[00:01] <lifeless> mkanat: sorry, more local interrupts
[00:02] <lifeless> mkanat: and we have a group dinner tonight
[00:02] <lifeless> mkanat: I'll pick this up tonight or tomorrow evening - I have caught up more fully with poolie on this
[00:02] <mkanat> lifeless: Okay.
[00:27] <dOxxx> mmm steak. bbl.
[00:28] <GaryvdM> dOxxx: I've read your finale patch. I looks very good. I think I should try make qbzr ext diff tools config similar to the new mergetool config.
[00:30] <GaryvdM> jam: bundel is at 37%, so I think it working :-)
[00:30] <GaryvdM> *bundle
[01:39] <dOxxx> I return, steakified.
[03:01] <peitschie> hi everone :)
[06:20] <Peng> Haha, test subject. Whatever it was, I'm probably not a good test subject ATM. Especially since I was on the road when you were discussing it. :P
[06:23] <mkanat> Peng: For loggerhead users who aren't LP.
[06:23] <Peng> Oh, you're here.
[06:23] <Peng> mkanat: Yeah...I kind of dropped off of the world a while back. I am back on IRC, but I'm not tinkering with Bazaar or Loggerhead again yet.
[06:23] <Peng> Honestly, I'm running lp:loggerhead from May. :D
[06:23] <mkanat> Peng: :-)
[06:24] <Peng> Plus I've hit old age -- you know, 19 -- and I've developed an aversion to breaking things that already work. :P
[06:25] <Peng> Which is not what's stopping me, but it's making me even less motivated in general. :P
[06:26] <fullermd> You'll get over that in a few years.
[06:26] <fullermd> Mostly, 'cuz you'll get to the age that you realize nothing ever gets as far as really 'working' anyway...
[06:26] <mkanat> Peng: Hahaha. :-)
[06:30] <Peng> fullermd: :(
[06:32] <fullermd> It's too early to get frowny.  That was the GOOD news!
[06:59] <Peng> fullermd: It's good news that nothing ever works?
[07:00] <Peng> fullermd: Don't worry, I'm not frowny. I just watched Evangelion, so I'm incapable of any emotions but excitement and vague confusion.
[07:00] <Peng> Woah, half an hour's gone by? Seems I'm also unable to track time properly. :P
[07:01] <mkanat> Peng: Hahahaha!
[07:02] <mkanat> Peng: Yes, that's pretty much like a Pan-Galactic Gargle Blaster.
[07:02] <mkanat> Peng: Did you just watch the *end* of Evangelion?
[07:02] <Peng> mkanat: No, 2.0! \o/
[07:03] <mkanat> OMG, didn't even realize that had happened!
[07:03] <spm> it ends? I got confused after about 6 or so episodes. "is there , in fact, a plot here that bears a vague resemblance to ... a plot" ;-)
[07:03] <mkanat> spm: Hahahaha. There's more of one than there is for FLCL....
[07:04] <mkanat> spm: Also, really, the first 12 episodes of Eva are where they're *nice* to you.
[07:04] <spm> wheee
[07:04] <mkanat> spm: That's where they make you like the characters so that later they can punish you for ever thinking they were good people.
[07:04] <Peng> Wait, 6 or so episodes? You didn't see Asuka?
[07:04] <spm> they failed then. I was starting to despise them all.
[07:04] <mkanat> spm: lol
[07:04] <spm> don't recall, this would ahve been 10 ish years ago?
[07:05] <mkanat> Anyhow, I'm off to the land of Hulu.
[07:06] <Peng> spm: It has a plot. Only Hideaki ANNO knows what that plot is, but we're pretty sure there is one. Anyway, who cares about that? Watch it for the crazy symbolism, psychological problems and cool fight scenes!
[07:07] <spm> heh. the cool fight scenes, sure. but he teenage angst. urgh.
[07:07] <spm> s/he/the/
[07:07] <Peng> Heh.
[07:07] <Peng> mkanat: Speaking of FLCL, FUNimation's DVD and Blu-ray release comes out in a month.
