[00:35]  * chandlerc[g] beats on bzr-svn
[00:35] <chandlerc[g]> errors in the middle of a branch... so weird
[00:35] <jelmer> chandlerc[g], ?
[00:35] <chandlerc[g]> setting it up on a clean box, all from development branches, and i'm getting "SSL negotiation failed: SSL error: parse tlsext" stuff
[00:35] <chandlerc[g]> looking on bugs now..
[00:36] <chandlerc[g]> jelmer: actually no bugs on this one... surprised
[00:36] <jelmer> chandlerc[g], does svn itself work ok?
[00:37] <chandlerc[g]> yep
[00:37] <chandlerc[g]> at least
[00:37] <chandlerc[g]> svn "ls"
[00:37] <chandlerc[g]> lemme do a full checkout
[02:52] <Peng_> beuno: FYI, your new theme branch of Loggerhead now has a conflict in NEWS with the trunk.
[09:12] <berto-> is there a way to get bzr log to show the commit's patch?
[09:13] <luks> no, there isn't
[12:03] <AnMaster> Does loggerhead have some irc channel?
[12:03] <bob2> not afaik, but the main devs are in here
[12:06] <AnMaster> ah... How do you use loggerhead with lighttpd?
[12:06] <AnMaster> I can only find instructions for apache
[12:14] <Jc2k> i think the main devs are in the NZ timezone so might be asleep right now
[12:21] <AnMaster> Jc2k, ah I will be asleep when they are up
[12:21] <AnMaster> As I live in Europe
[12:22] <bob2> proxying through lighttpd is probably simplest
[12:23] <AnMaster> bob2, ah
[12:23] <luks> isn't loggerhead now a WSGI application?
[12:23] <AnMaster> no clue, I hoped for fastcgi would work
[12:24] <bob2> not sure it is in trunk yet
[12:24] <AnMaster> anyway I'm not a python coder!
[12:24] <AnMaster> I'm just a bzr user who use lighttpd, and want a repo browser
[12:28] <Peng_> AnMaster: I'm just using mod_proxy.
[12:28] <Peng_> AnMaster: And Loggerhead's serve-branches.py.
[12:28] <AnMaster> Peng, hm...
[12:29] <AnMaster> Peng, is this last release of loggerhead? or do I need to get trunk?
[12:29] <AnMaster> I can't find "serve-branches.py" in the download (1.2.1, last it seems?)
[12:29] <Peng_> AnMaster: The trunk has some major improvements, and serve-branches.py is new in it.
[12:29] <AnMaster> Peng, so what does serve-branches do then, why do I want it?
[12:30] <Peng_> AnMaster: It's much simpler to run that start-loggerhead.py.
[12:30] <Peng_> than*
[12:30] <AnMaster> I see
[12:31] <AnMaster> I need to server multiple branches, from several projects, some are stored in shared repos
[12:31] <Peng_> AnMaster: apt-get install python-paste python-simpletal, edit serve-branches.py if you want to, then run it.
[12:31] <AnMaster> other are free standing ones
[12:31] <Peng_> AnMaster: Yeah, that's fine.
[12:31] <AnMaster> Peng, I don't use Debian
[12:31] <AnMaster> I use FreeBSD
[12:31] <Peng_> AnMaster: (They probably have to be under one directory.)
[12:31] <Peng_> AnMaster: Install Paste and SimpleTAL in some other way, then.
[12:32] <AnMaster> Peng, what are paste and simpletal for?
[12:33] <Peng_> AnMaster: Paste is a WSGI thingy, and SimpleTAL is a templating engine. Loggerhead uses both of them.
[12:33] <Peng_> (Well, the trunk does.)
[12:33] <AnMaster> ah
[12:33] <AnMaster> found both in ports
[12:34] <AnMaster> anyway I find it odd that in here everyone seem to assume you use Debian or Ubuntu
[12:34] <AnMaster> just wondering why
[12:35] <Peng_> I wasn't really assuming you did; just using it as an example.
[12:36] <AnMaster> well it isn't just you, both here and in #launchpad there is a certain level of "assume ubuntu"
[12:36] <luks> most people here really use debian or ubuntu, so I'd say it's normal assumtion
[12:37] <Peng_> AnMaster: You do know Bazaar, Launchpad, and Ubuntu are all owned by Canonical, right?
[12:38] <AnMaster> I know
[12:38] <AnMaster> but I like bzr even though I'm no ubuntu user
[12:38] <Peng_> :)
[12:38] <AnMaster> I use Gentoo Linux, Arch Linux and FreeBSD
[12:39] <AnMaster> all my servers run FreeBSD
[12:39] <AnMaster> socket.error: (48, 'Address already in use')
[12:39] <AnMaster> from the ./serve-branches.py
[12:39] <AnMaster> how do I tell it to use a different port hm?
