[00:10] <peitschie> hi poolie, fullermd :)
[00:11] <poolie> hi
[01:03] <mlh> something for the SCA folks: http://www.popularmechanics.com/technology/digital/fact-vs-fiction/medieval-knights-on-a-treadmill-put-historical-myths-to-the-tes
[01:03] <mlh> oops wrong chan
[01:05] <poolie> heh
[01:31] <thumper> poolie: hi
[01:31] <poolie> hi there
[01:31] <thumper> is there a way to get bzr to apply a patch file?
[01:32] <thumper> or do I just use patch?
[01:32] <poolie> "bzr patch" :)
[01:32] <thumper> gah, here I was looking for a merge option
[01:32] <poolie> from bzrtools i think
[01:32] <poolie> is this a bzr bundle, or just a plain patch?
[01:32] <poolie> if it's a bzr bundle with metadata, you want 'bzr merge'
[01:33] <thumper> it is a plain patch
[01:33] <thumper> which fails to merge
[01:33] <poolie> thumper, can you search your brain and social network for someone suitable for bit.ly/ubuntu-vcs-job
[01:33] <thumper> so I'll be editing the files manually by the look of it
[01:33] <poolie> try the external tool 'wiggle'
[01:33] <thumper> poolie: I mentioned the job to thomi
[01:33] <thumper> poolie: he is mildly interested
[01:34] <thumper> but committed to teach until Dec
[01:34] <poolie> yeah he could be interesting
[01:34] <thumper> wiggle?
[01:35] <AfC> Interesting to see Go in that job description
[01:38] <thumper> AfC: Go has been added to almost all the new canonical jobs
[01:38] <thumper> someone is becoming a fan
[01:38] <thumper> poolie: bit.ly is timing out for me :(
[01:38] <AfC> thumper: the same person who became a Qt fan? :)
[01:39] <thumper> AfC: perhaps
[01:39] <AfC> {sigh}
[01:39] <thumper> I couldn't possibly comment :)
[01:39] <thumper> that said, Go does look interesting
[01:39] <thumper> I've been trying to think of something interesting to write in it
[01:41] <AfC> thumper: http://www.dilbert.com/fast/1995-11-17/
[01:42]  * thumper sighs
[01:42] <thumper> it seems vodafone's 3G is having issues
[01:42] <thumper> some ports are fine
[01:42] <thumper> but 80 doesn't seem to be one of them
[01:47] <poolie> afc, more like http://twitter.com/#!/gniemeyer
[01:47] <poolie> also, http://browsertoolkit.com/fault-tolerance.png
[01:47] <poolie> and hi
[01:48] <AfC> poolie: and unto you, hi
[01:48] <AfC> hahahah
[01:48] <poolie> i love that
[01:48] <AfC> that's bloody awesome
[03:22] <mlh> go and qt huh?  has someone put them together?  I goog'd but didn't find anything
[03:23] <mlh> also, that ft comic is awesome; does anyone know where the original is?
[06:49] <thomi> poolie: got a second?
[06:49] <poolie> hi, sure
[06:49] <thomi> I've managed to get the system-wide config file working under Linux, with a few caveats:
[06:49] <poolie> hooray
[06:50] <thomi> first, as far as I can see, the only place in bzrlib that uses the config stack rather than the old config store API is when picking which editor to use for commit messages.
[06:50] <thomi> Second, I'm unsure how to go about writing tests for the new code.
[06:50] <poolie> it's mostly not updated yet
[06:50] <poolie> vincent's in the middle of working on this
[06:51] <poolie> hm
[06:51] <thomi> oh ok, cool.
[06:51] <poolie> but if you update some of them that are relevant to your work, i'm sure he won't object
[06:51] <poolie> so, what do we want to test?
[06:51] <poolie> i'd say it's mostly that when you ask for one of these stacks, you get something that includes the global config fie
[06:52] <poolie> *file
[06:52] <thomi> right
[06:52] <thomi> there are tests already for the stack stuff.. let me see....
