[01:26] <markh> jelmer: ping
[01:41] <jelmer> markh, pong
[01:42] <markh> hi jelmer - a couple of things - the 0.4 branch is still what I should be using for windows 1.7 builds?
[01:43] <jelmer> markh, yep
[01:43] <markh> and have you seen the thread on the mailing list talking about the crash?
[01:43] <markh> identifying the env var etc
[01:43] <jelmer> the DLL mess you mean?
[01:43] <markh> well - its not that bad as it turns out
[01:44] <markh> "As it happens, on my system APR_ICONV_PATH was set to c:\program files\subversion\iconv (perhaps created by a SVN installer)"
[01:44] <markh> from the most recent mail in the thread on the mailing list
[01:44] <markh> which leads to question 2 1/2 :) he suggests using svn 1.5 libs would be a good option
[01:45] <jelmer> using 1.5 instead of 1.4 would always be a good idea
[01:46] <markh> and I believe he is suggesting we just ensure to clobber that env var before loading svn
[01:46] <markh> bb in 20 mins...
[01:47] <jelmer> I'm not sure that's going to work in all situations
[01:47] <jelmer> I'm getting some sleep in a few minutes
[01:48] <jelmer> markh, Will be back tomorrow CET time
[03:51] <mwhudson> poolie: i don't want to be a pest, but are you going to land http://bundlebuggy.aaronbentley.com/project/bzr/request/%3Ce01316480809050416r72b3f857h4a664a2cc4e157b0%40mail.gmail.com%3E ?
[03:52] <poolie> !!
[03:52] <poolie> i did send it
[03:52] <poolie> hm that kind of sucks
[03:53] <mwhudson> i don't think it's in bzr.dev, it's possble bb is out of date
[03:57] <mwhudson> doesn't seem to have landed
[03:59] <poolie> it may have bounced
[03:59] <poolie> we'll need to get that into 1.7
[04:00] <poolie> thanks for reminding me
[04:00] <mwhudson> yes please :)
[05:03] <fullermd> Hm.  http://live.gnome.org/Bazaar says "For best performance, install 1.6b2."
[05:10] <jml> what do I need to do to serve bzr+http files from my apache?
[05:11] <jml> or is there a bzr plugin that will create a mini-server that will do the same?
[05:12] <poolie> jml, i think there is documentation on how to hook it up in mod_python
[05:12] <poolie> but also tim will know about the mini-codehost project
[05:12] <poolie> which may help
[05:13] <jml> well, it's mostly for testing purposes.
[05:13] <jml> so even something launched a toy web server with support would be fine for me.
[05:19] <jml> poolie: how do you spell NULL_REVISION for weave repos?
[05:20] <poolie> it should be the same, but old code might use None
[05:20] <jml> thanks.
[05:30] <Peng_> jml: bzr+http is very simple to set up. Take the script from the documentation, put it somewhere, then use mod_rewrite or something to point "foo/.bzr/smart" to it.
[05:30] <Peng_> It just uses WSGI, so you can use CGI or FastCGI or mod_wsgi or whatever you want.
[05:31] <jml> Peng_: the docs are in the bzr source tree?
[05:31] <Peng_> jml: http://doc.bazaar-vcs.org/bzr.dev/en/user-guide/index.html#serving-bazaar-with-fastcgi
[05:32] <jml> Peng_: thanks.
[05:53] <seyacat> hi bazaarmens
[05:53] <seyacat> please a question
[05:53] <MolePrince> Evening.  Does bzr support 'forests' well.  ie, I have a dozen projects that I want to maintain individual revisions/logs for, but be able to update/check them all out en masse if I work on a different terminal for instance.
[05:54] <seyacat> i make a copy of brach, and in copy i make a new file.  I use "bzr merge" to update the parent branch, but merge not copy new file
[05:55] <jml> MolePrince: the bzrtools plugin has 'multi-pull', which may be close to what you want.
[05:55] <MolePrince> jml: Thanks, I'll look into that.
[05:55] <jml> seyacat: did you do 'bzr add <new-file>' and then 'bzr commit' in the copy?
[05:56] <seyacat> yes i make "add" and "commit"
[05:56] <jml> seyacat: hmm. can you pastebin the exact commands you used, starting with the copy?
[05:59] <seyacat> where is pastebin?
[06:00] <seyacat> F:\bzrtest\copia>bzr status
[06:00] <seyacat> unknown:
[06:00] <seyacat>   prueba/newfile.txt
[06:00] <seyacat> F:\bzrtest\copia>bzr add prueba
[06:00] <seyacat> added prueba/newfile.txt
[06:00] <seyacat> F:\bzrtest\copia>bzr commit
[06:00] <seyacat> Committing to: F:/bzrtest/copia/
[06:00] <seyacat> added prueba/newfile.txt
[06:00] <seyacat> Committed revision 4.
[06:00] <seyacat> F:\bzrtest\copia>bzr merge
[06:00] <seyacat> Merging from remembered parent location F:/bzrtest/repo1/
[06:00] <seyacat> Nothing to do.
[06:00] <fullermd> MolePrince: You're looking for "nested trees".  In-progress feature.
[06:01] <seyacat> it can be a windows bzr problem?
[06:02] <fullermd> seyacat: No, you're doing it backward.  That bzr merge is trying to merge new stuff FROM repo1 INTO copia.  If you want to merge new stuff FROM copia INTO repo1, you need to cd ../repo1 ; bzr merge ../copia
[06:04] <seyacat> yes i can see ita work... and what is difference with pull command?
[06:05] <fullermd> Roughly speaking, 'pull' updates your copy of another branch.  'merge' merges in changes from the other branch into yours.
[06:11] <seyacat> aaa i cant make chain checkouts branches
[06:13]  * fullermd doesn't know what that means...
