[00:56]  * KombuchaKip hates Git
[01:26] <bignose> bzr-svn is breaking suddenly
[01:27] <bignose> AFAIK the Subversion server hasn't changed
[01:27] <bignose>   File "/usr/lib/python2.6/dist-packages/bzrlib/plugins/svn/fetch.py", line 505, in _open_directory
[01:27] <bignose>     base_file_id = self.editor._get_old_id(self.old_id, path)
[01:27] <bignose>   File "/usr/lib/python2.6/dist-packages/bzrlib/plugins/svn/fetch.py", line 805, in _get_old_id
[01:27] <bignose>     return ret.next()[1]
[01:27] <bignose> StopIteration
[01:27] <bignose> is this a known problem with an easy fix (and can I have a pony)?
[01:31] <lifeless> https://bugs.launchpad.net/bzr-svn/+bugs?field.searchtext=stopiteration
[01:31] <lifeless> shows nothing
[01:31] <lifeless> so I suspect not a known issue
[01:32] <bignose> lifeless: thanks
[01:39] <spiv> Morning folks.
[01:39] <bignose> can I ‘bzr branch’ from a local filesystem Subversion checkout?
[01:39] <spiv> I think that's meant to work.
[01:40] <bignose> ‘bzr help urlspec’ only talks about the Subversion server
[01:40] <bignose> and Bazaar doesn't recognise the Subversion checkout as a branch.
[01:57] <bignose> so ‘bzr branch foo.svncheckout/ foo/’ just gives “bzr: ERROR: Not a branch: …/foo.svncheckout/”.
[01:58] <bignose> spiv: so how can I branch from a working Subversion checkout?
[01:58] <bignose> (accessing the server causes Bazaar to crash.)
[01:58] <spiv> bignose: the same way you just tried to :/
[01:59] <spiv> bignose: I think you need to file a bug (or in your case, pseudo-file by mailing the list and/or jelmer)
[01:59] <bignose> shall do, thanks.
[03:52] <_mathrick> $ bzr unbind
[03:52] <_mathrick> bzr: ERROR: Unable to connect to SSH host lynx.imada.sdu.dk; [Errno 10060] A connection attempt failed because the connected party did not properly respond after a period of time, or established connection failed because connected host has failed to respond
[03:52] <_mathrick> uhh?
[03:52] <_mathrick> why does it want to connect to unbind?
[03:54] <lifeless> hmm, not sure.
[03:55] <_mathrick> lifeless: it's a colo-ified repo
[03:57] <mathrick> and I'm having trouble trying to find out what is bound to what exactly, since bzr info accepts no --local
[03:57] <mathrick> $ cat .bzr/branch/location
[03:57] <mathrick> sftp://mathrick@lynx.imada.sdu.dk/home/mathrick/public_html/repo/.bzr/branches/lipid-calc/
[03:58] <mathrick> oh, that'd explain
[04:04] <lifeless> thats not bound
[04:04] <lifeless> thats a lightweight checkout
[04:05] <mathrick> lifeless: well, I thought it was a lightweight checkout of .bzr/branches/lipid-calc which was then bound to the above
[04:05] <mathrick> which is why it was surprising to me that it'd want to connect to unbind
[04:08] <mathrick> and it sure worked they I'd expect it to before (when the server was not down), ie. both the local colo workspace and the remote one were updated on any commit/other change
[04:18] <mathrick> s/they/the way/
[04:39] <bignose> okay, this is looking less like a ‘bzr-svn’ problem

[05:03] <bignose> it seems the repository is being strange
[05:03] <bignose> could that be because of earlier ‘bzr-svn’ problems?
[05:05] <lifeless> I think thats unrelated, and fixed in trunk.
[05:05] <lifeless>  / 2.4
[05:09] <spiv> Actually, I think that is likely to be related to earlier bzr-svn problems.
[05:09] <spiv> FSVO of "earlier" :)
[05:10]  * spiv digs up the bug number
[05:14] <spiv> https://bugs.launchpad.net/bzr-svn/+bug/485601 is the one I'm thinking of
[05:15] <spiv> IIRC jelmer has corrected the problem in bzr-svn trunk (and maybe the current release?) but there isn't yet an easy way to repair affected bzr repos.
[05:37] <bignose> would those repository problems show up on ‘bzr check’?
