2008/10/05/#bzr.txt

[01:39] <unenough> what's the easiest way to create an archive of my repo? or at least the latest version of it?
[01:41] <jml> unenough: for what purpose?
[01:42] <unenough> backup
[02:00] <RAOF> unenough: bzr push sftp://mybackuphost/backup?
[05:20] <AfC> Um, er, help!
[05:21]  * RAOF waits for the rest :)
[05:21] <AfC> 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] <AfC> 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] <AfC> a) I didn't expect that
[05:22] <AfC> b) I don't really want it to do that.
[05:22] <AfC> So
[05:22]  * AfC wonders what to do.
[05:23] <RAOF> That's kinda wierd, yeah.  I dunno.
[05:25] <AfC> I mean, sure I could uninstall bzr-svn, but that's a bit much.
[05:27] <jml> AfC: ooh ooh, I know this
[05:27] <jml> gimme a sec
[05:28] <jml> AfC: bzr --builtin init .
[05:28]  * AfC just did `bzr init --no-plugins .` but that's ridiculous!
[05:28] <AfC> --builtin?
[05:28] <AfC> That's a new one on me.
[05:28] <jml> AfC: me too, I found out about it yesterday
[05:28] <jml> AfC: it's in 'help global-options'
[05:29] <AfC> jml: Ah. I was just searching help looking for that help page (and not finding it) :|
[05:35] <AfC> FUCKING HELL.
[05:36] <AfC> I just tried to `bzr add` a single file and then bzr-svn kicked off and tried to update its cache again.
[05:36] <AfC> This is not a bzr-svn branch, god damn it!
[05:37] <AfC> 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] <pygi> hi hi gour
[08:22] <gour> hello pygi
[08:37] <gour> 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] <gour> it looks bzr-svn cannot handle it (yet)
[09:02] <Quadduc> 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] <bob2> bzr whoami
[09:06] <Quadduc> bob2: Thanks. I thought I had already done that (using --branch), but it was reset when I re-created the branch.
[09:09] <Quadduc> However, the Bazaar plugin for Eclipse does not seem to respect whoami, and instead just uses my user name.
[10:07] <loswillios> hey guys
[10:08] <loswillios> 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] <loswillios> It is stalled after a while
[10:16] <Peng_> loswillios: Stalled? There's no network activity or anything? Are you sure? It might just be slow.
[10:17] <loswillios> Peng_: it says \ [----    ] bzr get http://bazaar.launchpad.net/~team-mms/mms/1.1.0 mms-1.1.0
[10:18] <loswillios> err, \ [---    ] Transferring 0/4 for a while and then there's nothing
[10:18] <bob2> the progress bar has stopped, but it is still downloading in the background
[10:18] <loswillios> nope, no progress bar at all
[10:19] <loswillios> just disappears and I have to kill the process
[10:19] <loswillios> not even ctrl-c works
[10:19] <loswillios> currently trying to reproduce on another machine which is faster
[10:19] <Odd_Bloke> loswillios: Try running with -Dhttp, that'll log all HTTP activity.
[10:19] <loswillios> but I left the command sitting for an hour or so
[10:19] <loswillios> ok
[10:20] <Odd_Bloke> loswillios: If it isn't printing to std{out,err}, then look in ~/.bzr.log
[10:21] <Odd_Bloke> (-EBASHISM)
[10:22]  * Odd_Bloke --> breakfast
[10:22] <loswillios> ok, currently running..
[10:23] <loswillios> is bzr get deprecated? I can't find it in the man page anymore
[10:23] <bob2> it has always just been an alias
[10:26] <loswillios> ah
[10:26] <Odd_Bloke> loswillios: It's probably listed in 'bzr help branch'.
[10:27] <Odd_Bloke> Right, actually to breakfast.
[11:22] <bob2> nice, trying to check it out oom'd my linode
[11:55] <clemente> 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] <jml> clemente: no, not really.
[12:22] <jml> clemente: the easiest thing to do is read the source for bzrlib/builtins.py
[12:22] <jml> clemente: although, for that thing specifically, you want something like WorkingTree.open('.').add('hello')
[12:23] <jml> clemente: there's also a hacking guide which might help
[12:25] <clemente> 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] <jml> clemente: cool.
