[00:10] sup [00:10] is there any simple way to get a git diff between 2 bzr revision trees [00:17] * ronny pokes lifeless [00:18] sorry, no [00:18] not that I know of anyway; bzr-git might have something [00:19] i dont think bzr-git will need it [00:20] ronny: what is a git diff? [00:21] ronny: there's some initial work there in "bzr send --format=git" to behave like git format-patch [00:21] is that what you mean? [00:22] jelmer: i want to get git style diffs betwen 2 revisions (ie the extra lines with rename/add/remove metadata), i never used gits format-patch [00:26] 'bzr diff' gives me that information. [00:26] wgrant: not in the same format as git [00:26] directory renames are more concise tho [00:38] where might I find the bzrlib.plugins.stats module? I'm assuming it's a plugin? Trying to install bzr-gtk from trunk and it says it needs that module. [00:40] cellofellow: the bzr-stats plugin [00:40] oh, duh [00:40] :) [00:43] lifeless: fwiw, looks like there is a eclipse package >= 3.4 in karmic :O [00:44] I'm running karmic... [00:44] lifeless: https://edge.launchpad.net/ubuntu/karmic/+source/eclipse [00:44] oh, this is interesting.. [00:45] Version: 3.4.1-0ubuntu2 yeah [00:45] lifeless: it's still old, but not "years" old [00:46] I was seeing.. [00:46] Version: 3.2.2-5ubuntu3 [00:46] must been new in the archive ;) [00:47] I haven't updated for a week I think [00:49] * verterok --> dinner [01:58] bzr-gtk nautilus plugin is very slow. :( [01:58] s/plugin/extension/ === lamont` is now known as lamont [02:09] howdy [07:02] igc1: hi? [07:03] poolie: Aaron's away until after 1.18 is intended to be released, right? [07:03] i think so [07:06] 'k. [07:13] Oh, rolled over another ten million on launchpadlibrarian. Neat. [10:03] Hi everyone. Does someone have a few minutes to help me with upgrading repository formats? I'm working on the Xibo project, which is developed cross-platform, and we keep having issues with line endings getting mixed up. I've upgraded to 1.17 and written a rules file, but I'm not sure how I go about upgrading the whole project (in Launchpad) and associated branches to use the new format. [10:03] I converted one over, and it shows Branch Working Format 5, but when I then branch from that it comes up as format 4? [10:10] I guess what I'm asking is do I have to upgrade the stable/development trunks first, then my branches? Or can I do stuff in any order? And what is it that causes future branches to use the newer format? [11:19] alexharrington: asking here never hurts, but if you don't get an answer I'd recommend mailing their list. [11:19] yeah - I'm posting on Launchpad Answers now - see if I get any joy there [11:19] thanks [12:04] alexharrington: actually, I said Bazaar's mailing list. [12:04] answers is fine === cprov-afk is now known as faces === faces is now known as cprov [13:26] hello. could anyone please tell me the file in the source in which unlocking is done? [13:26] (including the rename to unlock) [13:44] Noldorin: bzrlib/lockdir.py ? [13:44] Noldorin: combined with .unlock() on branch/repository/tre [16:08] I have a quick question about permissions after an initial branch. If I do the branch locally. bzr branch the umask is honored [16:09] however if I do it using the bzr+ssh method bzr branch bzr+ssh://user@host/path it doesnt honos the umask [16:09] s/honos/honor [16:09] so I have to manually chmod -R g+r /path on the server for every created branch :( [16:12] chmod -R g+w rather... [16:15] has no one here ever tried to have a multi-user bazaar server using the bzr+ssh method? [16:27] well, adding os.umask(002) right after the if __name__ == '__main__': might be a nasty hack but it works for now... bzr should really fix the bzr_ssh method to honor the umask as set in /etc/profile... perhaps its not running the shell in login mode so /etc/profile isn't being executed. This won't work for multi-user environments... [16:28] Proteuskor: please report a bug and your work around, no everybody will read the IRC logs... [16:29] I wonder if its a misconfig of my system still :) thats why I was asking here first [16:29] could be my strange ssh server (I use dropbear because of its reduced memory footprint) [16:29] oooh, is that you that plays with tiny servers ? :-D [16:30] s/that/who/ oops [16:30] indeed [16:30] hmm, well, at least, instead of your "nasty hack" (your words :) you can just define a wrapper and set BZR_REMOTE_SSH [16:31] err, BZR_REMOTE_PATH (see bzr help env-variables) [16:31] ok, let me look into that real quick :) [16:32] that way you eon't have to patch bzr every time (if it's due to your config) [16:32] ya, I like that idea better, thank you [16:32] although I'm not yet sure how to get my windows users to tickle that setting via the tortoise GUI [16:33] and remember that BZR_REMOTE_PATH can also be set in config files (don't remember the config variable name though) [16:33] lol, messages crossed on the wire [16:33] ok, I can prolly just distribute a config file to all of the users then :D [16:34] I am actually new to bzr, I used to use perforce, just because thats what the company was using when I joined [16:34] comming up the learning curve now [16:34] hehe, the pleasure is in learning a little bit every day :-D [16:39] i think the real fix is to get