[12:39] <AnMaster> $ ./serve-branches.py --help
[12:40] <AnMaster> gives same error
[12:40] <AnMaster> (not very userfriendly...)
[12:51] <Peng_> AnMaster: You should just edit serve-branches.py.
[12:51] <Peng_> AnMaster: It's designed to be simple and editable.
[12:51] <AnMaster> ah....
[12:52] <AnMaster> turbogears doesn't seem to install on FreeBSD :/
[12:52] <AnMaster> Peng, so I guess I just can't use loggerhead
[12:53] <Peng_> The Loggerhead trunk no longer uses TurboGears.
[12:53] <AnMaster> oh?
[12:53] <Peng_> They moved from TurboGears and Mako to Paste and SimpleTAL for simplicity and performance.
[12:53] <AnMaster> ah
[13:06] <Pilky> hey all, is there any single place that gives you a full run through of what it takes to make a repo specific post-commit hook?
[13:06] <AnMaster> Peng, is there any way to get it to write a pid file or so...
[13:07] <Pilky> I've been looking through the docs and I seem to be thrown all over the place
[13:07] <AnMaster> Peng, well I guess I'll just use daemontools then to run it. sigh
[13:08] <Pilky> first to the user reference which points to the user guide which points to another part of the user guide which points to the developer docs
[13:08] <Pilky> it's like you need 4 tabs open just to get even a basic understanding of how to write a hook
[13:09] <AnMaster> Pilky, I didn't even know bzr had commit hooks
[13:09] <AnMaster> beyond full plugins
[13:10] <Pilky> yeah, it does them as plugins, but you can set them up so they're called at various points
[13:10] <Pilky> it seems a hook in bzr is just a way to call a plugin after/before an action
[13:11] <Peng_> AnMaster: A PID file would probably be 4 lines in serve-branches.py.
[13:11] <Peng_> Wait, maybe 6. I dunno.
[13:12] <AnMaster> Peng, assume I know *no* python except that it lacks braces
[13:12]  * AnMaster is a C coder
[13:13] <Pilky> AnMaster: I'm in the same situation but I've just found this which looks good: http://www.diveintopython.org/
[13:14] <AnMaster> I don't really have time to learn python :P
[13:14] <Peng_> I think people have stopped recommending Dive into Python because of its age.
[13:14] <AnMaster> (maybe I should switch to darcs? I know *some* Haskell, not much but more than python. Not that I like darcs really...)
[13:15] <Peng_> AnMaster: What all does a PID file do? When the program starts, if it exists, error out; if it doesn't, create it? Then when it exits, delete it?
[13:15] <AnMaster> Peng, well it would lock on it
[13:15] <AnMaster> would be even better
[13:15] <AnMaster> open() with O_EXCL basically
[13:15] <Peng_> Oh.
[13:15] <AnMaster> to hold it as a lock file
[13:16] <AnMaster> that would be even more useful
[13:16] <Pilky> Peng_: what would you recommend? I'm just looking for something online that will help me learn a bit of python, I'm not expecting to be doing a huge amount
[13:16] <AnMaster> than just a pid file
[13:16] <AnMaster> Peng, any idea how that is done in python?
[13:18] <Peng_> Pilky: I'm not sure. I used Dive into Python. :P Maybe the official Python tutorial?
[13:18] <Pilky> heh
[13:18] <Pilky> well it seems to use Python 2.3 in here and I'm on 2.5 I've used older books to learn languages
[13:19] <Pilky> hell, my primary resource for Cocoa programming was a book from around 2004 up until about a month ago :P
 well it seems to use Python 2.3 in here and I'm on 2.5 I've used older books to learn languages <-- you can still learn C using K&R...
[13:23] <Peng_> Ehh.. I don't exactly know enough about PID files to want to hack it togehter.
[13:23] <AnMaster> you just would have to learn about prototypes later :P
[13:23] <Peng_> also, I'm a little tired and really lazy.
[13:23] <luks> AnMaster: but C didn't evolve much since K&R
[13:24] <luks> and K&R is the best book about C anyway :)
[13:24] <AnMaster> luks, oh yes a bit :P, for example variable length arrays in C99, or the _Complex type
[13:24] <Pilky> AnMaster: I did ;)
[13:24] <AnMaster> anyway I got no idea where to start with writing a pid file in python
[13:27] <nandersson> "Beginning Python" from WROX is a great book
[14:23] <lifeless> AnMaster: Peng_ uses loggerhead with lighttpd
[14:23] <lifeless> I think
[14:25] <AnMaster> lifeless, yes he helped me
[14:27] <AnMaster> lifeless, what about a pid file for it?