[12:29] <dwt_> Hey guys, I've got a problem with "bzr shell" from bzrtools
[12:29] <dwt_> that is, it kills itself with this error message:
[12:29] <dwt_> bzr: ERROR: [Errno 22] Invalid argument
[12:29] <dwt_> which according to the debugger occurs here:
[12:30] <dwt_> 3 dwt@dwcp ...ktive Projekte/otr-messaging/Adium-bzr % BZR_PDB=1 bzr shell
[12:30] <dwt_> bzr: ERROR: [Errno 22] Invalid argument
[12:30] <dwt_> **** entering debugger
[12:30] <dwt_> > /Users/dwt/.bazaar/plugins/bzrtools/shell.py(105)__init__()
[12:30] <dwt_> -> readline.read_history_file(self.history_file)
[12:30] <dwt_> (Pdb) l
[12:30] <dwt_> 100  	        ensure_config_dir_exists()
[12:30] <dwt_> 101  	        self.history_file = osutils.pathjoin(config_dir(), 'shell-history')
[12:30] <dwt_> 102  	        readline.set_completer_delims(string.whitespace)
[12:30] <dwt_> 103  	        if os.access(self.history_file, os.R_OK) and \
[12:30] <dwt_> 104  	            os.path.isfile(self.history_file):
[12:30] <dwt_> 105  ->	            readline.read_history_file(self.history_file)
[12:30] <dwt_> 106  	        self.cwd = os.getcwd()
[12:30] <dwt_> 107  	
[12:30] <dwt_> 108  	    def write_history(self):
[12:30] <dwt_> 109  	        readline.write_history_file(self.history_file)
[12:30] <dwt_> 110  	
[12:30] <dwt_> Now I'm not sure what might cause this - self.history_file looks great
[12:30] <dwt_> the fiel is there and readable too
[12:30] <jml> dwt_: this happens immediately on startup?
[12:30] <dwt_> (as the tests on the two previous lines should assure)
[12:30] <dwt_> yes
[12:31] <jml> dwt_: which version of bzr & bzrtools?
[12:31] <dwt_> bzr: 1.7.1
[12:31] <dwt_> bzrtools 1.7 (the latest)
[12:31] <dwt_> (I just downloaded it)
[12:32] <jml> dwt_: can you run the command with -Derror and paste the output to paste.ubuntu.com?
[12:32] <dwt_> sure
[12:33] <dwt_> http://paste.ubuntu.com/54222/
[12:34] <jml> thanks
[12:34] <dwt_> 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] <jml> yeah.
[12:35] <jml> dwt_: it might be worth trying: import readline; readline.read_history_file('/whatever/the/path/is')
[12:35] <dwt_> (Pdb) pp readline.read_history_file.__doc__
[12:35] <dwt_> 'read_history_file([filename]) -> None\nLoad a readline history file.\nThe default filename is ~/.history.'
[12:35] <dwt_> This however suggests that it is called correctly, right?
[12:36] <jml> dwt_: the error is almost certainly coming from some lower level file operation
[12:36] <jml> the actual Python call is correct.
[12:37] <dwt_> I*'l try the readline call myself
[12:37] <jml> dwt_: thanks
[12:37] <dwt_> hm, when I try this in an ipython shell
[12:37] <dwt_> it works perfectly
[12:38] <dwt_> interestingly
[12:38] <dwt_> when I try it in a normal python shell
[12:38] <dwt_> it bombs
[12:38] <jml> dwt_: ok, one more thing to try before I blame high solar electromagnetism
[12:38] <jml> oohh
[12:38] <jml> dwt_: incidentally, what's the value of self.history_file?
[12:39] <dwt_> (Pdb) pp self.history_file
[12:39] <dwt_> '/Users/dwt/.bazaar/shell-history'
[12:40] <jml> dwt_: ok. let's try strace :)
[12:40] <dwt_> oh dang
[12:40] <dwt_> that would be dtrace for me
[12:40] <dwt_> (mac leopard)
[12:40] <dwt_> as to why ipython can do it
[12:40] <jml> dwt_: I should have guessed from Adium.
