[02:27]  * igc lunch
[02:35] <unenough> is there a way to change a commit message after the commit?
[02:40] <poolie> unenough, uncommit then commit again
[02:40] <unenough> no other way?
[02:43] <poolie> unenough, sorry, not at present
[04:36] <indu> hi all
[04:36] <indu> good morning
[04:49] <indu> i have some doubts in bazaar, can i get some help here
[04:50] <indu> i want to to know how does the branches under the repo created using bzr init-repo are linked to the repo
[04:55] <spiv> indu: just by being in a directory under the repo.
[04:56] <indu> spiv, firstly i am not understanding one thing here
[04:57] <indu> spiv, in svn we create a repo, using svnadmin create repo
[04:57] <indu> and under that we create directories and make a structure
[04:57] <indu> later on, each can get a branch from those direcotries
[04:58] <indu> in bzr , where is the repo created ?
[04:58] <fullermd> indu: It might be clearer if you were to forget about [shared] repos in bzr for a while.  They're not necessary.
[04:58] <indu> spiv, ok as per your point, i was under that direcotry in that repo, and created branches using bar init
[04:58] <spiv> In bzr, the repo is created where you tell "bzr init-repo" to make it.
[04:58] <indu> bzr init
[04:59] <indu> and then under that bzr init direcory i created sub branches again using bzr init,
[04:59] <spiv> And then when reading/writing branches, bzr will check the parent directories all the way to the root until it finds a repository.
[05:00] <indu> spiv, how can I view those details, whethere while reading or writing its checking the parent direcotires or not
[05:00] <spiv> (but note that as fullermd says, there's no need to have a separate repo; if you don't create separate shared repo with "bzr init-repo", bzr will just make standalone branches which carry their own repo)
[05:00] <indu> as, removing any branch under the repo, is not showing any affect on the repo
[05:01] <spiv> On the filesystem, it's in the .bzr directories.
[05:01] <spiv> You'll probably find "bzr info"/"bzr info -v" an easier way to get that information than looking inside the .bzr directories, though.
[05:01] <indu> spiv, i need ur suggestion for creating the bzr structure , will exaplin u my project details
[05:02] <indu> spiv, my project is builiding linus distribution which is based on debian, and thus we are trying to develop a community for contribution
[05:02] <spiv> indu: Have you read http://doc.bazaar-vcs.org/latest/en/user-guide/index.html ?
[05:03] <indu> thus the source of the packages will be alphabatical order, a/acpid/acpid_x_xx_X
[05:03] <indu> b/boo/boo_x_xx_xx
[05:03] <indu> where x_xx_xxx are package names with version number
[05:04] <spiv> It sounds like you probably want a repository for each b/boo then.
[05:04] <spiv> And each b/boo/boo_x_... would be a branch.
[05:05] <spiv> Unlike SVN, in bzr a repository is just an optimisation.
[05:07] <spiv> So if you have several branches of a single project (or in your case, package), then it's good to have a repository so that the common history in those branches only needs to be stored once, rather than once for each branch.
[05:08] <spiv> So repositories save disk space and save on the amount of data to copy/transfer for many operations.  But they don't change the fundamental behaviour.
[05:09] <indu> spiv, what is this common history ?
[05:09] <spiv> indu: if you have boo_1.0 and boo_1.1, probably they have many revisions in common.
[05:10] <spiv> For instance, most software projects have a "trunk" branch, and feature branches and release branches are made from that branch.
[05:11] <spiv> So those branches share most of their history with the trunk.
[05:12] <spiv> For instance, the development branch of bzr and the 1.0 release branch of bzr share most of their history.
[05:13] <spiv> Does that make sense?
[05:14] <spiv> In svn, when you make a branch it basically just copies a snapshot of the files in that version.  In bzr, a branch keeps the full history, not just the snapshot.
[05:24] <indu> spiv, from where can i download the loggerhead source
[05:26] <spiv> indu: http://www.lag.net/loggerhead/
[05:27] <indu> spiv, ok
[05:27] <spiv> indu: although this branch may have some improvements that haven't made it into robey's version: https://code.launchpad.net/~mwhudson/loggerhead/production
[05:28] <indu> spiv, i am reading the user guide u sent, but before that
[05:28] <indu> spiv, as per your directiones
[05:28] <indu> i have to structure my repo like
[05:28] <indu> bzr init-repo a
[05:29] <indu> bzr init-repo a/aaa
[05:29] <indu> bzr init a/aaa/acpid-xxxx
[05:29] <indu> am i right ?