[06:13] <seyacat> im very very happy, i understand bzr at last,  its nicer than i think
[06:14] <seyacat> ty fullermd , you are a good suppot
[06:14] <fullermd> np  :)
[06:16] <poolie> well done
[06:29] <jml> the links to modpywsgi are broken :(
[06:39] <jml> ok, now I'm getting this error when I try to branch via bzr+http: http://paste.ubuntu.com/45643/
[06:41] <Verterok> jml: I think it's 1.6b4 problem
[06:41]  * jml updates that branch
[06:41] <Verterok> it's fixed in 1.6.1
[06:43] <jml> ok, now I'm just getting 'not a branch', which makes me happier.
[06:52] <poolie> markh, re bug 251701 do you know if paramiko is sure to be in the windows binaries now?
[06:53] <markh> hrm
[06:55] <markh> poolie: that surprises me.  it was in the first few builds I made, and it seems be in the last 1.6 build I made
[06:56] <poolie> it may be invalide
[06:56] <poolie> or user erro
[06:56] <poolie> i'd appreciate if you could look
[06:56] <vila> Good morning Bazaar
[06:59] <spiv> vila: Good morning.
[06:59] <jml> why do the http smartserver instruction say to do: Alias /srv/example.com/scripts/bzr-smart.fcgi /srv/example.com/scripts/bzr-smart.fcgi
[06:59] <jml> that seems... redundant.
[07:02] <markh> poolie: it is also certainly there in 1.6.1 and 1.7 builds
[07:03] <markh> I suspect that bug was user error, but just marking it as "fixed" would work.
[07:14] <jml> small doc patch on the list.
[07:14] <jml> massive review karma going for cheap
[07:15]  * markh just had the brain-wave he could close it, so did...
[07:17] <spiv> jml: because I suck at configuring apache
[07:17] <spiv> jml: I twiddle until it works, and then I twiddle some more to tidy up, and things mysteriously break, so I undo that and just go with messy-but-working and get on with more important things.
[07:18] <jml> oh right.
[07:18] <jml> spiv: well, my patch makes things a little better.
[07:39] <jml> Peng: so, I'm looking at a development copy of Launchpad, mirroring branches from a bzr+http server on my own laptop. it makes a lot of posts to .bzr/smart (as well as a few GETs of control and format files) -- I'm guessing that's bzr+http mirroring.
[07:45] <Peng_> jml: What?
[07:45] <jml> Peng_: you know about bug https://bugs.edge.launchpad.net/launchpad-bazaar/+bug/245918?
[07:45] <Peng_> Right. You got it working?
[07:45] <Peng_> Could someone do me a favor? Try to 'bzr branch http://bzr.mattnordhoff.com/bzr/imports/lighttpd/trunk/' outside of a shared repo with 1.7rc1 or bzr.dev. I get a format error (it's a rich-root-pack branch). nosmart+http works fine though, and so does branching locally.
[07:46] <jml> Peng_: well, I *think* it's working, but what I want to know is "how do I know when it's working"
[07:46] <speakman> I think I found a bug in bzr-svn: Can someone try to branch svn://uclibc.org/trunk/buildroot ?
[07:46] <spiv> POSTs to .bzr/smart is smart-over-http, yes.
[07:47] <jml> Peng_: I get http://paste.ubuntu.com/45655/
[07:47] <jml> spiv: thanks.
[07:51] <Peng_> jml: Yeah, that's what I get. That would be a bug, then.
[07:53] <Peng_> Actually, I have the same issue in 1.6.1.
[07:54] <Peng_> And "bzr info" says it's an unnamed format, so now I'm totally confused.
[07:55] <spiv> "bzr info -v" is helpful when it says unnamed.
[07:56] <spiv> It's often just format 5 branch with pack-0.92 repo.
[07:58] <Peng_> I have a format 7 branch in a rich-root-pack repo.
[07:58] <Peng_> Is that not how rich-root-pack is supposed to be?
[07:58] <Peng_> Which is branch 7?
[07:58] <mwhudson> 'supports stacking'
[07:58] <spiv> It allows stacking, IIRC.
[07:59] <Peng_> Oh, really? How did I get that?
[08:00] <Peng_> Hmm, my local copy is Branch6 as it should be.
[08:01] <Peng_> So...when does bzr decide to upgrade remote branches to support stacking?
[08:02] <Peng_> And then not upgrade the repo?
[08:02] <spiv> I would have assumed when --stacked is used.
[08:02] <spiv> But that doesn't explain your situation.
[08:02] <Peng_> spiv: It tries to upgrade the source branch?
[08:02] <spiv> Check your ~/.bzr.log?
[08:02] <Peng_> I may have tried to branch --stacked from it as a test, but I don't even remember doing that.
[08:02] <spiv> I'm not very familiar with this end of things.
[08:03] <Peng_> spiv: What do I grep for? "Upgrading random branch for fun"?
[08:03] <spiv> AFAIK bzr never silently tries to upgrade anything.
[08:03] <Peng_> Right.
[08:03] <spiv> Maybe search your bzr.log for --stacked?
[08:03] <Peng_> Oh, good idea.
[08:03] <Peng_> I am not in the mood for figuring out weird bzr issues. Hmph.
[08:04] <spiv> So far I'm guessing it's a weird user issue ;)
[08:04] <Peng_> Grep says I have never used --stacked in my life.
[08:05] <Peng_> At least in the history of my .bzr.log
[08:06] <Peng_> OK, here we go. According to .bzr.log.old, I twice tried to "bzr branch --stacked" that branch locally.
[08:06] <Peng_> But it upgrades the source branch? That makes no sense.
[08:06] <spiv> I doubt it.
[08:07] <spiv> Maybe you simply pushed the local --stacked branch up?
[08:07] <spiv> And bzr push then preserved the format (but not the stacking)?
[08:07] <Peng_> Both times I tried to "bzr branch --stacked", it yelled at me about IncompatibleRepositories.
[08:07] <spiv> In that case, no idea.
[08:07] <speakman> Is the bzr-svn stable, or should I expect vital bugs?
[08:07] <Peng_> :)
[08:08] <Peng_> I scared poolie away. :(
[08:08] <spiv> Maybe check .bzr.log on the server in case someone is did something strange on that end?
[08:08] <spiv> Peng_: nah, he just hopped on his bike and went home.
[08:08] <Peng_> spiv: I did check .bzr.log on the server.
