[01:42] <mwhudson> holy cow bzr-svn's "discovering revprop revisions" is slow
[02:31] <meoblast001> hi
[02:31] <meoblast001> i need some advice from you version control pros
[02:31] <meoblast001> if you guys like to be called that
[02:31] <meoblast001> i've been writing a program, it's about 20 commits in, written in C
[02:32] <meoblast001> i just finished moving it to C++, would you guys recommend starting a whole new repository or just committing it to the current
[02:33] <Peng> Are you, like, replacing the C version, or do they both exist at once?
[02:33] <Peng> And...do they share any code at all?
[02:33] <Peng> I mean, is it a rewrite from scratch, or a conversion?
[02:33] <meoblast001> i'm replacing the C version
[02:34] <meoblast001> it is a conversion, but i'm assuming >50% of the lines have been changed
[02:34] <meoblast001> probably >80%
[02:34] <meoblast001> regardless, they share the same concepts and design, just in C++ now
[02:34] <Peng> Um, I'm about to go to sleep, and I'm honestly not sure.
[02:35] <Peng> I could go either way. I'd probably keep it in the same branch.
[02:35] <spiv> Without thinking about it hard, a) I'd probably just keep using the same repo (after all, why not?), and b) I doubt it matters much anyway.
[02:35] <meoblast001> i know this will bloat the repository size a lot
[02:35] <meoblast001> but history is good
[02:35] <Peng> Yes, history is good.
[02:35] <Peng> meoblast001: And unless this is a Linux kernel-sized project, relatively large history is still not very much.
[02:36] <Peng> I mean, who cares if it's 120 KB instead of 70 KB?
[02:36] <meoblast001> yeah
[02:36] <spiv> I say point b) because 20 commits worth of history isn't really going to be that large in the long term so it's not going to bloat much at all, but at the same time hardly anyone is going to care about the first 20 commits in the long term either.
[02:37] <spiv> Hence why I default a), because that's the path of least resistance :)
[02:38] <spiv> (It'd be neat to have some way to measure how often old revisions *are* directly accessed...)
[02:41] <meoblast001> spiv: since you already know what i'm doing, i have another question for you
[02:41] <meoblast001> how would you word this commit
[02:41] <meoblast001> "Ported to C++"?
[02:41] <meoblast001> i can't think of words that can discribe this
[02:44] <spiv> meoblast001: sure.  Or "rewrote", or whatever you feel describes the change adequately.
[02:44] <meoblast001> spiv: is "port" an appropriate word for changing languages
[02:44] <meoblast001> and i didn't completely "rewrite" it
[02:44] <spiv> Appropriate enough.
[02:45] <spiv> I don't think commit messages are something to sweat over, especially not in the early days of a project.
[02:45] <spiv> You want something that will make some sense to you (or another person) that is looking back at 'bzr log', etc, for some reason.
[02:46] <spiv> Or at least enough sense that they can say "oh, ok, that's not the revision I'm looking for".
[02:47] <meoblast001> yeah, i'm sort of a perfectionist :P
[04:15] <meoblast001> oh, one of you told me once that you had a hook that a BZR server runs to have an IRC bot alert about commits
[04:15] <meoblast001> which of you was it
[04:15] <meoblast001> i lost the file
[04:15] <meoblast001> if i'm not here when whoever it was sees this, feel free to use memoserv
[07:26]  * shyam trying to get a copy of color-theme project from savannah.. loggerhead seems down at savannah..:(
[07:26] <shyam> bzr branch http://bzr.savannah.nongnu.org/r/color-theme says "http://bzr.savannah.nongnu.org/r/color-theme/ is permanently redirected to http+urllib://bzr.savannah.gnu.org/r/color-theme/ \n bzr: ERROR: Not a branch: "http+urllib://bzr.savannah.gnu.org/r/color-theme/"
[07:27] <shyam> what could be the problem?
[07:30] <shyam> i tried bzr branch sftp://username@bzr.savannah.nongnu.org/srv/bzr/color-theme where username is my savannah username though am not a member of color-theme project. that too said bzr: ERROR: Not a branch: "sftp://swathanthran@bzr.savannah.nongnu.org/srv/bzr/color-theme/".
[07:35] <spiv> shyam: add 'trunk' to the URL
[07:35] <spiv> shyam: if you look at it in a web browser there's no .bzr directory there, but there's a trunk directory which does have a .bzr directory
[07:36] <shyam> uh! thats the last one thing i didn't do! visit the url in a browser!
[07:37] <shyam> spiv: thanks!
[07:37] <spiv> I just tested, and http://bzr.savannah.nongnu.org/r/color-theme/trunk/ works for me
[07:37] <shyam> yeah, now it works here too
[07:39] <shyam> is it usual to have a trunk subdirectory?
[07:39] <shyam> it looks like for some other projects too..
[07:40] <spiv> It's usual to have, or at least allow for, multiple branches for a project.
[07:40] <spiv> So the structure of .../proj/branch is pretty common, and 'trunk' is a common name for the primary branch.
