[00:19] <jelmer> re
[01:26] <ubotu> New bug: #184237 in bzr "autopacking should be optional" [Undecided,New] https://launchpad.net/bugs/184237
[01:49] <AfC> What do I have to do to get bzr to ask me to GPG sign when I commit? Doing the `bzr sign-my-commits` for two years of work after the fact seems a bit after-the-fact.
[02:23] <jelmer> AfC: Set create_signatures=always in ~/.bazaar/bazaar.conf
[02:23] <AfC> jelmer: ah, sweet. Thanks
[02:23] <jelmer> or in the appropriate branch-specific place
[02:24] <AfC> [I heard some moron complaining that bzr lacked this feature that made git superior, and I was like "hey, we have that in Bazaar, I'm sure" so I figured I'd dog food it to make sure.]
[07:46] <AfC> So if the process that has locked a bzr branch etc no longer exists, could bzr just automatically break lock?
[09:01] <Peng> AfC left, but I'll respond anyway. I don't know whether I agree or not, but if bzr will automatically break locks, it should print a warning message.
[10:12] <pygi> quick question folks
[10:12] <pygi> I'm trying to expose bzr repository over webdav to allow anonymous branching/checkouts, and I get this when I try it:
[10:12] <pygi> Unable to handle http code 401: Authorization Required
[10:12] <pygi> which isnt fun at all:)
[10:15] <bob2> do your server logs elaborate on why (e.g. permissions)?
[10:16] <pygi> bob2, nothing there
[10:16] <pygi> just for an idea, bit does work if I set .htpasswd file in there :)
[10:17]  * pygi should just use smart server
[10:18] <rjek> I always get that from bzr unless I provide a username in the URL.
[10:18] <pygi> yea, I know :p
[10:18] <pygi> but heh :p
[10:18] <rjek> ie, I have to do http://rjek@foo/.../ rather than just http://foo/.../, and it will correctly ask for the password.  A faff.
[10:18] <pygi> I dont want a password :p
[10:20] <fullermd> Well, good, 'cuz I don't want to give you one   :p
[10:21] <pygi> funny folk :P
[10:21] <pygi> anyway, will just use it as is, before I get some time to setup a smart server :)
[10:24] <rjek> Well, if it's saying 401, then that's because your webdav server is set up to disallow anonymous acecss.
[10:24] <rjek> So it sounds like your server's misconfigured.
[10:30] <fullermd> What's with webdav, anyway?
[10:31] <Peng> For anonymous branching, what's wrong with plain http?
[10:56] <ubotu> New bug: #184292 in bzr "bzr bundle does not include tags" [Undecided,New] https://launchpad.net/bugs/184292
[11:56]  * AfC tries to work out what needs to be backported to Subversion 1.4.6 for the bzr-svn plugin to work
[11:56] <jelmer> AfC: There's a patch on the bzr-svn wiki page
[11:56] <AfC> Ah
[11:57] <AfC> jelmer: http://samba.org/~metze/subversion-1.4.0-metze-python-bindings.patch ?
[11:58] <jelmer> yep
[12:06] <AfC> Patch appears to apply cleanly to 1.4.6 ... building now
[12:33] <AfC> Damn. Doesn't seem to have worked. :(
[12:35] <jelmer> what problems did you run into?
[12:36] <AfC> jelmer: kind of you to ask. The answer is "well, none, really."
[12:36] <AfC> I took the existing package, added the patch line, and let 'er rip.
[12:37] <AfC> I did one thing wrong and ended up having to build it a few times, but no matter
[12:37] <AfC> But at the end of it all, I still have bzr-svn reporting it can't do it's thing, see README etc
[12:38] <jelmer> AfC: Did you run ./autogen.sh --release ?
[12:38] <jelmer> (see the instructions on the wiki)
[12:38] <AfC> Python bindings got built, etc so I'm not sure what to poke
[12:38] <AfC> Does the package run autogen? I doubt it
[12:39] <jelmer> autogen.sh --release reruns swig
[12:39] <jelmer> see the instructions on the wiki
[12:39] <AfC> Shit.
[12:39] <AfC> I was hoping just patching the sources and building would work.
[13:00] <jelmer> we could include the changes to the files generated by swig in the patch
[13:00] <jelmer> but that would probably mean having to put up a patch for each minor release of subversion
[13:00] <jelmer> because that's much more likely to cause conflicts
[13:01] <jelmer> AfC: what platform are you on?
[13:01] <AfC> Gentoo
[13:01] <AfC> I've been getting advice on from  the Desktop devs, but I just followed the wiki's instruction to run autogen and it still didn't work
[13:02] <AfC> The ebuild for subversion has a long history, and it looks a bit fragile. They'll need to figure it out from here; I give up
[13:02] <jelmer> there is a gentoo overlay that contains the patched subversion
[13:03] <jelmer> it's on launchpad somewhere
[13:03] <jelmer> though I'm not a gentoo user, so I haven't tried it
[13:03] <AfC> Hm. I'll look for that, then. We don't use overlays as a matter of policy, though.