[07:07] <mkanat> Sweet. 1080p of total nonsense. :-D
[07:07] <Peng> Upscaled, tho.
[07:08] <Peng> spm: Not a fan of the teenage angst? That could be a problem, yes...
[07:08] <mkanat> Okay, for reals now, though. Off to TV I go.
[07:08] <spm> later mkanat
[07:08] <mkanat> Later! :-)
[07:09] <spm> Peng: I wasn't a fan of it when I was a teenager. so... ;-)
[07:09] <Peng> Heh
[07:09] <Peng> spm: It's got adult angst too, you know. :D
[07:10] <Peng> Immature adults, but... :P
[07:10] <spm> that's not winning me over!
[07:13] <Peng> Hmm...
[08:32] <Erimos_Wolf> Morning @ all, verterok, are you there?
[11:18] <jdobrien> ssh: connect to host bazaar.launchpad.net port 22: Connection refused
[11:51] <verterok> Erimos_Wolf: Hi
[11:51] <Erimos_Wolf> hi
[11:52] <Erimos_Wolf> sorry for bothering you but i cannot explain the behavior of the eclipse plugin on my machine
[11:52] <verterok> Erimos_Wolf: np, I just wokeup :)
[11:52] <Erimos_Wolf> good morning :-)
[11:53] <verterok> Erimos_Wolf: do you have the latest version of bzr-xmloutput installed?
[11:53] <Erimos_Wolf> 0.8.6
[11:53] <verterok> Erimos_Wolf: and bzr version?
[11:54] <Erimos_Wolf> 2.2.2
[11:54] <Erimos_Wolf> got all that info from the bazaar manager
[11:55] <verterok> Erimos_Wolf: ok, so. I think the behaviour you'r seeing is caused by an old version of xmloutput (IN the nead future, bzr-eclipse will bundle a xmloutput version to avoid this kind of errors)
[11:55] <verterok> Erimos_Wolf: to use bzr-2.2.2, you need the latest trunk of bzr-xmloutput
[11:56] <verterok> Erimos_Wolf: is this windows?
[11:56] <Erimos_Wolf> jupp
[11:56] <Erimos_Wolf> i mean yes
[11:56] <Erimos_Wolf> :-)
[11:58] <verterok> Erimos_Wolf: sadly, I didn't make a release of xmloutput, so there is no installer for trunk
[11:58] <Erimos_Wolf> so iam not too dump to use the software ?
[11:58] <Erimos_Wolf> meant dumb
[11:59] <verterok> Erimos_Wolf: let me try to build one, as an alternative you can branch lp:bzr-xmloutput and replace the installed one with it
[11:59] <verterok> huh?
[11:59] <Erimos_Wolf> thx
[12:04] <verterok> Erimos_Wolf: it will take a while to build the installed, need to update my windows vm :/
[12:05] <Erimos_Wolf> non prob
[12:06] <Erimos_Wolf> take breakfast and a good cup of coffee
[12:11] <Erimos_Wolf> is this considered as a bug?
[12:25] <verterok> Erimos_Wolf: you can get a installer from trunk at: http://launchpad.net/bzr-xmloutput/trunk/trunk/+download/bzr-xmloutput-setup-0.8.7-r149.exe
[12:27] <Erimos_Wolf> unbelievable! Could you explain why this happend?
[12:27] <Erimos_Wolf> please
[12:30] <verterok> Erimos_Wolf: your xmloutput version is very old, and isn't compatible with the bzr version
[12:31] <Erimos_Wolf> ok. But i got it through the bazaar setup right?
[12:31] <verterok> Erimos_Wolf: ah, no idea...it's inscluded in the bzr windows installer?
[12:31] <Erimos_Wolf> jupp
[12:32] <Erimos_Wolf> i just followed the tutorial of bazaar, installed the 2.2.2 win standalone
[12:32] <Erimos_Wolf> got eclipse and installed your plugin (thats al i did)
[12:32] <verterok> Erimos_Wolf: ok, I think this is because the last "stable" release of xmloutput is 0.8.6, but is broken with latest bzr
[12:33] <verterok> Erimos_Wolf: I need to get some time to fix a few more bugs and do a release, in order to get it included in the windows installer
[12:33] <Erimos_Wolf> ok i understand
[12:34] <Erimos_Wolf> until that this "manual" installing of xmloutput is the best workaround?