[06:52] <poolie> for the purposes of test isolation, we might want to make sure we don't see the actual real global file
[06:52] <poolie> so i'd probably add something in osutils or config.py that says where it is
[06:53] <poolie> then that can be overridden by the test suite
[06:54] <thomi> So there's already a test that makes sure that the stack works as expected with mutliple config sources.
[06:54] <thomi> There's also tests that ensure that the config file stores work as expected.
[06:56] <thomi> Do we need a third test that explicitly tests with two files on disk? It seems like the functionality has already been covered by other tests; I'm not sure what I'd actually be testing...
[06:59] <poolie> so i think your incremental test mostly needs to be that the pre-canned class stacks are including something that points to the global one
[06:59] <poolie> i don't think so
[06:59] <poolie> we can add that if we find there's actually a problem
[07:02] <thomi> when you say "pre-canned class stacks" you're referring to "GlobalStack", "LocationStack" and "BranchStack"?
[07:02] <poolie> yes
[07:04] <thomi> hmmm, those Stack classes create the config store instances in their constructors, so unless I'm missing something (possible), I can't see how to write that test and keep it independant of the files on disk.
[07:05] <poolie> hm
[07:05] <poolie> it's ok if they read the file on disk during the test
[07:05] <poolie> we may want to make sure that other tests don't read the global config though, or it will tend to leak in
[07:05] <poolie> one way is to use overrideAttr to monkeypatch the method that gives you the file name
[07:05] <poolie> maybe you should push your branch and propose it and then we can look at that together?
[07:05] <thomi> ahhhh
[07:06] <thomi> ok, it's pushed, I'll propose it now, although you have to understand that it's really not finished yet ;)
[07:06] <poolie> that's totally ifne
[07:06] <poolie> *fine
[07:07]  * thomi waits for the diff to update...
[07:10] <thomi> Here we go:
[07:10] <thomi> https://code.launchpad.net/~thomir/bzr/add-global-config/+merge/69592
[07:13] <thomi> nuts - my editor automatically trims trailing whitespace, which makes the diffs in launchpad look a bit messy.
[07:21] <poolie> hm
[07:22] <poolie> that's ok, there's not meant to be any whitespace
[08:00] <Riddell> hello all
[08:05] <poolie> hello riddell
[08:53] <tmm1> is there a way to set the default repo/pack format to use?
[08:53] <tmm1> i'm dealing with a legacy remote and don't want to do the on-the-fly conversion all the time
[09:12] <jelmer> tmm1: hi
[09:13] <jelmer> tmm1: you can set an alias for init, e.g. "bzr alias init='init --format=pack-0.92'"
[09:33] <tmm1> jelmer: thanks
[14:39] <thevibe> bzr 2.1.1 break-lock not working - can't break the lock? how can I break a lock? Any ideas?
[14:49] <thevibe> so... any ideas?
[15:17] <Riddell> thevibe: what doesn't work about it?
[15:17] <Riddell> oh, too late
[15:24] <kjbbb> hey everyone. quick question. i created a branch with --no-trees. How do i get a working tree back?
[15:25] <jelmer> kjbbb: "bzr co"
[15:25] <kjbbb> ah, easy, thanks :)
[18:31] <lamont> if I want to alter a single tree on a machine such that "bzr uncommit" requires extra effort to do, maybe even blatting out a "no really, don't do that" message, is that trivial, or painful to do?
[18:38] <fullermd> There's that param to require the head to stay on the left path.  It's really aimed at keeping people from pushing over something changing the left path (e.g., 'merge X ; push X'), but it prevents uncommit too.
[18:38] <fullermd> No override for it other than commenting out the config param, AFAIK.  So that may be too heavy for you?
[22:36] <poolie> hi all
[23:07] <Riddell> good morning poolie
[23:07] <poolie> hi there, good evening
[23:08] <Noldorin> jelmer, heya
[23:13] <jelmer> Noldorin: hi
[23:13] <jelmer> hi Riddell, poolie
[23:13] <Noldorin> jelmer, i'm back (after much time) to trying to install bzr-git ;-)
[23:14] <Noldorin> thought i could use your assistance a bit
[23:14] <Riddell> how's Republika Srpska jelmer?