[05:29] <spiv> indu: the "bzr init-repo a" is wasted
[05:29] <spiv> indu: the repo at "a" would never be used.
[05:29] <indu> why?
[05:30] <spiv> But otherwise, that seems like a reasonable arrangement.
[05:30] <spiv> Because each branch belongs to exactly one repository.
[05:31] <indu> under aaa, there will be some more branches like, aaa-xxx, abc-xxx etc
[05:31] <indu> where all these are seperate packages
[05:31] <spiv> So when bzr looks at the branch in a/aaa/acpid-xxx, this is how it'll look for the repo:
[05:32] <spiv>  - first, it'll look for a repo in a/aaa/acpid-xxx.  There's no repo there.
[05:32] <spiv>  - then, it'll look in the next directory up: a/aaa.  There is a repository there.
[05:32] <spiv>  - Done!  It looks no further.
[05:32] <indu> ok
[05:32] <indu> one  more thing
[05:33] <indu> i tried created branches using "bzr init" , under "bzr inir-repo repo"
[05:33] <indu> sorry. bzr init-repo repo
[05:33] <indu> bzr init a
[05:33] <indu> mkdir a/aaa
[05:34] <indu> cd a,  bzr add aaa
[05:34] <indu> then I tried
[05:34] <indu> bzr init abc,
[05:35] <indu> from a remote, i tried getting this newly created branch using bzr branch file:///~/repo/a/abc
[05:35] <indu> this dint work, saying no valid working tree
[05:36] <fullermd> I don't think ~ expands in a file:/// URL...
[05:37] <fullermd> (of course, you're branching an empty branch there, so it doesn't do much)
[05:37] <indu> i just wrote that, instead of mentioning the whole path
[05:38] <spiv> I'm also not sure what you mean by "from a remote ... bzr branch file:///~/repo/a/abc"
[05:38] <fullermd> You don't need the file:/// spec on it; just "bzr branch ~/repo/a/abc foo" would work.
[05:39] <indu> not remotre, fullermd expalnation is correct in my case
[05:39] <indu> bzr branch ~/repo/a/abc is correct
[05:39] <indu> actually, i am using bzr-gtk
[05:40] <spiv> indu: what does "bzr info ~/repo/a/abc" tell you?
[05:40] <fullermd> Now, if you're still in ~/repo/a when you do that, it probably won't work, since it'll try to call the new branch 'abc', which would conflict with the real 'abc'.
[05:42] <indu> spiv, i will try some more and then ping you back, thankyou for the help
[05:42] <spiv> indu: you're welcome
[05:42] <indu> spiv, can you give me some more examples for a good web front end for bazaar
[05:43] <spiv> indu: loggerhead is the best I've seen IMO.  bazaar-webserve isn't as easy to use.  There's another one I think, but I forget what's called.
[05:44] <indu> ok
[05:45] <indu> how does this launchpad using bazaar ?
[05:46] <spiv> indu: do you mean, how do you use launchpad?
[05:47] <indu> spiv, no, i want to make a similar application like launchpad for my project also
[05:47] <indu> thus, i want to inbuilt bazaar into it
[05:47] <spiv> Oh, ok.
[05:48] <spiv> Launchpad does *lots* of stuff :)
[05:48] <indu> launchpad is not opensource , thus i cannot downlaod install  and use independently
[05:48] <indu> spiv, yeah, i knew :)
[05:49] <spiv> Launchpad uses loggerhead for its "codebrowse" feature.
[05:49] <indu> spiv, https://code.launchpad.net/?
[05:49] <indu> but the look is not like loggerhead
[05:50] <spiv> indu: http://codebrowse.launchpad.net/~robey/loggerhead/devel/changes
[05:50] <spiv> indu: (which is linked from https://code.launchpad.net/~robey/loggerhead/devel)
[05:51] <spiv> It's a bit hard to describe exactly how to write something like launchpad, because launchpad does so much.
[05:52] <spiv> But in general, because Launchpad is written in Python, it's fairly easy for it to use bzrlib.
[05:53] <spiv> (in so far as writing useful software can ever be called "easy"...)
[05:55] <fullermd> I have a friend who I've done some work with in the past who'd open discussion with "Look, it's real simple..."
[05:56] <fullermd> He'd tell you it's easy.  You just have to write the code, test it out real quick, then deploy it.  How long can it take to do 3 things?
[05:56] <fullermd> For some reason, I avoid doing work with him these days...
