[00:00] <disturbedsaint> will let you know if it solved it
[00:00] <markh> thanks
[00:00] <poolie> beuno: i'm a bit surprised by the bzr-pqm failure because afaics iter_changes does _not_ require 8 arguments in 1.9rc1
[00:01] <abentley> poolie: It's _iter_changes, but *that* shouldn't require 8 arguments either.
[00:01] <poolie> right
[00:02] <abentley> I don't know what happened there.  I've been treating the sympoms because that's an old version anyhow.
[00:02] <beuno> poolie, I was as well, but since slapping on the new version "just worked", I carried on without diving too deep into it
[00:03] <poolie> fair enough
[00:21] <lifeless> beuno: merged my fix for search?
[00:22] <beuno> lifeless, argh, no
[00:22] <beuno> I will now
[00:22] <lifeless> beuno: nag nag nag
[00:22] <beuno> it's on my starred emails
[00:22] <beuno> which have grown out of proportion
[00:23] <lifeless> beuno: FWIW seeing this has increased my 'do not do big layout cleanups' twitch factor
[00:24] <beuno> yeah
[00:24] <beuno> it did for me as well
[00:24] <beuno> also, it was a patch from someone
[00:24] <beuno> not something I actually did myself
[00:24] <lifeless> k
[00:24] <lifeless> well they don't understand python as much as they think they do :>
[00:25] <beuno> seems not
[00:26] <beuno> and, I should pay a little more attention, or just not merge those things
[00:27] <beuno> hrm
[00:27] <lifeless> more test coverage would reduce the risk
[00:27] <beuno> that change still doesn't fix the problem
[00:27] <lifeless> show me your diff?
[00:27] <lifeless> cause it WFM
[00:27] <beuno> yeah, tests are high up there next to "make it stop eating ram until it blows up"
[00:27] <beuno> which we kinda know what's doing it now
[00:28] <lifeless> bug 293750 might be fixed now poolie
[00:28] <beuno> lifeless, http://paste.ubuntu.com/67620/
[00:29] <lifeless> beuno: that should do it
[00:29] <lifeless> beuno: have you restarted? does command line search work?
[00:29] <beuno> command line works
[00:29] <beuno> I did restart
[00:29] <beuno> do you have latest trunk?
[00:29] <lifeless> what is the first exception
[00:30] <lifeless> rev 235
[00:30] <BasicOSX> python-2.6 unsupported in 1.10dev ?
[00:30] <lifeless> BasicOSX: EPARSE
[00:31] <beuno>   File "/home/beuno/bzr_devel/loggerhead/trunk/loggerhead/controllers/__init__.py", line 98, in __call__
[00:31] <beuno>     vals.update(self.get_values(h, revid, path, kwargs, headers))
[00:31] <beuno> TypeError: get_values() takes exactly 5 arguments (6 given)
[00:31] <BasicOSX> lifeless:  https://bugs.launchpad.net/bzr/+bug/293886 :-)
[00:31] <lifeless> BasicOSX: a regression I guess; no core devs I know of are usingn 2.6 regularly
[00:31] <lifeless> BasicOSX: or more accurately, we haven't announced full 2.6 support yet
[00:32] <lifeless> beuno: that exception is the symptom
[00:32] <BasicOSX> it's what I get for being bloody edge
[00:32] <lifeless> beuno: its hiding the error
[00:33] <lifeless> beuno: when I saw that I looked higher up in the log and could see the real exception, IIRC
[00:34] <beuno> lifeless, ah, yes
[00:34] <lifeless> beuno: I mean, you should fix that too; but its secondary
[00:35] <beuno> http://paste.ubuntu.com/67625/
[00:35] <lifeless> beuno: File "/home/beuno/bzr_devel/loggerhead/trunk/loggerhead/search.py", line 52, in search_revisions
[00:35] <lifeless> beuno: print query there
[00:36] <beuno> lifeless, [('a',)]
[00:38] <lifeless> beuno: at line 1221
[00:38] <lifeless> beuno: print key
[00:38] <lifeless> beuno: it may be you've just found a unrelated bug
[00:39] <lifeless> beuno: we can check this by doing a search for 'a' rather than a completion lookup for 'a'
[00:39] <lifeless> (line 1221 of search/index.py I meant)
[00:39] <beuno> lifeless, ('a',)
[00:39] <beuno> yeah, I guessed that  :)
[00:40] <lifeless> so type a and hit enter in the search box
[00:40] <lifeless> or search for something bigger
[00:40] <beuno> ah, see, I was trying the "find as you type"
[00:40] <beuno> searching works
[00:40] <lifeless> yes, I know :P
[00:40] <lifeless> ok
[00:40] <lifeless> now in the CLI
[00:40] <lifeless> try 'bzr search -s a'
[00:40] <lifeless> on the branch that was erroring
[00:41] <beuno> aha
[00:41] <beuno> same error
[00:41] <lifeless> ok
[00:41] <lifeless> lh is fixed
[00:41] <lifeless> push that anytime
[00:42] <beuno> prints (u'a',)
[00:42] <lifeless> file a bug on the search - and I need a copy of the .bzr/bzr-search folder for the thing that errored
[00:42] <lifeless> FWIW, 'bzr search -s a' works for me on an index of lh trunk
[00:43] <lifeless> I suspect it will be a bzrlib index change
[00:43] <lifeless> what bzrlib are you using?
[00:43] <beuno> lifeless, sent the tarball
[00:44] <beuno> I'm on 1.9dev
[00:44] <beuno> nightly PPA
[00:44] <lifeless> k, just tried with .dev
[00:44] <lifeless> worked on my index, I'll drop yours in
[00:44] <beuno> >>> bzrlib.__version__
[00:44] <beuno> '1.9dev'
[00:44] <lifeless> what branch does it index?
[00:45] <beuno> that's for lh's trunk
[00:45] <beuno> maybe it's an old format?
[00:45] <beuno> the search that is
[00:45] <lifeless> nah
[00:45] <lifeless> if it was normal search wouldn't work
[00:45] <lifeless> and the next format is btree based, but needs btree prefix searching implemented
[00:46] <beuno> lifeless, deleting the .bzr/bzr-search and indexing again works
[00:46] <lifeless> it'll be a corner case
[00:46] <lifeless> (duh)
[00:47] <beuno> ah, email is still sending
[00:47] <beuno> 3.2mb apparently
[00:47] <lifeless> cool
[00:47] <lifeless> though you could have attached it to the bug :>
[00:48] <beuno> ah, I could, yeah
[00:48] <beuno> it's almost 2am here, european time doesn't seem to mix well with me
[00:49] <lifeless> are you home?
[00:50] <beuno> I wish!
[00:50] <beuno> Spain
[00:50] <beuno> for a few days
[00:50] <beuno> then Washington DC
[00:50] <lifeless> holidays?
[00:50] <beuno> no no, LP UI world tour
[00:51] <lifeless> clearly I've missed some mails :)
[00:51] <lifeless> what are you doing?
[00:51] <beuno> Launchpad UI
[00:51] <beuno> we're gearing up for a big change in 3.0
[00:52] <lifeless> I got the UI bit :P
[00:52] <beuno> did you hear that we where sprinting in London for 2 weeks?
[00:52] <lifeless> what I mean is what are you doing at each location that can't be done from  home
[00:52] <lifeless> I knew about the EPIC
[00:52] <beuno> ah, sprints
[00:52] <beuno> with each team
[00:52] <beuno> go through aaaaaaaaaaaaaaaall the pages
[00:52] <lifeless> wow
[00:52] <beuno> improve existing, plan for future dominance
[00:52] <lifeless> you'll be fucked then
[00:52] <lifeless> :P
[00:53] <beuno> yes, I'm starting to guess that  :)
[00:53] <lifeless> when do you come to .au?
[00:53] <beuno> I've been in nz
[00:53] <lifeless> meh, sif that counts
[00:53] <lifeless> :)
[00:53] <beuno> hehe
[00:53] <beuno> long enough trip for me
[00:54] <bob2> .au jr
[00:54] <lifeless> lol
[00:56] <beuno> anyway, I'm off to sleep for a while
[00:56] <beuno> ping me if I can do anything else
[00:56] <beuno> that email is still sending  :/
[00:56] <beuno> g'night lifeless, and thanks for the patch + reminder
[01:08] <lifeless> fud
[04:31]  * arjenAU is trying to work out how to use tag so that it behaves as wanted
[04:32] <arjenAU> it appears to apply to the last committed rev# not the next
[04:32] <arjenAU> that's just so odd
[04:35] <lifeless> arjenAU: hmm?
[04:36] <lifeless> commit --tag tags the next commit
[04:36] <arjenAU> lifeless: ah
[04:36] <arjenAU> so what happens if I do bzr tag separately after doing bzr commit?
[04:36] <arjenAU> does it kinda get attached to that commit, or?
[04:36] <lifeless> you tag an existing revision
[04:36] <AfC> arjenAU: ie, the last one, so "yes"
[04:37] <arjenAU> lifeless: no I want to tag what i'm committing.
[04:37] <arjenAU> lifeless: but I understand what you're saying
[04:38] <arjenAU> so I should ideally just speify the --tag on the commit, that's clear.
