[01:14] <bob2> "Not checking SSL certificate for xmlrpc.launchpad.net" <- eh?
[01:18] <fullermd> XML and RPC have already exhausted the TLA support, so it can't do SSL at the moment.
[08:11] <mgz> morning
[15:46] <bobsapp> so i shelved some changes
[15:47] <bobsapp> but i never went back to them and i want to get rid of them.
[15:47] <bobsapp> in fact i fixed the bug before unshelving them.
[15:47] <bobsapp> can i somehow discard the shelf?|
[15:47] <mgz> `bzr shelve --list` then `bzr unshelve --destroy N`
[15:47] <bobsapp> many thanks!
[15:48] <mgz> the first command telling you what N to use (if you used a message when shelving)
[15:48] <mgz> otherwise you can `bzr unshelve --preview` to see which if you hae multiple shelves and no way of telling which is which
[15:49] <bobsapp> I think the problem is i was trying to do too much stuff at once.  Should have just not shelved those changes at all.
[15:49] <bobsapp> in the first place
[15:49] <bobsapp> In my manpage i dont have unshelve destroy
[15:50] <bobsapp> yeah i might have an older version of bzr
[15:50] <bobsapp> I have delete only which is the same thing though
[15:51] <bobsapp> it worked fine.
[15:52] <mgz> gah, sorry :)
[15:53] <mgz> one of shelve/unshelve uses destroy, one uses delete, and I always forget which is which
[17:28] <mgz> jelmer: can you have a look at torkvemada's branch? I got him to make the mp more verbose and add some tests based on one of your hook additions
[17:40] <jelmer> mgz: Yeah, bzr-git can create branches an dpush to them
[17:40] <jelmer> mgz: a somewhat annoying problem there is that dpush itself won't create new branches, unlike push
[17:40] <jelmer> mgz: but fixing that requires a bzr core change
[17:40] <jelmer> mgz: I'm not convinced about the commit hooks
[17:40] <jelmer> mgz: Commit until now was a pretty private object, passing it to a hook seems like a bad idea to me
[17:40] <mgz> wait, now I need to do my lines :)
[17:41] <mgz> is there a better way to do what he wants though?
[17:41] <jelmer> mgz: you're a shakespear man, you can handle it.
[17:41] <mgz> Commit.commit takes an impressive list of arguments
[17:41] <jelmer> mgz: pass specific arguments to start_commit I guess
[17:41] <jelmer> I guess it's just... message, revprops, committer and maybe the two timestamp bits?
[17:41] <mgz> hey, I said that! :)
[17:42] <jelmer> :)
[17:44] <mgz> when does Revision._check_properties get called... don't want a hook that can introduce borked properties...
[17:44] <jelmer> I guess we
[17:44] <jelmer> 'll ship 2.6 beta 2 in quantal?
[17:46] <mgz> seems that's going to happen
[17:46] <jelmer> so then it's perhaps not so bad to land a somewhat riskier change now
[17:47] <mgz> having multiple things trying to hit freeze window makes life painful
[17:48] <jelmer> yeah
[18:57] <cr3> if I "bzr checkout --lightweight lp:~cr3/checkbox/optical_write_test" and then do this in the directory "bzr st --short -r bzr+ssh://bazaar.launchpad.net/+branch/checkbox/", I get: bzr: ERROR: Server sent an unexpected error: ('error', 'GhostRevisionsHaveNoRevno', 'Could not determine revno for {bzr+ssh://bazaar.launchpad.net/+branch/checkbox/} because its ancestry shows a ghost at {tarmac-20120922215931-44bbvwzadme5c8rd}')
[18:58] <cr3> I don't understand what that means :(
[18:59] <jelmer> cr3: there is a reference to a revision that is not present in the repository
[18:59] <jelmer> cr3: that revision is necessary for determining the revno
[19:00] <jelmer> cr3: -r with a URL doesn't make sense
[19:00] <jelmer> -r takes a revision, not a branch
[19:01] <cr3> jelmer: odd, I got that from utilities/find-changed-files.sh in the launchpad source: rev=`bzr info | sed '/parent branch:/!d; s/ *parent branch: /ancestor:/'`
[19:01] <cr3> jelmer: maybe "parent branch" in some contexts is a revision, not a branch. that would explain it
[19:02] <jelmer> cr3: no, a parent branch is always a URL
[19:02] <jelmer> cr3: "branch:URL" is a revision though
[19:02] <jelmer> cr3: so perhaps it's just removing the word 'parent' there
[19:02] <jelmer> -rbranch:lp:bzr returns the tip revision of lp:bzr
[19:03] <cr3> jelmer: aha, it replaces "parent branch" with "ancestor", so you have "ancestor:URL" which works fine when I branch a project, not when I checkout though
[21:08] <w7z> that GhostRevisionsHaveNoRevno thing with lightweight checkouts of stacked branches is bug 1049124
[21:26] <jelmer> I guess it should just be saying NoSuchRevision?
[21:58] <w7z> jelmer: the cute thing is it actually has the rev... it's just in the stacked on branch, not the immediate branch
[22:40] <awilkins> jelmer, Ping?
[22:45] <jelmer> awilkins: hi
[22:46] <awilkins> Running into a problem pushing to a remote SVN repo - it looks like #394527 (with a slightly different message)
[22:46] <awilkins> "bzr: ERROR: At least one property change failed; repository is unchanged "
[22:47] <awilkins> It pushes about 8.5Mb and then coughs that up. The revision in question is not especially large but does rename a root level folder and nest it inside a new folder of the same name as the original
[22:48] <awilkins> So foo => foo/new-foo
[22:48] <awilkins> Alas, I have no access to the server logs or configuration, it's commercial SVN hosting
[23:05] <jelmer> awilkins: not sure if there's much I can help with there :-/
[23:07] <awilkins> jelmer, Yeah, was just hoping for a miracle :-)
[23:07] <jelmer> awilkins: I guess you could split up the commit so each change would be smaller
[23:07] <jelmer> that might not be very easy though
[23:07] <awilkins> jelmer, I have a feeling it's that one folder move, and all the paths inside being written as metadata in the revision properties
[23:08] <awilkins> The rest of it is pretty small changes to text files
[23:08] <jelmer> awilkins: that sounds about right
[23:08] <jelmer> for the bug you linked
[23:08] <awilkins> Yeah, I suppose I could ask if they'll raise that XML limit
[23:09] <awilkins> I'm trying to find an SSH port I can hit it through to circumvent it
[23:13] <awilkins> Bah, doesn't help that it's a horrible old build of CentOS in there
[23:13]  * fullermd upgrades pre-2a branches...