[12:34] <verterok> Erimos_Wolf: yes :(
[12:40] <Erimos_Wolf> Thank you for the installer and solving my problem :-)
[12:42] <Erimos_Wolf> So this is a dependency problem between the xmloutput and the plugin,which will be fix through the next release xmloutput?
[12:52] <verterok> Erimos_Wolf: eclipse plugin?
[12:52] <verterok> Erimos_Wolf: it's a dependency problem in the bzr windows installer
[13:09] <Erimos_Wolf> oh, ok.
[13:09] <verterok> Erimos_Wolf: is bzr-eclipse working as expected now?
[13:09] <Erimos_Wolf> yesss :-)
[13:09] <Erimos_Wolf> commiting is done without any problems (only local)
[13:11] <verterok> cool
[13:12] <Erimos_Wolf> before i made some changes but there weren recognized "No commits needed"
[13:14] <Erimos_Wolf> the bazaar manager said the opposite
[13:15] <Erimos_Wolf> this is why i came here yesterday
[13:38] <catphish> what protocols can bzr use to push?
[13:39] <catphish> is it just ssh or apache dav?
[13:44] <Lo-lan-do> Smart server over HTTP, too, I think.
[13:45] <catphish> that's what i want
[13:46] <catphish> can's see any onfo on running smart server on http though
[13:56] <james_w> catphish, http://doc.bazaar.canonical.com/bzr.2.2/en/admin-guide/other-setups.html#smart-server-over-http-s
[13:57] <catphish> oh yeah :)
[14:01] <catphish> does the smart server have an executable?
[14:02] <catphish> since my webserver is not python, it will need to accept a request and pipe it to a bzr executable
[14:03] <james_w> catphish, that would be the fcgi method wouldn't it?
[14:03] <catphish> is there an executable provided with stdio? or will i need to wrote my own?
[14:05] <james_w> the bzr command
[14:07] <catphish> so does the smart protocol simply contain bzr commands?
[14:09] <catphish> i'll sniff a few
[14:10] <maxb> It contains a custom message format
[14:14] <catphish> does that message format map to bzr CLI commands?
[14:15] <catphish> or is it more complicated?
[14:19] <catphish> maxb: is the http protocol documented by any chance?
[14:27] <maxb> catphish: Not that I know of, but you wouldn't want to reimplement it from scratch anyway
[14:28] <maxb> So, are you saying your webserver cannot do fastcgi?
[14:28] <catphish> yes
[14:28] <maxb> Nor has a WSGI adaptor?
[14:28] <catphish> most likely not
[14:28] <catphish> i will consider plugging the python in
[14:29] <catphish> but i would prefer to implement it myself, purely because that's what i've done for git and hg
[14:29] <catphish> rather than doing authentication then proxying, which is what i have had to do for svn
[14:29] <maxb> Well, it might not perform great, but you could run the smart server WSGI app as a CGI
[14:30] <maxb> plain CGI that is
[14:30] <maxb> http://henry.precheur.org/python/how_to_serve_cgi
[14:30] <catphish> that could work
[14:31] <catphish> ideally i'd like to take the http request and map it to a bzr command
[14:31] <maxb> "bzr command"?
[14:31] <catphish> perhsps i should have a read of the cgi adapter and see what's involved
[14:31] <maxb> "bzr command line invocation"?
[14:32] <catphish> either that or a bzr pipe interface
[14:32] <catphish> if there is such a thing
[14:32] <catphish> the ssh interface must execute something
[14:32] <maxb> erm, that's what running the smart server WSGI app as a CGI *is*
[14:32] <catphish> yes i suppose it is actually
[14:33] <catphish> what about the ssh smart server interface? how is that executed?
