[01:39] what's the easiest way to create an archive of my repo? or at least the latest version of it? [01:41] unenough: for what purpose? [01:42] backup [02:00] unenough: bzr push sftp://mybackuphost/backup? [05:20] Um, er, help! [05:21] * RAOF waits for the rest :) [05:21] I have a perfectly ordinary Subversion checkout of some upstream project. I want to use Bazaar to track the few files I care about and to shuttle my own patches back and forth. Nothing terribly unusual. [05:22] But I just did `bzr init .` in said checkout and suddenly bzr-svn kicked off starting to do its cache update etc as if I were going to use bzr-svn to pull from Subversion! [05:22] a) I didn't expect that [05:22] b) I don't really want it to do that. [05:22] So [05:22] * AfC wonders what to do. [05:23] That's kinda wierd, yeah. I dunno. [05:25] I mean, sure I could uninstall bzr-svn, but that's a bit much. [05:27] AfC: ooh ooh, I know this [05:27] gimme a sec [05:28] AfC: bzr --builtin init . [05:28] * AfC just did `bzr init --no-plugins .` but that's ridiculous! [05:28] --builtin? [05:28] That's a new one on me. [05:28] AfC: me too, I found out about it yesterday [05:28] AfC: it's in 'help global-options' [05:29] jml: Ah. I was just searching help looking for that help page (and not finding it) :| [05:35] FUCKING HELL. [05:36] I just tried to `bzr add` a single file and then bzr-svn kicked off and tried to update its cache again. [05:36] This is not a bzr-svn branch, god damn it! [05:37] Oh, that's just right royally fucked. `bzr --builtin add filename` doesn't skip the svn plugin. === luke-jr_ is now known as luke-jr [08:21] hi hi gour [08:22] hello pygi [08:37] i'd like to fetch pinax project via bzr-svn which contains several external apps referenced in svn.externals. since i'm not at familiar with svn, I wonder if bzr-svn can do the job or am i better using svn? [08:42] it looks bzr-svn cannot handle it (yet) [09:02] How do I make the "committer" field in a commit to a local branch use something other than my OS real name, user name and host? [09:03] bzr whoami [09:06] bob2: Thanks. I thought I had already done that (using --branch), but it was reset when I re-created the branch. [09:09] However, the Bazaar plugin for Eclipse does not seem to respect whoami, and instead just uses my user name. [10:07] hey guys [10:08] Using bzr-1.7 I can't check out http://bazaar.launchpad.net/~team-mms/mms/1.1.0 mms-1.1.0 [10:08] It is stalled after a while [10:16] loswillios: Stalled? There's no network activity or anything? Are you sure? It might just be slow. [10:17] Peng_: it says \ [---- ] bzr get http://bazaar.launchpad.net/~team-mms/mms/1.1.0 mms-1.1.0 [10:18] err, \ [--- ] Transferring 0/4 for a while and then there's nothing [10:18] the progress bar has stopped, but it is still downloading in the background [10:18] nope, no progress bar at all [10:19] just disappears and I have to kill the process [10:19] not even ctrl-c works [10:19] currently trying to reproduce on another machine which is faster [10:19] loswillios: Try running with -Dhttp, that'll log all HTTP activity. [10:19] but I left the command sitting for an hour or so [10:19] ok [10:20] loswillios: If it isn't printing to std{out,err}, then look in ~/.bzr.log [10:21] (-EBASHISM) [10:22] * Odd_Bloke --> breakfast [10:22] ok, currently running.. [10:23] is bzr get deprecated? I can't find it in the man page anymore [10:23] it has always just been an alias [10:26] ah [10:26] loswillios: It's probably listed in 'bzr help branch'. [10:27] Right, actually to breakfast. [11:22] nice, trying to check it out oom'd my linode [11:55] Hi, I know how to do self.run_bzr('add hello') in a test, but I would like to translate this to bzrlib's API calls. Is there a guide which tells how to translate each operation? [12:22] clemente: no, not really. [12:22] clemente: the easiest thing to do is read the source for bzrlib/builtins.py [12:22] clemente: although, for that thing specifically, you want something like WorkingTree.open('.').add('hello') [12:23] clemente: there's also a hacking guide which might help [12:25] jml: yes, thanks, that's a help. I also found