[13:04] <AfC> But the ebuild itself might be usable.
[13:04] <jelmer> https://launchpad.net/bzr-gentoo-overlay/
[13:05] <AfC> jelmer: thanks. I'll look at it tomorrow
[14:36] <zero-9376> hi can someone tell me if bzr-gtk will be available in gutsy soon? i have bzr 0.9 on my desktop and just upgraded my laptop and got bzr 1.0 but bzr-gtk was uninstalled in the process.
[14:43] <jelmer> zero-9376: You mean in the gutsy repository on bazaar-vcs.org ? afaik existing packages in released versions of ubuntu are not updated, except for major bug fixes
[14:45] <zero-9376> just looking in synaptic now version 0.90-1 is installed and 1.01 is availalbe (i have backports and proposed enabled)
[14:45] <zero-9376> i just wondered if anyone knew whether the gtk tools would be arriving as well
[14:58] <zero-9376> just added the bazaar-vcs repo and while bzrtools is available it still wants to remove bzr-gtk
[15:11] <jelmer> I'm not sure the bzr-gtk package at bazaar-vcs.org is kept up to date
[15:12] <jelmer> lifeless/poolie: ^
[15:12] <jelmer> this keeps coming up
[16:16] <ubotu> New bug: #184342 in bzr "Invalid url supplied to transport: "invalid port number  in url:" [Undecided,New] https://launchpad.net/bugs/184342
[17:13] <ryanakca> since you're removing the bzr bundle command, I guess its preferable not to use it ? https://lists.ubuntu.com/archives/bazaar/2007q3/029497.html
[18:00] <gustavonarea> Hello, everyone. At the organization I work for, we use subversion+trac in several projects and we're happy with both, but we're willing to give Bazaar a try. So I'd like to know if there's an *stable* alternative to Trac, compatible with Bazaar? I've only found dead projects
[18:03] <deepjoy> why switch from trac :-) http://hotdesign.com/seybold/everything.html
[18:03] <deepjoy> sorry wrong link
[18:03] <deepjoy> https://launchpad.net/trac-bzr
[18:06] <gustavonarea> deepjoy: thank you; I thought it was dead too, because of its website ( http://bazaar-vcs.org/ObsoletePlugins ). Is trac-bzr stable enough to be used in production environments?
[18:07] <gustavonarea> oh, sorry... here's its website:
[18:07] <gustavonarea> Trac Bazaar Plugin
[18:07] <gustavonarea> A plugin for Trac (>= 0.10) that provides a version control backend for Bzr
[18:07] <gustavonarea> As of version 0.10 Trac (http://projects.edgewall.com/trac/) does provide an abstract interface for version control systems. (http://projects.edgewall.com/trac/wiki/VersioningSystemBackend)
[18:07] <gustavonarea> This plugin allows the use of Bazaar-NG for projects.
[18:07] <gustavonarea> It's far from complete, and is based on Johan Rydberg's work for trac+darcs (http://progetti.arstecnica.it/trac+darcs/)
[18:08] <gustavonarea> Oh, excuse me... I pasted my whole clipboard
[18:09] <gustavonarea> I'm concerned about its stability because it'd serve this service: http://tracker.gnulinuxmatters.org/
[18:12] <ryanakca> is there a 1.1 of bzrtools ?
[18:13] <ryanakca> or ... would anybody know why bzr is still at 1.1~rc1-1 in Debian and 1.0 in Hardy, before I uupdate it?
[18:46] <Verterok> moin
[19:27] <pattern> is there a way to competely delete versioned files from a repository, so that all version history for that file is deleted from the depository as well ?
[19:29] <deepjoy> somebody asked this question and one of the contributor's replied that you can recreate the repository from scratch without that file with all other versions but cannot delete it from the current repository since that would compromise repository integrity
[19:29]  * deepjoy is an innocent bystander
[19:32] <pattern> ok, but if i recreate a repository without that file, how can i get the version history for all of the other files in to it?
[19:37] <deepjoy> by recreate I meant create patchsets and commit them 1 by 1 into the new repository
[19:37] <deepjoy> with all the information
[19:38] <pattern> ah
[19:38] <pattern> that sounds like fun
[19:39] <pattern> bazaar can't do that for me, then, i gather?
[19:39] <deepjoy> as I said I'm an innocent bystander
[19:40] <pattern> well, thanks for your help anyway, deepjoy
[19:40] <pattern> maybe someone else here can answer that last question
[19:45] <ubotu> New bug: #158883 in bzr-svn "ability to push non-lhs revisions" [Wishlist,Triaged] https://launchpad.net/bugs/158883
[20:17] <luislavena> hello everybody
[20:18] <Verterok> Hi luislavena
[20:20] <luislavena> wanted to know if bzr-svn still needs the patched version of svn bindings...