[08:08] <spiv> speakman: it's been stable for me
[08:08] <spiv> Peng_: dang.  There goes another idea.
[08:08] <Peng_> On my PC, I have never once used --stacked in the history of .bzr.log and .bzr.log.old.
[08:08] <speakman> spiv: could you please try to branch svn://uclibc.org/trunk/buildroot ?
[08:08] <Peng_> On the server, I twice tried to locally "bzr branch --stacked" that branch, and both times it failed.
[08:09]  * Peng_ blames poolie since he's not here. :)
[08:09] <spiv> speakman: I don't have the spare bandwidth for that atm, unfortunately :(
[08:10] <Peng_> How big is that branch? I have spare bandwidth, but not spare RAM. :)
[08:10] <spiv> speakman: are you getting an error?
[08:10] <spiv> Peng_: the current bzr-svn release is much friendlier RAM-wise.
[08:11] <spiv> (compared to early this year)
[08:13] <Peng_> True
[08:13] <Peng_> beuno: ping?
[08:14] <speakman> UnicodeDecodeError: 'utf8' codec can't decode bytes in position 85-87: invalid data
[08:15] <speakman> can I make bzr run under ipython, and use ipython's more informative traceback?
[08:17] <speakman> it seems to crash in the "insert_revprop" function in logwalker.py
[08:17] <speakman> self.cachedb.execute("replace into revprop (rev, name, value) values (?, ?, ?)", (rev, name.decode("utf-8"), value.decode("utf-8")))
[08:19] <a7p> Verterok: everyting fine - just tested it.
[08:19] <Verterok> a7p: great! :)
[08:19] <a7p> should I confirm the bug fixed on lp?
[08:20] <a7p> Verterok: thanks a fixing it.
[08:20] <Verterok> a7p: no, I marked it as "Fix committed", but if you encounter any problem, just reopen it :)
[08:21] <Verterok> a7p: np, this new build also comes with some extra goodies ;)
[08:21] <a7p> Verterok: I saw that, that's why I asked if futher confirment is needed.
[08:22] <a7p> Verterok: I am going to check it out as soon as I finished this (ugly) flash project I am suffering from at the moment.
[08:22] <Peng_> beuno: Never mind about the ping. I just filed bug 268867.
[08:22] <Verterok> a7p: ok, just let me know if you hit that bug again
[08:23] <a7p> Verterok: I will.
[08:30]  * Verterok is off to bed, g'night all
[08:35] <mwhudson> Peng_: i guess the breadcrumbs should key off SCRIPT_NAME somehow
[08:36] <Peng_> mwhudson: Oh, hi. :)
[08:37] <Peng_> mwhudson: They should do it the same way every other link does, cuz they get it right.
[08:39] <Peng_> I came to the computer to do one thing, decided to read the mailing list and pull new changes out of boredom, and here I am an hour later. Oops.
[08:41] <Peng_> OK, so how do I upgrade from branch7 to branch6?
[08:41] <Peng_> "bzr upgrade" doesn't know how. Should I init/pull (I hate that hack).
[08:41] <Peng_> ?
[08:42] <mwhudson> bzr upgrade --packs should do it i think
[08:43] <Peng_> "bzr upgrade --rich-root-pack errored out with "No converter".
[08:44] <spiv> speakman: file a bug on bzr-svn in launchpad, jelmer is pretty good at fixing those sorts of things
[09:12] <Peng_> Whew. Over 1 hour, 2355 bytes of IRC conversation, 1 bug and 1 long mailing list post later, I can get back to watching my depressing anime. Why did I decide to pull bzr.dev and get into all this? :(
[09:12] <Peng_> (Also, why did I spend 10 minutes counting my IRC conversation?)
[09:12] <Peng_> Anyway, good night everyone.
[09:15] <spiv> Peng_: good night :)
[09:15] <spiv> Peng_: thanks for the bug hunting!
[10:20] <awilkins> Hmm : bug: bzr rocks does not respect "-q" argument
[10:21] <awilkins> I'm looking for quiet optimism
[10:21] <awilkins> :-P
[10:37] <robsta> hi
[10:37] <robsta> all of a sudden bzr seems unwilling to play with bzr-playground
[10:37] <robsta> the error message is "bzr: ERROR: exceptions.AttributeError: 'ProtocolThreeDecoder' object has no attribute '_in_buffer'"
[10:38] <robsta> Bazaar (bzr) 1.6.1, from the ppa
[10:40] <robsta> can i go after this problem somehow?
[10:53] <visik7> hi
[10:53] <visik7> bzr-svn has some problem today
[10:53] <visik7>  bzr-svn/ra.c:919: undefined reference to `py_svn_log_entry_receiver'
[10:53] <visik7> what is it ?
[10:54] <lifeless> robsta: ttry 1.7 perhaps, that looks likea local error, bt will tell you
[10:54] <lifeless> visik7: no idea sorry, jelmer: may know
[10:54]  * lifeless waves bye
[10:54] <visik7> ok solved
[10:55] <visik7> sorry
[10:55] <robsta> lifeless: it may be related to the problem Jc2k just pinged you about, in #gnome-bzr
[10:57] <visik7> I've this error using latest svn-bzr and latest svn http://dpaste.com/77418/
[10:57] <visik7> sorry
[10:57] <visik7> latest svn-bzr and latest bzr
[10:58] <uws> visik7: That message is pretty clear, isn't it?
[10:59] <visik7> I've the latest version
[10:59] <uws> visik7: The bzr-svn you have is not compatible with the bzr version you have. you're running development code. stick to bzr 1.6
[10:59] <uws> visik7: "the latest version" is 1.6.1
[10:59] <uws> not 1.8dev
[10:59] <Jc2k> does anyone have the address for canonicals rt?
[11:00] <visik7> uws: my bzr is bzr.dev http://bazaar-vcs.org/bzr/bzr.dev/
[11:01] <lifeless> Jc2k: #canonical-sysadmin should
[11:01] <lifeless> *really* gone now, sorry
[11:01] <uws> visik7: What's the problem?