[05:43] <spiv> No, although if you made a fresh import (i.e. into a new bzr repo) then use 'bzr cross-check EXISTING_REPO FRESH_IMPORT' it would.  (requires ~spiv/bzr-crosscheck/experimental plugin, I should really make that be the trunk version and/or integrate it into bzr-repodebug)
[05:45] <bignose> spiv: are either of those in Debian Wheezy?
[05:45] <spiv> I don't know of anyone packaging those.
[05:46] <bignose> hmm, the ‘bzr check’ has failed on this repository.
[05:46] <spiv> bzr-repodebug is relatively new (although its contents aren't)
[05:46] <bignose> if I make a new repository and branch again from Subversion, would I be wasting my time?
[05:46] <spiv> Probably not.  How did check fail though?
[05:47] <bignose> NoSuchRevision: CHKInventoryRepository('file:///home/benf/projects/svn/.bzr/repository/') has no revision ('svn-v4:58802371-0b19-4c69-acb7-7d0c5b422186:trunk:3356',)
[05:48] <spiv> (If it is fallout from a now-corrected bzr-svn bug then making a fresh import would be the easiest way to continue working I think.)
[05:48] <spiv> Hmm, a bit odd.  Does the traceback give any hint as to whether it was looking for a revision or inventory when that happened?
[05:49] <spiv> (Unfortunately that error without context can mean either)
[05:53] <lifeless> there will be a backtrace in ~/.bzr.log
[05:56] <bignose> how much do you want to see?
[05:56] <bignose>   File "/usr/lib/python2.6/dist-packages/bzrlib/repository.py", line 2382, in get_inventory
[05:56] <bignose>     return self.iter_inventories([revision_id]).next()
[05:56] <bignose>   File "/usr/lib/python2.6/dist-packages/bzrlib/repofmt/groupcompress_repo.py", line 899, in _iter_inventories
[05:56] <bignose>     raise errors.NoSuchRevision(self, record.key)
[05:56] <bignose> NoSuchRevision: CHKInventoryRepository('file:///home/benf/projects/svn/.bzr/repository/') has no revision ('svn-v4:58802371-0b19-4c69-acb7-7d0c5b422186:trunk:3356',)
[06:00] <spiv> Ok, so it does think an inventory record is missing.  Have you used stacked repositories at all?
[06:00] <spiv> Anyway, I would certainly try a fresh import from SVN at this point

[06:01] <bignose> that's the full traceback
[06:01] <bignose> I will try a clean repository. but my goal is to merge some revisions from the existing repository :-(
[06:01] <spiv> There's a reasonable chance that you'll be able to merge them into the new repository.
[06:02] <spiv> I'm not sure how you'd have ended up with a revision record without a corresponding inventory record.  Perhaps just an old bug that's been fixed that I've long since forgotten about :)
[06:14] <bignose> new fun: the bound branch from Subversion into a new repository works
[06:14] <bignose> but branching from that give:
[06:14] <bignose> $ bzr branch ../svn/empowered/
[06:14] <bignose> inconsistent details in skipped record: StaticTuple('719@58802371-0b19-4c69-acb7-7d0c5b422186:trunk%2Fwdyt%2Fsource%2Fwdyt%2Fprofile_modules%2Fmanagement%2Fcommands%2Fsend_profile_reminders.py', 'ben@benfinney.id.au-20110701043735-hyhpvogab661xpcz') ('4655593 322707 1213090 1213211', ((('719@58802371-0b19-4c69-acb7-7d0c5b422186:trunk%2Fwdyt%2Fsource%2Fwdyt%2Fprofile_modules%2Fmanagement%2Fcommands%2Fsend_profile_reminders.py', 'ben@benfinney
[06:15] <bignose> Branched 3363 revision(s).
[08:16] <jam> morning all
[08:35] <sagaci> hi, just doing my first branch of lp:ubuntu/gnome-user-guide, it's up to  42616kB    84kB/s | Fetching revisions:Inserting stream:Estimate 5520/5533, how big is the initial revision size, should I save it for a rainy day or can someone shed light to the estimated total size..?
