[00:27] <cr3> spiv: dude, the upgrade has been going on for over an hour, might something be blocked?
[00:39] <mtaylor> lifeless: (or anybody) if I push a format 1.6 branch to launchpad - how do I tell it not to stack on something automatically
[00:39] <mtaylor> ?
[02:02] <spiv> cr3: probably it's just how long it takes to download all the data, upgrade it, and then upload it again.
[02:02] <spiv> cr3: we have some work to do there...
[02:15] <mtaylor> spiv: you guys need a launchpad pad button for the branches that lets people do things like "upgrade this branch to X format"
[02:15] <mtaylor> possibly save on some bandwidth
[02:20] <spiv> Yeah.  The launchpad-bazaar guys are working towards that.
[02:20] <spiv> It's currently hard for them to add a button to the web UI to affect the actual branches on disk, but they're fixing that.
[02:21] <spiv> But also, bzr itself should allow "bzr upgrade bzr+ssh://..." to work basically entirely server-side.
[02:24] <cr3> spiv: it's been a couple hours, I interrupted the process and I'll try to start it again from a machine that'll stay on
[02:32] <spiv> cr3: you may find it better to do the upgrade entirely locally, delete the remote .bzr with a tool like lftp, and then push.
[02:33] <cr3> spiv: can I actually delete a branch myself using something like lftp? I wouldn't mind reinitializing the whole branch
[02:33] <spiv> cr3: well, you can delete the branch entirely in the web UI now.
[02:34] <cr3> spiv: I'm looking, can't find how
[02:34] <spiv> cr3: but lftp will let you delete the .bzr directory recursively via sftp, and then push can just re-use that directory (--use-existing-dir)
[02:34] <spiv> cr3: there's a little red circle next to the branch name on the branch page, I think.
[02:35] <cr3> spiv: what if there's a pending backup.bzr directory on the server? will that be removed too?
[02:36] <spiv> if you delete through the web UI, it's all gone.
[02:39] <cr3> spiv: cheers! all done now and I didn't even have to ping tmhaddon :)
[03:40] <Leefmc> Question: I have a project where the client will be public OS, and the server will not be. I'll be using Bzr for each. The Client will be hosted on Launchpad, and the Server locally. With this said, should i create 2 repositories? Or would the normal method be to have each be in the same repo. ?
[03:41] <RAOF> It doesn't _really_ matter; repositories are basically an optimisation detail.
[03:41] <Leefmc> Ok
[03:42] <Leefmc> Just out of curiosity, what would most people do for something like that? Is the normal idea to keep only code forks/branches in repositories? And truely different code (ie, server and client) in different repositories?
[03:44] <RAOF> Not totally the same, but I keep gnome-do core and gnome-do-plugins in the same repository.
[03:45] <Leefmc> k thanks
[03:48] <markh> jam: ping
[03:57] <sohail> hey, howcome bzr log . only shows one entry (addition of directory)?
[09:22] <tca> Trying to work around bug #248540. From inside a shared repository, will "bzr checkout -r 898 solfege.dev 3.12 && cd 3.12 && bzr unbind" create a branch as good as just "bzr branch -r 898 solfege.dev 3.12"?
[09:28] <fullermd> I expect so.  But if the latter fails, I'd expect the former to as well...
[09:33] <tca> I would too, but it does not fail. Do you think a work-around for the ghost revision problem will be added to bzr?
[09:36]  * fullermd shrugs.
[09:36] <fullermd> Acting "as sensibly as possible" in the presence of ghosts is an explicit aim, so I'd assume so.
[10:17] <tca> It is safe to use the workaround in bug #248540 until it get resolved properly? Will doing so make the ghost situation in my repo worse?
[10:18] <LarstiQ> due to a failing firefox I can't browse atm
[10:18] <LarstiQ> tca: what is the workaround?
[10:23] <tca> The patch adding a try-except clause: https://bugs.launchpad.net/bzr/+bug/248540/comments/2
[10:28]  * LarstiQ still doesn't have a working browser
[10:29]  * fullermd imagines "bzr branch lpbug:[...]"    8-}
[10:39] <LarstiQ> tca: hmmm. Reading the comment James pasted it seems it should work
[12:22] <tca> LarstiQ: I remembered wrong, the try-except clause fixed some other problem some monts ago. But changing line bzrlib/repository.py:1652 from "parents = graph.get_parent_map([next_id])[next_id]" to "parents = graph.get_parent_map([next_id]).get(next_id, [])" works. Sorry for the confusion.
[12:23] <LarstiQ> ehm, parents being an empty list doesn't seem correct to me, without reading the code?