[11:02] <visik7> uws: I dunno I think I was on the devel branch of both bzr and bzr-svn
[11:02] <uws> visik7: perhaps bzr-svn lags behind a bit?
[11:03] <uws> visik7: it's called "devel branch" for a reason, namely that it is NOT stable or 'guaranteed' to work
[11:03] <visik7> oh
[11:03] <visik7> I misunderstood the error
[11:03] <visik7> nevermind :P
[11:03] <uws> visik7: ok
[12:00] <poolie> vila, hi
[12:00] <vila> poolie: hi !
[12:24] <poolie> night!
[12:27] <hersonls> I need register my RSA key fingerprint, where i do this?
[12:27] <hersonls> i dont remeber
[12:28] <Peng_> hersonls: Is this a Launchpad question?
[12:28] <hersonls> yes
[12:28] <Peng_> hersonls: There are instructions on the website, and you should ask in #launchpad anyway
[12:29] <hersonls> sorry :]
[12:30] <Peng_> hersonls: https://help.launchpad.net/YourAccount/ImportingYourPGPKey
[12:35] <Peng_> Huh, if I branch an old knit branch over "http", with the bzr+http smart server available, it does most things over plain http and is very quick. If I explicitly use "bzr+http", it's extremely slow, makes a huge number of requests, and doesn't use plain http.
[12:36] <Peng_> God, you guys *really* killed knits.
[12:37] <fullermd> I'm vaguely tempted to see what happens if I access a weave branch over bzr+ssh...
[12:37] <fullermd> Just out of sick, sick curiosity.
[12:38] <Peng_> Hahaha.
[12:41] <Peng_> Hey, it's done! It only took 7 minutes and made 1085 requests! That's not excessive, right?
[12:42] <fullermd> Heck, that's not even 3 a second.  What a yawner.
[12:42] <Peng_> You know, for a 2 MB branch with about 200 revisions?
[12:43] <fullermd> That's 5 per revision...  a fifth of 2 megabytes is about 400k...   who'd want to do more than 400k a request either?
[12:44] <fullermd> (I realize the math may not SEEM right, but...)
[12:45] <Peng_> fullermd: You're actually right. ...Wait, no you're not. The responses were 435 *bytes*, not 435 KB.
[12:45] <fullermd> So, see how awesome bzr is?  It's giving you 1000:1 compression!
[12:46] <fullermd> Imagine how much slower it would be if it did more per request.
[12:46] <Peng_> Say I file a bug, everyone ignores it for 4 months (ahem), and then when someone finally notices it, it's not reproducable anymore. How should it be resolved? Fixed or invalid?
[12:47] <fullermd> Ideally, the third choice.
[12:47]  * fullermd drags his "dammit, I need to write a BTS" out of the closet.
[12:50] <Peng_> It took precisely 128 days for someone to notice the bug. Nice coincidence. :)
[12:52] <Peng_> I wonder what Bazaar's average bug response time is?
[12:54] <fullermd> I'll bet the standard deviation is huge.
[13:51] <awilkins> Ranges from people who filed bugs so they have something to put a patch in, to bugs that were filed early on and never touched, I'll wager
[15:00] <proyvind> hoia
[15:00] <proyvind> anyone knows if it's possible to export just a directory of a repo?
[15:01] <proyvind> I have a directory in my git repo which I want to export and import to bzr
[15:03] <gimaker> Can anyone help me with an error I get when trying to branch a branch from LP: Permission denied (publickey).
[15:03] <gimaker> bzr: ERROR: Connection closed: please check connectivity and permissions (and try -Dhpss if further diagnosis is required)
[15:04] <gimaker> I can branch (bzr branch lp:peekabot) from my desktop, but not from my laptop
[15:04] <gimaker> and they use the same bzr version (1.3.1, Ubuntu 8.04)
[15:06] <ericvw> gimaker: Are you using public-key authentication?
[15:06] <gimaker> ericvw, I have my pubkey uploaded to LP, yes
[15:07] <gimaker> but it never even asks for the passphrase on my private key
[15:07] <gimaker> (it does on my desktop, though - where it works)
[15:07] <ericvw> how are you branching from your laptop?
[15:07] <gimaker> "bzr branch lp:peekabot trunk"
[15:08] <ericvw> Sorry, i am a bit unfamiliar with lp:peekabot syntax; could you clarify that for me?
[15:09] <gimaker> lp:peekabot is the same as lp:///~peekabot-devs/peekabot/trunk
[15:09] <gimaker> e.g. it's the "main branch of development"
[15:11] <ericvw> I don't `lp` is a URL identifier that bazaar supports
[15:11] <gimaker> it's supported if you have the Launchpad plugin, which is bundled with bazaar
[15:13] <ericvw> the only thing I can think of is maybe you have two different private keys going on for some reason.
[15:13] <gimaker> ericvw: yeah, that was it... I just figured it out too
[15:13] <gimaker> Uploaded my 2nd one and now it works ok!
[15:13] <ericvw> gimaker: Sorry, I am a bit new to Bazaar, but I know public-key quite well :D
[15:14] <gimaker> I still think it's lame that a valid key is needed to branch from a public branch, though :)
[15:15] <ericvw> is the lp:blah syntax just for the launch pad plugin?
[15:15] <Odd_Bloke> gimaker: It shouldn't be.
[15:16] <gimaker> Odd_Bloke: What should be? That I needed a pubkey to branch?
[15:16] <gimaker> ericvw: yes
[15:17] <ericvw> gimaker: Thanks, I should read up on that!
[15:17] <Odd_Bloke> A valid key shouldn't be needed to branch from a public branch.
[15:17] <Odd_Bloke> Though I suspect that's a known bug once you've launchpad-login'd.
[15:18] <gimaker> Odd_Bloke: exactly...
[15:19] <Odd_Bloke> If you remove 'launchpad_username' from ~/.bazaar/bazaar.conf, it should no longer be needed.