[14:27] <AnMaster> loggerhead that is
[14:27] <AnMaster> would be very useful
[14:44] <AnMaster> Peng, I can't get it to work
[14:44] <AnMaster> http://rage.kuonet.org/~anmaster/loggerhead
[14:44] <AnMaster> that results in "127.0.0.1 - - [26/Jul/2008:13:43:45 +0000] "GET /%7Eanmaster/loggerhead HTTP/1.0" 404 - "-" "Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.8.1.15) Gecko/20080708 Firefox/2.0.0.15""
[14:44] <AnMaster> :/
[14:45] <AnMaster> how do I make it not send the full path on
[14:45] <AnMaster> I can't use a specific subdomain
[14:45] <AnMaster> I need to use this directory
[14:49] <AnMaster> hm I asked in #lighttpd, lets see if I get any answer
[14:56] <AnMaster> lifeless, any idea?
[14:56] <AnMaster> using the serve-branches stuff
[14:58] <AnMaster> I guess no one knows :(
[15:08] <AnMaster> well it can't be solved on lighttpd side it seems
[15:18] <AnMaster> lifeless, Peng: tried webpath with start-loggerhead, doesn't work either
[15:18] <AnMaster> :(
[15:23] <jam> AnMaster: fn = os.open('path', os.O_EXCL | os.O_CREAT | os.O_TRUNC), something like that?
[15:23] <jam> os.write(fn, '%s\n' % os.getpid())
[15:23] <AnMaster> jam, that isn't the main issue now
[15:24] <AnMaster> the main issue here is that loggerhead refuses to work when a sub directory is forwarded
[15:24] <AnMaster> instead of a subdomain
[15:26] <jam> I don't know a lot about loggerhead, I would imagine it should be possible, though potentially require a bit of code tweaking
[15:26] <jam> I would have thought it would be the proxies job
[15:26] <AnMaster> jam, well the proxy is just mod_proxy of lighttpd
[15:26] <AnMaster> and it can't strip the leading bit it seems
[15:36] <jam> AnMaster: Also, in the copy of loggerhead I have here, it seems to create a pid file. (if that matters at all)
[15:37] <jam> Anyway, according to this page:
[15:37] <jam> http://discuss.joyent.com/viewtopic.php?id=3231
[15:37] <jam> You can only proxy by file-extension or by prefix
[15:37] <jam> not suffix
[15:38] <jam> http://trac.lighttpd.net/trac/wiki/Docs%3AModProxy
[15:38] <jam> So I'd have to agree with what you were finding
[15:43] <AnMaster> irc client crashed.... did I miss anything?
[15:45] <Peng_> AnMaster: Yes.
[15:45] <Peng_> AnMaster: Just a bit.
[15:46] <Peng_> AnMaster: Anyway, there's some middleware in Paste Deploy that can strip part of the path.
[15:49] <Peng_> AnMaster: Here's my serve-branches.py. It's got a horrible, complicated logging setup; the relevant part is the PrefixMiddleware part near the end. http://paste.pocoo.org/show/80377/
[15:49] <Peng_> AnMaster: You will need Paste Deploy installed, not just Paste
[15:49] <AnMaster> Peng, "paste deploy"?
[15:50] <AnMaster> ah... /usr/ports/www/py-pastedeploy
[15:52] <AnMaster> Peng, force port 80?
[15:52] <AnMaster> Peng, anyway I would prefer using the start-loggerhead because it can do pid file it seems
[15:52] <AnMaster> can I still use that?
[15:52] <AnMaster> Peng_, it can also handle the repos not being in same place (they aren't)
[15:55] <Peng_> AnMaster: I don't know what you would have to do with start-loggerhead.py; I've never used it.
[15:57] <jam> beuno: I have to say, setting up loggerhead here seems pretty easy (on win32 no less).
[15:58] <jam> The daemonize stuff doesn't work (obviously) but it was just a couple easy_installs, and tracking down simpletal
[15:58] <jam> oh, and you don't mention that you need "ConfigObj" installed as a dependency
[15:58] <jam> so Paste, ConfigObj, SimpleTAL, and things seem to work
[16:00] <jam> hmm.. the auto_publish_folder didn't seem to work right
[16:02] <jam> WAR [20080726-10:01:44.036] simpleTALES.Context: Exception occurred evaluating python path, exceptio
[16:02] <jam> n: type object 'branch' has no attribute 'url'
[16:12] <Odd_Bloke> Is LP running a recent Loggerhead, or is there a better place to point people who want to have a look at it in action?
[16:13] <jam> Odd_Bloke: I don't believe it is running the "latest"
[16:14] <jam> Odd_Bloke: This page is "shinier" http://bzr.mattnordhoff.com/loggerhead/imports/lighttpd/lighttpd-1.4.x/changes?remember=1117
[16:14] <jam> That is Peng's page
[16:15] <jam> I think LP *will* use the latest version, just doesn't yet
[16:15] <jam> beuno: I *am* trying to figure out the box with-an-arrow link
[16:15] <jam> It at least looks like it is redirecting you outside of the current domain
[16:15] <jam> when often it is *just* a link
[16:16] <jam> like, for *every* summary on the changes page
[16:18] <jam> anyway, off to another machine
[16:18] <Odd_Bloke> Peng_: Is pointing people to http://bzr.mattnordhoff.com/loggerhead/imports/lighttpd/lighttpd-1.4.x/changes?remember=1117 acceptable, or is it going to make your hardware break into floods of tears?