this http://bazaar-vcs.org/Integrating_with_Bazaar , but it tells about working with real files, and not directly with revisions, commits, inventories etc. [12:25] clemente: cool. [12:29] Hey guys, I've got a problem with "bzr shell" from bzrtools [12:29] that is, it kills itself with this error message: [12:29] bzr: ERROR: [Errno 22] Invalid argument [12:29] which according to the debugger occurs here: [12:30] 3 dwt@dwcp ...ktive Projekte/otr-messaging/Adium-bzr % BZR_PDB=1 bzr shell [12:30] bzr: ERROR: [Errno 22] Invalid argument [12:30] **** entering debugger [12:30] > /Users/dwt/.bazaar/plugins/bzrtools/shell.py(105)__init__() [12:30] -> readline.read_history_file(self.history_file) [12:30] (Pdb) l [12:30] 100 ensure_config_dir_exists() [12:30] 101 self.history_file = osutils.pathjoin(config_dir(), 'shell-history') [12:30] 102 readline.set_completer_delims(string.whitespace) [12:30] 103 if os.access(self.history_file, os.R_OK) and \ [12:30] 104 os.path.isfile(self.history_file): [12:30] 105 -> readline.read_history_file(self.history_file) [12:30] 106 self.cwd = os.getcwd() [12:30] 107 [12:30] 108 def write_history(self): [12:30] 109 readline.write_history_file(self.history_file) [12:30] 110 [12:30] Now I'm not sure what might cause this - self.history_file looks great [12:30] the fiel is there and readable too [12:30] dwt_: this happens immediately on startup? [12:30] (as the tests on the two previous lines should assure) [12:30] yes [12:31] dwt_: which version of bzr & bzrtools? [12:31] bzr: 1.7.1 [12:31] bzrtools 1.7 (the latest) [12:31] (I just downloaded it) [12:32] dwt_: can you run the command with -Derror and paste the output to paste.ubuntu.com? [12:32] sure [12:33] http://paste.ubuntu.com/54222/ [12:34] thanks [12:34] Do I read the debugger output right that calling readline.read_history_file() immediately bombs? (it seems to be a wrapped c-library) [12:35] yeah. [12:35] dwt_: it might be worth trying: import readline; readline.read_history_file('/whatever/the/path/is') [12:35] (Pdb) pp readline.read_history_file.__doc__ [12:35] 'read_history_file([filename]) -> None\nLoad a readline history file.\nThe default filename is ~/.history.' [12:35] This however suggests that it is called correctly, right? [12:36] dwt_: the error is almost certainly coming from some lower level file operation [12:36] the actual Python call is correct. [12:37] I*'l try the readline call myself [12:37] dwt_: thanks [12:37] hm, when I try this in an ipython shell [12:37] it works perfectly [12:38] interestingly [12:38] when I try it in a normal python shell [12:38] it bombs [12:38] dwt_: ok, one more thing to try before I blame high solar electromagnetism [12:38] oohh [12:38] dwt_: incidentally, what's the value of self.history_file? [12:39] (Pdb) pp self.history_file [12:39] '/Users/dwt/.bazaar/shell-history' [12:40] dwt_: ok. let's try strace :) [12:40] oh dang [12:40] that would be dtrace for me [12:40] (mac leopard) [12:40] as to why ipython can do it [12:40] dwt_: I should have guessed from Adium. [12:40] its actually using it's own implementation of readline [12:40] ¿ [12:42] dwt_: so, if you can paste the last few bits of strace (or dtrace) output to the pastebin, I'd be much obliged. [12:42] I'l try [12:42] But you have to give me some minutes [12:42] as I'm not at all up to speed with dtrace [12:43] dwt_: with strace, it's just "strace bzr shell" [12:43] my OS X-fu is rusty, I'm afraid. [12:43] yeah, as far as I remember with dtrace [12:43] it should be "sudo dtruss bzr shell" [12:43] I'm working on it. :) [12:44] so when I try this: [12:44] sudo dtruss python -c "import readline; readline.read_history_file('/Users/dwt/.bazaar/shell-history')" [12:44] I'm getting tons of output [12:44] Hope you can read it. :) [12:44] http://paste.ubuntu.com/54225/ [12:45] are those the last lines? [12:45] yep [12:46] sigaction looks like it died from a signal, right? [12:46] well, there's a syntax error there. [12:47] damn [12:47] what happens when you drop the 'dtruss' [12:47] Just a second [12:49] where exactly do you means I should drop the dtruss? [12:49] as far as I can see I only have it in the