marioxcc: right, that is true freedom: "do what you want but cite me"
[22:36] <C-S> marioxcc: it does not force another person to continue the same way
[22:36] <C-S> marioxcc: anyway, that becomes off topic now :-)
[22:37] <marioxcc> yeah