[15:20] <gimaker> Odd_Bloke: well, now that I uploaded a second pubkey it should be ok anyway :)
[15:21] <gimaker> Though it's still being annonying and demand that I type my passphrase twice after spewing ut an annoying "Server is too old for streaming pull, reconnecting.  (Upgrade the server to Bazaar 1.2 to avoid this)" message :)
[15:21] <Peng_> gimaker: You should upgrade your client.
[15:21] <gimaker> ...which is obviously wrong, seeing as LP runs 1.6rc3 :)
[15:22] <Peng_> Or...I don't know. I think upgrading the client is the right solution.
[15:22] <gimaker> Peng_: I prefer to stick with what's in Ubuntu
[15:22] <Peng_> gimaker: What's in Ubuntu?
[15:22] <gimaker> 1.3.1
[15:22] <Peng_> gimaker: Well, I'm 99% sure upgrading the client will fix that
[15:23] <gimaker> Either way, it's no biggie and I should stop whining already
[15:23] <ericvw> You can add the Bazaar repository into sources for apt-get to have it get the latest and greatest!
[15:23] <Peng_> Yeah. See https://launchpad.net/~bzr/+archive if you want to do that.
[15:24] <gimaker> ericvw: does that work with all plugins as well? I use bzrtools, bzr-svn and rebase on a regular basis
[15:25] <gimaker> s/all plugins/all plugins available in stock ubuntu/
[15:25] <Peng_> gimaker: A couple plugins are available in the PPA, but not that many.
[15:25] <ericvw> I believe bzrtools is
[15:25] <Peng_> gimaker: I think it's just bzrtools and bzr-gtk now. bzr-svn used to be.
[15:25] <jelmer> John uploads bzr-svn to PPA now as well
[15:26] <Peng_> Ah.
[15:26] <Peng_> That's good. :)
[15:27] <gimaker> maybe when I have some more time on my hands (I'm leaving for the first field-test of our robot in half an hour... a bit stressed atm :)
[15:32] <gimaker> jelmer: does using the new dpush mean I won't have rebase as soon as someone else does a (non-conflicting) commit on the SVN-repository?
[15:33] <jelmer> gimaker, No, you will still have to rebase or pull
[15:33] <Gioacchino> hey all
[15:34] <Gioacchino> I need help configuring bzr-eclipse
[15:34]  * beuno pokes awilkins 
[15:35] <gimaker> jelmer: ok. bummer :) rebase everytime someone else commits it is then!
[15:35] <jelmer> gimaker, That's not specific to bzr-svn though, you would have to do the same thing if you were using a native bzr branch that other people committed to as well
[15:37] <Gioacchino> I need help configuring bzr-eclipse someone help me ?
[15:38] <Gioacchino> I should put  lp:~gmazzurco89/mastro/work  in branch location ?
[15:38] <gimaker> jelmer: hmm? surely you must jest. native bzr branches I can just pull/merge.
[15:39] <jelmer> gimaker, there shouldn't be anything preventing you from pulling/merging in the current situation either
[15:40] <Gioacchino> some one can help em ?
[15:41] <gimaker> jelmer: merging was fine, but I couldn't push without rebasing - since I had local commits "in between" the two latest revisions in the SVN repos. I thought maybe there was some fancy new feature that automated that for me.
[15:41] <jelmer> gimaker: You would have to rebase or merge again too if you were in a native bzr branch
[15:43] <gimaker> jelmer: yeah, I'd have to merge (or rebase, but that hardly seems necessary). But for svn-branches rebase is the only option, right?
[15:44] <rockstar> jam, friendly reminder about TWiB today.  I've got alerts going everywhere to make sure I don't forget.
[15:45] <Gioacchino> statik can you help me please ?
[15:45] <Gioacchino> statik someone tell me you use eclipse
[15:45] <jelmer> rockstar, Hi!
[15:45] <rockstar> jelmer, hello
[15:46] <jelmer> gimaker, No, merge should work too for bzr-svn but not if you use dpush
[15:47] <jelmer> gimaker, well, it would work but it would do strange (though correct) things to your history
[15:47] <Gioacchino> can some body help em please ??'
[15:47] <jelmer> rockstar: were you looking into creating a mirror of the KDE svn repo?
[15:47] <jelmer> Gioacchino, sorry, none of the bzr-eclipse folks appear to be around
[15:48] <jelmer> Gioacchino, Verterok is the person you'd want to talk to
[15:48] <Gioacchino> ok
[15:49] <gimaker> jelmer: really? What I did was: branch from svn, local commit, local commit, someone else commits to svn, local commit, bzr push. bzr push failed saying the branches had diverged, so I merged from the SVN-repo and then push failed with some error message about "changing the order of commits already in the SVN repository"
[15:49] <rockstar> jelmer, I was, but we put it off because it was looking like it would be a while before they started the process of choosing a new VCS.
[15:49] <gimaker> It also suggested rebase to solve the problem
[15:49] <jelmer> gimaker: Ah, you're committing directly to the repository root?
[15:50] <gimaker> jelmer: yeah. does that make a difference?
[15:50] <jelmer> gimaker, Yep, that means you can not change the order of revisions on the mainline
[15:50] <jelmer> gimaker, and if you "bzr merge" and then push that means the mainline changes
[15:50] <gimaker> jelmer: which makes sense :)
[15:50] <jelmer> gimaker, since some revisions from upstream that you merged will disappear from mainline
[15:50] <jelmer> rockstar, Ah, ok
[15:51] <jelmer> rockstar, bzr-svn now supports branching KDE branches pretty quickly fwiw
[15:51] <rockstar> jelmer, I think we could probably set it up easily if we had a box.
[15:51] <rockstar> jelmer, I saw that actually.
[15:52] <gimaker> jelmer: gotta run, but thanks for the help
[15:52] <jelmer> rockstar, some of the Kubuntu folks were asking about mirrors earlier so I think there's some interest in it
[15:52] <jelmer> gimaker, you're welcome
[15:52] <rockstar> jelmer, well, then let's find a box.
[16:13] <pinotree> hello
[16:14] <ree> Hi! Does bzr join --reference actually works somewhere?
[16:14] <pinotree> is there a way to "export" some local commits, as in diff+commit log?
