[00:12] Jc2k, re [00:14] Jc2k, k [00:17] :) [01:46] Merry Christmas, #bzr. :) [02:00] merry christmas Odd_Bloke :-) [04:05] Hello colleagues, and Merry Christmas! [04:07] I keep getting 'ImportError: no module named commands' with TortoiseBzr from bzr-setup-1.10-1.exe [04:08] on several systems, most of them were working well with bzr-setup-1.9.exe === Mario__ is now known as pygi === mark1 is now known as markh === mark1 is now known as markh [13:17] merry christmas :) === magcius_ is now known as magcius [13:44] merry christmas. [13:44] merry christmas :-) [13:44] hi Jc2k [13:44] *blinks* [13:45] i was just going to ping you jelmer :) [13:45] push and clone against server.py are working again [13:49] ah cool [13:49] * jelmer merges [13:49] hopefully pull too, really need tests gah :p [14:00] Is bzr daemonless? [14:02] magcius, not sure what you mean with daemonless [14:02] Does there need to be a server running like traditional VCS toold? [14:02] tools* [14:02] magcius, you can pull / push to a remote server without a bzr-specific daemon necessarily running on that server [14:03] jelmer, so it uses WebDAV? [14:03] magcius: But you can optionally use a server on the remote side, for better performance [14:03] magcius, the main "dumb" transport that seems to be used is sftp [14:03] magcius, ftp is also supported out of the box, and there is a plugin for webdav [14:03] Ah. Okay. [14:06] Jc2k: I'm working on getting packs fixed, and on supporting fetch from remote in bzr-git [14:06] jelmer: okie dokie. i'm tempted to move the pack generation out of server.py and make client.py use it [14:07] Jc2k: Go for it :-) [14:09] jelmer: im a bit cautious of moving anything to pack.py as i need to go via the repo object.. dont like the bouncy-around-dependencies :\ am i just being over sensitive, or shall i add packwriter.py/packbuilder.py? [14:09] Is there a bzr emacs ... err.. hook? [14:10] magcius: there is the emacs dvc mode iirc, and it does support bzr [14:10] Jc2k, yeah, same here. What do you need the repo object for? [14:11] asabil, not according to this here HOWTO [14:11] magcius: http://www.xsteve.at/prg/emacs_dvc/dvc.html#table [14:12] jelmer: if i'm building a pack, i walk the dag to work out what objects to send. currently i use the repo both for the tree() /commit() methods and because in theory i might be generating a pack out of loose objects [14:13] Jc2k, perhaps just extract out a function that works out what objects to send based on a list of shas and a callback for get_object ? [14:14] it would need def pack_contents(want, have, get_commit, get_tree).. [14:15] i guess i could also follow the bzr pattern and have a wrapper on the repo object that fills in get_commit and get_tree for me [14:16] hello. for example, i create branch from other place. does exist some way to check for available changes in origin branch without 'bzr update' run? [14:18] Jc2k, You can just use get_object() rather than get_commit / get_tree; they do the same thing [14:18] ia, "bzr missing " [14:21] jelmer: oh, thank you very much. that's exactly what i looking for :-) === magcius_ is now known as magcius === magcius_ is now known as magcius === Mario__ is now known as pygi [16:36] Jc2k, updating my samba tree is actually already pretty quick === magcius_ is now known as magcius === magcius_ is now known as magcius === magcius_ is now known as magcius [18:38] for bzr, how can I get the last checkin to a specific directory? [18:38] with svn, I'd just do svn log /path/to/dir | head -n 1 or something [18:40] sohail, do you want the checkin time or the revision number? [18:40] magcius, I want the revision number [18:40] I tried bzr revno /path/to/dir but that gives me the revno for the whole repo not that directory [18:41] sohail, changes to the dir, or anything in it? [18:41] beuno, anything in it actually [18:41] I think each revision is a branch, so it can't distinguish non-changed files [18:41] sohail, right, I don't think you can do that with the bzr client [18:41] maybe bzrlib, if you're into python [18:41] :-O [18:42] oh well, I need it to overcome a slight design miscalculation in my build automation but I'm still surprised [18:42] I'd rather fix that than figure out bzrlib :-) [18:42] thanks for your help! [18:42] :) [18:43] file