[08:38] <jam> sagaci: I'll check, just a sec
[08:39] <sagaci> thanks
[08:41] <jam> sagaci: are you sure that is spelled correctly: bzr: ERROR: Not a branch: "bzr+ssh://bazaar.launchpad.net/%2Bbranch/ubuntu/gnome-user-guide/"
[08:41] <sagaci> jam, it just finished at around 90MiB
[08:41] <sagaci> but thanks anyway
[08:41] <sagaci> sorry, gnome-user-docs
[08:42] <sagaci> *facepalm*
[08:42] <jam> hmm... I see it for natty,  but "lp:ubuntu/gnome-user-guide" didn't seem to work
[08:42] <jam> ah, thanks
[08:42] <jam> well, I'm glad it worked for you
[08:42] <sagaci> yeah, just didn't want it to be ~1GiB and 4 hours later
[08:43] <jam> yeah
[08:45] <deni> question
[08:46] <deni> can i have a repo in say /home/user1/repo and create a symlink to it in folder /home/user2/repo2 and have user2 user repo2 ?
[08:46] <deni> i know i know this is completely wrong....on so many levels
[08:47] <deni> but i have a ssh jail on one user and i want him to be able to just user that one repo that is located outside of the jail folder
[08:47] <bob2> symlinks don't let you escape chroots
[08:47] <lifeless> your jail would prevent that working
[08:47] <lifeless> nevermind what bzr might support :)
[08:48] <lifeless> you could use a bind mount
[08:48] <deni> hmmm
[08:48] <deni> user2 is jailed in /home/user2
[08:49] <deni> the link is located is that folder.....wouldn't it seem like he never left /home/user2
[08:49] <lifeless> sadly no, thats not how symlinks work
[08:49] <lifeless> they get dereferenced (followed) in userspace
[08:49] <deni> why can i cd to repo2 and write to it then? :D
[08:50] <lifeless> because your shell + libc are collaborating to confuse you
[08:50] <deni> also pwd says /home/user2/repo2/ and not /home/user1/repo
[08:50] <deni> be as it may....i can still write to that folder
[08:50] <lifeless> while you are jailed ?
[08:50] <deni> yes
[08:51] <maxb> What kind of jail is this?
[08:51] <maxb> Doesn't sound very secure :-)
[08:51] <deni> aarrghh..nope..just me being stupid
[08:51] <deni> wait a secx
[08:51] <deni> *sec
[08:51] <lifeless> python -c 'import os;os.symlink("/etc", "foo")'
[08:51] <lifeless> if you run that, while jailed
[08:51] <lifeless> and ls foo shows you your /etc
[08:52] <lifeless> then your jail is, uhm, not working well :)
[08:52] <deni> lifeless: the jail is working fine, like i said it was just me being stupid (i was not logged in as the jailed user :D)
[08:52] <lifeless> there you go
[08:53] <deni> well i can always branch the repo in the jail and then have user2 push to it and then just merge from time to time
[08:53] <maxb> The important thing to understand about symlinks is that they are *just* a pathname stored on the disk, that's all - and all of the dereferencing happens at the priviledge level of the user accessing them
[08:53] <deni> lifeless: maxb: tnx
[08:53] <lifeless> deni: or use a bind mount
[08:53] <lifeless> deni: which will do what you wanted
[09:01] <Rakesh> hello friends
[09:01] <Rakesh> please help me in configuring bzr in windows
[09:04] <bialix> Rakesh: ask
[09:04] <Rakesh> helloo
[09:05] <Rakesh> i installed bzr on my machine
[09:05] <Rakesh> and its configured with some RSA key pair
[09:05] <bialix> its = what?
[09:05] <Rakesh> bzr
[09:06] <Rakesh> i mean we have RSA public key used in server
[09:06] <bialix> I don't think bzr is configured with RSA, are you talking about your SSH client software?
[09:06] <Rakesh> and i ned ma private key in ma system to get connected
[09:07] <bialix> what is your SSH client software?
[09:07] <Rakesh> hmm:(
[09:07] <bialix> I'd recommend you to use pageant from PuTTY package to hold your private key
[09:07] <Rakesh> actually i created it from my Ubuntu machine
[09:08] <Rakesh> and configured my bzr client in there
[09:08] <bialix> on bzr it won't work the same way as on Ubuntu
[09:08] <bialix> maybe you want install Cygwin to have SSH client similar to Ubuntu
[09:09] <bialix> or just use PuTTY
[09:09] <bialix> you need 2 utilities: puttygen and pageant
[09:10] <bialix> using puttygen you have to import your private key and then save it in putty format.
[09:10] <bialix> use pageant as ssh-agent software, load your putty-private key there
[09:11] <bialix> also it will be better to ensure you don't have ssh.exe or plink.exe in your path, otherwise you'd better set env variable BZR_SSH=paramiko
[09:11] <bialix> Rakesh: does it make sense for you?