[16:19] <AmanicA> this logger head looks sweet
[16:20] <AmanicA> I like the way it expands when you click th triangle
[16:22] <AmanicA> and the mouse-overs rocks
[16:33] <AmanicA> can you guys get into bundlebuggytoday?
[16:33] <AmanicA> can you guys get into bundlebuggy today?
[16:35] <jelmer> no, it seems down
[16:35] <jelmer> and also not replying by email to my merge requests
[16:36] <AmanicA> cool so its not just me
[16:58] <Peng_> My Loggerhead runs the trunk, or other branches. The different theme comes from https://code.launchpad.net/~beuno/loggerhead/new_theme_trunk
[16:59] <Peng_> Odd_Bloke: Heh. I don't know how much traffic I can handle, but it should be fine, as long as it doesn't get on Slashdot or something.
[16:59] <tacone> hello, is there a way to enable diff's syntax highlighting for bzr diff ?
[17:01] <AnMaster> hey how do you check revisions signed with bzr sign-my-commits?
[17:04] <AnMaster> also what is the rich-root stuff good for? I can't find details on it (tried google)
[17:06] <tacone> mmh, found a way with vim: bzr diff | vim -u /usr/share/vim/vim71/macros/less.vim -
[17:09] <Peng_> AnMaster: Rich roots aren't really useful for anything yet. In the future, I think they'll be necessary for things like subtrees and splitting a subdirectory out into its own branch (which is possible now, I guess).
[17:14] <AnMaster> subtrees?
[17:14] <AnMaster> what are they?
[17:14] <AnMaster> Peng, I just use pack-0.92 format here
[17:15] <AnMaster> Peng_, also what about the checking of signatures?
[17:15] <AnMaster> signatures is one feature I would like to see expanded
[17:15] <AnMaster> along with adding support for cherrypicking
[17:16] <AnMaster> once those are in I will consider bzr perfect for my needs
[17:18] <Peng_> AnMaster: I dunno about verifying signatures.
[17:18] <AnMaster> and subtrees?
[17:18] <Peng_> Ehh. /me handwaves.
[17:18] <AnMaster> you mean like svn:external?
[17:18] <AnMaster> or?
[17:19] <Peng_> Yeah, basically.
[17:19] <AnMaster> ah
[17:19] <Peng_> Subtrees are still experimental (and have been for ages).
[17:19] <AnMaster> anyway what does "handwaves" mean in this context? I'm not a native English speaker
[17:21] <Peng_> I meant that I was avoiding trying to explain something, but I don't think I was using it entirely correctly.
[17:21] <Peng_> It's in the Jargon File: http://catb.org/jargon/html/H/handwave.html
[17:22] <AnMaster> are there any plans for cherry picking support in bzr?
[17:22] <AnMaster> like the advanced cherry picking of darcs
[17:25] <Peng_> Due to the model of Bazaar (and many other VCSes), cherrypicking is kind of difficult.
[17:25] <Peng_> I don't know if there are any current plans.
[17:25] <Jc2k> tacone: install bzrtools and then do bzr cdiff
[17:25] <AnMaster> what a pitty :/
[17:37] <tacone> sorry, how to apply a patch generated with bzr diff > ../my.patch ?
[17:37] <tacone> patch -p0 < ../my.patch doesn't seem to work
[17:38] <Odd_Bloke> tacone: Are you trying to apply the patch to a different bzr branch?
[17:38] <tacone> Odd_Bloke: to more recent revision of the same branch
[17:38] <jelmer> AnMaster, cherrypicking is already supported
[17:38] <tacone> I am quite sure conficts shuold not happen.
[17:39] <jelmer> AnMaster, tracking cherrypicks is not and support for that is planned for the future
[17:39] <Odd_Bloke> tacone: Try using 'bzr merge -c<revision>' instead.
 AnMaster, tracking cherrypicks is not and support for that is planned for the future <-- GREAT!
[17:39] <Ryan52> I saw somebody using a program like gitk on a bzr branch. What is that?
[17:39] <tacone> I don't have a revision of the changes to apply, only that .patch.
[17:40] <Jc2k> Ryan52: bzr-gtk perhaps?
[17:40] <Odd_Bloke> jelmer: ZOMG commit spam. :)
[17:40] <Ryan52> Jc2k: ah. thanks.
[17:40] <jelmer> Odd_Bloke: (-: Sorry about that
[17:40] <Odd_Bloke> jelmer: How's bzr-git coming along?