[06:02] <indu> ok. i will try for my own
[06:55] <indu> i need help in loggerhead configuration
[07:46] <igc> night all
[07:47] <indu> how can I make bazaar use apache web server for publishing the branches and repo
[07:47] <fullermd> igc: Say, before you go...
[07:48] <fullermd> igc: Did you see bug 178591?
[07:48] <ubotu> Launchpad bug 178591 in bzr "running `bzr diff` in a subdirectory does not show changes in the	parent directory" [High,Confirmed] https://launchpad.net/bugs/178591
[08:57] <glyph> Is this expected behavior?  http://rafb.net/p/8u9eU837.html
[08:57] <mwhudson> glyph: are you pulling into a repository?
[08:58] <glyph> mwhudson: The only command I didn't list there was "mkdir PyDoctor"
[08:58] <mwhudson> oh oops
[08:58] <mwhudson> pydoctor is still a dirstate-tags branch
[08:58] <mwhudson> so if you rm -rf .bzr and init-repo --dirstate-tags, it should worl
[08:58] <mwhudson> work
[08:59] <fullermd> No, it's not.
[08:59] <fullermd> lp:pydoctor is redirected to http://bazaar.launchpad.net/~mwhudson/pydoctor/dev/
[08:59] <fullermd> Standalone branch (format: dirstate-with-subtree)
[09:00] <mwhudson> oh right, bzr-svn
[09:00] <mwhudson> that means the error makes more sense
[09:12] <GoClick> When the page says Requires PyBaz which in turn requires Baz does that mean it requires bazaar or baz something else?
[09:13] <mwhudson> nothing relevant should really require pybaz any more
[09:13] <mwhudson> which page are you referring to?
[09:13] <GoClick> Interesting
[09:13] <GoClick> The install faq recommends Bzrtools
[09:14] <GoClick> which states "Requires PyBaz, which, in turn, requires baz."
[09:14] <GoClick> I know nothing about bazaar I just know we use svn now and hate it and I've been trying for 5 days to get it to compile in our environment but can't and #svn is jerky
[09:15] <GoClick> so why should I try so hard to make software I hate work
[09:17] <mwhudson> ah
[09:17] <fullermd> bzrtools only requires pybaz if you use baz-import, which it doesn't sound like you care about.
[09:17] <mwhudson> bzrtools only requires pybaz for one specific command
[09:18] <spiv> indu: mwhudson knows about loggerhead configuratino
[09:18] <GoClick> Oh ok
[09:18] <indu> ok spiv , thanx
[09:18] <GoClick> I was just trying to get everything installed so I could read about switching from svn to bzr
[09:19] <mwhudson> GoClick: ah, ok
[09:19] <mwhudson> GoClick: the page should be clearer, perhaps
[09:19] <GoClick> It's not so bad I guess, It's 2:19am and I need to go to bed I have to work tomorrow
[09:19] <GoClick> :P
[09:19] <GoClick> Wait I'm working right now
[09:19] <GoClick> my life is hell
[09:20] <GoClick> Actually it's not that bad it's just stressed right now because someone decided we had to run RHEL4 on our new server instead of Ubuntu and I've never used RHEL so I'm like "wtf" with everything
[09:20] <GoClick> night guys/gals whatever thanks
[09:21] <indu> mwhudson, hi,
[09:22] <indu> mwhudson, i want to know how i can use loggerhead, as well as, configuring apache web server for accessing repository
[09:22] <mwhudson> indu: have i talked you you before about this?
[09:22] <indu> mwhudson, no
[09:23] <mwhudson> oh, ok
[09:23] <mwhudson> so the latter part should be easy
[09:24] <indu> i have a repo under ~/boss-repo
[09:24] <indu> and i created a link for this under the /var/www (I am in Debian OS)
[09:26] <indu> then I added the lines like,  http://rafb.net/p/PcN0Q142.html  in apache2.conf file
[09:26] <indu> is it correct ?
[09:28] <mwhudson> hang on
[09:28] <mwhudson> when you say you want to configure apache to access the repo
[09:28] <mwhudson> do you mean using bazaar or using a web browser and loggerhead?
