[08:39] <mgz> morning
[08:40] <jelmer> hey mgz
[08:41] <mgz> is it samba thing this week jelmer?
[08:41] <jelmer> mgz: yes
[08:41]  * jelmer waves from gÃ¶ttingen
[08:43] <james_w> hi jelmer
[08:43] <jelmer> hey James
[09:20] <vila> hi all
[09:21] <lifeless> mgz: hi
[09:22] <lifeless> mgz: did you see my query about .buffer on python3 ?
[09:23] <mgz> hey lifeless, no, I didn't
[09:24] <mgz> where should I be looking?
[09:31] <lifeless> here
[09:31] <lifeless> I addressed it to wgz, couple days back
[09:38] <wgz> I see it.
[09:42] <wgz> ...I'm not completely clear on what's okay with the new io library either, but do need something other than TextIOWrapper
[09:43] <wgz> as I understand it, python 3 now turns off the newline translation itself, and does it in the text io instead
[09:43] <lifeless> yah, have a look at what I put in 0.0.8
[09:43] <wgz> and .buffer does look okay to me.
[09:43]  * wgz pulls
[09:43] <lifeless> (or trunk) and let me know how it plays on windows
[09:46] <wgz> hm, I'm not crazy about type checking the read(0) return
[09:46] <wgz> but I guess it's a neat hack
[09:48] <wgz> I get some failures on trunk but they all seem to be tags related.
[09:48] <lifeless> :(
[09:48] <lifeless> you have testtools 0.9.15 ?
[09:49] <wgz> ...nope, apparrently not, will update
[09:52] <wgz> so, the other option it seems io wants for detecting the stream type would be `isinstance(stream, io.RawIOBase)` for byte streams
[09:53] <wgz> but that's assuming people who write their own little classes subclass it, which people are not in the habit of doing with python
[09:54] <wgz> or `isinstance(stream, io.TextIOBase)` which then promises a buffer attribute
[09:55] <wgz> ...which I now note it says " is not part of the TextIOBase API and may not exist on some implementations" in the docs
[09:55] <wgz> the io module has been a pain whenever I've need to do anything with it
[09:56] <wgz> possibly should call detach() rather than just using the buffer?
[10:15] <lifeless> wgz: I'm not sure, happy to take recommendations
[10:16] <lifeless> wgz: there was a pycon presentation I found
[10:16] <mgz> I shall try poking people/things this evening and see what falls out
[10:16] <lifeless> that talks about using buffer (its presenting the io module)
[10:16] <mgz> ah, I'll look for that as well later.
[10:18] <lifeless> http://www.dabeaz.com/python3io/
[10:18] <lifeless> its on slideshare or something too
[10:18] <lifeless> which is where I found it
[17:46] <LeoNerd> I have one bzr branch that stores some C library code; it has a src/ and an include/ directory
[17:46] <LeoNerd> Is there some way I can "mount" those two dirs in the same place in a different branch's checkout?
[18:48] <bsd> LeoNerd: you can check out a branch within a branch
[20:32] <LeoNerd> bsd: Yes, but just some subdirs of it..?
[20:34] <bsd> LeoNerd: check out 'bzr help view'
[20:34] <LeoNerd> Ah OK thanks
[21:09] <KombuchaKip> Are there any plans to port the nautilus-bzr integration to Thunar, the native Xfce file manager?
[21:13] <jelmer> KombuchaKip: not that I'm aware of
[21:14] <KombuchaKip> jelmer: Thank you.
[21:15] <jelmer> KombuchaKip: the nautilus-bzr is also pretty crappy atm, so I'm not sure if it's worth porting
[21:16] <jelmer> KombuchaKip: it might be a good idea to extend rabbitvcs to support bzr
[21:17] <KombuchaKip> jelmer: Yes, that's another option.
[23:22] <Noldorin> can bzr push remember FTP passwords?