[17:41] <jelmer> Odd_Bloke, nicely
[17:41] <jelmer> Odd_Bloke: viewing history works
[17:41] <tacone> Odd_Bloke: I guess I got it. it worked for everything but a erased (with bzr rm file). the revision to be patched has it and it's more big than before.
[17:41] <Odd_Bloke> jelmer: Nice. :)
[17:41] <jelmer> now working on supporting "bzr branch <git-url>"
[17:42] <tacone> ok, solved
[17:42] <Odd_Bloke> jelmer: \o/
 now working on supporting "bzr branch <git-url>" <-- nice!
[17:51] <beuno> jam, oh, LH on windows...  that's unexpected...
[17:57] <beuno> jam, the box with the arrow was originally intended to let you know it was a link you could click on, but that may of derailed a bit
[17:57] <beuno> may be too confusing now...
[18:02] <beuno> cool, patches for LH!
[18:02] <beuno> nothing better than going to sleep and waking up with patches in your inbox
[18:42] <jelmer> beuno, fwiw I also filed some bugs to keep you busy ;-)
[18:42] <jelmer> I filed an ITP (cced to the list) for loggerhead as well but upload to NEW is not likely to happen soon
[18:44] <beuno> jelmer, heheh, I saw  :)   I'm going through the remaining patches, and RC-bugs  :)
[18:45] <beuno> jelmer, I saw, very cool, thanks.  When does the freeze disallow you from getting it into Lenny?
[18:45] <jelmer> beuno, I think the freeze has already happened
[18:46] <beuno> jelmer, ah, could be. I got dizzy after a few mails bounced back and forth
[18:46] <jelmer> yeah, full freeze is planned for mid july so I guess that's already past
[18:48] <beuno> heh, yeah, I just though the release would get delayed for at least a year  :)
[18:54] <beuno> Peng_, synced my branch with trunk, no more conflicts
[19:11] <Peng_> beuno: Do you pushed to your branch on LP? I don't see anything new.
[19:13] <beuno> Peng_, I'm stupid, I keep pushing somewhere else. I just pushed and added --remember  :)
[19:15] <rocky> don't suppose there's a quick way to get running with bzr 1.5 + bzr-svn on hardy heron ?
[19:15] <jelmer> rockstar: grab the bzr-svn from intrepid
[19:16] <rocky> jelmer: ah cool
[19:16] <rocky> it's funny... all my paying projects are stuck on svn and every several months i try using bzr to access those svn repos but somehow get turned off
[19:16] <rocky> don't suppose bzr-svn supports svn externals yet?
[19:17] <jelmer> rocky, no, because bzr doesn't have nested trees yet (which we'd have to map externals to)
[19:17] <rocky> gotcha
[19:17] <Peng_> beuno: Heh, OK.
[19:18] <fullermd> jelmer: 'minds me...   are stock svn 1.5.x bindings solid now, or are there still patches needed?
[19:18] <jelmer> fullermd: bzr-svn >= 0.4.11 works with svn 1.4 and 1.5 out of the box
[19:19] <Peng_> Oh yay, you resolved the conflict the same way I did.
[19:19] <fullermd> Awesome.
[19:21] <jelmer> beuno, loggerhead really seems to've come a long way since the original version
[19:24] <rocky> jelmer: there's no releases of bzr-svn 0.4.11 yet right?
[19:24] <jelmer> rocky, that's right, it's still in development
[19:25] <jelmer> rocky, 0.4.11 won't work with bzr 1.5 anyway (only with >= 1.6)
[19:25] <rocky> i can't seem to find a link for the download of bzr-svn intreprid ... i still find launchpad a bit hard to navigate
[19:25] <jelmer> rocky: see http://packages.ubuntu.com/bzr-svn
[19:25] <rocky> ah perfect, thanks
[19:28] <rocky> where's the best place to look to get a high-level list of functional changes between 1.4 and 1.5 of bzr ?
[19:30] <Peng_> Well, NEWS contains a long list, including many less-major things.
[19:30] <rocky> guess i'm more looking for a release announcement sorta deal
[19:30]  * rocky keeps googling
[19:32] <jelmer> I think NEWS is what you're looking for
[19:32] <jelmer> you may want to skip some of its sections though such as INTERNALS
[19:32] <rocky> is http://packages.ubuntu.com/bzr-svn resolving for anyone? can't seem to get to it
[19:33] <jelmer> looks like its down atm
[19:33] <LarstiQ> rocky: NEWS is (or was) also what gets verbatim mentioned in release announcements
[19:34] <zbrown> How does one apply a merge directive?
[19:35] <jelmer> zbrown: bzr merge <file> or bzr pull <file>
[19:35] <zbrown> ah ok
[19:35] <zbrown> jelmer: is there a way to apply a merge directive without actually making it a revision?