bugs, everyone lines bugs in bzr [18:44] I assume you mean to say likes bugs [18:44] ah, yes [18:44] I do [19:13] im looking for configuration eclipse + bazaar plugin help... === asac_ is now known as asac [21:20] lifeless: around? [21:58] bzr-svn is giving me headaches. I'm trying to push and it tells me the branches have diverged and I need to merge first, which is certainly false. I merge anyway and it says "nothing to do," and I loop back to the beginning :(. Any help? [21:58] bewst1, did you commit your merge? [21:59] bewst1, or do you perhaps have any local changes? [21:59] bewst1, what version of bzr-svn are you using? [22:01] jelmer: I don't know what it means to "commit a merge." I did commit one file that I have changed, but I still have some changed files in my working copy that I don't want to push. So, yes, I have local changes [22:01] bewst1, and the branch you're pushing to has no changes you don't have locally? [22:01] That's correct [22:02] double checking... [22:02] Why does everything seem to go so slowly? [22:03] Yep, pull says "no revisions to pull" and merge says "nothing to do." [22:03] bewst1, and you're pulling from the same location that you're pushing to? [22:03] bewst1, what does "bzr missing " say? [22:06] yes, pull and push locations match [22:06] bewst1, what version of bzr-svn are you running? [22:06] bzr missing says I have 4 extra revisions. [22:06] How to check the bzr-svn version? [22:07] bewst1, "bzr plugins" should print it [22:07] 0.4.13 [22:07] It's what Intrepid gives me by default [22:08] I don't know what this "4 extra revisions" business is; I'm going to look at it [22:08] Oh, it also says I'm missing 13 revisions [22:09] bewst1, that would explain the divergence [22:10] not sure why pull doesn't do anything though [22:10] How is that possible? Those revisions are basically the whole history of the file [22:10] Maybe I did a lightweight branch? [22:10] bewst1, did you change the branching scheme recently? [22:10] bewst1, is the branch public? [22:10] Sorry, n00b. What is a branching scheme? [22:11] The svn branch is not public [22:11] bewst1, if you don't know I guess you haven't changed it :-) [22:11] the bzr branch is in a local repo on my machine [22:11] I don't know from public bzr branches [22:12] The svn branch is... a branch of some svn trunk. The 13 branches missing appear to be those that were made on trunk before the svn branch [22:12] Does that help? [22:14] The 4 "extra" revisions appear to be the ones made in the svn branch plus the one I committed to my bzr repo and am trying to push [22:14] bewst1, if you run "bzr missing --show-ids" are the revision ids for the revisions the same? [22:15] Sorry, the same as /what/? [22:16] bewst1, the revisions in the svn branch as reported by "bzr missing" [22:17] since "bzr missing" would report some revisions twice in the "missing" and the "extra" listings [22:17] ?? without the --show-ids, no ids are reported. Do you mean "are the revisions numbers the same?" [22:18] bewst1, no, the revision ids [22:18] bewst1, if you run "bzr missing --show-ids", it should report a list of revisions that's missing locally and a list of revisions that's extra locally [22:18] jelmer: yes [22:18] bewst1, for each revision it should print the revision id [22:18] ues [22:18] yes [22:18] compare that id with what? [22:19] bewst1, if I understand you correctly, some revisions are listed in both the first and the last list [22:19] first? last? Sorry, I don't know what you mean [22:19] bewst1, some revisions are in both the "locally missing" /and/ the "locally extra" list, right? [22:20] no [22:20] not sure I follow you then [22:20] The "locally extra" list contains revisions made in the svn branch, plus the one I committed to my local bzr repo. The "locally missing" list contains revisions made in the svn trunk before the svn branch point [22:21] Taken together that appears to be *all* of the revisions of the one file that I changed in the local repo [22:22] Sorry, I should have said "directory," not "file" [22:22] bewst1, can you pastebin the "bzr missing" output? just the revision-id for each commit should be sufficient [22:26] http://dpaste.com/102508/ [22:27] The only revision made in the bzr repo is on line 1 [22:27] Lines 