[09:30] <indu> i want to use web browser to browse the branches/repository created using bazaar, then later, i even want to use loggerhead to view the repository , as it good web absed front end
[09:30] <mwhudson> ok
[09:30] <mwhudson> well, you'll be wanting to put some port numbers in that apache config file
[09:31] <indu> mwhudson, i was having 8080 as the port
[09:31] <mwhudson> that sounds more likely
[09:31] <indu> so the lines were like         ProxyPass http://127.0.0.1:8080/
[09:31] <indu> and         ProxyPassReverse http://127.0.0.1:8080/
[09:32] <mwhudson> yez
[09:39] <indu> this is showing an outout in web browser as,
[09:41] <indu> mwhudson, are messages receievd by the channel
[09:42] <mwhudson> indu: i only saw
 this is showing an outout in web browser as,
[09:42] <indu> mwhudson, /boss/ not found
[09:43] <indu> mwhudson, 404 Not Found
[09:43] <indu> The path '/boss/' was not found.
[09:43] <mwhudson> well, that suggests that the apache config is wrong somehow
[09:44] <mwhudson> i'm no expert in apache, unfortunately :/
[09:44] <indu> mwhudson, ok, then help in loggerhead ?
[09:44] <fullermd> error log might say something.
[09:45] <indu> fullermd, the apache log is saying, [Wed Jan 02 15:03:35 2008] [error] [client 127.0.0.1] client denied by server configuration: proxy:http://127.0.0.1:8080/
[09:45] <mwhudson> oh
[09:45] <mwhudson> maybe you need an "allow from all" in there
[09:46] <indu> lemme try
[09:48] <indu> mwhudson, same o/p, but log file content is not the same
[09:49] <indu> now the log shows. nothing on the failure
[09:53] <indu> mwhudson, how to configure loggerhead
[09:54] <mwhudson> take the loggerhead.conf file in the branch and change the bits that are relevant to your situation
[09:54] <indu> mwhudson, deos loggerhead have any relation with apache?
[09:55] <indu> mwhudson,  the port number of loggerhead, should be different of that of apache or same ?
[09:55] <mwhudson> it is common to use apache in front of loggerhead
[09:55] <mwhudson> so that the user's browser talks to apache, and apache talks to loggerhead
[09:56] <mwhudson> so loggerhead should use the port number you put in the ProxyPass directive
[09:56] <indu> mwhudson, i just typed, localhost:8080, in my browser and it showing an output as
[09:56] <indu> mwhudson, Bazaar branches in loggerhead, (which is the description in loggerhead.conf file) and under that the bazaar and loggerhead logos
[09:57] <mwhudson> well, that sounds good
[10:00] <indu> mwhudson, oops, i have mistaken, the line  "Bazaar Branches in Loggerhead " is not in my loggerhead.conf file
[10:00] <indu> mwhudson, its displayed frm somewhere else
[10:00] <mwhudson> oh, then maybe you're not using the loggerhead.conf file you think you are?
[10:01] <indu> exaclty :)
[10:02] <indu> mwhudson, ok now i corrected
[10:02] <indu> mwhudson, what should be the values for [xxx] and [[xxx]]
[10:03] <mwhudson> indu: what ever you like, they are just names
[10:04] <indu> mwhudson, then where does i give the repo path ?
[10:04] <mwhudson> as the folder = '..' value inside the [[xxx]] section
[10:08] <indu> mwhudson, can you give me an example, of any site, which used loggerhead as the fron t end, as i want to see how does it look
[10:13] <indu> mwhudson, is it needed that for each branch created there should be an entry in the loggerhean.conf file
[10:16] <indu> mwhudson, and this loggerhead is not dependent on apache, as even after stopping the apache web server I am able to see the loggerhead page in the browser
[10:17] <lifeless> 09:28 < lifeless> mwhudson: consider relative urls for codebrowse, kthxsavebybandwidth
[10:17] <lifeless> 09:29 < lifeless> e.g. consider http://codebrowse.launchpad.net/~lifeless/squid/squid3-trunk/files/cvs-1%3Ahno-20071220081046-d59dmpbuvvyioekf?file_id=src-19960222132353-t43j7l5jst5z-28
[10:19] <mwhudson> lifeless: there's a bug report for that already
[10:20] <lifeless> mwhudson: kk
[10:20]  * mwhudson feels like he is not, in fact, over his cold
[10:20] <lifeless> :(
[10:22] <indu> mwhudson, starting loggerhead is starting succesfully, but when I am browsing , the following message appears on the terminal as the traceback
[10:22] <indu> mwhudson,  http://rafb.net/p/PDM8p311.html
[10:23] <mwhudson> indu: do you have more than one loggerhead process running?
[10:24] <indu> mwhudson, i check the output of ps -A, i dint find any
[10:25] <indu> mwhudson, how can i check
[10:25] <mwhudson> that's what i'd have used
[10:27] <indu> mwhudson, now got it.