[13:04] <tca> From the line below, you see that parents can be an empty list. If it is empty, iter_reverse_revision_history returns None. But I don't claim to have any understanding of the code, thats why I was hoping someone that could fix #248540.
[13:43] <jollyr0ger> hi to all
[13:44] <jollyr0ger> there's some one that can tell me how to update a branch that I have on my desktop?
[13:45] <jollyr0ger> for example my copy is updated at rev 40, and on the serve (launchpad) is updated to rev 50. how can I update my copy to the last aviable?
[13:45] <jollyr0ger> thanks
[14:33] <Peng_> jollyr0ger: "bzr pull" if it's a branch, "bzr update" if it's a checkout.
[14:33] <Peng_> jollyr0ger: You should read the user guide or whatever.
[16:35] <duckx> hy
[16:51] <Leefmc> Question: I believe i pushed a branch into an actual repository.. how can i remove it?
[16:52] <Leefmc> (ie, i have a repository of repositories. init-repo. Normally i push into that directory, so bzr push bzr+ssh://user@host/repository/my_trunk_code. But instead of pushing into a subdirectory inside user@host/repository, i pushed directly into the repository directory.
[17:05] <Peng_> Does bzr officially support Python 2.6?
[18:20] <mkanat> Do you guys know of a bzr->cvs mirror, so that if we switch to bzr, people who checked out from our old cvs servers can still "cvs up"?
[18:56] <nosklo> I want to commit using bzr-svn, and authentication is not working, how can I make it work?
[18:59] <nosklo> hmm I did it using svn+http syntax... I got a warning telling that it is deprecated, but it is how I got authentication to work on svn. Is that right? Any other ideas/workarounds? is this a bug?
[20:27] <nosklo> hmm I did it using svn+http syntax... I got a warning telling that it is deprecated, but it is how I got authentication to work on svn. Is that right? Any other ideas/workarounds? is this a bug?
[20:29] <LarstiQ> nosklo: did you try http://user@.. ?
[20:32] <nosklo> LarstiQ, good idea, I tried user@host and it works, it seems to trigger svn... thank you. Any way to save password? user:passwd@host won't do...
 Do you guys know of a bzr->cvs mirror, so that if we switch to bzr, people who checked out from our old cvs servers can still "cvs up"?
[20:35] <sohail> why would you want to do that?
[20:35] <sohail> sounds like a lot of work
[20:36] <mkanat> sohail: We ship tarballs that include the CVS directories.
[20:36] <mkanat> sohail: hundreds of thousands of tarballs.
[20:37] <mkanat> That is, the tarballs get over 100,000 downloads.
[20:42] <sohail> mkanat, I think you can have CVS send a message or something redirecting them to bzr but I think a cvs mirror is too much work unless those 100K downloads pay you or something
[20:42] <mkanat> They don't pay me, but I do care about them. :-)
[20:42] <mkanat> Well, a few of them pay me. :-)
[20:44] <sohail> mkanat, then you have an incentive to make it painless :-)
[20:44] <sohail> can't say I know how
[20:52] <LarstiQ> nosklo: ~/.bazaar/authentication.conf
[20:52] <LarstiQ> mkanat: old and experimental
[20:53] <LarstiQ> mkanat: https://edge.launchpad.net/bzr-cvs
[20:54] <mkanat> LarstiQ: Awesome, thanks.
[20:54] <LarstiQ> mkanat: looking at bzr svn-serve might give more ideas
[20:54] <LarstiQ> oh
[20:54] <LarstiQ> maybe that was not the cvs thing I had in mind
[20:55] <mkanat> LarstiQ: Yeah, kind of doesn't look like it. :-)
[20:55] <LarstiQ> ho hum
[20:55] <mkanat> LarstiQ: I imagine I could also just write a hook that commits to CVS after every checkin, yeah?
[20:56] <LarstiQ> mkanat: yeah, roundtripping would be hard I think, but for a read-only cvs mirror that might not be an awful setup
[20:56] <mkanat> LarstiQ: Yeah, that's all it will be.
[20:56] <LarstiQ> mkanat: looking at bzr svn-serve might give some hints also
[20:56] <mkanat> LarstiQ: Okay, I'll check it out.
[20:57] <LarstiQ> after that, I'd say ask on the list
[20:57] <mkanat> Okay.
[20:58]  * LarstiQ swears he's seen a launchpad project for a bzr cvsserve somewhere, meh
[20:59] <mkanat> mwhudson: Does serve-branches read any config file?