[14:33] <maxb> The ssh interface executes bzr serve --init --allow-writes --root=/something
[14:33] <catphish> that sounds like what i want
[14:33] <catphish> in hg for example i run "hg serve --stdio"
[14:34] <catphish> then map the http requests to ssh interface commands and pipe them in
[14:34] <maxb> whoops, I mean bzr serve --inet --allow-writes --directory=/something
[14:35] <catphish> of course, that's the raw smart server
[14:36] <maxb> huh, doesn't mercurial have a http interface of its own? Why are you so determined to reimplement things?
[14:36] <maxb> In fact, if you wrote a WSGI adapter for your server, you could use it for both hg and bzr :-)
[14:36] <catphish> maxb: because i have a ruby web server that implements my ACLs
[14:37] <catphish> and i don't want to then run apache in addition to host the repos
[14:37] <catphish> and proxy to it
[14:38] <catphish> you're right that i could implement a WSGI adapter, then do the authentication before passing through to that
[14:38] <catphish> but i didn't
[14:39] <catphish> and most SCMs have a web protocol that easily maps to their SSH protocol
[14:39] <catphish> so i convert the http requests and pipe them through
[14:40] <catphish> i realise i'm reinventing in places
[14:41] <catphish> also, because my repositories are located on an array of storage hosts and my web server is on an array of frontend hosts, doing it this way means i can create an rpc interface between the web server and the storage server
[14:44] <catphish> looking at the http protocol and the smart protocol, it looks like they might actually be identical
[14:44] <catphish> which makes this very easy
[15:56] <catphish> what does "No repository present" mean? (error from smart protocol)
[15:57] <catphish> i am specifying a branch inside a repository, would it not be finding the parent repository for some reason?
[15:58] <thumper> catphish: where is your remote branch?
[15:58] <catphish> its inside a repository
[15:59] <catphish> i'm specifying it to bzr serve "bzr serve --allow-writes --inet --directory=/path/to/repo/branch"
[15:59] <catphish> i then issue: "bzr message 3 (bzr 1.6)\n\000\000\000\034d16:Software version5:2.1.1es\000\000\000\034l20:BzrDir.open_branchV31:.ee"
[16:00] <catphish> and get back "bzr message 3 (bzr 1.6)\n\000\000\000\034d16:Software version5:2.1.1eoEs\000\000\000;l5:error47:No repository present: \"filtered-162975116:///\"ee"
[16:00] <thumper> holy shit... bzr 1.6 is old
[16:01] <catphish> oh yeah i didn't spot that
[16:01] <catphish> i'm running bzr 2.1.1
[16:01] <catphish> i wonder what's happening there
[16:01] <thumper> what is the server running?
[16:01] <thumper> 2.1.1 too?
[16:01] <catphish> its the same machine right now so yes
[16:02] <catphish> i'll look into that first!
[16:04] <catphish> i guess its probably falling back to an old protocol version
[16:04] <catphish> i'll try to fix that
[16:08] <catphish> actually i think that is the latest version of the protocol
[16:31] <catphish> it appears that my problem is that because i specified --directory to point to the branch, the client is not being allowed access to the repository above
[16:32] <fullermd> Yes, --directory acts much like a chroot.
[16:34] <catphish> that makes sense, but if i point the directory to the repository, the client dies because it's not a branch
[16:34] <catphish> am i missing something in between?
[16:34] <fullermd> The client has to supply a path below the --directory served.
[16:35] <catphish> actually that seems kinda obvious now that you say it
[16:35] <catphish> how would i do that over http? i enter the url as: bzr+http://localhost/repo.bzr/branch
[16:36] <catphish> how does the client know that branch is the branch?
[16:36] <catphish> or do i need to do something active in my webserver?
[16:37] <catphish> as far as i know, the client does not specify a branch location at all right now, it just relies on --directory to put it in a branch
[17:01] <maxb> catphish: --directory should be specified by your *server* to the root of the bazaar storage to which the client is allowed to see
[17:02] <catphish> maxb: i understand that part
[17:02] <catphish> however, what i am stuck with is how to instruct the client to select a particular branch
[17:04] <catphish> quite simply, the client does not bother to do so at all
[17:05] <catphish> so i am thinking that i need to inject a command
[17:12] <catphish> maxb: do you happen to know the command to specify the branch?