2-4 are revisions made to the svn branch [22:27] The rest are revisions made to the svn trunk [22:29] jelmer: nothing jumps out at you? [22:29] bewst1, it does actually [22:30] bewst1, the branching scheme has changed underneath you [22:30] possibly when we upgraded to svn 1.5? [22:31] I'm still not sure what a branching scheme is ;-) [22:31] bewst1, (fwiw, this bug is fixed in bzr-svn 0.5, which no longer has branching schemes) [22:31] bewst1, it should be possible to work around it though [22:32] Why don't I just upgrade? [22:32] I'd rather get something that works than get involved in workarounds. [22:32] bewst1, 0.5 is still in development, and requires bzr 1.10 [22:32] bewst1, the workaround shouldn't be too bad [22:33] listening [22:33] bewst1, in ~/.bazaar/subversion.conf, set "branching-scheme = trunk2" in the section [bbad15d5-7c10-0410-ba9e-f711d5c1d17a] [22:33] How is it possible that the branching scheme, which is not a concept I even know about, has changed? [22:33] bewst1, bzr-svn detects it based on the URL used [22:34] Oh... well, there's a little problem with the svn+ prefix which I have occasionally ommitted and then added back [22:34] bzr-svn complained that it was deprecated, but then wouldn't do something unless I supplied it. [22:35] bewst1, yeah, that's a known bug (in bzr itself) [22:36] OK, so I'll try this workaround [22:37] maybe also set "branching-scheme-mandatory = True" [22:38] jelmer: same result without mandatory. Will try that next [22:39] btw, it changes the branching scheme back to what it was when it fails [22:40] It's not failing... yet. But damn it's slow [22:41] Finally it asked for my svn password twice and failed with an internal error. [22:41] internal error? [22:42] http://dpaste.com/102509/ [22:43] I've gotta say, this first encounter with bzr is a bit of a nightmare. [22:44] bewst1, wait.. weren't you trying to push to boost-consulting/site/branches/seo ? [22:44] I know DVC is supposed to speed up development, but I could've gone a long way further with SVN by now. Is it worth trying to figure this out, or should I come back when 1.5 is out? [22:45] bewst1, I would recommend trying again when bzr-svn 0.5 is out [22:45] Well, I was in a subdirectory of seo called utils, and was trying to push that [22:45] * LarstiQ would say trying to back on svn is the main source of pain here [22:45] bewst1, that prevents all this mess with branching schemes [22:45] bewst1, you can't push subdirectories in bzr [22:46] bewst1, at least that explains why it was switching to a different branching scheme [22:46] Oh, I totally messed this up. I'm an idiot. Let me try again. [22:46] you probably want to push to boost-consulting/site/trunk I think? [22:47] bewst1, you're not an idiot, but bzr-svn 0.4 exposed some internals that really should be hidden from the user.. [22:47] I was even pushing to the trunk. Gah. But now it says no new revisions to push [22:48] Do I need to revert the branching scheme? I didn't save it :( [22:48] bewst1, if you push to trunk it should use the right branching scheme [22:48] jelmer: I didn't want to push to trunk! [22:48] I was being a fool [22:48] bewst1, are you sure the revision isn't in svn yet? [22:48] Will look. [22:50] it is! [22:51] OK, will try some more. Any idea what the release date for 0.5 /might/ be? [22:51] bewst1, roughly a month from now I hope [22:51] That might be worth hanging on for [22:51] there is a release candidate out that people found a few bugs in, I plan to fix those and release another RC in a week or two [22:52] kewl; thanks [22:52] unless any significant bugs come up in that RC, I will release that as 0.5 [22:53] Does bzr-svn interoperate properly with svn 1.5 branch/merge tracking? [22:53] bewst1, somewhat [22:53] bewst1, only for full merges (no cherrypicking) [22:53] bewst1, since none of the main DVCS'es do cherrypicking tracking [22:53] (yet) [22:53] Sorry, what constitutes a full merge? [22:54] bewst1, if you "svn merge" a branch, it will track that [22:54] bewst1, if you "svn merge -c" a specific revision only or a revision range, it won't [23:00] jelmer, OK, thanks. So if you think you might want to do something like that, just create separate bzr branches for each such separate change? [23:07] bewst1, for example