[19:36] <zbrown> that is, just flat apply the patch
[19:36] <jelmer> zbrown: Not sure why you'd want to do that
[19:36] <jelmer> "patch -p0 < FILENAME" should work though
[19:36] <zbrown> jelmer: patch was broken
[19:36] <zbrown> it needs completing
[19:36] <jelmer> zbrown: In that case, "bzr merge" to merge the merge directive
[19:36] <jelmer> then fix it followed by commit
[19:36] <zbrown> oh so merge won't commit?
[19:37] <jelmer> no, it'll just apply the changes, you have to commit it explicitly
[19:37] <zbrown> ah ok
[19:37] <jelmer> (and this sort of situation is exactly why it does't commit)
[19:37] <zbrown> ah :)
[19:37] <zbrown> I work mostly in git so still learning some of these things :)
[19:39] <beuno> jelmer, yeah, I thought it would take longer to get to where we are
[19:40] <rocky1> packages.ubuntu.com offline for everyone here? (sorry, was DC'd)
[19:40] <zbrown> rocky: seems that way.
[19:41] <beuno> jelmer, it's actually usable by most people now!
[19:41] <jelmer> beuno, :-)
[19:43] <Peng_> beuno: Shouldn't the SQL cache be deleted on shutdown?
[19:43] <Peng_> I think it used to usually be deleted, but it isn't anymore.
[19:44] <beuno> Peng_, ideally, no. It takes quite a few CPU cycles to generate it, not sure why we wouldn't want to take advantage of it already have been created
[19:45] <Peng_> Ideally, then, you wouldn't use mkdtemp to pick it. :P
[19:46] <rocky> jelmer: don't suppose there's a good link somewhere that i'm missing for svn users who haven't really used bzr on how to get going with bzr-svn ? or do you just refer people to standard bzr docs ?
[19:46] <Peng_> Hmm, I bet I could monkeypatch it to do that.
[19:46] <jelmer> rocky, see the FAQ included with bzr-svn for bzr-svn things
[19:46] <jelmer> rocky, for everything else, the regular bzr docs apply
[19:47] <beuno> Peng_, agreed. That's some of the left-overs ToDo's for serve-branches
[19:49] <AmanicA> rocky: If youre new to bzr you might want to checkout the following too: http://bazaar-vcs.org/BzrForSVNUsers
[19:53] <rocky> thanks
[19:53] <rocky> is there a simple mirror for packages.ubuntu.com that is up atm?
[19:58]  * Peng_ writes a stupid monkeypatch in serve-branches even though it would be easier to just edit loggerhead/apps/filesystem.py.
[20:00]  * beuno cheers Peng_ 
[20:01] <jelmer> rocky, you may be able to use the debian package
[20:01] <jelmer> rocky, http://packages.debian.org/bzr-svn
[20:01] <jelmer> woot, "bzr branch git-url" works \o/
[20:02] <Peng_> beuno: My monkeypatch can't stop the mkdtemp directory from being created, so I'm automatically deleting it. :D
[20:02] <jelmer> not at a workable speed though
[20:03] <beuno> Peng_, hehe, ok, we may work a bit on top of that, but it'll get us started  :)
[20:04] <rocky> is it just me or is there no bzr deb for hardy at:  https://launchpad.net/~bzr/+archive
[20:04] <rocky> do i need to get the intrepid one for that too?
[20:05] <rocky> um, actually, there is no intrepid one for that
[20:05] <beuno> rocky, there was a problem with the package, you can find them here: https://edge.launchpad.net/~bzr-beta-ppa/+archive
[20:05] <beuno> latest betas/rcs
[20:05]  * rocky needs a "yes rocky, you're stupid, so here are the links you need for bzr 1.5 on hardy and bzr-svn"  ;)
[20:05] <Peng_> Someone uploaded 1.6b3 for hardy (but no other series), then it was deleted, and 1.5 was never added back (I dunno if it's even possible). Right?
[20:06] <beuno> I don't think you can upload an older package, no. We'll probably have to wait til 1.6 is out
[20:06] <berto-> if i wanted to make a plugin to hook into the commit command, where would i learn how to do that?
[20:06] <rocky> ugh ... well if i upgrade to bzr 1.6beta3 (which does seem to have a deb for hardy) that means i'm out of luck with a bzr-svn plugin version to match up right?
[20:07] <ToyKeeper> Yay, just a page of shell to migrate svn tags to bzr.
[20:07] <beuno> berto-, http://doc.bazaar-vcs.org/bzr.dev/en/user-guide/index.html#using-hooks
[20:07] <beuno> is a good place to start
[20:07] <jelmer> ToyKeeper: bzr-svn 0.4.11 will also support importing tags natively
[20:07] <jelmer> ToyKeeper, fwiw
[20:07] <ToyKeeper> Yeah, but that's in the future.  :)
[20:07] <jelmer> rocky, you can run the development version of bzr-svn
[20:08] <Peng_> beuno: BTW, I wrote a one-line patch to call the sql_dir "/tmp/loggerhead-cache-XXXXXX" instead of "/tmp/tmpXXXXXX". Interested?