[12:40] <dwt_> its actually using it's own implementation of readline
[12:40] <dwt_> ¿
[12:42] <jml> 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] <dwt_> I'l try
[12:42] <dwt_> But you have to give me some minutes
[12:42] <dwt_> as I'm not at all up to speed with dtrace
[12:43] <jml> dwt_: with strace, it's just "strace bzr shell"
[12:43] <jml> my OS X-fu is rusty, I'm afraid.
[12:43] <dwt_> yeah, as far as I remember with dtrace
[12:43] <dwt_> it should be "sudo dtruss bzr  shell"
[12:43] <dwt_> I'm working on it. :)
[12:44] <dwt_> so when I try this:
[12:44] <dwt_> sudo dtruss python -c "import readline; readline.read_history_file('/Users/dwt/.bazaar/shell-history')"
[12:44] <dwt_> I'm getting tons of output
[12:44] <dwt_> Hope you can read it. :)
[12:44] <dwt_> http://paste.ubuntu.com/54225/
[12:45] <jml> are those the last lines?
[12:45] <dwt_> yep
[12:46] <dwt_> sigaction looks like it died from a signal, right?
[12:46] <jml> well, there's a syntax error there.
[12:47] <dwt_> damn
[12:47] <jml> what happens when you drop the 'dtruss'
[12:47] <dwt_> Just a second
[12:49] <dwt_> where exactly do you means I should drop the dtruss?
[12:49] <dwt_> as far as I can see I only have it in the right location
[12:51] <jml> I mean, run this: sudo python -c "import readline; readline.read_history_file('/Users/dwt/.bazaar/shell-history')"
[12:51] <dwt_> sudo python -c "import readline; readline.read_history_file('/Users/dwt/.bazaar/shell-history');"
[12:51] <dwt_> Traceback (most recent call last):
[12:51] <dwt_>   File "<string>", line 1, in <module>
[12:51] <dwt_> IOError: [Errno 2] No such file or directory
[12:51] <dwt_> Interesting
[12:51] <dwt_> Now where's that syntax error.
[12:52] <jml> so, you aren't getting the same output as when you run it as your user.
[12:52] <dwt_> yes
[12:52] <dwt_> interestingliy
[12:52] <jml> bummer.
[12:52] <dwt_> all the while the fine sits there just fine
[12:52] <dwt_> 127 dwt@dwcp ~ % ll /Users/dwt/.bazaar/shell-history
[12:52] <dwt_> -rw-------  1 dwt  dwt  2443  7 Aug 10:18 /Users/dwt/.bazaar/shell-history
[12:52] <jml> o.O
[12:53] <bob2> on HFS?
[12:53] <dwt_> yes
[12:55] <dwt_> do you guys get
[12:55] <dwt_> why on earth its a different error message from doing this directly
[12:55] <dwt_> vs. calling it via bzr shell?
[12:55] <dwt_> when I do it directly it's: IOError: [Errno 2] No such file or directory
[12:56] <jml> dwt_: even without the 'sudo'?
[12:56] <dwt_> 1 dwt@dwcp ~ % python -c "import readline; readline.read_history_file('/Users/dwt/.bazaar/shell-history');"Traceback (most recent call last):
[12:56] <dwt_>   File "<string>", line 1, in <module>
[12:56] <dwt_> IOError: [Errno 2] No such file or directory
[12:56] <dwt_> while with bzr I get
[12:56] <dwt_> 1 dwt@dwcp ~ % bzr shell
[12:56] <dwt_> bzr: ERROR: [Errno 22] Invalid argument
[12:56] <bob2> no idea, but case-insensitive filesystems make me suspicous
[12:57] <dwt_> Even though it seems readline is called exactly the same
[12:57] <jml> bob2: good call!
[12:57] <jml> dwt_: what does, python -c "open('/Users/dwt/.bazaar/shell-history')" say?
