=== twohey_ is now known as twohey [02:06] just out of curiosity, is it easier to apply patches in bzr than git? [02:16] meoblast001: If you have the patch plugin installed, which you should on most platforms, it's just "bzr patch FILE-OR-URL [02:17] no wonder i don't use git [02:17] i just gave up on trying to patch a file [02:17] apply a patch rather [02:22] meoblast001: Did bzr patch work for you? [02:22] i don't have anything to patch with bzr.. i just know i'm going to have to use it in the future [02:23] and wanted to make sure it wasn't as much of a PITA as it is with git [02:23] meoblast001: You may want to run "bzr help patch" just to check that you have it installed. [02:23] i have Ubuntu so i'm assuming i do [02:26] meoblast001: Hmm. It's not in bzr core. [02:26] yeah, i have it [02:32] I'm having trouble pushing my branch to an SFTP server. [02:33] I'm following the instructions here: [02:33] http://doc.bazaar-vcs.org/bzr.dev/en/mini-tutorial/index.html#publishing-your-branch-with-sftp [02:33] When I push the branch, everyone seems to complete successfully. But when I ssh to the SFTP server to look at my branch, only the .bzr directory is there. None of my files are there. [02:33] Does anyone have any idea why this might be? [02:52] damijit: bzr does not push a copy of the working tree [02:52] damijit: There are plugins to do so, though [03:10] davidstrauss: Thanks, I'm looking at this: https://launchpad.net/bzr-push-and-update/. My problem is that bzr is not install on the SFTP server (it is my school's server), so "ssh [remote] bzr update [branch]" fails. Am I out of luck, or is there some way around this? [03:36] hey y'all [03:36] I'm helping to teach a course in comp sci and I'm trying to use svn to deal with homework handins and assignment/lab handouts [03:36] or bzr [03:37] basically [PUT VCS HERE] [03:37] we have the students pair program through the year but then we have them switch partners a few times through the year [03:37] we (ideally) want the students to see the same source when working together, but i'm not sure how we'd do that [03:37] I guess we could have one repo for all the handouts and push to each user account then let the students deal with sharing their assignment repos [03:56] yeah, I'll probably just make a webapp [09:10] vila: I just ran selftest with up-to-date bzr.dev. Still get the errors that seem related to nonexistent revs (and I forgot --no-plugins, so I get a stack of e.g. qbzr too, but that's another matter) [09:11] vila: But the other stuff you fixed seems fixed here too. [11:24] fullermd: thanks for the feedback, I'd really like to understand the difference between our setups, [11:26] so of course 1) Can you re-rerun with --no-plugins (or better BZR_PLUGIN_PATH=-site ./bzr selftest... which is more precise) 2) try again with python25 or 3) teach me how to install python-2.6 while retaining python-2.5 in FreeBSD [11:27] 2 & 3 are relevant only if you can reproduce the failures, but a mail or better a bug with the failures would be really nice [11:29] fullermd: I just upgraded the botnet to stop using '--no-plugins' which is one more step on the road to test a set of predefined plugins, but testing against several python versions may be relevant if your intuition is right there [13:00] Does automatic repacking run for each 10/100/1000/etc. revisions per branch, or per repository? [13:04] Lo-lan-do: repo [13:10] Great, thanks [13:18] Hm. Pushing a branch to a 2a repo running bzr 1.16 through bzr+ssh:// yields a "bzr: ERROR: Server sent an unexpected error: ('error', "'LocalTransport' object has no attribute 'startswith'")" errol [13:18] error [13:18] Does that ring a bell? [13:20] The resulting branch doesn't seem to be broken [14:00] Lo-lan-do: no idea, sorry [14:01] Lo-lan-do: seems like something is trying to use LocalTransport as a string [14:14] np [14:40] hello [14:41] is it possible for me to use bzr with github (and git branches in general), or does it screw things up? [14:43] Stavros: no, that works [14:43] Stavros: I'm using it for some projects [14:44] jelmer: oh, amazing [14:44] which plugin are you using it with? [14:44] i wanted to use github but i didn't want to switch away from bzr [14:44] does it discard any metadata or anything? [14:46] push isn't supported? what? :( [14:46] Stavros: I'm using bzr-git [14:46] Stavros: you can use dpush [14:46] oh, hmm, what's that? [14:46] Stavros: it throws away metadata that can't be stored in git [14:47] (He's *developing* bzr-git :-) [14:47] * Lo-lan-do grovels [14:47] Lo-lan-do: so are you now >-) [14:47] haha [14:47] Nah, I'm trying to understand it currently. [14:48] Got your mail, btw, I'll try to fix what you mention, but