[10:27] <indu> mwhudson, tell me onw more please, do i need to mention for every new branch i create under my repo, in the loggerhead.conf file?
[10:27] <mwhudson> indu: no, there's some kind of "auto folders" thing
[10:28] <mwhudson> indu: but i've never used it, so i don't know how it works
[10:28] <indu> mwhudson, i am trying to use the same but its not fruitful
[10:28] <indu> mwhudson, i am trying
[10:30] <lifeless> mwhudson: let me know what I need to do to help you with my bug report; needless to say this is rather important to get squid to migrate ;)
[10:31] <mwhudson> lifeless: well, my approach would be to do print-based debugging to try to see where the '/' is coming from
[10:32] <indu> mwhudson, yes wuto_publish worked for me
[10:33] <indu> mwhudson,  i am seeing something surprising here,
[10:33] <indu> mwhudson,  i have a repo, boss-repo, created using bzr init-repo
[10:34] <indu> and under that i created i initiated a branch using bzr init a
[10:34] <indu> and under a/ , bzr init axyz
[10:35] <indu> I am seeing in this loggerhead auto_publish_folder, the a, and axyz under the same tree, but not like, axyz under the a/ directory
[10:36] <mwhudson> well, (a) don't put branches in branches (b) auto_publish probably isn't recursive
[10:39]  * fullermd weeps over log -v speed/memory.
[10:40] <mwhudson> lifeless: browsing squid will loggerhead seems to work fine here :/
[10:43] <mwhudson> lifeless: although, the [logging] section in your config isn't doing what you probably think it is...
[11:41] <lifeless> mwhudson: I have no idea whats doing what; its undocumented voodoo as far as I can tell without reading code.
[11:42] <lifeless> mwhudson: I can reproduce this (obviously). The get_history is being given a path or something of None
[11:42] <mwhudson> lifeless: my voodoo + painkillers suggestion is to remove the [logging] section
[11:42] <lifeless> mwhudson: it may be a version of turbogears thing. had TERRIBLE trouble installing it
[11:42] <mwhudson> lifeless: i have a not-so secret plan to get rid of the turbogears dependency
[11:43] <lifeless> as its not documented to need a specific version I ended up with the latest stuff
[11:44] <mwhudson> well, tg should be able to interfere at this level (i would hope)
[11:44] <mwhudson> should NOT
[11:44]  * mwhudson really needs to get away from the computer
[11:51] <lifeless> seems to be constructing a branchview with garbage parameters
[12:09] <lifeless> night all
[12:11] <lifeless> mwhudson: I could waste a lot of time learning loggerhead as a once off.
[12:11] <lifeless> mwhudson: I'd like instead to get you to directly look at the problem if thats possible.
[12:12] <mwhudson> lifeless: well, unless i can reproduce it, how can i
[12:12] <mwhudson> ?
[12:13] <lifeless> mwhudson: ssh to the box that its failing on ?
[12:13] <mwhudson> lifeless: that could work
[12:14] <lifeless> I'll ask for an account for you. mwhudson as the name ?
[12:14] <mwhudson> lifeless: mwh would be preferable
[12:14] <mwhudson> my key is on lp
[12:16] <lifeless> url for it please
[12:17] <mwhudson> https://edge.launchpad.net/~mwhudson/+sshkeys
[12:31] <lifeless> mwhudson: in my homedir, ~loggerhead has the checked out copy, and I'm turning it off so you can run up your own if you get an account overnight
[12:31] <mwhudson> lifeless: will i get an email if i get an account?
[12:31] <lifeless> mwhudson: I've included your canonical address in the mail; I expect courtesy would have them inform you
[12:32] <mwhudson> ok, thanks
[12:33] <lifeless> you'll need to copy it to your home dir to play with, naturally.
[13:00] <Stavros> can someone help me with bzr-svn? it won't check out my branch :/
[13:00] <Stavros> "Unsupported protocol"
[13:05] <indu> mwhudson, still i am facing problem with my bazaar with apache configuration
[13:50] <Stavros> bzr-svn, anyone?
[14:07] <jelmer> hi Stavros
[14:07] <Stavros> hi
[14:07] <jelmer> Stavros: What version of bzr-svn are you using and what platform are you running it on?
[14:08] <Stavros> i am using the latest one on a mac
[14:08] <Stavros> but i got errors with it in linux as well, where i had to delete its config file to make it browse the server correctly
[14:09] <jelmer> Stavros: If it's http or https that's failing, the build of python-subversion you are using is probably not built with http/https support
[14:10] <Stavros> i am trying svn+http... hmm
[14:10] <Stavros> i'll get another one, then
[14:10] <jelmer> Stavros: What errors did you get on linux?