[12:57] <dwt_> nothing
[12:58] <jml> weird weird weird
[12:58] <dwt_> dwt@dwcp ~ % python -c "print open('/Users/dwt/.bazaar/shell-history')"
[12:58] <dwt_> <open file '/Users/dwt/.bazaar/shell-history', mode 'r' at 0x62458>
[12:58] <dwt_> dwt@dwcp ~ % python -c "print open('/Users/dwt/.bazaar/shell-history').read()"
[12:58] <dwt_> ls
[12:58] <dwt_> diff
[12:58] <dwt_> shell open AnimatedGrid.xcodeproj/
[12:58] <dwt_> help shell
[12:58] <dwt_> open AnimatedGrid.xcodeproj/
[12:58] <dwt_> exit
[12:58] <dwt_> diff
[12:58] <dwt_> ci  -m  "added a new provider and refactored the storage model to use a two dimensional array"
[12:58] <dwt_> push
[12:58] <dwt_> help push
[12:58] <dwt_> bzr info
[12:59] <dwt_> bzr push parent
[12:59] <dwt_> bzr st
[12:59] <dwt_> ls
[12:59] <dwt_> ls parent/
[12:59] <jml> yeah, we don't need it all :)
[12:59] <dwt_> ...
[12:59] <dwt_> I don't get this
[12:59] <bob2> python -c "import os ; print os.listdir('/Users/')"
[13:00] <dwt_> dwt@dwcp ~ % python -c "import os ; print os.listdir('/Users/')"
[13:00] <dwt_> ['.localized', 'dwt', 'Shared']
[13:00] <dwt_> I mean what can be the reason for this?
[13:00] <dwt_> the path is as case-perfect as it can be
[13:00] <dwt_> its also there
[13:01] <dwt_> the file is readable
[13:01] <dwt_> but still libreadline does not like it
[13:01] <dwt_> at all.
[13:01] <dwt_> maybe I should try write a c-program that does the same thing
[13:01] <dwt_> and see if that fails
[13:02] <dwt_> I am considering just rm-ing the history file
[13:02] <jml> dwt_: at the least, I think you should file a bug against bzrtools
[13:03] <jml> dwt_: well, try moving it out of the way first
[13:03] <dwt_> I'l do
[13:03] <dwt_> well, moving it out of the way
[13:03] <dwt_> seems to allow me to invoke the shell again
[13:03] <jml> (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] <jml> dwt_: hmm
[13:04] <jml> dwt_: I am perplexed!
[13:04] <dwt_> reinvoking the shell also works
[13:04] <dwt_> with history support
[13:04] <jml> hmm.
[13:05] <dwt_> one interesting difference
[13:05] <dwt_> the newly created file starts with this line:
[13:05] <dwt_> _HiStOrY_V2_
[13:05] <dwt_> while the old one doesn't
[13:06] <jml> dwt_: I wonder... have you upgraded Python recently?
[13:06] <dwt_> not that I can think of
[13:06] <dwt_> ha
[13:06] <dwt_> prepending the _HiStOrY_V2_ to the old file
[13:06] <dwt_> makes it load again
[13:07] <dwt_> so that's the reason
[13:07] <jml> that sounds like either corruption or version incompatibility
[13:07] <dwt_> well, I can try to check my backups
[13:07] <dwt_> but I guess that it was never there
[13:07] <dwt_> I did update bzr recently
[13:07] <dwt_> not python
[13:08] <jml> dwt_: what about OS X? Had leopard for a while?
[13:08] <dwt_> yep
[13:08] <dwt_> the last update was from 10.5.4 to 10.5.5
[13:08] <dwt_> that might have caried a python update of sorts
[13:09] <dwt_> I didn't however remember to read that in the release notes
[13:09] <jml> or a libreadline update...
[13:09] <dwt_> then again, apple is known to be sloppy at this stuff
[13:09] <dwt_> yep
[13:09] <dwt_> neither anything about libreadline
[13:09] <jml> hmm.
[13:09] <dwt_> so, should I file a bug still?
=== luke-jr is now known as luke-jr_
[13:10] <jml> dwt_: no, probably not.
=== luke-jr_ is now known as luke-jr
[13:10] <dwt_> ok...
[13:10] <jml> dwt_: I'll mention it to abentley when I next speak with him
[13:10] <dwt_> super
[13:10] <jml> (or he'll see his name in his IRC logs!)
[13:10] <dwt_> I mean, its just a prepend of "_HiStOrY_V2_"
[13:10] <dwt_> and everything works again.