I won't be at home for a week. [14:50] oh yay, this plugin is awesome [14:50] no more pulling stuff on linux and copying to windows [14:50] yay [14:50] now let's see if it works [14:51] ah, it doesn't :/ [14:52] jelmer: are you trying to read a NamedTemporaryFile anywhere? [14:53] Stavros: We're doing something with temporary files [14:53] jelmer: that doesn't work under windows [14:53] Stavros: though I don't think we're using that specific class [14:53] well, it's failing for me and i'm not getting a traceback :/ [14:53] See also https://bugs.edge.launchpad.net/bzr-git/+bug/382125 [14:53] Launchpad bug 382125 in bzr-git "Can't branch from remote repository on Windows" [Undecided,New] [14:54] but if you want to do it in windows you need to make a file with mkstemp() [14:54] I'm not on Windows, so not entirely sure how to debug that [14:54] is it in dulwich or in your code? [14:55] dulwich is my code too :-) [14:55] oh [14:55] I suspect the code that's problematic for you is in bzr-git [14:55] can you check to see if you're using tempfile? i'm going to comment on that bug but i don't want to be irrelevant [14:56] as far as I can tell we use tempfile.mkstemp everywhere [14:56] hmm [14:56] let me grep the code then [14:58] that's odd, you're right [14:58] I'll be back later today [14:59] ah, okay, thanks for your help [15:00] it could be that we're trying to open that file twice or something [15:00] yeah, might be missing a close() [15:07] so, someone branched off my revision 62 and made various changes to various files. i changed one file they hadn't touched, committed, and then merged their changes. bzr produced conflicts in unrelated files i hadn't touched. is this normal? [17:14] i am trying to understand pipes. i have one pipe upstream where i created a file. now i switched to another pipe by bzr switch-pipe and i thought the file would be inherited, but it is not .... [17:21] vila: is there a way to reorder pipes? [17:29] Hi. I've shelved local changes, merged (but have forgotten to commit the merge) and unshelved. Can I separate my local changes now again? [17:35] I'll create a diff against the parent branch, then revert and patch. Looks like what I need. [18:02] this does not work as expected.. gives me a diff of change 7935 (tip) only: bzr diff --old lp:b2evolution -r7909 --new lp:b2evolution -r7934 [18:02] how do I get the changes between 2 revisions in a remote branch? [18:03] oh.. I see.. r7909..7934 should do. [18:05] seems like when using "-r" only the latter gets used? Shouldn't it error out then? [18:12] Is there any way to specify which text editor bzr uses for comments? Right now, it looks like the default is nano. I would prefer for it to use vi. [18:13] RobOakes: I recommend to set $EDITOR and/or $VISUAL [18:13] RobOakes: bzr respects those. [18:13] as do most tools. [18:13] That's good to know. Thanks for the tip. [18:14] RobOakes: otherwise, `bzr help configuration` mentions you can set the editor explicitly. [18:14] * LarstiQ back to studying [18:29] LarstiQ: btw, I'm pretty sure it was nano because that is what 'alternatives' uses as the default editory [18:29] something like /usr/bin/editor ... [18:29] right [18:29] we fall back to vi after that iirc === jam1 is now known as jam [19:32] if a file is moved, are things logged as a deletion and creation, or a move? [19:33] as a move [19:33] it seems quite inefficient to me to delete a file and create a new one if both contain the same contents [19:33] luks: so it doesn't create a completely new copy of a file? [19:34] you mean in the repository, right? no, it just changes name for the file entry [21:24] moin moin [22:18] jam: hi; wondering if remove_expensive_references should discard the null node as well [22:59] ow [22:59] I just did a wrong: bzr add [22:59] how can I undo it plz ? [23:00] note: Haven't committed yet [23:00] bzr revert [23:00] ow shit [23:00] it fucked all up >_< [23:01] bzr rm --keep the-file-you-added [23:01] well it re-added my files anyway [23:01] thanks guys [23:08] vila: (1) yields nothing unexpected; usual list of fails. [23:38] morning [23:39] * fullermd waves at igc. [23:39] hi fullermd! [23:41] good morning [23:48] hi poolie [23:49] hi lifeless, poolie [23:52] * lifeless watches memory usage go past 5G [23:54] * fullermd . o O ( `bzr memtest86` ) [23:54] fullermd: meliae - johns py_memory_dumps' final name [23:54] fullermd: I'm tuning it to let me analyse a dump I got from the fast-import of netbeans, when that when up the kneebend @56K revisions [23:55] 28012 robertc 20 0 5352m 5.2g 1036 R 100 90.2 22:16.97 python [23:55] so this is a dump, of the memory graph, of a 6.4GB runaway process [23:55] I have 6GB of ram on this box :P