[14:11] <Stavros> invalid branch type, i think. it was looking in the config and seeing another URL for the same server and refused to checkout the one i wanted
[14:27] <dlee> Where do I go for the svn trunk I need for bzr-svn to work?  I run svn and bzr under Windows+Cygwin in case that matters.
[14:28] <jelmer> dlee: There are no prebuilt packages for cygwin afaik
[14:28] <Stavros> well, i installed the python-subversion bindings from tigris and it still doesn't work :/
[14:28] <jelmer> Stavros: There are prebuilt packages of Subversion 1.5?
[14:29] <Stavros> apparently, http://pysvn.tigris.org/project_downloads.html
[14:29] <jelmer> dlee: So your best bet would be to build Subversion from source, trunk can be obtained through subversion itself
[14:29] <jelmer> Stavros: You need python-subversion, not pysvn
[14:29] <Stavros> damn
[14:29] <jelmer> Stavros: that's a completely different project
[14:29] <Stavros> ah
[14:30] <Stavros> do you know the webpage? google doesn't turn up anything official-sounding
[14:30] <jelmer> Stavros: http://subversion.tigris.org
[14:31] <Stavros> oh
[14:31] <dlee> My Cygwin includes svn 1.4.5.  I know that's not enough, but I think it came from a package.  Hmm... guess I need to find the svn source and see what happens. I'll look in Cygwin setup first though to see if there's a devel version of svn.
[14:31] <jelmer> Stavros: However, there are no binary packages for Subversion 1.5 yet afaik
[14:31] <Stavros> jelmer: for the mac?
[14:31] <Stavros> i'm confused :/ don't i just need the svn bindings?
[14:32] <jelmer> Stavros: Yes, you need just the svn bindings
[14:32] <Stavros> can't i use pre-1.5 ones?
[14:32] <jelmer> no, you need 1.5 (or patch the 1.4 ones)
[14:32] <Stavros> ah
[14:32] <jelmer> there were some people on the Bazaar mailing list trying to build a package of 1.5, but I don't think that's done yet
[14:33] <Stavros> hmm
[14:33] <Stavros> how did it work on linux?
[14:33] <jelmer> Most distributions ship with patched versions of Subversion 1.4
[14:34] <Stavros> ah
[14:34] <Stavros> hm
[14:34] <Stavros> i should just install linux on this thing
[14:35] <Stavros> ah never mind, i'll just do it on the server and pull
[14:41] <Stavros> any idea why branching over ssl doesn't work on the mac?
[14:42] <Stavros> nm, got it
[15:43] <mtaylor> jelmer: good morning... I have a bzr-svn problem :)
[15:43] <jelmer> hey mtaylor
[15:43] <mtaylor> jelmer: I was doing an svn-import, and it ran for hours (big repos)
[15:43] <mtaylor> and I got this:
[15:43] <mtaylor> NoSuchId: The file id "7924@3c33494c-61f7-0310-86b9-b90697347e9d:branches%2Fdevelopment-2.0:server%2Fmerlin%2FWEB-INF%2Fsrc%2Fcom%2Fmysql%2Fetools%2Fscratch" is not present in the tree <Inventory
[15:43] <mtaylor> this is using (I think) the latest bzr-svn
[15:44] <jelmer> latest as in 0.4 branch?
[15:44] <mtaylor> but I'm happy to upgrade if I'm using an old version by mistake
[15:44] <mtaylor> I think so... lemme check
[15:44] <jelmer> is this repository public?
[15:44] <mtaylor> jelmer: no, unfortunately
[15:46] <mtaylor> jelmer: should I bother reporting a bug with the whole traceback? or is it useless without a repos to look at?
[15:46] <jelmer> mtaylor, please file a bug. the traceback may be useful
[15:47] <mtaylor> jelmer: k. will do.
[15:47] <jelmer> at the very least, it's an indication there's still a bug left in the pull code
[15:55] <mtaylor> jelmer: the traceback seems to include a dictionary of the entire repos or something... is that important?
[15:55] <mtaylor> (it's big)
[15:56] <jelmer> it may be useful, if it doesn't contain any sensitive information
[15:59] <mtaylor> no. I just wanted to make sure before I bothered pasting it all in
[16:18] <mgedmin> can I get the bzr log for a subdirectory?