[04:38] <lifeless> arjenAU: so you want to record that you want to commit on the next commit
[04:38] <lifeless> arjenAU: I don't think thats built-in, so you need to use --tag on the commit, yes
[04:38] <arjenAU> lifeless: exactly. see most htings operate as a thing that gets committed later. bzr tag adds onto the previous commit. that's kinda unexpected
[04:39] <arjenAU> bzr commit --tag works as expected
[04:45] <Petrach> A question: When a repository is out of date, i.e. bzr update returns "working tree is out of date, run 'bzr update'"
[04:45] <bob2> that would be a working tree being out of date, not a repository
[04:46] <Petrach> is there an easy way to determine how foar out of date the repository is
[04:46] <Petrach> ?
[04:49] <Petrach> Is it possible to determine what revision the state of the working tree corresponds to?
[04:49] <bob2> bzr info should tell you
[04:50] <Petrach> That tells me what revision I would get if I run bzr update
[04:50] <lifeless> Petrach: the wt has the revision id embedded in it, but not a revision number
[04:51] <lifeless> Petrach: we should expose this I guess
[04:51] <Petrach> Where would I have to look for that?
[04:52] <lifeless> python -c 'import bzrlib.workingtree; t = bzrlib.workingtree.WorkingTree.open("."); t.lock_read(); print t.get_parent_ids()'
[04:54] <Petrach> Ah, ok. Thanks
[04:54] <lifeless> (sorry that its crude)
[04:56] <Petrach> It's comprehensible enough, thanks
[04:57] <lifeless> log --show-ids on the branch will let you find that rev
[04:57] <lifeless> (unless someone uncommitted on the branch)
[05:06] <poolie> hi all
[05:07] <lifeless> welcome back
[05:16] <lifeless> poolie: so, want to talk split inv
[05:37] <NfNitLoop> echo "Obama" > president;  bzr commit;
[05:37] <NfNitLoop> ;)
[06:05] <arjenAU> NfNitLoop: hehe
[06:05] <arjenAU> hmm lp not scanning branches at the mo it seems
[06:07] <spm> arjenAU: no - am trying to figure out wtf is broken atm.
[06:10] <arjenAU> spm: no worries, was just noting
[06:21] <poolie> hi arjen
[06:25] <arjenAU> hey poolie
[06:25] <arjenAU> poolie: in case i haven't raved enough - bzr rocks
[06:29] <poolie> that's always nice to hear
[06:30] <arjenAU> not even craving back bk, this is just good. doing the weirdest cross-branch merges and it hasn't failed me yet
[06:32] <poolie> arjenAU: i was going to propose to do a tutorial at the developer conference
[06:32] <poolie> i think i need to send that tomorrow
[06:33] <arjenAU> poolie: euh, OSDC? won't be any tutes.
[06:33] <poolie> also, some canonical people including myself are going to be in brisbane next week, do you want to catch up?
[06:33] <poolie> no, the mysql conference
[06:33] <arjenAU> poolie: oh right. sounds good.
[06:33] <arjenAU> poolie: ehm in Melb Sun eve- Wed eve, then US from Fri arvo. other than that, fine ;-)
[06:35] <spm> arjenAU: and to think I felt guilty about calling myself a Qlder still; even after moving from brisvegas to canberra 18ish years ago. At least when I lived there, I *lived* there. ;-)
[06:35] <poolie> heh
[06:35] <poolie> where were you, spm?
[06:36] <spm> In brisbane? jindalee - since about 71/72 ish.
[06:36] <spm> I can *just* remember scenes from the 74 floods
[06:39] <spm> poolie: what about yourself? I believe we went to *cough* .. HIGHLY... rival high schools. :-)
[06:40] <poolie> in brisbane from about 1980 to 1999, and I went to BGS
[06:44] <poolie> arjenAU, so i suppose you'll be pretty busy, but maybe we could meet for lunch or during the day on Thursday?
[06:55] <poolie> lifeless: if you're still around, do you have an easily obtainable test branch in your new format?
[06:58] <lifeless> poolie: the only precanned branch i have is the 18GB one on banchmarks
[06:59] <lifeless> generally I take plugins and upgrade a copy
[06:59] <poolie> in that robertc directory somewhere?
[06:59] <lifeless> yah
[06:59] <poolie> like of the bzr-gtk plugin?
[06:59] <lifeless> but bzr branch bzrtools foo; cd foo; bzr upgrade --development3
[06:59] <lifeless> is quite fast
[07:01] <lifeless> poolie: and with that, I'm going to call 'day'
[07:02] <poolie> i'm going to head off soon too
[07:02] <poolie> cheerio
[07:02] <spm> arjenAU: looks like we're scanning again
[07:02] <spm> poolie: oh yes - huge rivals @ highschool - GT. :-)
[07:09] <vila> hi all
[07:10] <poolie> hi vila
[08:34] <arjenAU> poolie: sure - you've got my mobile, right?
[08:34] <arjenAU> poolie: whereabouts will you guys be?
[10:25] <dissonans> what's the correct way to revert an (uncommitted) merge?
[10:26] <beuno> dissonans, bzr revert?
[10:26]  * LarstiQ nods
[10:26] <LarstiQ> but, that will also revert any local changes you had
[10:27] <LarstiQ> dissonans: if that is ok, then just `bzr revert` right ahead
[10:27] <beuno> hiya LarstiQ
[10:27] <LarstiQ> dissonans: if you want to keep the file changes, but not the recording of the merge, supply --forget-merges to revert
[10:27] <LarstiQ> beuno: heya :)
[10:28] <LarstiQ> beuno: still in London?
[10:28] <beuno> LarstiQ, almost. In Spain now for the rest of the week
[10:28] <beuno> then Washington DC
[10:28] <dissonans> LarstiQ: actually, I'd like to keep some changes from before the merge
[10:28] <dissonans> but I guess bzr can't discern those from those introduced by merge?
[10:28] <LarstiQ> dissonans: correct
[10:29] <LarstiQ> dissonans: you could perhaps try to do something with merging the reverse changes
[10:29] <LarstiQ> but that would be tricky
[10:30] <beuno> or
[10:30] <beuno> shelve
[10:30] <beuno> and go through the changesw
[10:33] <LarstiQ> that would work if the merge didn't touch the same hunk as you changed yourself
[10:33] <LarstiQ> and/or you had to resolve conflicts
[10:38] <beuno> yeah, I don't think there's a perfect and simple solution
[10:50] <luks> that's why such a situation requires merge --force
[11:38] <thrope> how can I install bazaar 1.8 with tortoisebzr? the 1.8 win32 setup lite installer doesn't seem to include it, but I can't find a non-lite one
[11:38] <thrope> also is there any documentation/tutorial online for tortoisebzr
[11:38] <thrope> I have to collaborate with some not so technical windows users and I would really like to use bazaar
[11:40] <poolie> thrope: hi, i think you should try the 1.9rc1-2 installer instead
[11:40] <poolie> several problems have been fixed in tbzr packaging
[11:40] <thrope> ok
[11:40] <thrope> the trouble is I use 1.8 on other platforms
[11:40] <thrope> will it be ok to mix them
[11:40] <poolie> it will
[11:41] <poolie> unless you specifically choose the 1.9 format for something
[11:41] <poolie> i'm going to bed but if you hit problems ask someone here or on http://answers.launchpad.net/bzr
[11:41] <thrope> hmmm - it doesnt work anyway - get a dll error trying to init with tortoisebzr from 1.9 installer
[11:41] <poolie> really
[11:41] <poolie> do you have the -2 version?
[11:42] <poolie> it was just uploaded today
[11:42] <thrope> yes
[11:42] <thrope> i just downloaded it now
[11:42] <thrope> dll load failed with errorcode 193
[11:42] <thrope> anyway I think I will leave it for now
[11:44] <poolie> please send a mail or file a question
[11:44] <poolie> so we can track it
[11:44] <poolie> possibly rebooting after installing would help that
[11:44] <poolie> anyhow, good night and good luck
[11:46] <davi_> is it possible to lock a branch so that no reads are possible?
[11:50] <LeoNerd> chmod? ;)
[11:55] <AfC> rm -r ? :)
[14:39] <Tak> hmm, how does one close a bug in launchpad?
[14:40] <Tak> I've already set status to "Fix Committed" ...
[14:42] <LarstiQ> Tak: that depends on how the project wants to do things.
[14:43] <LarstiQ> Tak: fix committed could mean that the bug is fixed in a particular branch, and Fix released means it's in trunk.
[14:44] <LarstiQ> Tak: or fix committed could mean the bug is fixed in trunk, and fix released that is in a an actual release of the project.
[14:46] <jrydberg> it would be nice with a bug tracking software that knew about merges
[14:46] <Tak> I more meant, what needs to happen for the bug not to show up in the open bugs list
[14:46] <jrydberg> you mark a bug as fixed in one branch.  when you merge that into trunk, the tracking software automaticly sets the bug as fixed in the trunk
[14:55] <LarstiQ> jrydberg: yeah, either launchpad already does that or it's supposed to.
[14:55] <LarstiQ> jrydberg: but what we really need, imo, is a different bug status to distinguish between merged to trunk and committed on a branch.
[14:55] <luks> launchpad just links branches and bug reports, no?
[14:55] <luks> it doesn't change the bug reports
[14:56] <LarstiQ> Tak: either of those two statuses should do it afaict, if not, the Fix released certainly would.
[15:02] <Tak> apparently "Fix Released" does and "Fix Committed" does not
[15:28] <gorgapor> question: i deleted a file a few revisions ago, and now i want it back. what's the best way to resurrect this file and keep its history from before?