[23:14] <Noldorin> so firstly
[23:14] <Noldorin> how do i install dulwich so that it gets included by the precompiled version of bzr on Windows?
[23:15] <jelmer> Riddell: it's great, much nicer than I had expected
[23:16] <jelmer> Noldorin: you'll have to compile dulwich yourself, the standalone installer doens't include it
[23:16] <Noldorin> jelmer, sure...are there instructions anywhere? i see none
[23:18] <jelmer> Noldorin: I don't think there are; if you're ok with running just the python source (not the optimized compiled versions) it should be a matter of putting dulwich somewhere in your $PYTHONPATH
[23:18] <Noldorin> jelmer, no INSTALL file in the package...
[23:18] <Noldorin> just dloaded 0.7.1
[23:19] <jelmer> Noldorin: The installation should be like any other Python package
[23:19] <Noldorin> python setup.py install
[23:19] <Noldorin> yes
[23:20] <Noldorin> jelmer, that works but remember i am running the Windows bzr binaries...which ignores PYTHONPATH i believe
[23:20] <jelmer> Noldorin: you can't use the standalone python installer if you do that
[23:20] <jelmer> I think
[23:21] <Noldorin> jelmer, so i cannot use the standalone python installer + bzr-git at all?
[23:21] <jelmer> Noldorin: as far as I understand, no
[23:22] <Noldorin> right
[23:22] <Noldorin> got it
[23:22] <Noldorin> jelmer, time to uninstall the standaloen version and re-install the python 2.7 version :-)
[23:22] <Noldorin> jelmer, shouldn't affect how i use it, right?
[23:24] <jelmer> Noldorin: I don't think so, but I have almost zero experience using bzr on windows
[23:24] <Noldorin> ok sure
[23:24] <Noldorin> yeah think you've said before
[23:24] <Noldorin> jelmer, it sounds logical though
[23:27] <Noldorin> jelmer, any x64 builds of the windows installer?
[23:27] <Noldorin> bzr-2.3.4-1.win32-py2.7.exe
[23:27] <Noldorin> but i want x64
[23:28] <jelmer> no idea, sorry
[23:28] <Noldorin> sure
[23:28] <jelmer> Noldorin: the bzr-windows mailing list might also be of use
[23:28] <Noldorin> yeah
[23:28] <Noldorin> found something there already
[23:46] <Noldorin> jelmer, ok i've installed bzr-python (with a bit of trouble on my x64 box)...where is its location on disk?
[23:47] <Noldorin> also no tortoisebzr i guess? :-(
[23:48] <jelmer> Noldorin: sorry, I have no idea about those things
[23:48] <Noldorin> ok sure
[23:49] <jelmer> Noldorin: the bzr-windows list might be a better bet, or jam/bialix when they get back from holidays
[23:49] <Noldorin> jelmer, no-one else here eh??
[23:50] <jelmer> Noldorin: I'm not sure if there is anybody else here at the moment familiar with the internals of the windows installer
[23:51] <poolie> i think you can still catch jam today before he leaves but you'd better send mail
[23:51] <Noldorin> ok sure
[23:52] <Noldorin> poolie, nickname is 'jam' too?
[23:52] <poolie> yes
[23:52] <Noldorin> thanks
[23:54] <Noldorin> poolie, the code for building the windows installer is in the standard bzr branch?
[23:54] <Noldorin> maybe i can just modify that
[23:54]  * jelmer gets some sleep
[23:55] <Noldorin> g'night
[23:56] <poolie> Noldorin: i think it's in a separate project https://launchpad.net/bzr-windows-installer
[23:56] <Noldorin> ok cool
[23:56] <Noldorin> poolie, might be my best bet if i want dulwich/bzr-git on windows?
[23:57] <poolie> yes
[23:57] <Noldorin> ta
[23:57] <poolie> if you get that and update it to include them it could well work
[23:57] <Noldorin> yeah
[23:57] <Noldorin> that's the plan
[23:57] <Noldorin> and also make the installer x64 if it's not too hard!