[23:25] <Noldorin> also, what's the bzr equivalent of "git checkout" ?
[23:26] <mwhudson> i guess switch, mostly, but well
[23:26] <mwhudson> git checkout does a few things
[23:27] <Noldorin> mwhudson, i thought switch was a loom plugin though?
[23:27] <mwhudson> no
[23:27] <Noldorin> that's what it says in bzr help :-S
[23:28] <mwhudson> loom wraps switch i think
[23:28] <mwhudson> but it's in core
[23:28] <Noldorin> oh right
[23:28] <Noldorin> okay
[23:28] <Noldorin> might be nice to show that in bzr help, but i got it now :-)
[23:29] <Noldorin> mwhudson, incidentally, bzr switch gives me a maximum recursion depth error
[23:29] <Noldorin> bzr 2.6
[23:31] <jelmer> Noldorin: bzr-loom?
[23:32] <Noldorin> jelmer, 2.3.0dev
[23:32] <jelmer> Noldorin: uninstalling it might fix it
[23:32] <Noldorin> okay
[23:32] <mgrandi> ok, so i'm really confused on how i'm supposed to resolve conflicts in bzr. I'm using this external tool, which seems to want to resolve all the changes into the '.BASE' file, but then bzr doesn't have an option to use the .BASE file, only OTHER and THIS
[23:32] <Noldorin> jelmer, cheers.
[23:32] <mgrandi> how should i be using these third party tools then? or what file should i be editing?
[23:32] <Noldorin> jelmer / mwhudson any idea bout saving the the FTP password btw?
[23:33] <mwhudson> Noldorin: i think it's possible, maybe involving ~/.bazaar/authentication.conf
[23:33] <mwhudson> but i certainly don't know
[23:33] <Noldorin> ah
[23:33] <Noldorin> will have asearc, ta
[23:33] <jelmer> Noldorin: you can stick it in netrc or authentication.conf
[23:33] <mwhudson> Noldorin: bzr help authentication maybe?
[23:33] <Noldorin> jelmer, netrc? i'm on windows
[23:35] <Noldorin> jelmer, but yeah authentication.conf sounds good. cheers
[23:35] <Noldorin> where does it belong on windows?
[23:35] <Noldorin> win7
[23:36] <Noldorin> oh wait; think i've found it
[23:37] <mgrandi> *pokes about merging conflict question*
[23:39] <jelmer> mgrandi: sorry, I have no idea about that
[23:39] <mgrandi> well how does everyone resolve conflicts is what i'm asking
[23:39] <mgrandi> what file do you edit?
[23:39] <Noldorin> jelmer, btw, i am thankfully git-free at the moment, so no need for bzr-git fix yet...
[23:40] <Noldorin> jelmer, if someone wanted a project (like a mirror of mine) in git, could they pay you to fix up bzr-git though? ;-)
[23:40] <mgrandi> whats wrong with bzr-git? seems to work fine for me o.o
[23:41] <Noldorin> it's broken for commits where files get moved
[23:41] <Noldorin> and some other slightly unusual cases.
[23:41] <Noldorin> and some other slightly unusual cases.
[23:42] <mgrandi> (the only featured i'd like to see is native push support rather then dpush with rebasing the tree)
[23:42] <Noldorin> mgrandi, i'm pretty sure that's impossible due to the differences in how bzr and git work with commits
[23:43] <mgrandi> jelmer mentioned its possibly by just storing information in a magic file that git can't represent natively =P
[23:43] <Noldorin> information that git can't represent natively in a magic file, you mean?
[23:44] <mgrandi> yeah.
[23:44] <mgrandi> or something along those lines
[23:44] <Noldorin> yeah, might work... he knows how it works, naturally
[23:46] <jelmer> native push support can indeed be done with a magic file where metadata that can't be represnted in git is stored
[23:50] <Noldorin> can someone pay you to do that too? :-)
[23:50] <Noldorin> heh