[16:19] <mgedmin> bzr log dir only shows those revisions that actually changed dir, but not those that changed files inside it
[16:19] <mgedmin> I fail to find any --recursive options
[16:19] <mgedmin> is this use case not supported?
[16:20] <dato> mgedmin: it doesn't work at the moment, and I *believe* there is an open bug about it
[16:21] <mgedmin> https://bugs.launchpad.net/bzr/+bug/97715
[16:21] <ubotu> Launchpad bug 97715 in bzr "bzr log DIR should show changes under dir" [Undecided,New]
[16:23] <mgedmin> what's the bzr-speak for "please update this working dir to the state it was in revision X" (in svn: svn up -r X)
[16:23] <mgedmin> ?
[16:23] <dato> mgedmin: bzr revert -r X
[16:27] <mgedmin> revert sounds scary
[16:27] <mgedmin> can I then revert back to the latest version?
[16:28] <mtaylor> jelmer: so once I've gotten that error... is there any way to recover my repos? or am I just screwed for the moment
[16:28] <jelmer> mtaylor: recover in what sense? continue the import ?
[16:29] <mtaylor> jelmer: yes... or be able to use it in any way
[16:29] <jelmer> continuing the import is impossible until we fix the bug that causes bzr-svn to barf
[16:29] <mtaylor> ah
[16:29] <mtaylor> ok
[16:29] <dato> mgedmin: yes
[16:31] <mgedmin> thanks
[19:16] <lifeless> moin
[19:25] <dlee> Do we have an easy way to add the commits from one bzr project to another existing bzr project?  If so, does it matter whether or not the first project's commit dates intersect those of the one being added to it?
[19:29] <lifeless> what do you mean 'add' ?
[19:33] <LeoNerd> Hrm... apt-get update is throwing me:
[19:33] <LeoNerd> Failed to fetch http://bazaar-vcs.org/releases/debs/feisty/./Packages.gz  MD5Sum mismatch
[19:33] <LeoNerd> Anyone any ideas?
[19:33] <lifeless> possibly someone is rsyncing right now (I'm not)
[19:34] <lifeless> alternatively you got one file partly downloaded and stuck in a proxy cache or some such
[19:34] <LeoNerd> I've been getting it every day for the last two weeks... Don't think it's just a transient thing
[19:35] <lifeless> release.gpg hasn't been updated
[19:35] <lifeless> but packages.gz has been
[19:35] <lifeless> I have some code to write, but I'll do adminny things this evening
[19:35] <LeoNerd> OK
[20:01] <ubotu> New bug: #179927 in bzr "bzr help formats line-breaks inside urls" [Undecided,New] https://launchpad.net/bugs/179927
[21:19] <paulproteus> I hope that git lets me fulfill the dream of using distributed version control without also making me hate my life.
[21:19] <paulproteus> Er, I mean bzr.
[21:19] <paulproteus> Git is failing at letting me live that dream.
[21:21] <fullermd> To dreeeeeeeeam, the impossible dreeeeeeeeam...
[21:29] <igc> morning
[21:32]  * fullermd waves to igc.
[21:32] <igc> hi fullermd - just looking at that bug now
[21:32] <igc> it was me
[21:32]  * fullermd nods.
[21:32] <thumper> morning igc
[21:32] <igc> morning thumper
[21:32] <fullermd> Just making sure you saw it, in the flood of messages I guess you're all dealing with after deserting us for a few weeks   :)
[21:45] <dato> revert not making backups of something is always a bug, right?
[21:46] <LarstiQ> dato: not if there is no information lost
[21:46] <dato> yeah, that's what I meant
[22:00] <lifeless> vila: remember what bug # spiv had where squid was involved ?
[22:04] <vila> #172701 ?
[22:05] <vila> bug #172701
[22:05] <ubotu> Launchpad bug 172701 in bzr "Large readv requests can fail if there is a squid proxy" [High,Fix released] https://launchpad.net/bugs/172701
[22:05] <ubotu> New bug: #179953 in bzr "revert after pull of a revision that removes a file looses wt	changes to that file" [Undecided,New] https://launchpad.net/bugs/179953
[22:06] <dato> LarstiQ: ^
[22:08] <lifeless> vila: thanks
[22:08] <lifeless> vila: got someone seeing similar symptoms with 1.0
[22:21] <vila> lifeless: similar like issuing single range requests ? Because I fix another cause for that (see BB about bug #179368)
[22:21] <ubotu> Launchpad bug 179368 in bzr "bug in lighttpd 1.4.18 makes bzr loop or issue too many requests" [Medium,Fix committed] https://launchpad.net/bugs/179368
[22:35] <lifeless> vila: interesting
[22:35] <lifeless> he's offline now, I'll drop him a mail
[22:37] <dato> lifeless: hey, regarding that packaging thread, I only thought it'd be nice to say how things looked on this side of things, wasn't the intention to come and blab about how you should be doing your stuff.