[15:31] <rockstar> gorgapor, a reverse merge would be in order.
[15:32] <LarstiQ> does that bring back the file with the same fileid as well?
[15:32] <gorgapor> does revert -r do the trick?
[15:33] <gorgapor> ok, how do i do that?
[15:34] <gorgapor> i tried revert -r123 deleted_file, but it didn't work
[15:38] <NfNitLoop> gorgapor: I think it may be:   bzr merge -r <n>..<n-1> .    (where n is the revision that deleted the file you want)
[15:38] <NfNitLoop> but I don't have a repo handy to test that.
[15:38] <gorgapor> okay i'll try that out
[15:43] <gorgapor> so, that seems to bring back the entire changeset
[15:44] <gorgapor> is there a way to just bring back the one file?
[15:55] <Peng_> lifeless: ping?
[15:55] <Peng_> eh
[15:57] <Peng_> Hmm, searching in Loggerhead is busted in one branch, but not another.
[15:59] <beuno> Peng_, I patched trunk
[15:59] <beuno> did you upgrade?
[15:59] <beuno> also, found a bug in bzr search index
[16:00] <beuno> so you could delete de .bzr/bzr-search
[16:00] <beuno> and re-index
[16:04] <Peng_> beuno: Yeah, I upgraded.
[16:05] <beuno> Peng_, try re-indexing the branch
[16:09] <Peng_> Yeah, that fixed it.
[16:09] <Peng_> Huh.
[16:10] <Peng_> Oops, I didn't keep a backup of the bad indexes, if anyone was interested.
[16:11] <beuno> I have a b0rked index already
[16:11] <beuno> attached it to a bug
[16:12] <Peng_> ok :)
[16:13] <Peng_> Oh, LP is approaching 300,000 bugs.
[16:13] <beuno> yeah, seems software is REALLY buggy  :p
[16:15] <Peng_> OK, I just checked all of my other branches, and there weren't any more with broken indexes. Just the two.
[16:29] <emmajane> beuno, I think the upload plugin is your genius invention?
[16:29] <beuno> emmajane, well, vila did most of the heavy lifting
[16:30]  * emmajane nods.
[16:30] <beuno> and, by heavy lifting, I mean code  :)
[16:30] <emmajane> hehehe
[16:30] <beuno> I just complained a lot
[16:30] <emmajane> Do you know if there's a way to toggle between two different servers?
[16:30] <beuno> ah, interesting
[16:30] <beuno> well, why wouldn't you be able to?
[16:30] <beuno> it remembers the upload server-side
[16:30] <emmajane> I upload to testing, and then when it's working I need to upload to the live server.
[16:31] <emmajane> without having to put in the sftp:// nonsense again.
[16:31]  * emmajane is lazy. :)
[16:31] <beuno> ah, multiple locations
[16:31] <emmajane> yeah
[16:32] <vila> emmajane: you can use --remember for testing (since it's most often used) and look at the bookmarks plugin for your lazt fingers :0
[16:32] <beuno> well, bzr, I *think*, has something to alias locations
[16:32] <beuno> that's it
[16:32] <beuno> bookmarks
[16:32] <beuno> see, vila is the real brains
[16:32] <emmajane> vila, awesome, thanks. :)
[16:32] <vila> beuno: hi :)
[16:32] <vila> emmajane: happy to help (c)
[16:32] <beuno> hi vila!
[16:34] <vila> emmajane: on the other hand, if you control the test server and have ssh access to it, why not install your working tree there and use sshfs if you need to access it from a remote workstation ?
[16:34] <emmajane> vila, the test server is my dreamhost account. the live machine is a college mainframe superfancy machine. I'm very happy to not tie dreamhost to the college.
[16:35] <vila> emmajane: ok
[16:35] <emmajane> vila, I had to get Very Special Permission to even be allowed to connect to the college machine. :)
[16:36] <emmajane> vila, in a normal world where deployments happened from testing to live, it would make sense to do what you suggested. :)
[16:36] <vila> evrything that suits a user needs make sense :)
[16:37] <emmajane> :)
[16:41] <emmajane> hm. I have bookmarks installed, but the documentation is a little... vague.
[16:42] <emmajane> will it just remember places that I've uploaded files to? Or do I need to enter the locations manually?
[16:42]  * beuno suspects manually
[16:43]  * emmajane reads the actual plugin python file to figure it out
[16:44] <emmajane> beuno, let's pretend I wanted to actually add documentation for the plugins that I use (because if I can't figure it out, I'm sure there are others)... Where would I put this information?
[16:45] <emmajane> Launchpad doesn't seem to have a place for project docs.
[16:46] <beuno> emmajane, branch the code, addd the docs, push the branch, file a merge proposal
[16:46] <beuno> I'll be happy to walk you through it
[16:47] <emmajane> beuno, hrm. and then normal people would have to find it in the --help at the command line?
[16:47] <beuno> emmajane, well, what other way would there be?
[16:47] <emmajane> beuno, I usually check the wiki, to be honest. :)
[16:47] <luks> you can create a wiki page on bazaar-vcs.org
[16:47] <luks> and make sure it's linked from the launchpad page
[16:48] <emmajane> luks, none of the plugins seem to have documentation online... or am I just lucky in the ones I choose...?
[16:48] <emmajane> http://bazaar-vcs.org/BzrPlugins <--- there's usually a terse description but that's about it.
[16:48] <luks> emmajane: why not make something better than what other people do :)
[16:48] <emmajane> hehe
[16:49] <emmajane> luks, that I can handle. :)
[16:49] <NfNitLoop> Hrmmm.  I'm curious.  How would bzr-svn handle working with a svn repo if someone went and changed a log message in the repo history?
[16:50] <NfNitLoop> I mean, I'm guessing at the very least, you wouldn't see that new message in `bzr log`, but I'm hoping it wouldn't break interoperability. :p
[16:50] <vila> emmajane: 'bzr help bookmark' for usage and I'm pretty sure it stores them in .bazaar.conf
[16:51]  * emmajane nods to vila 
[16:51] <LarstiQ> NfNitLoop: good question
[16:51] <emmajane> the pluginname with and without the S still trips me up.
[16:51] <emmajane> help bookmarks != help bookmark
[16:52] <emmajane> help tags != help tag
[16:52] <james_w> emmajane: do they have a "See also:" at the bottom to link them?
[16:52] <emmajane> help bookmarks doesn't say "see also: bookmark"
[16:53] <emmajane> tags does, bookmarks doesn't.
[16:53] <james_w> emmajane: that can be fixed at least :-)
[16:53] <emmajane> james_w, patches welcome? ;)
[16:53] <james_w> it's an easy one
[16:53] <emmajane> "easy"
[16:53] <emmajane> sounds like a good hour of procrastination...
[16:53] <emmajane> I'm in!
[16:54] <james_w> hmm, though I'm not sure what "help bookmarks" displays
[16:54] <james_w> probably the module help, and I'm not sure if you can link from that
[16:54] <james_w> if not then a bug report on bzr is in order :-)
[16:54] <emmajane> the relevant bit is: From:     plugin "bookmarks"
[16:54] <emmajane> See also: plugins/bookmarks
[16:55] <emmajane> It's missing See also: bookmark
[16:55] <emmajane> I think?
[16:55] <james_w> yeah
[16:56] <vila> emmajane: did you *try* bzr help plugins/bookmarks ? :)
[16:56] <vila> I think that the one that needs love :)
[16:57] <emmajane> vila, it needs love, I agree. :)
[16:57] <emmajane> and I see where to edit that...
[16:57] <vila> great :)
[16:57] <emmajane> it /seems/ as though it would be the same information though. How come plugins/* is different than just the plugin name?
[16:58] <emmajane> hrm.
[16:58] <luks> because the plugin name happens to be also a command name
[16:58] <emmajane> actually.
[16:59] <emmajane> help bzr bookmarks != help bzr bookmark != help bzr plugins/bookmarks
[16:59] <emmajane> and help bzr plugins/bookmark is not found.
[16:59] <luks> because there is no plugin "bookmark"
[16:59] <emmajane> that's enough to confuse a person right there.
[16:59] <emmajane> luks, correct.
[17:00] <emmajane> is there a way to make bookmarks/bookmark relate to each other?
[17:00] <james_w> "help bookmarks" is also known as "help commands/bookmarks"
[17:00]  * emmajane nods
[17:00] <emmajane> help plugins/bookmarks just reads from the actual .py file.
[17:00] <james_w> there are different categories of help, commands/*, plugins/*, and topics/*
[17:00] <emmajane> that popping noise? that was my brain exploding. :)
[17:01] <LarstiQ> iirc help plugins/foo uses the module/package docstring
[17:01]  * LarstiQ trots off to jitsu
[17:01] <emmajane> LarstiQ, it appears that way, yes.
[17:01] <james_w> yeah, LarstiQ is right
[17:01] <emmajane> LarstiQ, have fun beating up people. :)
[17:01] <LarstiQ> emmajane: thanks, I'll try not to get too sore myself ;)
[17:01] <emmajane> LarstiQ, duck and cover! :)
[17:02] <emmajane> so I have two problems that need fixing. See also: for the help commands/* is not referencing relevant commands. AND help plugins/* could be more useful.
[17:02]  * emmajane wonders if that even made sense.