[17:12] <maxb> It is not part of the command, it is part of the protocol
[17:12] <catphish> yes, but the protocol sends commands
[17:12] <catphish> right now my client sends: "BzrDir.open_2.1" then "BzrDir.open_branchV3"
[17:13] <catphish> no mention of *which* branch to open
[17:13] <maxb> yes there is, in the parameters
[17:15] <catphish> i don't see it
[17:19] <catphish> it definitely doesn't send a branch
[17:19] <catphish> it sends just the 2 commands i mentioned (this is over http)
[17:23] <maxb> l20:BzrDir.open_branchV31:.e
[17:23] <maxb> command and parameter
[17:24] <catphish> oh, the parameter is just '.' ?
[17:25] <catphish> that makes much more sense
[17:26] <catphish>  do you know if bzr+http will even send anything other than '.'?
[17:26] <catphish> *ever
[17:28] <maxb> Use the source
[17:29] <maxb> or stop reimplementing it :-/
[17:29] <catphish> i'm not reimplementing the client :)
[17:29] <catphish> but i am finding my way through some of the code
[17:29] <catphish> just hoped you'd know
[17:30] <catphish> sadly i'm not a native python speaker
[17:53] <catphish> i can't seem to find a way to make the client differentiate between the base url and the branch name
[17:54] <fullermd> There isn't one.
[17:56] <catphish> so am i correct in saying that the only way to implement multiple branches in bzr+http is to have the http server rewrite the BzrDir.open_branchV3 command?
[17:57] <catphish> rather than using BzrDir.open_branchV3('.') which the client always sends
[17:58] <catphish> or is there a better way to tell the client what the correct base url is?
[18:03] <catphish> unfortunately I can't find a repo anywhere that does this correctly
[18:03] <catphish> for example "bzr clone bzr+http://pyzgoubi.bzr.sourceforge.net/bzr/pyzgoubi/branches/0.4.x"
[18:03] <catphish> shows the same error on SF
[18:03] <fullermd> Well, I dunno what's what on the level you're looking.  But from a client side, a URL is atomic.
[18:03] <fullermd> It's all just one opaque thing.
[18:04] <catphish> to summarize, the problem i have is that when using bzr+http the command that the client sends to the server is always BzrDir.open_branchV3('.')
[18:06] <catphish> assuming that the branch is the root of the repository at the other end
[18:07] <fullermd> Other way around.  The client doesn't care anything about a repository; that's all under the covers.
[18:08] <catphish> the problem is that on the server side, I pipe the request to "bzr serve --allow-writes --inet --directory /path/to/repo/branch"
[18:08] <catphish> which has the unfortunate side-effect of not allowing access to "/path/to/repo"
[18:09] <catphish> whereas if i pipe it to "bzr serve --allow-writes --inet --directory /path/to/repo" then the client tries to open branch '.' which obviously fails
[18:11] <catphish> the only solution i can come up with is for the web server to rewrite the "BzrDir.open_branchV3:." command to be "BzrDir.open_branchV3:branchname" based on the url
[18:11] <catphish> but i wanted to know if that's the correct solution
[18:11] <catphish> since i don't have sufficient python knowledge to read the python http smart server
[18:13] <catphish> i was really hoping to be able to pass the http body through to the 'bzr serve' unmolested
[18:14] <catphish> but i guess that's not quite going to cut it
[18:21] <catphish> well i have to go now, will test some things and see
[18:48] <poolie> i put up a tuolumne patch that will correctly represent 0 new bugs
[19:02] <catphish> excuse my stupidity one more time, but where is bzr-smart.fcgi?
[19:13] <catphish> ah got it
[19:51] <cody-somerville> hmm... I'm using the bzr pipeline plugin
[19:52] <cody-somerville> and the the top pipe (ie. the upstream branch I'm maintain patches against) has somehow turned into what appears to be the last pipe
[19:53] <cody-somerville> or wait a tick, it looks like a bunch of stuff from different pipes has made it into it
[20:00] <cody-somerville> and interesting... these revisions I see haven't been piped.