[16:14] <pinotree> (like git am does)
[16:14]  * pinotree is not a git fan
[16:15] <Odd_Bloke> pinotree: What do you want to do with said exported commits?
[16:16] <pinotree> emailing them to another bzr user
[16:16] <Odd_Bloke> pinotree: 'bzr send'.
[16:17] <pinotree> i saw that; i was wondering if there was a way to just get the "result", instead of being sent directly
[16:18] <jelmer> pinotree, "bzr send -o <filename>"
[16:19] <pinotree> d'oh, i misread the documentation... you're right, jelmer, thanks
[16:19] <pinotree> i understood like not specifying -o would give the result to stdout, stupid me
[16:19]  * pinotree headdesks
[16:23] <ree> Anyone is using bzr join --reference and can confirm it works?
[16:29] <pinotree> hmm but bzr send does not include the commit log, does it?
[16:30] <jelmer> pinotree, it does
[16:31] <pinotree> oh?
[16:31] <jelmer> Pieter, it includes all revision metadata
[16:31] <jelmer> s/pieter/pinotree/
[16:31] <Pieter> hah!
[16:31] <jelmer> Pieter, sorry!
[16:31] <Pieter> :)
[16:31] <jelmer> ree: It does work, but is experimental
[16:31] <pinotree> jelmer: is it the bundle?
[16:32] <jelmer> pinotree, It's the base64-encoded data at the bottom of the genreated file
[16:32] <pinotree> oh i see
[16:32]  * pinotree is a bzr noob
[17:02] <lalo> quickie -- how do I upgrade a branch with subtrees?  1.6.1 doesn't know the old format, 1.5 doesn't know the new format
[17:03] <pinotree> jelmer: another question.. if i ask for a diff of a revno range (eg revno:x..revno:x+4), does the bundle include the commit logs for all of the commits?
[17:03] <jelmer> pinotree, yes
[17:04] <lalo> or is there a way to discard the information that the branch has subtrees?  I don't mind losing that info, the sub-branch only has one revision which I can replicate
[17:04] <jelmer> lalo: 1.6.1 doesn't k now the old format?
[17:05] <lalo> bzr: ERROR: Unknown repository format: 'Bazaar development format 0 with subtree support (needs bzr.dev from before 1.3)\n'
[17:06] <jelmer> lalo: Ah, a development format... I guess you could upgrade to "pack-0.92-subtree" with bzr 1.5
[17:06] <lalo> or maybe there's another format that supports subtrees and that's what I'm missing... but I think I tried all formats that 1.5 supports
[17:07] <lalo> aha, succeeds
[17:07] <lalo> yeah and 1.6 can read it :-D awesome, thanks
[17:07] <lalo> guess I assumed 0.92 had to be older than development?
[17:07] <jelmer> yes, it is older
[17:08] <jelmer> you'd never want to use development unless you're interested in hacking on bzr
[17:08] <jelmer> and losing data once in a while
[17:09] <lalo> so pack-0.92-subtree is the current "supported" format with subtrees?
[17:10] <Verterok> Gioacchino: hi
[17:14] <jelmer> lalo, yep
[17:15] <jelmer> lalo, you're actually using subtrees?
[17:15] <lalo> just started, as an experiment
[17:16] <lalo> I have projects that need subtrees much more seriously than this one
[17:16] <Verterok> Gioacchino: maybe you are hitting a problem related to Bug #121936
[17:16] <Verterok> Gioacchino: try to use a ssh-agent/pagent to handle ssh keys
[17:18]  * jelmer would love to see subtrees move out of "experimental" status
[17:21]  * lalo does grep -i development `find ~SRC -wholename \*/.bzr/repository/format`
[17:21] <lalo> will take a while with all the crud I have tho :-P
[17:22] <abentley> lalo: There are no supported subtree formats.  It is not a supported feature.
[17:23] <lalo> ok
[17:24] <lalo> yeah I figured it out -- I started using development-subtree because of all the bzr-svn branches I have, then one day a stupidity bug bit me and I thought "I should try subtrees myself"
[17:24] <lalo> ah well
[17:26] <abentley> lalo: Good god, why would you use developement-anything?
[17:27] <lalo> development-repo.html says, possibly incorrectly, that bzr-svn requires development-subtree
[17:27] <abentley> lalo: That is false in serious ways.
[17:28] <lalo> "The development-subtree format is required for the bzr-svn plug-in but should otherwise not be used until the subtree feature is complete within bzr."
[17:28] <fullermd> Poor wording.
[17:28] <jelmer> lalo, It does say that you only need development-subtree if you want to use bzr-svn *and* assist in testing development formats
[17:29] <lalo> well, I do have an interest in testing development formats :-P
[17:29] <lalo> but up until today my "testing" has mostly been "it works perfectly, never had any grief" -- only problem I ever had was today, that the new bzr doesn't read the old development formats, and that was easily solved
[17:30] <abentley> lalo: Development format are "if it breaks, you get both pieces".
[17:30] <lalo> and all the stuff I have on dev format is either well backed up, or stuff I don't really care if I lose it (like checkouts of third-party svn stuff from sourceforge and google-code)
[17:31] <lalo> if my nxhtml branch is lost, boo-haa, I'll rm -r it and branch another :-P
[17:33] <abentley> lalo: so your answer to "Why would you use development-anything" should be "I have an interest in testing development formats".
[17:34] <lalo> well I thought I did, but "testing" was so uneventful that I ended up forgetting about it
[17:34] <lalo> which I suppose is a nice thing
[17:34] <jelmer> clearly we need less stable "experimental" formats
[17:34] <lalo> absolutely
[17:35]  * lalo goes write a "blow up randomly" plugin
[17:35] <fullermd> They have that; it's called VSS.
[17:35] <lalo> I could use libephem to crash only on full moon nights
[17:35] <jelmer> :-)
[17:37] <lalo> actually bzr is one of the very few projects I stopped contributing to because it's doing so well that I felt my time would be better spent elsewhere :-D
[17:37] <lalo> (not entirely true -- there's also the fact that it's doing so well, every time I try to do something, there are too many new concepts to wrap my brain around)
[17:39] <Gioacchino> thanks Verterok
[17:47] <Verterok> Gioacchino: using an ssh-agent solved your problem?