[17:03] <james_w> I don't understand what you mean for the first bit
[17:03] <luks> emmajane: if you can write the documentation, I'm sure I can fix these semantic issues
[17:03] <emmajane> james_w, the See also: at the bottom of bzr help bookmarks doesn't say "see also: bookmark" the way it does for tags/tag
[17:03] <james_w> aha
[17:03]  * emmajane grins at luks. "Deal!"
[17:04] <emmajane> james_w, which is different than the documentation being vague in the actual plugin which is used for bzr help plugins/bookmarks
[17:04] <james_w> emmajane: for that you need to add see_also = ["bookmark"] somewhere in cmd_bookmarks
[17:04] <emmajane> knowing nothing about Python, /me looks
[17:05] <james_w> just under """List bookmarks.""" would do
[17:05] <james_w> keeping the same indentation
[17:05] <emmajane> yeah
[17:06] <emmajane> white space makes you able to fly.
[17:06] <emmajane> I know that part thanks to xkcd.
[17:06] <james_w> :-)
[17:06] <emmajane> will it append, or overwrite the See also: plugins/bookmarks?
[17:06] <emmajane> I don't want to lose any information.
[17:07] <james_w> append I believe
[17:07] <emmajane> kay
[17:07] <james_w> you'll have to try it I'm afraid
[17:07] <emmajane> heh
[17:07] <emmajane> after making my edits do I need to do anything fancy to test? or just run the command again?
[17:08] <emmajane> woo. fail.
[17:08] <james_w> :-(
[17:09] <emmajane> Unable to load plugin 'bookmarks' from '/home/emmajane/.bazaar/plugins'
[17:09] <emmajane> Plugin 'bzrlib.plugins.bookmarks' has no docstring.
[17:09] <james_w> try "bzr -Derror help bookmarks"
[17:09] <luks> most likely broken indentation :)
[17:09] <emmajane> hrm. I let vim do the indentation...
[17:10] <luks> using tabs maybe?
[17:10] <james_w> you might need ":set expandtab"
[17:10] <emmajane> james_w, same error message
[17:11] <emmajane> I redid the edits with :set expandtab on
[17:11] <emmajane> it seems to work now. :)
[17:11] <james_w> heh :-)
[17:11] <emmajane> although the see also was ignored. :(
[17:11] <james_w> boo
[17:12] <emmajane> yeah
[17:13] <emmajane> the other two plugins I have installed don't use variants, so they don't have a see_also for me to compare against
[17:13] <james_w> change it to _see_also
[17:13] <james_w> that seems silly to me
[17:14] <emmajane> hurrah
[17:14] <emmajane> that worked. :)
[17:14] <emmajane> and it appends.
[17:14] <james_w> cool
[17:14] <emmajane> From:     plugin "bookmarks"
[17:14] <emmajane> See also: bookmark, plugins/bookmarks
[17:14]  * emmajane thinks that's nifty.
[17:15] <emmajane> ok.
[17:15] <emmajane> so.
[17:15] <emmajane> I should do something with this change.
[17:15] <james_w> so was there more to fix, or just documentation to improve?
[17:15] <emmajane> seems greedy just to keep it on my own machine.
[17:15] <james_w> commit it!
[17:16] <emmajane> I want to make changes to the help at the top (which is triggered by help plugins/bookmarks) too.
[17:16] <emmajane> but I haven't actually been able to /use/ the plugin yet because I couldn't find all the help files. ;)
[17:18] <emmajane> first commit made for the See also: stuff.
[17:18] <emmajane> I'm going to actually test out the bookmarks and then I'll bug y'all again when i'm ready to push the changes. :)
[17:54] <emmajane> ok. the help information is updated.
[17:54] <emmajane> Now I suppose I need to push the changes back to the server?
[17:56] <james_w> yeah
[17:56] <james_w> bzr push lp:~emmajane/bzr-bookmarks/fix-documentation
[17:56] <james_w> or similar
[17:57] <emmajane> that creates a new branch, right?
[17:57] <james_w> yeah
[17:57] <emmajane> fail
[17:57] <emmajane> bzr: ERROR: Transport operation not possible: http does not support mkdir()
[17:57] <emmajane> I don't have write permissoins I guess.
[17:57] <james_w> bzr launchpad-login emmajane
[17:58] <james_w> that will make the "lp:" thing use a writeable transport for you
[17:58] <emmajane> it's thinking
[17:58] <emmajane> which is better than failing.
[17:59]  * emmajane wonders if that was the right password....
[17:59] <emmajane> it's still thinking...
[17:59] <beuno> depends if it fails at the end, then it's worst!
[17:59] <emmajane> and new branch created.
[18:01] <emmajane> erm.
[18:01] <emmajane> now what?
[18:03] <luks> now you ping me about merging the branch? :)
[18:03] <emmajane> luks, ping!
[18:03] <emmajane> luks, would you be able to take a look at the changes I made to bzr-bookmarks? :)
[18:04] <emmajane> hm. I should also update my email to have the Ubuntu address I guess....
[18:04] <emmajane> identity is such a hassle.
[18:06] <luks> emmajane: pulled, reviewed, pushed -- thank you!
[18:06] <emmajane> WOO!
[18:08] <emmajane> luks, thanks :)
[18:09] <james_w> emmajane: unfortunately launchpad won't yet detect that, so you should mark your branch as merged in launchpad
[18:09] <emmajane> erm
[18:11] <emmajane> Propose for merging into another branch ?
[18:11] <NfNitLoop> so I just commented on a bzr-svn bug that appears to still be a bug despite having been marked "fixed":  https://bugs.launchpad.net/bzr-svn/+bug/290664
[18:11] <NfNitLoop> do I need to do anything other than that?   (Mark it as not-fixed somehow?)
[18:12] <beuno> emmajane, that creates a merge proposal
[18:12] <beuno> which is useful if it hasn't alaready been merged by the magic of IRC
[18:12] <emmajane> hehe
[18:12]  * emmajane makes the merge proposal.
[18:13]  * emmajane resists complaining about how confusing she finds LP.
[18:14] <beuno> emmajane, are you on the beta-testers team of LP?
[18:14] <emmajane> beuno, I'm not.
[18:14] <beuno> because we *just* re-wroked the whole merge-proposals bit
[18:15] <beuno> and it's 100x times simpler now
[18:15]  * beuno adds emmajane compulsively to the beta testers team
[18:15] <emmajane> 100x is an awful lot :)
[18:15] <emmajane> hehe
[18:15] <luks> I wonder how you do measure that :)
[18:15] <emmajane> luks, in beers, I'm sure. :)
[18:15] <beuno> luks, how much work it took  :)
[18:15] <emmajane> luks, how many new gray hairs the entire team has ;)
[18:15] <luks> heh
[18:16] <beuno> emmajane, you're in
[18:16] <emmajane> beuno, cool.
[18:16] <emmajane> how does this change my life?
[18:16] <beuno> emmajane, new things faster
[18:16] <beuno> isntead of every 1 month or so
[18:17] <beuno> (my typing is horrible today)
[18:17] <emmajane> beuno, fancier tag clouds. that's what it means. ;)
[18:17] <beuno> emmajane, well, to me it means more feedback on UI changes faster  :)
[18:17] <emmajane> beuno, hehe.
[18:18] <emmajane> beuno, are they all reported as bugs?
[18:18] <emmajane> cos baby I gots feedback if you want it. ;)
[18:18] <luks> it would be nice if the beta launchpad didn't insist on using edge. urls
[18:18] <beuno> emmajane, yes, we live in bugs
[18:19] <james_w> beuno: you don't know what you have started :-)
[18:19] <emmajane> beuno, what you've started is a total release of james_w from listening to me complain about one of my favourite topics. ;)
[18:19]  * beuno stops and thinks
[18:19]  * emmajane chuckles.
[18:19] <beuno> uhm
[18:19] <beuno> maybe the beta team isn't a good fit for you (?)
[18:19] <emmajane> james_w, you're going to have to buy beuno a LOT of beer at the UDS. ;)
[18:20] <james_w> always :-)
[18:20] <beuno> I may need beer now, from what I'm seeing happening
[18:20] <emmajane> :)
[18:20] <beuno> emmajane, are you going to UDS?
[18:20] <emmajane> nah
[18:20] <emmajane> I'm not a programmer. ;)
[18:21] <beuno> good, you can help balance things out!
[18:21] <emmajane> by not attending. ;)
[18:22] <beuno> :)
[18:22] <beuno> anyway, I'm looking forward to your feedback
[18:22] <james_w> beuno: how's Madrid?
[18:22] <emmajane> I'm starting REAL easy. :)
[18:22] <beuno> and, now, I'm running away from the computer before I make my life harder in other new creative ways
[18:22]  * emmajane grins at beuno 
[18:22] <james_w> :-)
[18:22] <emmajane> beuno, first one submitted.
[18:22] <beuno> james_w, much better weather than london
[18:22] <emmajane> beuno, vvvveeerrrrry easy
[18:23] <beuno> emmajane, if that's true, i may even fix it quickly
[18:23] <beuno> :)
[18:23] <james_w> beuno: of course :-)
[18:23] <emmajane> beuno, :)
[18:23] <beuno> now, I'll be back in a few hours to go through new bugs  :)
[18:23] <emmajane> :)
[18:24] <beuno> good talking to both of you
[18:24] <emmajane> beuno, laters :)
[18:24] <emmajane> thanks again for your help today
[18:26] <NfNitLoop> Hrmm, how do I enable a debug flag in bzrlib.debug?