[07:41] <shyam> oh ok.. but other projects do have a .bzr at bzr.savannah.nongnu.org/r/project-name/
[07:42] <spiv> Do they have other directories too?
[07:42] <shyam> yeah
[07:43] <spiv> A .bzr directory doesn't necessarily mean there's a branch there, it can also be a "shared repository" for the branches inside that directory.
[07:45] <spiv> (And the upcoming bzr 2.1 release should give a slightly more informative error if you do attempt to branch from a location that iss a repository and not a branch)
[07:46] <shyam> yeah i saw a group of bug lists while scroogling for it..
[08:01] <shyam> spiv: what is missing here? how to make it work on bzr.savannah.nongnu.org/r/color-theme ?
[10:15] <chx> heya.
[14:11] <gerard_> hey
[14:18] <gerard_> how about I just add a "text" parameter to assertFileEqual?
[14:19] <gerard_> I really don't care for \r\n most of the time
[14:23] <gerard_> I guess I'll leave it until I run into problems on windows
[14:23] <gerard_> #&^$*&^@# line endings
[14:23] <gerard_> ;)
[15:06] <bialix> gerard_: usually in test suite files created with binary flag and LF endings
[15:06] <gerard_> bialix: oh good
[15:32] <alefteris> hi all! I did a bzr pull a file didn't change because of wrong permissions, now I fixed the permissions, but can't get the file up to date, tryed bzr pull and bzr revert without luck
[15:33] <alefteris> nevermint, bzr update did the trick
[16:24] <gerard_> mmm, test driven development
[16:24]  * gerard_ likes it
[17:45] <justdave> if I have a branched checkout and I move an existing tag, is there a way to get that tag move to go upstream with the push?  It pushes all my revisions, but gives me an error that there's a tag conflict instead of pushing the tag change.
[17:46] <jpds> justdave: bzr push --overwrite ?
[17:47] <justdave> that works, thanks
[17:48] <justdave> if I try to push and there's been other revisions upstream since I last pulled, will it error that I'm out of sync, attempt to merge, or overwrite the other revisions?
[17:48] <Peng> "bzr push --overwrite" overwrites.
[17:48] <justdave> yeah, that's why I was asking.  :)  So I want to be careful not to use that unless I know I have an otherwise up-to-date copy. :)
[17:49] <Peng> Yeah. I'm not sure if there's a (simple) way to just overwrite tags while preserving revisions.
[18:44] <RenatoSilva> if I merge 5 revisions, and use --forget-merges, will I get 5 new revisions in the main level or will them all be compacted as 1 single revision?
[18:47] <RenatoSilva> RenatoSilva: the latter
[18:47] <RenatoSilva> RenatoSilva: oh thanks
[19:59] <lifeless> justdave: Peng: there isn't a 'push tags only' UI at the moment.
[19:59] <Peng> Other than "scp -p .bzr/branch/tags". O:)
[20:00] <mkanat> I wonder if scp works when the shell is bzr serve.
[20:00] <mkanat> Probably does, I suppose.
[20:00] <mkanat> There's always hitchhiker, if not.
[20:01] <lifeless> scp won't
[20:01] <lifeless> sftp will
[20:01] <lifeless> (sftp is a subsystem)
[20:02] <lifeless> although, its possible scp uses the sftp subsystem but I don't think it does
[20:02] <lifeless> besides which, bzr won't work if bzr serve is the shell
[20:36] <chx> heya
[20:45] <gerard_> ahhh
[20:46] <gerard_> I FINALLY understand the inner workings of bzr update
[20:46] <gerard_> it only took two weeks
[20:46] <gerard_> :p
[20:53] <slestak> hi guys, I am trying to use the push_and_update plugin for a new repo based on the BazaarForWebDevs use case.
[20:53] <slestak> i am running into the same issue as Sean from the bzr answers page
[20:54] <slestak> i have don the initial import and update, however, new pushes are failing
[20:54] <slestak> i have tried with --use-existing-dir but that didnt help
[20:57] <slestak> new pushes say: bzr: ERROR: [Error 2] The system cannot find the file specified
[20:58] <slestak> i have tried with the cached path, i have tried specifying the path with and without trailing slashes
[20:58] <slestak> i have also tried with the plugin installed only locally and also with the plugin installed locally and remotely
[21:03] <chx> oh nice, you have a parent_location in branch/branch.conf so nothing mandates a shared repo to be kept under one directory. Nice!
[21:05] <chx> is it enough to edit just that one file and then I can move the directory freely?
[22:17] <justdave> 15:02:16 < lifeless> besides which, bzr won't work if bzr serve is the shell
[22:17] <justdave> ours is set up that way and it does.
[22:17] <justdave> although technically, the shell is a python stub that launches bzr serve
[22:18] <justdave> so whether scp would work would depend on that python stub I suppose
[22:19] <justdave> it's run via ForceCommand in sshd_config, and if you're root it runs bash, and any other user gets bzr serve
[23:26] <RyNy_> Hi, I issued a commit command after adding a directory and now Bazaar can't access the lock file for other commands.