[20:20] <luislavena> I'm on Windows and subversion build process is still painful
[20:21] <Verterok> luislavena: I think there is a patched version of the python bindings for windows
[20:21] <luislavena> Verterok: yes, but they are badly built with VC8, which could show more problems than solutions.
[20:22] <luislavena> Verterok: know as cross-runtime linking (subversion links to MSVCRT, and the bindings link to MSVCR80).
[20:22] <luislavena> that affects malloc and free calls.
[20:23] <Verterok> luislavena: if that is the case, I think the only solution is to patch the python-bindings using MSVCRT
[20:24] <Verterok> but I don't know much about windows build process
[20:24]  * Verterok is a *nix guy
[20:25]  * luislavena now that being a windows-user make hima  second class citizen :)
[20:25] <luislavena> s/hima/him a
[20:26] <jelmer> luislavena: thre were updated bindings for Windows, but I think the link has gone dead
[20:26] <jelmer> luislavena: you may want to subscribe to bug 161082
[20:26] <ubotu> Launchpad bug 161082 in bzr-svn "very hard to get going on windows" [Medium,In progress] https://launchpad.net/bugs/161082
[20:26] <luislavena> jelmer: you mean the ones at here: http://bazaar-vcs.org/BzrForeignBranches/Subversion
[20:27] <luislavena> I just wonder (without a practical background) how Hg guys managed to use the standard bindings.
[20:27] <jelmer> luislavena: no, different ones, posted as comment to the bug I just mentioned
[20:27] <jelmer> luislavena: They don't use bindings at all afaik
[20:27] <jelmer> they just invoke the svn binary
[20:28] <Verterok> luislavena: I'm just telling you that my knowledge of windows and how it  build+link is practically 0 :)
[20:28] <luislavena> Verterok: maybe you can read a ongoing articles series of 'X for Windows'...
[20:29] <luislavena> Verterok: http://blog.mmediasys.com/2008/01/17/ruby-for-windows-part-1
[20:29] <luislavena> as commented on the bug report, ruby for windows faces the same situation (can blow up in your face) :P
[20:30] <Verterok> luislavena: thanks for the link
[20:47] <pmezard> jelmer: we use the bindings in convert extensions. what the problem with them ?
[20:49] <jelmer> pmezard: but you also invoke the svn binary directly, no?
[20:50] <pmezard> for the "push" part yes. the bindings are used to retrieve svn revisions not to write them.
[20:51] <jelmer> pmezard: There are a couple of bugs in the bindings that you don't hit that way
[20:51] <jelmer> pmezard: (luislavena was asking why hg didn't require certain patches to the bindings)
[20:52] <pmezard> jelmer: yes i understood that, and i was wondering why we would need such patches
[20:52] <jelmer> pmezard: There's a huge memory leak in the bindings
[20:52] <pmezard> jelmer: in the get_log part ?
[20:52] <pmezard> or in general ?
[20:52] <jelmer> not just there
[20:52] <jelmer> in general
[20:52] <jelmer> and there are several other issues
[20:53] <jelmer> you will need a patch in order to be able to do commits using the bindings if you don't want to have a local subversion working copy
[20:53] <pmezard> weird. is the svn api that complicated that swig bindings are hard to generate ?
[20:53] <pmezard> ok
[20:54] <jelmer> pmezard: The swig bindings are usually ok
[20:54] <jelmer> but there are a couple of places where they were autogenerated but never tested
[20:54] <jelmer> for example, where double pointers are used to return values but there's no swig typemap
[20:55] <jelmer> s/double pointers/double pointer arguments/
[20:55] <pmezard> interesting (i don't know swig, but it makes more sense why brian called the binaries for the push part). thanks.
[21:18] <Peaker> Hey, bzr 1.0-1~gutsy1's Ubuntu package seems to be incompatible with bzrtools/bzrgtk Ubuntu packages.. is that right?
[21:19] <jelmer> Peaker: Yep, the bzr-gtk package on bazaar-vcs.org hasn't been updated
[21:19] <jelmer> lifeless: ^
[21:20] <Peaker> oh, it seems a little ironic given that Ubuntu and bzr are both Canonical products :)
[22:29] <pattern> does anyone know if there are any automated tools that will delete files from a repository, and then rebuild the repository from scratch, and move all the version history from the old repository to the new repository minus the deleted files?
[22:36] <ubotu> New bug: #184457 in bzr-svn "bzr crash on commiting pending merges" [Undecided,New] https://launchpad.net/bugs/184457
[22:39] <jelmer> pattern: There's only something like that for removing complete revisions
[23:05] <pattern> ok
[23:05] <pattern> thanks, jelmer
[23:48] <welterde> hi