right location [12:51] I mean, run this: sudo python -c "import readline; readline.read_history_file('/Users/dwt/.bazaar/shell-history')" [12:51] sudo python -c "import readline; readline.read_history_file('/Users/dwt/.bazaar/shell-history');" [12:51] Traceback (most recent call last): [12:51] File "", line 1, in [12:51] IOError: [Errno 2] No such file or directory [12:51] Interesting [12:51] Now where's that syntax error. [12:52] so, you aren't getting the same output as when you run it as your user. [12:52] yes [12:52] interestingliy [12:52] bummer. [12:52] all the while the fine sits there just fine [12:52] 127 dwt@dwcp ~ % ll /Users/dwt/.bazaar/shell-history [12:52] -rw------- 1 dwt dwt 2443 7 Aug 10:18 /Users/dwt/.bazaar/shell-history [12:52] o.O [12:53] on HFS? [12:53] yes [12:55] do you guys get [12:55] why on earth its a different error message from doing this directly [12:55] vs. calling it via bzr shell? [12:55] when I do it directly it's: IOError: [Errno 2] No such file or directory [12:56] dwt_: even without the 'sudo'? [12:56] 1 dwt@dwcp ~ % python -c "import readline; readline.read_history_file('/Users/dwt/.bazaar/shell-history');"Traceback (most recent call last): [12:56] File "", line 1, in [12:56] IOError: [Errno 2] No such file or directory [12:56] while with bzr I get [12:56] 1 dwt@dwcp ~ % bzr shell [12:56] bzr: ERROR: [Errno 22] Invalid argument [12:56] no idea, but case-insensitive filesystems make me suspicous [12:57] Even though it seems readline is called exactly the same [12:57] bob2: good call! [12:57] dwt_: what does, python -c "open('/Users/dwt/.bazaar/shell-history')" say? [12:57] nothing [12:58] weird weird weird [12:58] dwt@dwcp ~ % python -c "print open('/Users/dwt/.bazaar/shell-history')" [12:58] [12:58] dwt@dwcp ~ % python -c "print open('/Users/dwt/.bazaar/shell-history').read()" [12:58] ls [12:58] diff [12:58] shell open AnimatedGrid.xcodeproj/ [12:58] help shell [12:58] open AnimatedGrid.xcodeproj/ [12:58] exit [12:58] diff [12:58] ci -m "added a new provider and refactored the storage model to use a two dimensional array" [12:58] push [12:58] help push [12:58] bzr info [12:59] bzr push parent [12:59] bzr st [12:59] ls [12:59] ls parent/ [12:59] yeah, we don't need it all :) [12:59] ... [12:59] I don't get this [12:59] python -c "import os ; print os.listdir('/Users/')" [13:00] dwt@dwcp ~ % python -c "import os ; print os.listdir('/Users/')" [13:00] ['.localized', 'dwt', 'Shared'] [13:00] I mean what can be the reason for this? [13:00] the path is as case-perfect as it can be [13:00] its also there [13:01] the file is readable [13:01] but still libreadline does not like it [13:01] at all. [13:01] maybe I should try write a c-program that does the same thing [13:01] and see if that fails [13:02] I am considering just rm-ing the history file [13:02] dwt_: at the least, I think you should file a bug against bzrtools [13:03] dwt_: well, try moving it out of the way first [13:03] I'l do [13:03] well, moving it out of the way [13:03] seems to allow me to invoke the shell again [13:03] (the bug should include the -Derror traceback, the python -c 'import readline; ... ' error, the ls -l command and the python -c 'print open(...)') [13:04] dwt_: hmm [13:04] dwt_: I am perplexed! [13:04] reinvoking the shell also works [13:04] with history support [13:04] hmm. [13:05] one interesting difference [13:05] the newly created file starts with this line: [13:05] _HiStOrY_V2_ [13:05] while the old one doesn't [13:06] dwt_: I wonder... have you upgraded Python recently? [13:06] not that I can think of [13:06] ha [13:06] prepending the _HiStOrY_V2_ to the old file [13:06] makes it load again [13:07] so that's the reason [13:07] that sounds like either corruption or version incompatibility [13:07] well, I can try to check my backups [13:07] but I guess that it was never there [13:07] I did update bzr recently [13:07] not python [13:08] dwt_: what about OS X? Had leopard for a while? [13:08] yep [13:08] the last update was from 10.5.4 to 10.5.5 [13:08] that might have caried a python update of sorts [13:09] I didn't however remember to read that in the release notes [13:09] or a libreadline update... [13:09] then again, apple is known to be sloppy at this stuff [13:09] yep [13:09] neither anything about libreadline [13:09] hmm. [13:09] so, should I file a bug still? === luke-jr is now known as luke-jr_ [13:10] dwt_: no, probably not. === luke-jr_ is now known as luke-jr [13:10] ok... [13:10] dwt_: I'll mention it to abentley when I next speak with him [13:10] super [13:10] (or he'll see his name in his IRC logs!) [13:10] I mean, its just a prepend of "_HiStOrY_V2_" [13:10] and everything works again. [13:10] (strange though) [13:11] ok, thanks a lot for your help [13:12] http://osdir.com/ml/lang.r.mac/2005-09/msg00031.html [13:12] we aren't the only ones :) [13:13] great [13:13] so they stumped it. [13:13] Ah well [13:13] on with the real hacking. :) [13:13] Thanks again [14:34] how do I change the email address associated with bzr whoami? [14:35] bzr whoami "Your Name " [14:37] thanks === lamont` is now known as lamont [15:15] hi everyone. I 've had recently a pretty weird behaviour from bzr, after a merge, one of the revision had been "skipped", I had it in the history but none of the modifications had been applied [15:15] can you see it in the output of 'bzr log'? [15:16] yes [15:16] perhaps a later revision undid it [15:17] it's what i thought, but none of the following revisions modified the files that should have bee reverted [15:17] ...that should have been reverted for that behaviour to be normal [15:19] any ideas? [15:19] have a look at the 'bzr blame affectedfile' output [16:04] hi all [16:04] who is actually working on bzr-git ? [16:30] asabil: you can see it in the log; the most active seems to be jelmer [16:31] clemente: oh ok thanks [16:31] jelmer: ping ? [16:37] asabil, hi [16:37] hi jelmer, could you please take a look at my bzr-git fixes branch ? [16:45] asabil, thanks, emrged [16:47] :) === fta_ is now known as fta [18:21] How can I tell Bundle Buggy that a patch has been superseded by another? [18:51] clemente: click "Superseded" [18:54] jelmer: I suppose I need an account for that [18:54] clemente: ah, could be [18:54] clemente: it will also mark patches as superseded if the new bundle includes a superset of revisions of the old one IIRC [18:57] jelmer: yes, but in my case it didn't work, because I took my own bundle, merged it in another branch, and commited it again. Only to have the same tree in other computer [18:58] maybe there are better ways to copy revisions which can preserve history [18:59] (Revision history, but whatever Bundle Buggy is tracking) === tretle_ is now known as tretle === fta_ is now known as fta [20:42] how can I get bzr log to show the files affected by each revision? [20:43] thrope_: bzr log -v [20:44] also if I do -r-5..-1 on a file that hasnt had a change in those revisions there isn't any output - what options should I use to get the last 5 revisions of that file [20:44] jelmer: right - sorry I don't know I missed that in the help [20:44] *how [20:46] thrope_: "bzr log -l 5 some_file"? [20:46] thrope_: --limit [20:47] oh right - thanks [20:49] trouble is when I use it with --forward I get the first 5 instead of the last 5 [20:50] makes sense to me [20:51] so how I can I get the last 5 revisions of a file in --forward order [20:52] I'm trying to find an alias for a screenfull of revisions with the most recent at the bottom [20:52] thrope_: you could do something like jam's shortlog plugin [20:53] what is that? [20:54] hmm, can't seem to find it right now [20:54] wasn't shortlog from uws? [20:54] jelmer: that's lastlog [20:54] ah [20:55] which also might be useful [20:55] thrope_: a custom logformatter [20:55] ah I tried googling but I think shortlog is too general a term [20:57] yeah, "bzr lastlog" is mine [20:58] trivial plugin but very useful for me [20:58] note that you can do roughly the same with aliases in ~/.bazaar/bazaar.conf [20:58] [ALIASES] last = log -r -10:-1 --forward --short # for example [20:58] only difference is that my plugin handles branches with < 10 revisions correctly [20:59] and that you can give the number on the command line (e.g. bzr lastlog 5) [20:59] * uws gone now