[22:38] <lifeless> dato: no worries
[22:38] <lifeless> I've been on leave; I'll be doing mail stuff this evening
[22:38] <dato> very well
[23:27] <Keybuk> if I have "trunk" in my working directory, and the URL of someone else's branch with some changes in it, how do I compare the two?
[23:28] <Peng> Keybuk: "bzr missing $URL".
[23:28] <Keybuk> Peng: doesn't show the differences
[23:28] <Keybuk> (just the revision info)
[23:28] <Peng> Keybuk: Oh, you want to see diffs too?
[23:28] <Keybuk> yeah
[23:29] <Kinnison> bzr diff branch:someurl..-1
[23:29] <Kinnison> ?
[23:30] <Keybuk> Kinnison: ERROR: Path(s) are not versioned
[23:30] <Kinnison> :-(
[23:31] <fullermd> Er.  -rbran[....]
[23:31] <Kinnison> good point
[23:31] <fullermd> It helps   :)
[23:31] <Keybuk> why doesn't just "bzr diff URL" work?
[23:32] <Kinnison> or bzr diff --old . --new http://blahblahblah
[23:32] <Kinnison> Keybuk: because that means bzr diff <file called http://blahblahblah>
[23:32] <Keybuk> Kinnison: no such args
[23:33] <Kinnison> does bzr help diff show them?
[23:33] <Keybuk> Kinnison: but I *can* do "bzr diff ../OTHERBRANCH"
[23:33] <Keybuk> so why doesn't URL work there?
[23:33] <Kinnison> not a clue
[23:33] <Keybuk> bug I suspect
[23:33] <Kinnison> file it then I guess :-(
[23:34]  * Kinnison is sorry he can't help
[23:34] <fullermd> Heck, --old isn't even in my slightly-outdated bzr.dev...
[23:34] <fullermd> bzr diff ../OTHERBRANCH probably isn't doing what you think it is...
[23:34] <Keybuk> it gives me exactly what I think I should get
[23:34] <fullermd> URL "works", it just doesn't mean much of anything.
[23:35] <Peng> Oh yay. This autopack was smart enough to leave the 200 MB, 100 MB and 70 MB packs and repack the rest.
[23:35] <Keybuk> why does diff against the checkout of the URL work when local but not remote?
[23:35] <fullermd> bzr diff ../OTHERBRANCH does the same thing as (cd ../OTHERBRANCH ; bzr diff)
[23:36] <fullermd> It doesn't compare OTHERBRANCH against $PWD.
[23:36] <Keybuk> bzr diff . ../OTHERBRANCH
[23:36] <Keybuk> that says in the help that it compares the two
[23:36] <Keybuk>     Show the differences between the two working trees:
[23:36] <Keybuk>         bzr diff bzr.mine bzr.dev
[23:36] <fullermd> That does, yes.
[23:37] <Keybuk> sorry, just a typo above
[23:37] <Keybuk> bzr diff . URL doesn't work
[23:37] <fullermd> You need to use the branch: revspec for full URL's to remote branches.
[23:37] <Keybuk> why?
[23:37] <Keybuk> if a path works, so should a URL
[23:37] <Keybuk> otherwise that's just stupid
[23:37] <Kinnison> presumably because noone has thought to add that
[23:37] <fullermd> Ambiguity resolution, probably.
[23:38] <Kinnison> if the branch: revspec works, then file a bug to suggest it try urlparse as a fallback?
[23:42] <hmeland> Keybuk: In bzr.dev, the "bzr diff bzr.mine bzr.dev" is no longer allowed, you have to do "bzr diff --old bzr.mine --new bzr.dev"; I think this change was done to disambiguate trees/branches from files/directories within a tree/branch.
[23:45] <ubotu> New bug: #179972 in bzr "bzr diff URL doesn't work, but bzr diff /path/to/branch does" [Undecided,New] https://launchpad.net/bugs/179972
[23:47] <hmeland> ... and with bzr.dev, e.g. "cd mytree; bzr diff --old URL" works nicely, AFAICT.
[23:57] <abentley> Keybuk: Which version of bzr are you using?