[18:26] <NfNitLoop> --debug <flag> ?
[18:26] <luks> -D<something>, I think
[18:26] <NfNitLoop> no, that's not it, hrmm.   I'll try that.
[18:27] <NfNitLoop> luks: ah, that did it, thanks.
[18:39] <NfNitLoop> So if I'm trying to branch from an svn repo that has had...  very poor standards compliance w.r.t. tags/branches....
[18:39] <NfNitLoop> is there an option to tell bzr-svn to only grab /trunk and ignore everything else?
[18:40]  * emmajane stops at three new bugs for beuno 
[18:40]  * emmajane goes back to work-work now. :)
[18:40] <NfNitLoop> I see the bzr svn-branching-scheme command, but didn't find a list of branching schemes after some goodling.
[18:40] <NfNitLoop> googling*
[18:41] <Peng_> NfNitLoop: "bzr branch svn://.../trunk/"?
[18:41] <Peng_> Though that would still try to fetch tags.
[18:46] <NfNitLoop> Peng_: right, that's what I'm doing, and it's trying to get tags.
[18:46] <NfNitLoop> and someone thought it would be fun to put a *file* in /tags/
[18:46] <NfNitLoop> (I wouldn't even rely on the directories in /tags/, to tell the truth.)
[19:10] <Peng_> NfNitLoop: Oh. I think that's been fixed recently.
[19:11] <Peng_> NfNitLoop: What did it do when it came across the file in tags/?
[19:11] <Peng_> Well, from 0.4.14, there's "Ignore tags that happen to be files.".
[19:21] <mickbeaver> I'm having difficulty pushing to a usb flash drive on Windows Vista 64. Using both the win32 and the cygwin version yield the same result.
[19:21] <mickbeaver> bzr: ERROR: Could not acquire lock "[Errno 13] Permission denied"
[19:21] <mickbeaver> Any tips for starting to debug this?
[19:31] <disturbedsaint> markh, the chages you made seem to have fixed the would recurse to death error
[19:32] <disturbedsaint> none of those in the log from today, been browsing the directories that are under bzr control every now and then (in explorer and different file-open dialogs)
[19:41] <NfNitLoop> Peng_: it threw an exception and died.
[19:42] <NfNitLoop> Peng_: and I'm using 0.4.14... is there some option I need to specify to "Ignore tags that happen to be files"? :p
[19:43] <Peng_> NfNitLoop: Oh. I dunno then.
[19:46] <Peng_> Sorry I can't help. :\
[19:57] <jelmer> NfNitLoop, check the 0.4 branch
[19:57] <jelmer> NfNitLoop, it's got another fix
[19:58] <Spaz> w3
[19:58] <Spaz> oops
[20:35] <NET||abuse> we have an svn repo in the office here, and i was looking to work with something like bzr or git as my primary client, but i am a beginner with bzr.. is it a reasonable thing to do to take a checkout of the trunk into a bzr branch, operate away on it, maybe pull the branch between my laptop and desktop a few times (few days of development work)
[20:36] <NET||abuse> then merge that work back into the svn trunk?
[20:39] <NfNitLoop> NET||abuse: Yes.
[20:39] <NET||abuse> ok, just looking to learn the workflow
[20:39] <NfNitLoop> NET||abuse: in fact, that's exactly what I'm working on advocating at my workplace at the moment.
[20:39] <NfNitLoop> NET||abuse: the nice thing is there is no "the workflow"...  bzr supports quite a few. :)
[20:40] <NfNitLoop> http://bazaar-vcs.org/Workflows
[20:41] <NET||abuse> ok,, well, in my case i just need to learn the commands, and "the" workflow i'm interested in is, we have central svn repo, i do bzr branch svn+ssh://me@host/path/to/repo/
[20:41] <NfNitLoop> path/to/repo/trunk
[20:41] <NfNitLoop> yes. :)
[20:42] <NET||abuse> then i do some work on some files, which i actually just did, edited one file, delete a directory tree that was obsolete. then i do bzr commit, write my log, then bzr merge, it says mergin from remembered location svn+ssh:///blah.. Nothing to do
[20:42] <NET||abuse> ??? nothing??
[20:42] <NfNitLoop> (others:) I just installed and read up on `bzr help rebase`, but `bzr rebase` in my local, changed repo claims that there is no work to do?
[20:42] <NfNitLoop> NET||abuse: That's merging from "upstream" into your local repository.
[20:43] <NfNitLoop> NET||abuse:  you want to "push" back to the remote repository.
[20:43] <NET||abuse> .. oh.
[20:43] <NET||abuse> ok,, umm, little confused by the workflow description,, bzr comit --local? bzr unbind/commit/bind
[20:43] <NET||abuse> looking at the centralized with local commits
[20:44] <NfNitLoop> If you do "branch", you by default have local commits.
[20:44] <NfNitLoop> and are working with an "unbound" branch.
[20:44] <NET||abuse> ah, ok
[20:44] <NfNitLoop> which means that when you "commit", it goes into only your branch.
[20:44] <NET||abuse> so bzr pull is like an svn style checkout?
[20:44] <NfNitLoop> if you have a "bound" branch, it means that any commits you do go into your local branch *and* the upstream.
[20:44] <NfNitLoop> no, `bzr checkout` is. :)
[20:44] <NET||abuse> ... umm, ok
[20:44] <NET||abuse> ah, but i did bzr merge?
[20:44] <NfNitLoop> `bzr pull` is like... svn update.
[20:45] <NET||abuse> hmm, the forks in the process, because it's not a linear cyle of workflow, it's a bit confusing..
[20:47] <NET||abuse> hmm, are there any dangers to the central repo that i should be away of for now though, big nono's that i shouldn't do when operating bzr on the central svn repo?
[20:47] <NET||abuse> umm, away=i meant aware
[20:48] <NfNitLoop> NET||abuse: Yes.  I've been reading up on those, actually.
[20:48] <NfNitLoop> a big one:   SVN only supports linear history.
[20:48] <NET||abuse> ooh, do share.. i realllly don't wanna screw up our svn repo.
[20:49] <jelmer> NET||abuse, you should have a look at the bzr-svn FAQ
[20:49] <NfNitLoop> so, while you could do "bzr commit (some stuff); bzr merge (from trunk): bzr commit (more stuff);"...
[20:49] <NET||abuse> jelmer: will do .
[20:49] <NfNitLoop> ... that would make your `svn log` look strange.
[20:49] <NfNitLoop> so you probably want to use `bzr rebase` instead of `bzr merge`.
[20:50] <NET||abuse> bzr rebase... hmm
[20:50] <NfNitLoop> NET||abuse: it's a plugin.
[20:50] <NfNitLoop> https://launchpad.net/bzr-rebase
[20:50] <NET||abuse> i need a cheetsheet with the sequence of commands detailed and explained..
[20:51] <Odd_Bloke> Didn't Ian make one of those a while ago?
[20:52] <lifeless> its in the docs
[20:53] <NET||abuse> which docs?
[20:53] <lifeless> NfNitLoop: NET||abuse: I don't think y ou want rebase
[20:54] <lifeless> rebase is rarely the correct tool if you are just collaborating
[20:54] <NET||abuse> hmm,
[20:54] <lifeless> svn supports merge as of 1.5
[20:54] <NET||abuse> well, it's 5 of us building django site
[20:55] <NET||abuse> 4 coders(1 is an intern) and 1 graphics guy who does some html/css also.
[20:55] <lifeless> standard offline work for svn with bzr is 'bzr branch; hack, commit, hack commit, bzr merge <svn> until happy; bzr push <svn> || bzr dpush <svn>'
[20:55] <lifeless> push and dpush do slightly different things, I imagine the bzr-svn docs explain
[20:55] <NET||abuse> lifeless,, that's really good, thanks.. think i needed it sort of just summarized like that.
[20:55] <lifeless> jelmer: you pinged me yesterday
[20:56] <jelmer> lifeless, Hi!
[20:56] <NET||abuse> lifeless: what's the <svn> bit, do i need to specify teh repo again?
[20:56] <lifeless> NET||abuse: it should remember it, I was being clear
[20:56] <jelmer> lifeless, Yeah, that was about branch.nick on bound branches - I commented in lp already (and you replied)
[20:56] <lifeless> NET||abuse: also, after a 'bzr merge <svn>' you need to 'bzr commit' to save the result of the merge
[20:56] <lifeless> jelmer: cool
[20:57] <james_w> hey jelmer
[20:57] <jelmer> 'evening James
[20:57] <james_w> jelmer: I saw you said you fixed ptabtools, but you didn't upload anything
[20:57] <NET||abuse> lifeless: ahh, ko.. will any changes from the svn repo be checked for conflicts
[20:57] <lifeless> NET||abuse: of course
[20:57] <lifeless> back in a little bit, breakfast time
[20:57] <jelmer> james_w: I did, but it hasn't shown up on revu for some reason
[20:57] <NET||abuse> so, bzr merge, does a sort of, svn up and svn commit..
[20:57] <james_w> jelmer: ah, I'll ask an admin to have a look
[20:57] <NET||abuse> the bzr commit after the merge does what though?
[20:58] <Odd_Bloke> NET||abuse: A merge can cause conflicts or cause tests to stop working, so you'll want to check that before committing it.
[20:58] <Odd_Bloke> So you need to commit separately.