[20:08] <beuno> berto-, http://doc.bazaar-vcs.org/latest/en/user-reference/bzr_man.html#hooks is also good
[20:08] <ToyKeeper> jelmer: any idea if it'll handle svn branches too?  Those seem more difficult.
[20:08] <jelmer> ToyKeeper, heh, true
[20:08] <jelmer> ToyKeeper, older versions already handle svn branches
[20:08] <jelmer> ToyKeeper, e.g. bzr svn-import
[20:08] <beuno> Peng_, yes, that's an improvment over what we have now. Send it over to the list please [loggerhead/merge]
[20:09] <Peng_> beuno: OK
[20:09] <ToyKeeper> Ah, okay.  I only tried that with 'bzr branch /blah/svn/repo'
[20:09] <rocky> starting to wonder if it's worth all this or if i should just use hardy's native bzr 1.3 and whatever bzr-svn version comes with it
[20:09] <berto-> beuno: thanks!
[20:09] <jelmer> rocky, I would recommend just using 1.3 unless there are particular bugs you're hitting that are fixed in later releases
[20:10] <rocky> i don't think there are
[20:10] <rocky> just figured if i'm learning to use this system i should be using the latest
[20:26] <berto-> is there an easy way to get the parsed command line from a plugin?
[20:29] <Odd_Bloke> jelmer: I'm about to file an ITP on GitPython, have you looked at packaging it at all?
[20:30] <jelmer> Odd_Bloke, nope, not yet
[20:30] <jelmer> berto-, why would you need that?
[20:30] <Odd_Bloke> jelmer: Cool, I'll take a look at it.
[20:31] <berto-> jelmer: i'm looking into writing a pre-commit hook that checks for implicit commits.  i've grown accostomed to git where you either add something then commit, or explicitly specify the file[s] being committed.
[20:32] <berto-> jelmer: so, i wanted to check the args part of the command from a pre_commit_hook
[20:33] <jelmer> berto-: a pre_commit_hook can't modify what is being committed because of the moment in the commit process it's being called
[20:33] <berto-> jelmer: i dont' want to modify, just raise a BzrError if no file was specified.
[20:33] <berto-> it's really a pretty simple addition i want to make.
[20:33] <jelmer> berto-: start_commit_hook can modify what is being committed but it doesn't receive the list of files to commit atm
[20:33] <berto-> right now 90% of the time i commit, say doh, uncommit and then properly commit just the files i want.
[20:35] <berto-> jelmer: but i dont' want to modify, just reference, which is why i wanted to check the command line.  if the args part of options, args = optparse.parse_args([...]) is empty i'll just raise BzrError.
[20:35] <jelmer> berto-: well, sys.argv should contain the command line arguments
[20:35] <berto-> jelmer: right, but all the options as well.
[20:35] <berto-> i guess i can just exclude any that are prefixed with '-'.
[20:35] <jelmer> berto-: right, so that's why you'd want start_commit_hook
[20:36] <jelmer> but that needs to be modified to receive the list of specified files
[20:36] <berto-> start_commit_hook rather than pre_commit_hook ?
[20:36] <jelmer> yeah
[20:36] <berto-> is there a page that describes all the hook invocations; like, what is the difference between pre_ and start_ ?
[20:38] <jelmer> berto-: I'm not sure
[20:39] <jelmer> start_commit_hook is part of MutableTree
[20:39] <berto-> jelmer: hmm, start_commit_hook does not work, but pre_commit does.
[20:39] <jelmer> berto-: Doesn't work in what way?
[20:39] <berto-> when i try start_commit_hook, i get: Unable to load plugin 'explicitcommit' from '/Users/berto/.bazaar/plugins'
[20:40] <jelmer> berto-: Are you sure you have the right name?
[20:40] <berto-> i thought start_commit_hook was the right name.
[20:41] <jelmer> bzrlib.mutabletree.MutableTree.hooks['start_commit'] rather than bzrlib.branch.Branch.hooks['pre_commit']
[20:41] <berto-> jelmer: this works: http://dpaste.com/67613/
[20:41] <jelmer> berto-, you'd need mutabletree.MutableTree where you have branch.Branch at the moment
[20:41] <jelmer> and 'start_commit' where you have 'pre_commit'
[20:42] <jelmer> and "from bzrlib import mutabletree" rather than "from bzrlib import branch"
[20:42] <berto-> jelmer: hmm, alright.  sorry if you already explained this, but why is start_commit_hook better than pre_commit ?
