=== mmrazik is now known as mmrazik|otp
[07:58] <mgz> morning!
[08:15] <jelmer> moin
=== mmrazik|otp is now known as mmrazik
[08:58] <christiank> mgz: I will be in this channel today in case you have further questions regarding the bug we discussed yesterday.
=== mmrazik is now known as mmrazik|lunch
[09:23] <mgz> christiank: thanks. I won't be able to do anything else on it till this evening.
[09:57] <christiank> mgz: Thanks for letting me know. I will try to be on this channel tomorrow as well then.
=== mmrazik|lunch is now known as mmrazik
[10:40]  * christiank is away: Lunch
[11:44]  * christiank is back (gone 01:04:14)
[13:40] <mark06> why has bzr set a wrong time for my commit? http://bazaar.launchpad.net/~renatosilva/pidgin/windev/revision/12
[13:40] <mark06> right time was 1h earlier, 23h39 GMT-3
[14:11] <delinquentme> https://gist.github.com/ff315efc13d0582ed809  <, I've got this guy and I was assuming that bzr would make a fast forward commit here
[14:11] <delinquentme> but it didint
[14:11] <delinquentme> SO
[14:16] <delinquentme> OK another user has branched my redesign branch ... they've made edits .. and pushed up
[14:16] <delinquentme> i had subsequenly made another commit
[14:16] <delinquentme> we're both on revno 150
[14:17] <delinquentme> bzr merge gives me "Nothing to do."
[14:22] <mgz> mark06: I wouldn't trust bzr for an alibi, but I'd be surprised if it was an issue there rather than with your system time
[14:22] <mgz> mark06: pastebin the section of your .bzr.log (find it with `bzr version`) that corresponds to the commit?
[14:23] <mgz> delinquentme: are you trying to merge the right thing?
[14:24] <mgz> create a local copy of his version of your branch, merge that into your branch, then push the result
[14:25] <delinquentme> mgz, so i should branch his branch
[14:25] <delinquentme> merge
[14:25] <delinquentme> and then push to mine?
[14:26] <mgz> yup, though you can do the merge in either direction
[14:27] <delinquentme> ok so mgz im a bit confused here ... so I'm running a merge between two branches right?
[14:27] <mgz> yes.
[14:27] <delinquentme> however I thought with bzr branches correspond to file
[14:27] <delinquentme> files*
[14:27] <delinquentme> right?
[14:27] <delinquentme> so if there are two branches .. there then would be two files?
[14:27] <mgz> ...that doesn't mean anything to me
[14:28] <delinquentme> I mean that if you've got a project
[14:28] <delinquentme> that project has a primary branch ( canonically named trunk ) as well as b1 b2 b3
[14:29] <mgz> right, but that's just convention
[14:29] <mgz> there's nothing special about any of those
[14:29] <mgz> if I have a change on trunk, someone else pushes a change to trunk, I can merge his change and push my branch with:
[14:31] <mgz> `bzr merge -d trunk lp:PROJECT && bzr commit trunk && bzr push -d trunk lp:trunk` (with 'trunk' being my copy of the branch in the cwd)
[14:35] <mgz> lp:PROJECT in both places, rather
[14:36] <fullermd> And then he can show up on #bzr wondering why bzr lost his commit  ;p
[14:37] <mgz> well, yes, but the merge should be pretty obvious :)
[14:37] <mgz> I'm not going into append_revisions_only fun
[14:52] <drostie> I don't know the technical term for this, maybe 'underlying object structure' or so -- can anyone point me to a doc on the 'physics' of how bazaar conceptualizes versions/branches and stores them and so on?
[14:52] <drostie> I'm trying to build an accurate mental model of what's going on 'behind the scenes' in various revision control systems.
[14:59] <mark06> mgz: I'll need to do some test as I don't use .bzr.log, but this is later as the PC with problem is not at hands
[15:00] <mgz> drostie: see <http://doc.bazaar.canonical.com/bzr.2.5/en/user-guide/core_concepts.html> and the bits under <http://doc.bazaar.canonical.com/bzr.2.5/developers/> and the docstrings in the source as needed
[15:01] <mark06> mgz: that was bzr in windows + mingw msys, any clue? even if date command mistakenly swicthed to DST, I wouldn't think bzr would reply on it
[15:01] <mark06> mgz: or is bzr sensitive to TZ env var?
[15:01] <mgz> mark06: .bzr.log happens automatically, or do you mean you redirect it to /dev/null
[15:02] <mark06> s/reply/rely
[15:02] <mgz> probably windows knew the correct tz adjustment then but mingw was confused, I'd guess
[15:02] <mark06> mgz: yes, BZR_LOG=nul (under Windows)
[15:02] <mark06> mingw is not confused
[15:02] <mark06> everything is allright, only bzr is doing abd things here
[15:03] <mgz> bzr uses whatever the python gets from the system, which is different on native windows an emulated nix
[15:04] <mark06> the only exception would be some mistake in TZ so that DST has been activated earlier (which I need to check later), but even so it's crazy bzr relying directly on TZ or even date, no?
[15:04] <drostie> mgz: thanks, will look at it! :D
[15:05] <mgz> if `python -c "import datetime;print datetime.datetime.now()"` prints the right thing, I think that's what we go from
[15:06] <mgz> or, specifically, look at local_time_offset/format_date in bzrlib.osutils
[15:10] <mark06> I'll check later mgz, thanks
[15:25] <diddly> hi, i'm trying to set up a centralized repo following http://wiki.bazaar.canonical.com/SharedRepositoryTutorial but I don't think the SGID part is working, directories don't inherit the g+w bit
[15:26] <fullermd> sgid doesn't make them inherit g+w, it makes [SysV-ish filesystems] inherit the group ownership.
[15:27] <diddly> then i'm misunderstanding the instructions I think?
[15:28] <fullermd> No, you're probably more stepping outside of them.
[15:29] <fullermd> And/or you're hitting the bug where bzr doesn't propagate perms on newly created branches.
[15:30] <diddly> fullermd: so i'll need to manually set it when branches are created?
[15:30] <diddly> (ie on the server-side)?
[15:31] <fullermd> Simplest solution.  In some cases you can also get away with playing umask games, etc.
[15:31] <diddly> yea, I want to avoid that, it would probably make worse issues :)
[15:31] <diddly> what bug id# are you referring to?
[15:32] <fullermd> Dunno.  Lemme see if there's one filed.
[15:32] <diddly> fullermd: ohh, no thats ok I thought you had already looked it up
[15:33] <fullermd> Only insofar as it's a long-standing shortcoming that I know about.
[15:34] <fullermd> bug 403250 talks about it, looks like.
[15:34] <ubot5> Launchpad bug 403250 in Bazaar "branching in a shared repo doesn't inherit permissions" [Medium,Confirmed] https://launchpad.net/bugs/403250
[15:36] <diddly> fullermd: wow thank you much for helping me with this!
=== mmrazik is now known as mmrazik|otp
[16:43] <delinquentme> get the last commited version of a file?
[16:44] <delinquentme> IE undo the edits made in the working tree?
[16:44] <mgz> `bzr revert file`
[16:44] <mgz> `bzr cat file` if you just want to look at it
=== mmrazik|otp is now known as mmrazik
=== yofel_ is now known as yofel
[22:44] <zyga> vila: ping
[23:14] <zyga> anyone around, could I get a pair of eyeballs on https://code.launchpad.net/~zkrynicki/bzr/find_ancestors/+merge/127610