[13:10] <dwt_> (strange though)
[13:11] <dwt_> ok, thanks a lot for your help
[13:12] <jml> http://osdir.com/ml/lang.r.mac/2005-09/msg00031.html
[13:12] <jml> we aren't the only ones :)
[13:13] <dwt_> great
[13:13] <dwt_> so they stumped it.
[13:13] <dwt_> Ah well
[13:13] <dwt_> on with the real hacking. :)
[13:13] <dwt_> Thanks again
[14:34] <tretle> how do I change the email address associated with bzr whoami?
[14:35] <bob2> bzr whoami "Your Name <your@example.org>"
[14:37] <tretle> thanks
=== lamont` is now known as lamont
[15:15] <aidos> 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] <bob2> can you see it in the output of 'bzr log'?
[15:16] <aidos> yes
[15:16] <bob2> perhaps a later revision undid it
[15:17] <aidos> it's what i thought, but none of the following revisions modified the files that should have bee reverted
[15:17] <aidos> ...that should have been reverted for that behaviour to be normal
[15:19] <aidos> any ideas?
[15:19] <bob2> have a look at the 'bzr blame affectedfile' output
[16:04] <asabil> hi all
[16:04] <asabil> who is actually working on bzr-git ?
[16:30] <clemente> asabil: you can see it in the log; the most active seems to be jelmer
[16:31] <asabil> clemente: oh ok thanks
[16:31] <asabil> jelmer: ping ?
[16:37] <jelmer> asabil, hi
[16:37] <asabil> hi jelmer, could you please take a look at my bzr-git fixes branch ?
[16:45] <jelmer> asabil, thanks, emrged
[16:47] <asabil> :)
=== fta_ is now known as fta
[18:21] <clemente> How can I tell Bundle Buggy that a patch has been superseded by another?
[18:51] <jelmer> clemente: click "Superseded"
[18:54] <clemente> jelmer: I suppose I need an account for that
[18:54] <jelmer> clemente: ah, could be
[18:54] <jelmer> clemente: it will also mark patches as superseded if the new bundle includes a superset of revisions of the old one IIRC
[18:57] <clemente> 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] <clemente> maybe there are better ways to copy revisions which can preserve history
[18:59] <clemente> (Revision history, but whatever Bundle Buggy is tracking)
=== tretle_ is now known as tretle
=== fta_ is now known as fta
[20:42] <thrope_> how can I get bzr log to show the files affected by each revision?
[20:43] <jelmer> thrope_: bzr log -v
[20:44] <thrope_> 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] <thrope_> jelmer: right - sorry I don't know I missed that in the help
[20:44] <thrope_> *how
[20:46] <Peng_> thrope_: "bzr log -l 5 some_file"?
[20:46] <LarstiQ> thrope_: --limit
[20:47] <thrope_> oh right - thanks
[20:49] <thrope_> trouble is when I use it with --forward I get the first 5 instead of the last 5
[20:50] <LarstiQ> makes sense to me
[20:51] <thrope_> so how I can I get the last 5 revisions of a file in --forward order
[20:52] <thrope_> I'm trying to find an alias for a screenfull of revisions with the most recent at the bottom
[20:52] <LarstiQ> thrope_: you could do something like jam's shortlog plugin
[20:53] <thrope_> what is that?
[20:54] <LarstiQ> hmm, can't seem to find it right now
[20:54] <jelmer> wasn't shortlog from uws?
[20:54] <LarstiQ> jelmer: that's lastlog
[20:54] <jelmer> ah
[20:55] <LarstiQ> which also might be useful
[20:55] <LarstiQ> thrope_: a custom logformatter
[20:55] <thrope_> ah I tried googling but I think shortlog is too general a term
[20:57] <uws> yeah, "bzr lastlog" is mine
[20:58] <uws> trivial plugin but very useful for me
[20:58] <uws> note that you can do roughly the same with aliases in ~/.bazaar/bazaar.conf
[20:58] <uws> [ALIASES]    last = log -r -10:-1 --forward --short          # for example
[20:58] <uws> only difference is that my plugin handles branches with < 10 revisions correctly
[20:59] <uws> and that you can give the number on the command line (e.g. bzr lastlog 5)
[20:59]  * uws gone now