[20:58] <jelmer> james_w: nevermind
[20:58] <jelmer> james_w, I seem to have uploaded a binary package..
[20:59] <james_w> ah, that'll be it
[20:59]  * jelmer bangs his head against the wall and uploads a source package
[20:59] <NET||abuse> hmm, does the commit send the revision to itself and the remote repo at the same time?
[20:59] <NET||abuse> just not sure at what point files are being written to the remote repo
[20:59] <jelmer> NET||abuse, if you have a bound branch the files are going to the remote repo on "bzr commit"
[21:00] <james_w> jelmer: actually, I've just noticed, your version number should be -0ubuntu1, not -1ubuntu1
[21:00] <jelmer> NET||abuse, if the branch is not bound, the changes are local only and you have to push them to the remote repo using "bzr push" (or dpush)
[21:00] <lifeless> NET||abuse: when you 'push' they are written to the svn repo
[21:00] <lifeless> NET||abuse: using the workflow summary I gave
[21:01] <jelmer> james_w, Thanks, I'll fix that as well.
[21:01] <lifeless> NET||abuse: if you use 'bzr co' instead of 'bzr branch' then ignore all bzr stuff and just use your exact same svn commands
[21:02] <NET||abuse> lifeless: thanks.. ok, i think i get it fairly completely now.. this is awsome.
[21:02] <NET||abuse> i will get it in a linear sense, then in a day i'll try pulling code between my 2 workstations (laptop/desktop) without touching the central svn and do a real branch style work flow
[21:04] <james_w> jelmer: -0ubuntu1 not -1ubuntu0 :-)
[21:04] <NET||abuse> lifeless: thank you very much, NfNitLoop: you too
[21:04] <NET||abuse> ok, i'm late for a dinner.
[21:04] <james_w> jelmer: and is there a need for bzrtags?
[21:05] <NfNitLoop> lifeless: I don't want rebase?
[21:05] <NfNitLoop> If you do a 'push' back into svn with non-linear history, bzr does a replace on that branch in svn, which breaks history.
[21:06] <jelmer> NfNitLoop, it's not the non-linear history that's the problem
[21:06] <NfNitLoop> which would make my coworkers unhappy. :p
[21:06] <NfNitLoop> jelmer: no?
[21:06] <jelmer> NfNitLoop, it's the fact that the mainline sometimes changes
[21:06] <NfNitLoop> oh, that's what I meant, was I misusing terminology?
[21:07] <jelmer> NfNitLoop, non-linear history means merge revisions and indented revisions in "bzr log"
[21:07] <NfNitLoop> in short:  if Joe has committed 10 changes to svn, and suddenly my 'bzr push' blows them away (from the perspective of svn log) he's going to be unhappy.
[21:07] <jelmer> NfNitLoop, changing of mainline can also happen if you e.g. uncommit and then commit another revision
[21:07] <NfNitLoop> jelmer: ah, but that's what happens when I merge in from svn, no?
[21:08] <NfNitLoop> I mean, I just did "bzr merge" when someone had committed a few changes to svn and I had local changes committed.
[21:08] <NfNitLoop> and I got a merge revision w/ indented revisions.
[21:08] <jelmer> NfNitLoop, yes, if you merge from a svn location and then push that means the mainline of the svn mainline changes
[21:09] <NfNitLoop> so to avoid doing that I should... use rebase?
[21:09] <NfNitLoop> or something else?
[21:10] <jelmer> james_w: there were actually older debian changelog entries (from my private debian repo) in there as well - I've removed those now
[21:10] <NfNitLoop> (I'm curious... because I'm eventually going to be making the case for using bzr to coworkers.)
[21:11] <james_w> jelmer: also, I was wondering how that Makefile rule ever worked :-)
[21:11] <jelmer> NfNitLoop, yes, bzr rebase was written with that use case in mind
[21:11] <NfNitLoop> jelmer: Ok.  So I should just always 'bzr rebase' instead of 'bzr merge'?
[21:11] <jelmer> james_w, the shared library on e? :-)
[21:11] <NfNitLoop> (I tried merge then rebase, but that seems to make it think there's no rebase work to do?)
[21:12] <jelmer> NfNitLoop, if you don't want to change the existing mainline in svn, then yes
[21:12] <NfNitLoop> Ok. :)
[21:12] <NfNitLoop> jelmer: thanks!
[21:13] <jelmer> james_w, It probably works if the shared library is already there
[21:13] <jelmer> james_w, so it's sort of a chicken-egg problem :-)
[21:13] <james_w> jelmer: heh, yeah, that'll probably do it.
[21:15] <lifeless> NfNitLoop: alternatively
[21:15] <lifeless> NfNitLoop: have a second branch which is a checkout of the svn mainline; cd to that, update, merge your branch, commit
[21:16] <lifeless> NfNitLoop: it will not alter the mainline at all, ever.
[21:16] <lifeless> NfNitLoop: and has less to do that rebase
[21:18] <NfNitLoop> lifeless: but that way I lose my local revision history, no?
[21:18] <NfNitLoop> all of my local revisions will appear as a single merge.
[21:18] <NfNitLoop> that then gets pushed to svn?
[21:19] <lifeless> unless you use dpush IIRC
[21:19] <lifeless> bbiab
[21:19] <lifeless> I think my point was that regular bzr does this too
[21:19] <jelmer> no, dpush' behaviour doesn't differe there
[21:19] <lifeless> jelmer: hmm, would be good to allow some way to set the branch flag to rpeserve mainline, for svn branches
[21:20] <lifeless> jelmer: and perhaps have it on by default for svn branches
[21:20] <jelmer> lifeless, problem is, you don't want to preserve the mainline of the local branch
[21:20] <jelmer> but that of the branch you're merging
[21:21] <NfNitLoop> lifeless: regular bzr does that too, but still maintains the history.  In bzr I can see that, yes, that merge actually merged in X number of changesets (each still containing its original commit message)
[21:21] <jelmer> NfNitLoop, bzr-svn can preserve that history too but it wouldn't show up in the same svn branch
[21:21] <NfNitLoop> jelmer: *nod*
[21:22] <jelmer> NfNitLoop, just like it doesn't show up in the same bzr branch, but bzr has a nice way to show merged revisions (by indenting them), svn doesn't
[21:22] <NfNitLoop> mostly, I'm getting at this:   If we use bzr to commit to /trunk, and bzr "breaks" 'svn log' on trunk... bzr will get blamed for it.
[21:22] <NfNitLoop> fair or not. :)
[21:23] <NfNitLoop> so I'm just thinking of the best workflow to document & recommend to coworkers.
[21:28] <NfNitLoop> I think I may compromise and go w/ lifeless's recommendation of merging into a local mirror of trunk then committing that.
[21:28] <NfNitLoop> one less plugin to install.  one less plugin for me to explain to newbies. :p
[21:42] <lifeless> jelmer: I mean to honour the dont-alter-history setting on the svn side:)
[21:43] <jelmer> lifeless, we already have that - the append_revisions_only setting
[21:43] <jelmer> I guess it would indeed be nice to have bzr merge honor it
[21:43] <lifeless> jelmer: how do you set that on a svn branch?
[21:43] <lifeless> jelmer: I don't mean 'on a branch pulled via bzr-svn'
[21:43] <lifeless> I mean literally on the svn branch
[21:43] <jelmer> lifeless, you set it in locations.conf
[21:44] <lifeless> jelmer: that won't default to on
[21:44] <lifeless> jelmer: which is what I proposed above
[21:44] <lifeless> jelmer: nor do other users see it
[21:44] <jelmer> I don't think this sort of thing should be a default
[21:44] <lifeless> jelmer: which is what I was assuming when I said it would be nice to allow this
[21:44] <jelmer> unless it's a default for bzr branches as well
[21:45] <lifeless> jelmer: back in a few, local interrupt
[21:45] <jam> lifeless: I'm heading out to pick up my son. If I'm not back for the standup, can you have poolie/whoever call me on my cell phone?
[21:54] <lifeless> jam: sure
[21:54] <lifeless> jelmer: I think that branches in an svn repository have some expectations held by svn client user
[21:54] <lifeless> *users*
[21:54] <lifeless> jelmer: that flag makes bzr meet those expectations; I don't see why it would be an issue to have it default on for SVNBranch
[21:55] <jelmer> lifeless, I don't think it's a bad idea to suggest this setting to users
[21:56] <jelmer> but I don't think having a non-changing mainline is particularly expected by all svn users
[21:56] <jelmer> especially now with merge tracking support in 1.5
[21:56] <lifeless> jelmer: the problem is every user needs to read the right docs and set the right setting
[21:56] <lifeless> jelmer: what proportion would you say?
[21:58] <jelmer> lifeless, or e.g. people who embed a svn branch in their bzr project (by-value nested trees)
[21:58] <lifeless> jelmer: I don't see why it would affect those people
[21:58] <jelmer> lifeless, or e.g. people who push their bzr work to a svn repository but have their primary branch in bzr
[21:59] <jelmer> lifeless, bzr merge will do funny things to them if they merge in a new copy of the svn branch
[21:59] <jelmer> lifeless, since it will change to use the mainline from the remote svn branch
[21:59] <lifeless> it will ?
[21:59] <jelmer> well, that's what this setting would do, no?