[20:00]  * cody-somerville wonders if this is some sort of bug in sync-pipe
[20:06] <dOxxx> cody-somerville: there's a new version of bzr-pipeline, have you tried that?
[20:07] <cody-somerville> nope
[20:07] <cody-somerville> is it good?
[20:08] <cody-somerville> any release notes/changelog available?
[20:08] <dOxxx> Changes:  - bzr pump supports --show-base, --reprocess and --merge-type.  - Pipes can no longer be accidentally named ":next", etc.  - Pipeline imposes smaller start-up cost when not used.
[20:08] <dOxxx> so, from the sounds of it, may not fix your problem
[20:09] <dOxxx> check if you can find a similar bug at https://bugs.edge.launchpad.net/bzr-pipeline, otherwise file a new one.
[20:09] <dOxxx> you could also try posting on the bzr mailing list.
[20:15] <maxb> bzr-pipeline 1.1 is ready in the proposed PPA. probably fine to promote now
[20:57] <jelmer> maxb: Didn't you fix bug 622188 ?
[20:58] <maxb> I think so
[20:58]  * maxb checks it got merged
[20:59] <maxb> yes, in 0.7.5
[20:59] <jelmer> maxb: Great, thanks for checking!
[20:59]  * jelmer marks as fix released
[21:05] <mkanat> Hey poolie. :-)
[21:29] <maxb> jelmer: btw, what is the plan for bzr-svn version numbers now?
[21:29] <maxb> lp:bzr-svn seems slightly confused about whether it is 1.0.5 or 1.1.0
[21:31] <jelmer> I'll move to 1.1
[21:31] <jelmer> thanks for the reminder
[21:34] <maxb> This could get confusing :-)
[22:20] <kire> I am trying to run a bazaar smart server over https, but when I try to push something I get an error like "ERROR: Cannot lock LockDir .... ERROR: Cannot lock LockDir"
[22:21] <kire> ehm, last part is "Transport operation not possible: http does not support mkdir()"
[22:22] <jelmer> kire: that suggests the smart server is not working
[22:26] <kire> do you have to use bzr+https?
[22:29] <jelmer> you shouldn't have to
[22:29] <kire> ah, I should upgrade my server version probably, seems to be 2.0, are there any decent repositories for debian lenny?
[22:30] <jelmer> I don't think there are any recent backports
[22:30] <jelmer> I'm planning to do one when squeeze gets released
[22:30] <kire> and the pain of terrible ubuntu support of my server provider hits me again :p
[22:31] <kire> okay, I'll see if I can get something to work, thanks anyway :)
[23:05] <maxb> jelmer: I have one pending MP concerning ~bzr-svn/bzr-svn/1.0, do you want it resubmitted against ~bzr-svn/bzr-svn/1.1 ?
[23:16] <dOxxx> eugh... damn 2.4 compatibility.
[23:17] <dOxxx> can't use str.format ><
[23:22] <mgz> damn 2.5 compatibility too then.
[23:22] <dOxxx> yah
[23:22] <dOxxx> so now I'll have to fake it
[23:25] <catphish> just in case anyone missed this earlier, and knows of an easy solution, is there any way to make "bzr serve --directory /path/to/repository/branch" work? by default it does not appear to work because it denies access to the repository directory above
[23:27] <mkanat> catphish: The branch would have to be standalone.
[23:27] <mkanat> catphish: You could use "bzr reconfigure" to make it standalone.
[23:27] <catphish> mkanat: i'm hoping that will work as a last resort, but i'm trying to avoid it
[23:28] <mkanat> catphish: I'm pretty sure there's no way to avoid it.
[23:29] <catphish> i hope to serve /path/to/repository instead and have something between the client and the server prepend the branch name to requests
[23:30] <catphish> unfortunately the http url scheme does not differentiate between repo and branch, it assumes the url points to the root of a branch, and always assumes the branch is '.'
[23:32] <catphish> anyway, went over this earlier and i don't think there's a simple solution other than inserting the branch into the request between receiving the http request and passing it to the smart backend