the ssh transport to use login mode so /etc/profile is processed before calling the specified server. That way the systems configured umask is honored... but there may be a reason they didnt want to use login mode. If the comment at the top of transport/ssh.py is right I could email robey pointer and just ask. I dun want to file a bug if there is a good reason for the current behaviour [16:40] Proteuskor: robey doesn't work on bzr these days, you'd better file a bug, you'll get a far better tracking of the issue, really [16:40] ok [16:43] Hello guys [16:44] is there a way to override make AC_PACKAGE_VERSION in autoconf reflect the current revno in a bzr branch? [16:44] to make* [16:47] AirBender: hacking around: bzr version-info --custom --template "{revno}" [16:49] vila: yeap, but do you know the procedure to make AC_PACKAGE_VERSION read a file with the revno, or execute this bzr command to get the number? [16:49] I know nothing about autoconf except using it :) [16:50] because in AC_INIT I can't put a macro I think... [16:50] me too lol [16:51] but I'm pretty sure there's a simple way to do this... I don't like to change manually the revno every time in configure.ac [16:53] Did you ask Google? :-) [16:53] http://doc.bazaar-vcs.org/bzr.dev/en/release-notes/NEWS.html#in-development has two bzr 1.18 releases [16:54] weigon: the more the merrier !! [16:55] please add at least a 3rd :) [16:56] vila: yes I have, but may be I'm searching in the wrong way... [16:59] well, most of the time yo found in the last place you search so... keep digging :) === Xavura is now known as Xavur[a] === Xavur[a] is now known as Xavura [19:36] is there any way to have access control with `bzr serve` ? [20:05] is there a known bug with 2a and patch files? [20:05] I just tried to apply a merge bundle someone sent me, and bzr died with AttributeError: 'AbsentContentFactory' object has no attribute 'get_bytes_as' [20:07] mmm actually no [20:07] the branch has the problem [20:07] it dies with the same error when doing bzr check [20:18] so I branched the bad branch with 1.18rc [20:18] and the new branch looks ok (according to check) [20:19] however now the merge command is failing with "bzr: ERROR: Bad bzr revision-bundle: "Can't convert to target format" " [20:19] whatever the hell that means [20:25] johnjosephbachir: not at present, but I just started working on it [20:26] denys: cool. so it seems like the only read/write hosting solution with access control that doesn't require making OS users is webdav, yeah? [20:28] johnjosephbachir: and ssh to some extent [20:47] No, you could use the smart server over HTTP. [20:47] (and of course over ssh doesn't necessarily require system users, since you can putz with how sshd authenticates) [21:01] Hi. I'm a git user. A few months ago I wanted to contribute to a project using bzr, and installed some old version of bzr (maybe 1.6?). Today I updated my bzr to 1.18rc1, because I want to contribute to a different project. [21:02] that's ok [21:02] After the upgrade, upon running bzr status in the repository I branched, I get an uncaught TypeError. I read a bit and found this bug to match what happened to me: https://bugs.launchpad.net/bzr/+bug/334202 [21:02] Ubuntu bug 334202 in bzr "internal error occured on bzr status" [Undecided,Fix released] [21:03] The workaround works and indeed I have bzr-loom 1.4dev in the output of bzr plugins. [21:03] The question is: I'm not sure how to remove the plugin altogether, and, more importantly, since I'm not well versed in bzr I don't know if bzr-loom is a very important and often used plugin or not. [21:03] Recommendations? [21:03] run `bzr plugins -v` [21:04] it will tell you where plugin located [21:04] And I just rm -fr it? [21:04] go there and delete loom directory [21:04] yep [21:04] Excellent, very nice. [21:04] bzr-loom provided special workflow, so in most cases you don't need it [21:05] And indeed bzr works now. [21:05] OKie, thanks a bunch. [21:05] :-) [21:05] You know what, while I'm here, another question. [21:05] shoot [21:06] I might have my terminology a bit git-ish, I hope you'll understand me. The feature I probably like best in git is 'git rebase -i'. [21:06] -i means interactive? [21:06] After doing a series of local commits on my repo, and before I 'pushed' the changes to the upstream repo, I can see my local changes, edit them, edit their commit messages, reorder them, squash them into one commit, etc, etc. [21:06] Yep. [21:06] Is there something similar to play with in bzr? [21:07] no, there is rebase but it has no -i [21:07] OK, never mind. I'll do without for now. [21:07] Thanks again, see you around bialix. [21:07] np [21:20] fullermd: is usage of the bzr server over http mentioned/described/explained somewhere? [21:34] denys: Windows or Linux? [21:37] awilkins: access control, you mean? [21:38] HTTP server [21:38] awilkins: linux [23:22] help! on windows, bzr add on a specific subdir fails with ERROR: "blabla..." is not a working copy [23:24] lifeless: ping? [23:24] sorry to bother you again :) [23:30] crap. bzr does that because the subdir is versioned by .svn [23:30] but who cares! [23:31] did I ask bzr to give a damn about svn? [23:31] the error is very confusing. [23:32] sinelaw: Do you maybe have bzr-svn installed?