[23:26] <RyNy_> Get this error: Unable to obtain lock file:///.bzr/branch/lock
[23:26] <RyNy_> Any ideas/help? The directory that I added was on LAMP and holds drupal.
[23:27] <RyNy_> I'm using Terminal on my Mac to work with Bazaar on my server.
[23:27] <RyNy_> The drupal folder that I added is 30MB in size.
[23:29] <Peng> The branch is locked? Is bzr still running?
[23:32] <RyNy_> I think so. I just broke the lock and tried the same commit and it seems to be hanging again.
[23:33] <Peng> That'd explain it, yes. :P
[23:34] <RyNy_> Do you know what's going on? Why is it hanging?
[23:34] <RyNy_> This is the error I get:
[23:34] <RyNy_> bzr commit -m "Added a bunch of modules"
[23:34] <RyNy_> Committing to: /
[23:34] <RyNy_> added var
[23:34] <Peng> That's not an error. :P
[23:35] <RyNy_> Is it just processing?
[23:35] <RyNy_> How long does it take to add a directory that's 30mb in size?
[23:35] <Peng> Dunno. Not that long.
[23:35] <RyNy_> Before I broke the last lock it was running for 2 hours.
[23:36] <Peng> Yeah, not that long...
[23:36] <RyNy_> So, something weird is happening?
[23:36] <Peng> Yeah. No idea what, though. Maybe an strace would help? Or various -D options?
[23:37] <RyNy_> don't know those commands ...
[23:37] <RyNy_> idea: can having 2 .bzrignore files screw things up?
[23:37] <Peng> Bazaar only cares about /.bzrignore. I imagine it completely ignores others.
[23:38] <Peng> RyNy_: Does .bzr.log contain anything interesting? Sorry, I don't know much about debugging this.
[23:38] <Peng> RyNy_: What version of bzr?
[23:39] <RyNy_> Bazaar 2.0.4
[23:42] <RyNy_> This is the most recent entry in the log
[23:42] <RyNy_> Sat 2010-02-06 23:39:22 +0000
[23:42] <RyNy_> 0.080  bzr arguments: [u'version']
[23:42] <RyNy_> 0.114  looking for plugins in /root/.bazaar/plugins
[23:42] <RyNy_> 0.230  looking for plugins in /usr/lib/python2.6/dist-packages/bzrlib/plugins
[23:42] <RyNy_> 0.287  encoding stdout as sys.stdout encoding 'UTF-8'
[23:42] <RyNy_> 0.425  opening working tree '/'
[23:42] <RyNy_> 0.472  Traceback (most recent call last):
[23:42] <RyNy_>   File "/usr/lib/python2.6/dist-packages/bzrlib/commands.py", line 842, in exception_to_return_code
[23:42] <RyNy_>     return the_callable(*args, **kwargs)
[23:42] <RyNy_>   File "/usr/lib/python2.6/dist-packages/bzrlib/commands.py", line 1037, in run_bzr
[23:42] <RyNy_>     ret = run(*run_argv)
[23:42] <RyNy_>   File "/usr/lib/python2.6/dist-packages/bzrlib/commands.py", line 654, in run_argv_aliases
[23:42] <RyNy_>     return self.run(**all_cmd_args)
[23:42] <RyNy_>   File "/usr/lib/python2.6/dist-packages/bzrlib/commands.py", line 1052, in ignore_pipe
[23:42] <RyNy_>     result = func(*args, **kwargs)
[23:42] <RyNy_>   File "/usr/lib/python2.6/dist-packages/bzrlib/builtins.py", line 3510, in run
[23:42] <RyNy_>     show_version(to_file=self.outf)
[23:42] <RyNy_>   File "/usr/lib/python2.6/dist-packages/bzrlib/version.py", line 42, in show_version
[23:42] <RyNy_>     src_revision_id = src_tree.last_revision()
[23:42] <RyNy_>   File "/usr/lib/python2.6/dist-packages/bzrlib/workingtree.py", line 1789, in last_revision
[23:42] <Peng> Eek
[23:42] <RyNy_> sorry
[23:43] <Peng> That's just from running "bzr version" (and apparently it failed somehow).
[23:43] <RyNy_> Sorry, don't understand
[23:49] <Peng> That log snippet isn't about the command that failed. It's just from running "bzr version".
[23:54] <RyNy_> thanks, running commit again. It's hanging again. Here's the log. Look fine:
[23:54] <RyNy_> 0.474  preparing to commit
[23:54] <RyNy_> [32216] 2010-02-06 23:51:14.878 INFO: Committing to: /
[23:54] <RyNy_> 0.482  Selecting files for commit with filter None
[23:54] <RyNy_> [32216] 2010-02-06 23:51:14.943 INFO: added var
[23:54] <RyNy_> Is there a way to "unadd" something from Bazaar
[23:55] <Peng> RyNy_: bzr rm (which has an argument to not delete stuff), or bzr revert if it hasn't been committed yet.
[23:58] <RyNy_> Thanks, so to remove a directory "var"
[23:58] <RyNy_> bzr rm var
[23:58] <Peng> Uh-huh. Or "bzr revert var" if it hasn't been committed yet.