[21:59] <lifeless> no
[21:59] <lifeless> it would only affect the svn branches
[21:59] <lifeless> not anything thts been copied into bzr
[21:59] <lifeless> the setting doesn't propogate
[22:00] <jelmer> yes, but what happens if you "bzr merge <remote-url>" ?
[22:00] <lifeless> you merge its content
[22:00] <jelmer> and if you then push it, the revision changes !?
[22:00] <lifeless> that setting is totally unrelated to merge
[22:00] <lifeless> when you push, then the push will fail if that setting is set
[22:00] <lifeless> because it enforces svn-like behaviour
[22:00] <poolie> hello all
[22:01] <jelmer> so you're basically suggesting append_revisions_only defaulting to yes?
[22:01] <lifeless> jelmer: for SVNBranches only
[22:01] <lifeless> jelmer: I'm suggesting two things: allow setting it in svn, not just in locations.conf (so that the admin can choose a setting); and defaulting it on for SVNBrach
[22:01] <lifeless> *SVNBranch*
[22:01] <lifeless> hi poolie
[22:02] <lifeless> poolie: jam says use his mobile for the call
[22:02] <lifeless> poolie: if he's not answering on skype
[22:02] <jelmer> defaulting it to on would make sense, but we need to make sure to indicate properly why the push is failing
[22:02] <lifeless> jelmer: sure, improving the error text shouldn't be hard
[22:08] <poolie> lifeless, jam, spiv, call in 2m
[22:12] <jam> poolie, lifeless: I'm back around, is the call still going?
[22:22]  * NfNitLoop reads scrollback. 
[22:22] <NfNitLoop> I like the idea of making the preserve-history option default to true.
[22:23] <NfNitLoop> I've been using bzr since... 0.6?  ...  and have never even heard of locations.conf :p
[22:24] <NfNitLoop> and breaking 'svn log' history is a definite bad impression if someone isn't expecting it, and isn't familiar enough with distributed vs. non-distributed SCMs to know why it happened.
[22:25] <markh> are all the bzr-devs aware of the dvcs/bzr discussions on python-dev?
[22:25] <spiv> markh: yeah
[22:26] <jelmer> NfNitLoop, The thing I'm not sure about is having inconsistent behaviour between svn and bzr branch
[22:27] <NfNitLoop> it's completely consistent while you're in bzr-land.
[22:27] <NfNitLoop> the only inconsistency is when you try to push it back into svn.
[22:28] <NfNitLoop> and things do work differently in svn-land.
[22:29] <markh> I decided its best to not throw "windows support" into the mix for that thread at this time ;)
[22:29] <jelmer> NfNitLoop, it's consistent in both, except svn doesn't have a good way to display it before svn 1.5 and svn users aren't used to it
[22:30] <NfNitLoop> jelmer: aah, I'm still in 1.4 land, so I'm not entirely sure how 1.5 support looks for all of that.
[22:31] <NfNitLoop> though...
[22:31] <NfNitLoop> my initial reading of 1.5 features makes me very wary for using it for anything more complicated than one-off branches.
[22:31] <NfNitLoop> there are different commands for trunk->branch and branch->trunk merging....
[22:31] <NfNitLoop> and once you've merged back to the trunk, you have to throw away your branch.  (!?)
[22:32] <jelmer> yes, just like you do with bzr
[22:34] <NfNitLoop> You don't have to throw it away, though...
[22:34] <NfNitLoop> you could keep using the branch and then merge back into trunk again and again.
[22:34] <NfNitLoop> (whether or not that's a great workflow)
[22:35] <NfNitLoop> So it looks like svn is just tracking r# when the branch is created, and tracking each merge from trunk into branch... then replaying diffs to merge back into main -- NOT actually keeping all revision IDs of what's been merged.
[22:36] <NfNitLoop> (This is my guess after reading the svn book page on branching & merging in 1.5)
[22:36] <NfNitLoop> (but I admit it wasn't all that thorough.)
[22:36] <jelmer> it does keep track of what revisions were merged in 1.5
[22:36] <NfNitLoop> Oh? Ok.
[22:36] <NfNitLoop> Makes me wonder why you have to throw your branch away then.
[22:37] <jelmer> well, you don't have to - you can have a single feature branch
[22:37] <uws> jelmer: fwiw, gnome svn has been upgraded to svn 1.5 repositories recently
[22:39] <NfNitLoop> So, when bzr-svn pushes into svn1.5, does it actually show merge history?
[22:39] <NfNitLoop> I don't have a 1.5 install around to play with. :p
[22:40] <jelmer> yes, it should be able to show which revisions were merged (if they are also present in the svn repository)
[22:41] <NfNitLoop> aah.  but not if they're bzr-local.
[22:42] <lifeless> jelmer: I don't think that inconsistency is bad
[22:42] <lifeless> jelmer: because svn branches are by definition not bzr branches
[22:43] <jelmer> lifeless, it makes behaviour less predictable
[22:43] <lifeless> jelmer: you can't tell a-priori whether a bzr branch has that setting either
[22:43] <jelmer> lifeless, e.g. it means we'd need a hack to be able to run the bzr repository testsuite (if it'll ever be generic enough) against svn repositories
[22:44] <jelmer> lifeless, you can know that if you create a bzr branch yourself it doesn't
[22:44] <lifeless> jelmer: I don't think we'd need a hack for that;
[22:44] <lifeless> jelmer: I think we'd just call the api setting to set it off for any test that needs it off
[22:46] <NfNitLoop> re: "predictable" behavoir:  breaking my svn history was unpredictable.  ;)
[22:46] <NfNitLoop> well, at least, I didn't see it coming.
[22:52] <lifeless> spiv: so
[22:52] <lifeless> spiv: what inner loop was it in?
[22:52] <emmajane> Instruction question... https://wiki.ubuntu.com/Training/KnowledgeBase#Launchpad%20and%20Bazaar Step # 4, Create your own branch doesn't seem to make sense to me. Can someone confirm that I'm just full of fail? :)
[22:53] <lifeless> emmajane: looks the same as 3 to me
[22:54] <emmajane> lifeless, Ok.
[22:54] <emmajane> lifeless, would that even work?
[22:54] <emmajane> I'm pretty sure it's "wrong" to include Step #4, but I'd hate to alter good instructions. ;)
[22:54] <lifeless> no idea; I'd suggest stepping through the docs end to end; thats what I do to validate docs
[22:54]  * emmajane nods. kay.
[22:54] <lifeless> it certainly looks odd to me
[22:55] <emmajane> maybe I'm not completely crazy then. Cool. :)
[22:55] <emmajane> lifeless, thanks :)
[22:55] <lifeless> np
[22:55] <spiv> lifeless:  http://rafb.net/p/58nFen92.html
[22:56] <jam> So... the installer for bzr-setup-1.9rc1-3 is built and tested on my machine, I just have to get an upload to LP to finish
[22:57] <jam> It takes about 8 minutes to upload 15MB on my connection
[22:57] <jam> and it has timed out once already.
[22:57] <markh> yeah, that sucks :(
[22:58] <markh> well - I don't see timeouts I don't think - just occasional launchpad errors
[22:58] <markh> jam: any idea about that bug re the qt binaries failing to load?
[22:59] <jam> markh: for whatever reason in the -2 build QtGui4.dll is only 6MB instead of 10MB
[22:59] <jam> my best guess is that a copy got cancelled
[22:59] <markh> right - weird!
[22:59] <jam> and that the build process though the half-copy was up-to-date
[22:59] <markh> right
[22:59] <jam> the new build script I wrote nukes the target
[22:59] <jam> and rebuilds everything
[22:59] <jam> so I should be immune to that
[22:59] <markh> ideally it would also nuke the 'build' dir from every sub-project too
[23:00] <jam> markh: hmm... I suppose
[23:00] <jam> I actually could just nuke them completely
[23:00] <markh> or exact same thing could happen during *its* build
[23:00] <jam> The script already is smart enough to download the latest releases, etc.
[23:00] <lifeless> spiv: uhm
[23:00] <lifeless> spiv: my likne 2183 is empty space :P
[23:01] <lifeless> *line*
[23:01] <markh> a re-download each build might be going a little too far though?
[23:01] <jam> markh: We have shared repos for that
[23:01] <lifeless> markh: why? reproducable :>
[23:01] <jam> so it is mostly just nuking the working dir
[23:01] <jam> and then connecting to LP, seeing it has everything, and checking out a WT
[23:01] <markh> right - I thought you also meant all the python packages (qt, crypto, etc)
[23:02] <markh> the bzr based stuff sounds reasonable
[23:02] <jam> anyway, I'm done for tonight. You can find the script in C:\home\shared\bzr\releases\build_release.py if you want to give it a look
[23:02] <jam> and I know the build-release for bzr uses "build_ext -i -f" to force a rebuild of everything
[23:02] <jam> I'm not sure if we would want to do that for the others
[23:03] <markh> I don't quite trust distutils --force
[23:04] <markh> (in the same way I don't trust --dry-run to not do anything in all cases ;)
[23:04] <jam> markh: as an aside, the icon overlays aren't working for me
[23:04] <jam> is that a known problem for Vista?
[23:05] <markh> in the binary version?  It should work - but most likely is .bzr.log is reporting it would "recurse to death" - in which case a fix has been pushed
[23:05] <markh> (and assuming its a 32bit vista)
[23:07] <jam> markh: I don't see tbzr putting much of anything into .bzr.log
[23:07] <markh> jam: right - so no icon in the taskbar at all?