[20:42] <jelmer> berto-: start_commit is run before the tree is generated that is going to be committed
[20:43] <berto-> jelmer: ah, cool. less processing is done.
[20:43] <jelmer> berto-: and pre_commit is not intended to receive the list of files to commit, it operates on a tree
[20:43] <jelmer> berto-: start_commit doesn't receive the list of specified files at the moment, but it will in the future
[20:44] <berto-> jelmer: got it.
[23:28] <bpeterson> lightweight checkouts don't store any of the revision data with them, correct?
[23:29] <jelmer> bpeterson: yep
[23:29] <bpeterson> so that's why just running bzr diff is so slow?
[23:31] <jelmer> yeah, if the branch you've got a checkout is remote somewhere it can be pretty slow
[23:32] <bpeterson> hmm, is this the same of stacked branches?
[23:33] <james_w> not quite
[23:33] <james_w> a stacked branch splits the storage of the revisions for a branch.
[23:34] <james_w> create a heavyweight checkout where the local branch is stacked at a depth of one revision. This would be like svn, with the exception that the local size will grow as you commit.
[23:34] <bpeterson> I don't want to carry the whole revision history around, but I would like bzr diff to be fast as a branch
[23:35] <bpeterson> good that seems nice
[23:35] <james_w> that last line should have started "what you could do is"
[23:35] <james_w> so you wouldn't have the whole of the revision history, but you would have some
[23:35] <bpeterson> I didn't think you could stack things at different revision levels
[23:36] <james_w> you could recreate it periodically to prevent the local storage growing without bound
[23:36] <bpeterson> well mostly I just don't want to download the whole history
[23:37] <james_w> a stacked branch provides a fallback when revisions can't be found locally.
[23:37] <james_w> so if you create an empty stacked branch then bind it to the parent and then either commit, or fetch the last revision, you get fast "bzr diff", but you have to hit the network for "bzr log", like svn.
[23:38] <bpeterson> that's ok
[23:38] <james_w> it's the sort of thing that would be good for a plugin once we have stacking
[23:39] <james_w> the only problem would be thinking up a name for that arrangement :-)
[23:40] <bpeterson> so it's not yet possible to ask Bazaar to pull a given number of revisions initially from the branch it's stacked on (i.e. 1000)
[23:40] <bpeterson> ?
[23:41] <bpeterson> for some reason I can't get Bazaar to stack on rich-root-pack repos
[23:42] <jelmer> bpeterson: I think there was a fix that went into bzr.dev recently to help with this
[23:42] <jelmer> what version of bzr are you running?
[23:42] <bpeterson> bzr.dev, the latest
[23:43] <jelmer> hmm, that's probably an unknown bug then..
[23:44] <bpeterson> perhaps it hasn't been merged?
[23:44] <jelmer> I'm pretty sure it's been merged since it broke bzr-svn :-)
[23:45]  * bpeterson goes to file a bug report
[23:45] <jelmer> bpeterson: thanks :-)
[23:50] <a7p> hi everyone
[23:50] <a7p> anybody using bzr under MacOS X? -- I've got a couple of problems and would like to know if other people share them, so I can file proper bug reports.
[23:51] <bpeterson> a7p: it works fine for me
[23:53] <a7p> bpeterson, ah, great, could you try to checkout lp:eclipse-bzr ... that one fails reproducible (works fine with under Linux).
[23:53] <a7p> bpeterson, and do you get the progress display?
[23:54] <bpeterson> it says: no such project eclipse-bzr
[23:54] <Peng_> a7p: Are you using the same version of bzr on both?
[23:55] <jelmer> a7p: did you log into launchpad?
[23:55] <bpeterson> I can't find this project eclipse-bzr
[23:56] <a7p> Peng, no, I am not, on Hardy I tryed 1.3.1 (which is default) and on macosX I tryed 1.5 (which is the Download-Package as well as the version distributed via macports)
[23:56] <Odd_Bloke> jelmer: python-git package at http://packages.daniel-watkins.co.uk/
[23:56] <bpeterson> jelmer: https://bugs.launchpad.net/bzr/+bug/252212
[23:56] <a7p> bpeterson, sorry ... bzr-eclipse is what I mean
[23:56] <jelmer> Odd_Bloke, w00t!
[23:57] <a7p> jelmer, it works fine with some other launchpad branches
[23:57] <jelmer> Odd_Bloke, in order for python-git to be usable for bzr-git, the patches in my git repo are necessary as well
[23:57] <a7p> my Macos installation is also much slower than the linux-ones
[23:57] <bpeterson> knit-repos are deprecated now
[23:57] <a7p> jelmer, and I registered my launchpad account with bzr.
[23:58] <a7p> bpeterson, okay, so it can't work due to the too new version I am using under macos ...
[23:58] <bpeterson> no
[23:58] <bpeterson> It's working so far for me