[17:47] <Gioacchino> Verterok:  I am looking for how to use ssh agent
[17:48] <Verterok> Gioacchino: are you on *nix or windows?
[17:48] <Gioacchino> linux
[17:48] <Gioacchino> kubuntu
[17:49] <Verterok> Gioacchino: there is a nice utilily: sshaskpass (gtk), that promtps you a GUI asking your passphrase
[17:49] <Gioacchino> mm gtk :P
[17:50] <Gioacchino> I install it
[17:50]  * Verterok don't know the package name in *buntu
[17:51] <Gioacchino> is this? under X, asks user for a passphrase for ssh-add
[17:51] <Verterok> Gioacchino: also there is ksshaskpass (KDE) ;) http://hanz.nl/p/program
[17:52] <Verterok> but the site seems to be down :(
[17:52] <Gioacchino> yes..
[17:52] <Gioacchino> than I have not probelm with gtk..
[17:52] <Gioacchino> eclipse is also gtk :P
[17:53] <Gioacchino> than it now work ?
[17:53] <Verterok> Gioacchino: for the moment, there are soem efforts to port SWT to QT
[17:53] <Gioacchino> eclipse to QT ?? WOW I am very happy :D
[17:54]  * Verterok is waiting too :)
[17:55] <Gioacchino> sshaskpass look very  ugly
[17:55] <jelmer> Verterok, what's the status of having xml output supported by bzr by default rather than in a plugin?
[17:55] <Gioacchino> but work
[17:55] <Gioacchino> very thanks ;)
[17:56] <Verterok> Gioacchino: the gtk version?
[17:56] <Verterok> Gioacchino: np, glad to help
[17:56] <Gioacchino> in repo I found  Xsshaskpass
[17:56] <Gioacchino> without gtk
[17:56] <Gioacchino> now I am tryng gnomesshaskpass
[17:57] <Verterok> jelmer: it ca n be included ATM, but I'ld like to avoid the current code duplication
[17:57] <Gioacchino> gnomeask is better
[17:57] <jelmer> Verterok: Ok, just curious if it's worth packaging bzr-xmloutput for Debian or waiting
[17:58]  * beuno votes for packaging
[17:58] <Gioacchino> Verterok in eclipse cvs I have the option to add to .cvsignore the file how to with eclipse bazar ?
[17:59] <Gioacchino> or I should edit manually .bazar ?
[17:59] <Verterok> jelmer: I don't know the debian process, but I think it worth the packaging right now
[17:59] <Verterok> Gioacchino: you hitted a not-yet-implemented (tm) feature
[17:59] <Verterok> Gioacchino: for the moment, you could use: 'bzr ingore <your file>' from the CLI
[18:00] <Gioacchino> ok
[18:00]  * Verterok waves beuno
[18:00] <Gioacchino> you know if soon is implemented this feature?
[18:00]  * beuno waves back at Verterok 
[18:00]  * Verterok is wainting a review of his resibmitted LH branch ;)
[18:01] <Verterok> s/resibmitted/resubmitted/
[18:01] <jelmer> Verterok, ok
[18:01] <beuno> Verterok, logging?
[18:01] <jelmer> Verterok, I'll wait for loggerhead and bzr-search to be accepted into Debian, then I'll have a look
[18:02] <jelmer> Verterok, Is it possible to build the eclipse plugin without actually starting eclipse?
[18:02] <Verterok> jelmer: the main blocker to include it in bzr, is a refactoring to some commands (status, ls, etc). So xmloutput can plugin without duplicate code
[18:02] <jelmer> Verterok, ah, ok
[18:02] <Verterok> jelmer: yes :D, I have a headless build setup
[18:03] <Verterok> jelmer: it's still in early stages (a.k.a needs a lot of manual config to setup)
[18:04] <jelmer> Verterok, Ah, ok - that would be a requirement for debian packaging as well
[18:04] <Verterok> jelmer: it only requirement is a eclipse install with PDE (plugin development environment)
[18:05] <Verterok> beuno: yeap, logging
[18:06] <beuno> Verterok, I'll get to it soon-ish. I'd also recommend poking mwhudson, he may have good insights on it as well
[18:06] <jelmer> Verterok, Ah, ok. That's doable
[18:06] <Verterok> Gioacchino: it could be implemented soon, please file a bug :)
[18:07] <Verterok> bzr-eclispe is bug-driven :p
[18:07] <Gioacchino> than is more fast if I submit this bug ;)
[18:07] <Peng_> jelmer: What's the story with bzr-svn and bzr.dev? Last I checked, neither 0.4 or trunk were compatible with bzr 1.8dev.
[18:08] <Verterok> Gioacchino: indeed
[18:08] <Gioacchino> hihihi Verterok
[18:08] <jelmer> Peng_, they are, I added a compatibility marker for 1.7 last night
[18:08] <Peng_> jelmer: On both?
[18:09] <jelmer> Peng_, yes
[18:09] <jelmer> not sure if I've pushed those changes yet
[18:09]  * Verterok is sseeking LH reviewers for his logging branch
[18:09] <Verterok> mwhudson: ping ^
[18:09] <jelmer> let me push just to be sure
[18:09] <Peng_> jelmer: I don't think you have.
[18:12] <Peng_> jelmer: Thanks
[18:34] <abeaumont> hi, i'd like to merge 2 repos, is this possible?
[18:34] <pygi> hi phanatic :p
[18:36] <abeaumont> i mean, i have two independent repos, one svn and one bzr, i want to use svn through bzr-svn and integrate the other bzr repo with the svn one throug bzr-svn, is it possible?
[18:45] <jelmer> abeaumont, yes
[18:46] <jelmer> abeaumont: If you have a branch of the svn repository, you should be able to copy the bzr branch into it and then merge it into the svn branch using "bzr join"
[18:47] <abeaumont> jelmer: ahhh, join is the command i was looking for, thanks!
[18:56] <Gioacchino> how to remove a folder from launchpad branc ?