[23:07] <jam> I guess I see:
[23:07] <jam> 0.172  return code 3
[23:07] <jam> 134.107  opening working tree 'C:/Users/jameinel/dev/bzr/bzr.dev'
[23:07] <jam> 134.138  opening working tree 'C:/Users/jameinel/dev/bzr/bzr.dev'
[23:07] <jam> It does have a taskbar icon
[23:07] <jam> though it takes a *long* time to show up
[23:07] <jam> and if I kill it
[23:07] <jam> it will eventually show up again
[23:07] <markh> yeah - it will come back next time an icon overlay or menu is requested
[23:08] <markh> (that is also when it *first* appears)
[23:08] <jam> ah, well something is causing it to not actually get results
[23:09] <markh> fyi, tbzrcache.exe can be run explicitly from a cmdline to see what it is doing.  '-v' and --log-level=debug might be useful options.  You will need to shutdown the existing one before a new one will start.
[23:09] <markh> (tbzrcachew.exe is a gui version that is automatically run)
[23:09] <markh> but I'll grab the binary and have a play too
[23:11] <lifeless> spiv: hello?
[23:12] <spiv> lifeless: I think that's line 2111 in your branch
[23:13] <lifeless> sure, I'm just puzzled at the difference :P
[23:13] <spiv> lifeless: I had bzr.dev merged in from an earlier attempt to get apples-to-apples comparisons on network behaviour
[23:13] <lifeless> ah right
[23:13] <lifeless> push that somewhere if there were conflicts :P
[23:13] <lifeless> I can benefit from that
[23:14] <spiv> lifeless: but to be sure nothing wacky was going on, I just tried reproducing it with your pristine branch, and got a traceback instead!
[23:14] <lifeless> spiv: remember my tip may be bust
[23:14] <lifeless> go back 1
[23:14] <spiv> Ah, ok.
[23:14]  * spiv does that
[23:14] <lifeless> anyhow
[23:14] <lifeless> so this function is currently full-inventory every time
[23:17] <spiv> Yeah, going back one fixed that.
[23:18] <lifeless> spiv: that function could be better cast like the loop I put into fetch
[23:18] <lifeless> to delta all the inventories in a chain
[23:19] <lifeless> and only do a full inspect of the first one if the basis parent of the first is inaccessible
[23:20] <lifeless> spiv: I think that would be quite managable to do and fix the major inefficiency
[23:27] <mDuff> Is a working bzr-svn package available for current Ubuntu (Intrepid)? The one apt finds for me is too old for the Ubuntu-installed bzr 1.8.
[23:28]  * mDuff is playing dumb user for the moment.
[23:29] <mDuff> ...blerg, probably ought to upgrade to the New Shininess (bzr 1.9), and break from sticking with the Ubuntu packages anyhow.
[23:30]  * mDuff finds the bzr PPA on Launchpad... oooh, didn't know 'bout that.
[23:32] <lifeless> mDuff: :)
[23:34] <mDuff> ...odd; "apt-get install bzr-svn" is still trying to install an older one, rather than the new package from the bzr-beta PPA.
[23:35] <mDuff> ...oh, the bzr-svn in the beta PPA still depends on bzr <<1.7~ :(
[23:40] <emmajane> jelmer, downloading is push?
[23:40] <jelmer> emmajane, no, but it's the same bug
[23:40] <emmajane> kay :)
[23:40] <jelmer> emmajane, It's a bug in the progress bar handling
[23:40] <emmajane> ahhh
[23:41] <emmajane> do you work on olive?
[23:41] <emmajane> I tried using this afternoon for the desktop course and none of us could get it to even open a downloaded branch for the course. :(
[23:41] <Odd_Bloke> emmajane: How were you trying to open the branch?
[23:41] <emmajane> Odd_Bloke, by clicking on the directory in Olive.
[23:42] <emmajane> Odd_Bloke, massive freeze.
[23:42] <emmajane> Odd_Bloke, no problems for other projects though
[23:42] <jelmer> emmajane, no, though I work on most other things in bzr-gtk
[23:42]  * emmajane nods
[23:42] <Odd_Bloke> emmajane: Right, I think you're meant to run it from within the branch.
[23:42] <emmajane> I know our team would *love* to have something other than the CLI for the project.
[23:43] <jelmer> olive could use some attention :-/
[23:43] <Odd_Bloke> Yeah.
[23:43] <emmajane> jelmer, mostly I know how to submit bug reports. :/
[23:43] <emmajane> Odd_Bloke, how do you run it from within the branch? it's got a button under applications to start it ;)
[23:43] <jelmer> emmajane, that's still very useful
[23:43] <Odd_Bloke> emmajane: From the CLI. >.<
[23:43] <emmajane> jelmer, three today. :/
[23:43] <emmajane> Odd_Bloke, nono, there's a button in the applications menu.
[23:44] <emmajane> Odd_Bloke, you sholdn't have a button if it doesn't work. :)
[23:44] <Odd_Bloke> Right, I'm saying there shouldn... yeah.
[23:45] <Odd_Bloke> I'm too CLI-centric to want Olive, TBH.
[23:45] <Odd_Bloke> I don't even used gvim. :p
[23:45] <emmajane> I have a button for gvim.
[23:45] <emmajane> control-S is wrong though, so I never use it
[23:45] <Odd_Bloke> :D
[23:46] <emmajane> jelmer, let me know if there's anything else I can do to help out with it.
[23:46] <jelmer> emmajane, thanks, will do :-)
[23:47] <emmajane> btw, it freezes if i open it from the command line as well... and then about 10 seconds later it kicks in
[23:47] <Odd_Bloke> emmajane: How big is the branch you're running it on??
[23:47] <emmajane> 500M
[23:47] <Odd_Bloke> That probably explains the pause.
[23:47] <Odd_Bloke> Though doesn't really justify it.
[23:48] <emmajane> lots of pauses as I open folders and move around, but it's not freezing now
[23:48]  * emmajane tries to remember which dir has 125 PNGs in it...
[23:48] <emmajane> found it. :)
[23:48] <Odd_Bloke> Leave Britney^WOlive alone!
[23:49]  * emmajane waits a long time.
[23:49] <emmajane> hehe
[23:49] <emmajane> and there we go
[23:50]  * emmajane wonders if beuno hates her yet too :)
[23:50] <lifeless> spiv: does that make snese?
[23:50]  * emmajane promises to go back to work tomorrow and stop reporting bugs. :)
[23:51] <beuno> emmajane, you're one of my favorite people now!
[23:51]  * emmajane chuckles.
[23:52] <emmajane> beuno, i think I stopped at four.
[23:52] <beuno> emmajane, you did. 3 of them are assigned to me already  :)
[23:52] <emmajane> hehe
[23:53] <beuno> you will have to start hiding once I start to produce mockups, cauise I'll hunt you down and squeeze all the information I can out of you!
[23:53] <emmajane> beuno, you're welcome to ping me whenever you'd like!
[23:53] <emmajane> I would love to not be complaining about launchpad anymore.
[23:53]  * beuno double checks he's logging the conversation
[23:54] <emmajane> :)
[23:54] <beuno> emmajane, we have a big re-design ahead of us
[23:54] <emmajane> beuno, good :)
[23:54] <beuno> I want to stop complaining as well  ;)
[23:54] <emmajane> heheeh
[23:55] <james_w> beuno: I used your merge proposal redesign the other day
[23:55] <james_w> beuno: I like it
[23:55] <beuno> james_w, ah!  great to hear!
[23:55] <beuno> I still have loads of things to improve on it
[23:55] <beuno> but it's a pretty massive change  :)
[23:55] <james_w> I would have your babies right now if it came with a multi-branch loggerhead thing to review with though
[23:56] <james_w> and built-in PQM
[23:56] <beuno> hahahahaha
[23:56] <beuno> well
[23:56] <beuno> PQM + Launchpad is coming soon
[23:56]  * emmajane chuckles.
[23:56] <james_w> woo!
[23:56] <beuno> code is there, needs a proper UI
[23:56] <james_w> that's your job!
[23:56] <beuno> and, diffs in merges is very far ahead as well
[23:56]  * james_w cheers
[23:57] <beuno> so I'll plan for my babies around Jan/Feb
[23:57] <james_w> I think you'll be receiving crates rather than beers from me at UDS
[23:57] <beuno> haha
[23:57] <Odd_Bloke> When/where is UDS?
[23:57] <lifeless> dec san fran
[23:57] <beuno> I'll make sure I eat often
[23:57] <lifeless> right after fosscamp
[23:57] <lifeless> http://www.fosscamp.org/
[23:57] <james_w> it took me ages to find the thing to mark one as merged though
[23:58] <beuno> james_w, they get marked automagically
[23:58] <james_w> oh, great!
[23:58] <james_w> I was just going to point out that it should be possible :-)
[23:58] <beuno> once LP scans either the target or proposed branch, it finds the revid, and bingo
[23:58] <beuno> same will happen with branch statuses
[23:59] <james_w> woo!
[23:59]  * james_w dances
[23:59] <beuno> there is a patch floating around
[23:59] <beuno> yeah, we did a lot of work in the 1 week sprint
[23:59] <james_w> now just for bug statuses and I'll be happy
[23:59] <beuno> plus a ton of notes  :)
[23:59] <beuno> what about bug statuses?
[23:59] <james_w> like Fix Committed when it's in the development focus