[20:59] <LarstiQ> mkanat: https://edge.launchpad.net/bzrcvsserve
[20:59] <LarstiQ> _old_
[20:59] <mkanat> LarstiQ: Ah, but maintained by jdahlin. :-)
[20:59] <mkanat> So I can ask him about it.
[21:00] <LarstiQ> cool :)
[21:00] <LarstiQ> mkanat: may I ask what project this is?
[21:00] <mkanat> LarstiQ: Bugzilla
[21:00] <mkanat> LarstiQ: https://bugzilla.mozilla.org/show_bug.cgi?id=470570
[21:01] <mkanat> Nice, ubottu.
[21:02] <LarstiQ> mkanat: sweet!
[21:03] <mkanat> :-)
[21:06] <mkanat> mwhudson: I'm asking because it seems to be giving me "localhost:8080" urls by default when I ProxyPass it through Apache.
[21:17] <mkanat> mwhudson: Ah, figured it out, had to install paste-deploy and that magically fixed it.
[21:48] <mkanat> If I pull from a remote branch, will it overwrite my local tags if there are no tags in the remote branch?
[21:54] <Peng_> mkanat: No, it merges tags.
[21:54] <mkanat> Okay, great.
[21:54] <Peng_> mkanat: If tags conflict, they might get overwritten though. Try it. :)
[21:55]  * Peng_ confirms it
[21:55] <mkanat> Okay. There's no tags in the remote branch, so I'm safe. :-)
[21:56] <Peng_> Seems it'll overwrite conflicting tags, but it does warn you when it does so.
[21:56] <mkanat> Ah, great.
[22:00] <mkanat> Oh, so I asked this before, but if I have like 2000 commits with timestamps that are actually wrong by 4 hours, is there any way to fix that?
[22:00] <mkanat> It's actually more like 6500 commits.
[22:00] <Peng_> No.
[22:00] <Peng_> Commits, including timestamps, are immutable.
[22:01] <mkanat> I figured.
[22:01] <Peng_> You could ask the government to redefine your time zone 4 hours different. Then it would be right, sort of, I think?
[22:01] <Peng_> :)
[22:01] <mkanat> LOL
[22:13] <fullermd> Even easier would just to be start sleeping in 4 hours later.  A good plan on its own merits, IMO.
[22:14] <mkanat> lol
[22:27] <mkanat> Is there a way to have branches all use a single repository even though they all reside in deeper subdirectories?
[22:29] <Peng_> mkanat: When using shared repos, it'll look all the way up the directory tree until it finds a repo. You don't have to do anything special.
[22:30] <mkanat> Peng_: Oh. So I don't have to create repos as subdirectories of my current repo, or do I, for that to work?
[22:32] <mkanat> Peng_: That is, the root is a repo, and then bugzilla/ is not a repo (just an empty directory), and then bugzilla/3.2 is a branch, and it'll use the root repo?
[22:32] <fullermd> Yah.
[22:32] <mkanat> Ohhhh. I never knew that.
[22:32] <fullermd> In a degenerate case, you could make a shared repo on / and EVERY branch would use it.
[22:33] <mkanat> Heh. :-)
[22:33] <fullermd> (assuming you don't make them standalone, or make other repos lower that grab them, etc)
[22:34] <fullermd> It steps as far as it has to; first $BRANCH, then $BRANCH/../, $BRANCH/../../, $BRANCH/../../../, etc.
[22:34] <mkanat> And it only makes a standalone if there are no repos anywhere?
[22:35] <mkanat> (In the ancestry, that is.)
[22:36] <fullermd> By default, yah.  You can use reconfigure to force it standalone, or presumable use other such tricks.
[22:36] <mkanat> Ah, okay. :-)
[22:36] <mkanat> Has it always worked that way?
[22:38] <fullermd> Well, it didn't in 0.7   8-}
[22:38] <mkanat> That might have been when I started using bzr.
[22:38] <Peng_> That's when shared repos were added, right?
[22:38] <fullermd> [Shared] repos only existed since the metadir formats in 0.8.
[22:38] <mkanat> fullermd: Ah, I thought that might have been the joke you were making, also. :-)
[22:38] <mkanat> Okay. :-)
[22:39] <fullermd> I don't remember exactly when reconfig came on.  Post-0.15 I'm pretty sure.  Pre-1.0?
[22:40] <fullermd> Yah, 0.92.  But its various options have trickled in since.
[22:45] <mkanat> Is signing commits the only way to guarantee that people are who they say they are in whoami?
[22:46] <mkanat> Or does it even check that?
[23:49] <alperkanat> anyone here using loggerhead standalone ?