[09:12] <Rakesh> sorry, i ment SSH key
[09:12] <Rakesh> its not RSA key :(
[09:12] <bialix> I know, I told you about SSH key
[09:12] <Rakesh> yea , thanks
[09:13] <Rakesh> ok  i have the kay pair and i configured bzr in Ubuntu
[09:13] <Rakesh> and what i need is i want to configure it for my windows machne too
[09:14] <Rakesh> and  whan i tried to configure it there it asks for my private key password
[09:14] <Rakesh> and its not detecting the password
[09:14] <bialix> Rakesh: re-read my advices starting from "or just use PuTTY" sentence
[09:15] <bialix> are you using GUI tools or command-line?
[09:15] <Rakesh> GUI
[09:15] <bialix> that's could be a problem then
[09:16] <Rakesh> :(
[09:16] <bialix> it seems you have Cygwin's ssh and ssh-agent instaled in your system, right?
[09:17] <Rakesh> not actually
[09:17] <Rakesh> I have the key which is generated  from ubuntu
[09:18] <Rakesh> ssh-keygen -t rsa -b 4096 -f ~/.ssh/rakesh
[09:18] <Rakesh> this is the way i created there
[09:18] <Rakesh> and i need to use he same key in windows
[09:19] <Rakesh> since that public key is used to register my acount in remote server
[09:19] <deni> lifeless: mount --bind works like a charm. tnx
[09:19] <bialix> Rakesh: you wrote: "and whan i tried to configure it there it asks for my private key password"? so I assume you have ssh-agent software
[09:20] <Rakesh> what this ssh-agent doing :)
[09:20] <bialix> Rakesh: if this is not your computer, then maybe you need to ask your system administrator about installed software
[09:21] <bialix> ssh-agent holds your private key so you don't have to type the password every time
[09:21] <lifeless> deni: cool
[09:22] <Rakesh> bialix: okey , i will try once more and if angin the problem persist wil inform the admin
[09:22] <Rakesh> bialix:  thanks for your supoprt
[09:22] <Rakesh> :)
[11:03] <Kamping_Kaiser> dont' remember if i asked this here - is ther a way to remotely rename branches?
[17:01] <hellnest_> hello i've got this invalid http Missing the Content-Range header in a 206 range response
[17:01] <hellnest_> i'm trying to clone a bzr repo
[17:58] <ndurner> Hi
[18:00] <ndurner> The patch https://code.launchpad.net/~ndurner/bzr/bzr-ftp checks for the return code 502, but now I have encountered a server that returns 452 (IIRC) at work
[18:02] <ndurner> 452 is not compliant to RFC 452, which states "insuffient storage"
[18:02] <ndurner> Now, what's the right thing to do here?
[18:03] <ndurner> Interpret any 4xxx and 5xx return code as "server does not implement this"?
[18:03] <ndurner> Or check for 502 and 452 explicitely?
[18:04] <ndurner> * RFC 959
[20:34] <ndurner> it's 451, not 452
[20:35] <ndurner> "Append/Restart not permitted"
[23:06] <jo-erlend> how big overhead does bzr introduce? I was considering placing my entire ~ in revision control. But I have some folders that contain rather large sets of changing data, such as Videos where I download to, and VM_disks. I would like to know when those files have been added, changed and removed, but I would not want to compare their content. What do you think? Will this cause lots of overhead, or should it be doable?
[23:12] <bob2> not doable
[23:13] <bob2> there'sa  thing for git called git-annex, possibly someone has done something similar for bzr
[23:13] <jo-erlend> bob2, why not?
[23:14] <bob2> ?
[23:14] <bob2> because multigigabyte files are expensive for bzr to deal with
[23:15] <bob2> (or any other tool, too, hence git-annex et al)
[23:21] <jo-erlend> ok. But I can exclude those directories?
[23:35] <lifeless> jo-erlend: yes you can exclude them
[23:36] <lifeless> jo-erlend: bzr will need enough memory to hold the entire directory listing in memory, + 2 times the size of your largest file that you choose to version control.
[23:38] <jo-erlend> oh, ok! That makes sense and that makes it rather clear why I cannot put those directories in vcs :)
[23:52] <lifeless> jo-erlend: entire ~ is probably fine; we've tested up to 1M files
[23:52] <lifeless> jo-erlend: though it may get a bit slow :) - and any subdirs that are also bzr projects won't get versioned [because they are their own projects]