[19:13] <abeaumont> jelmer: is there a limitation in the number of trees joinable? i'm trying to join a second tree but it fails
[19:25] <fattymattyo> is it possible to run a bazar server rather than svn?
[19:26] <fattymattyo> without using launchpad... it would be for a company website
[19:26] <fattymattyo> sorry if I'm asking a stupid question
[19:27] <nDuff> fattymattyo, it's pretty straightforward to run your own bzr server... what beyond that are you asking?
[19:28] <nDuff> fattymattyo, you don't *really* even need to run a dedicated "bzr server"; bzr can use dumb filestorage (a shared directory, sftp, WebDAV, etc) in place of a dedicated server, though the latter will perform better..
[19:29] <fattymattyo> nDuff I would probably want to put it on the same machine as the webserver
[19:29] <nDuff> fattymattyo, you can have the webserver just share a directory over plain WebDAV (no bzr-specific plugins or configuration needed) if you're going for the simplest-possible setup.
[19:30] <fattymattyo> I just wasn't clear if I could use it on its own without hooking up to launchpad
[19:30] <nDuff> fattymattyo, yes; you certainly can.
[19:30] <fattymattyo> thanks nDuff, I'll go look up how to do it, just wanted to make sure I wasn't trying to do something that's impossible.
[19:31] <Peng_> fattymattyo: Are you using *nix?
[19:31] <fattymattyo> I am, and the server will be unix
[19:31] <fattymattyo> one of the users checking in/out will be mac and the other will be you using a yet to be determined OS
[19:31] <Peng_> fattymattyo: Then the simplest thing to do is use plain old HTTP for read-only access and ssh or sftp for read-write access.
[19:32] <fattymattyo> I really want them to be able to read and write to bzr
[19:32] <fattymattyo> I want to use it as a version control system the way its meant to be
[19:32] <Peng_> fattymattyo: You can use SFTP without installing anything (other than OpenSSH, obviously) on the server. bzr+ssh requires installing bzr too.
[19:32] <nDuff> *nod*; bzr will work as a revision control system over HTTP or SSH/SFTP.
[19:32] <nDuff> ...you just give it a http or sftp address when checking things in/out.
[19:33] <fattymattyo> are there any bzr packages for windoze?
[19:33] <nDuff> fattymattyo, yes, bazaar-vcs.org has packages to download for win32
[19:33] <fattymattyo> awesome
[19:33] <fattymattyo> you guys are great
[19:54] <rockstar> jam, TWib?
[20:07] <jam> rockstar: probably have a phone call and a bunch of email to catch up on... but yeah, we need to do something
[20:08] <rockstar> Okay, when's good for you?
[20:10] <fm> hm anybody using bzr on opensuse factory?
[20:10] <fm> the dependencies of the package say python < 2.6 is this really required?
[20:10] <fm> i have python-2.6b3-2
[20:26] <lifeless> fm: dunno, I'm not aware of 2.6 breaking bzr, but I haven't explicitly tried
[20:31] <fm> lifeless: running the shiped version i get http://pastebin.ca/1200339
[20:31] <fm> that is bzr-1.5-26
[20:44] <vila> lifeless: It looks like pqm is hung on success: merge lp:~vila/bzr/bzr.integration/ http://bazaar-vcs.org/bzr/bzr.dev ?
[20:45]  * vila wonders if lifeless is really up at that time...
[21:32] <vila> fm: sha module deprecated surely rings a bell
[21:34] <vila> Months ago I tried python2.6, there was a couple of minor fixes needed, if you're still interested let's try to talk about it tomorrow (it's 22h30 here and I will leave soon)
[21:34]  * vila failed to notice: 'fm has quit (Read error: 104 (Connection reset by peer))', bah, may be he'll read the logs
[22:01] <ozzloy> " This branch was probably created by bzr 0.15pre.  Create an empty file to silence this message."  what file am i supposed to create?
[22:04] <ozzloy> i tried creating the directory "branch" with the file "tags" but that doesn't suppress the message
[22:35] <james_w> ozzloy: .bzr/branch/tags ?
[22:35] <ozzloy> i'll give that a shot
[22:36] <ozzloy> nope
[23:40] <jam> lifeless: PQM appears wedged.
[23:40] <jam> Looking at http://pqm.bazaar-vcs.org/
[23:40] <jam> I see it claim that it successfully completed the merge
[23:40] <jam> but now it is just sitting their
[23:40] <jam> vila: what did you do to it :)
[23:55] <jml> I've got a bzr riddle for you all.
[23:56] <jml> given a standalone branch, how can I make another empty standalone branch with the same format?
[23:56] <beuno> cp?   :)
[23:56] <jml> _empty_
[23:56] <beuno> ah, ah
[23:57] <spiv> jml: e.g. a format 5 branch with a pack-0.92 repo?
[23:57] <jml> spiv: yes.
[23:57] <jml> spiv: if that's what the first branch is.
[23:57] <spiv> jml: I did this just yesterday.
[23:57] <beuno> cp and uncommit til your fingers bleed?   :)
[23:57] <jml> spiv: but the first branch might also be a loom or a weave branch or something pathological.
[23:57] <jml> beuno: full marks for creativity :)
[23:57]  * beuno stops making noise
[23:58] <spiv> jml: bzr init-repo --pack-0.92 temp-repo --no-trees && bzr init --dirstate-tags temp-repo/temp-branch && bzr branch temp-repo/temp-branch empty-branch
[23:58] <mwhudson> spiv: that's not a standalone branch
[23:58] <mwhudson> oh i see
[23:58] <spiv> jml: you can use "bzr info -v" in the empty branch to verify it worked.
[23:58] <jml> how would you do that in API terms?
[23:58] <spiv> mwhudson: yeah.  I should add "rm -r temp-repo" to that.
[23:59] <jml> (please don't say cmd_init_repo().run(...))
[23:59] <spiv> jml: oh, API terms.  That's easier, in a sense.
[23:59] <spiv> (less concise to type on IRC, though)
[23:59] <mwhudson> jml: we can init ok, right?