[00:00] mgedmin: MySql is a very big branch - I need to upgrade my copy to 1.9, because thats what it's using [00:00] 1.9 = btree index [00:00] how many megs in ~/.bzr ? [00:01] mysql moving to bzr was very good PR for bzr [00:02] hai marius [00:02] we're fast now [00:03] with the 2a format, which will be default in 2.0 [00:03] mgedmin: ../.bzr/repository = 643300 kb [00:03] The thing that makes qlog slow is 67003 revisions [00:04] mgedmin: SmileyChris said "and now it's the pre-unbound revision" - So I don't think he has lost anything, just the merge that bzr up did. [00:04] lifeless: hi! [00:05] awesome to hear that! [00:05] Hi lifeless [00:05] hi garyvdm [00:11] zope 3 svn currently has 103359 revisions, but then it contains a multitude of projects [00:11] still, 6-digit revnos scare me -- they start to look more like shortened hashes [00:34] sup [00:35] m i license-technical ok, if my code is cappable of running a directly imported bzr, but the backend-loader chooses to run it in another process instead [00:35] i taught anyvc to run its backends in different processes and on different computers [00:36] ronny: I'm not sure I understand the question [00:36] lifeless: ie not getting gpl imposed [00:37] lifeless: i want to have fontends for anyvc that are bsd-licenses, so i need to get the gpl'd backends out of the process [00:38] IANAL [00:38] I suggest you bring this up on the list [00:39] http://www.gnu.org/licenses/gpl-faq.html may have something to say [00:46] i belive im ok [00:47] night === JaredWigmore is now known as JaredW [04:52] hello. where can i read about what's coming in 2.0? [07:02] hey guys, i'm wondering if there's a flag i can pass to make bzr merge into an existing directory... [07:04] i'm sorry, i meant 'bzr branch' [08:44] james_w: do you know why config-manager was removed from debian? [08:44] (and ubuntu in intrepid) [08:44] all I've found so far is 'abandoned upstream', which is utterly utterly false. [08:53] lifeless: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=501890 [08:53] Debian bug 501890 in config-manager "config-manager: should this package be orphaned?" [Important,Closed] [08:54] yes, found that since asking - thanks [08:54] so I guess no maintainer in Debian would be more accurate [08:54] I'm a bit miffed [08:54] I wasn't asked [08:54] no beer for anand [08:54] * lifeless ITP's the silly thing [12:22] hi [12:25] i'm quite new to cvs-managing projects. i'm looking for a way to create/manage a bzr repository via ftp on a shared hosting account without shell access. is this possible? [12:26] bzr init ftp://username@host/path/to/branch [12:26] you should give the getting started guide a read [12:27] i did, but i guess i got lost somewhere :) didn't hope it would be that easy. thx! === Pilky_ is now known as Pilky [14:16] is there a way to get the status of the remote branch? "bzr status ftp://user@domain/path/branch" doesn't work for me, although init did [14:18] could it be the ftp plugin won't work with that command? [14:20] No remote protocols handle WT's, so bare status won't work. [14:20] Though I think historical status should... [14:21] (of course, if you init'd it remotely, there wouldn't be a WT anyway, so even if it worked, it wouldn't ;) [14:23] i guess i didn't fully grasp the concept ... *sigh* looks like there's no way around the deeper docs :( [14:25] Sooner or later, the docs always win... [14:26] yeah, "i fought the docs, and the docs won" [14:26] * fullermd wanders off for a few hours. === Pilky_ is now known as Pilky [17:03] hello [17:04] i have a branch i was working on a feature in, and i merged it to head, where did these changes go [17:04] ? [17:04] log-wise, i mean === abentley1 is now known as abentley [18:56] hi again === abentley1 is now known as abentley [18:59] asked before, but i think i can clarify a bit what i'm trying to achieve: [18:59] i'd like to directly work with a remote branch via ftp. that is, init the branch remote, add all files remote, commit remote. there are several reasons not to mirror the branch and work locally on it. environment is a shared hosting space without shell access. [19:00] is this possible? [19:01] In a word, no. [19:01] darn [19:02] is there an in-three-words, "you could instead ..."? [19:02] Well, sorta, but it mostly boils down to having or faking shell access. [19:03] I suppose if somebody smoked enough crack to write a ftpfs implementation, you could fudge it that way. You'd probably have to REALLY like pain, though. [19:04] Basically, the only thing you can do via remote protocols is ship revisions around. You have to make 'em somewhere localish. [19:04] well, the server does speak python. would that help? [19:04] And while you COULD theoretically do that without having a local copy of a branch or local files and all, it would have to be such a contrived setup... [19:05] e.g. would it be possible to set up bazaar as cgi and commit via a web interface? [19:05] You could write a web interface that would do such things. I'm not aware that anybody has. [19:06] If the CGI could write into $WHEREVER and execute /usr/local/bin/bzr (or, if you wrote python, load bzrlib and do stuff directly), it would be workable. [19:06] I dunno about _usable_, really, but... [19:07] The best answer, really, is to address the "reasons not to mirror the branch and work locally on it", IMO. [19:08] That's the point of the D in DVCS after all ;) [19:09] it's the differences between the local and the remote environment and the dependance upon some db tables that would have to be fetched each time [19:10] i don't even need anything fancy like branching or merging. a simple add, commit and revert will be all i need [19:12] The quick&dirty way would be to just throw together a bargain-basement-hax0r style CGI where you just type in the command lines and it runs them. [19:12] sounds tempting :) [19:12] (of course, it goes without saying that whatever you do needs to at least PRETEND to be secure ;) [19:13] like, don't post the link on boards, right? :D [19:13] You'd probably be happier in the long run with local dev environments, though. I do web work, and essentially all my dev is on my workstation. [19:13] It means I have to make a copy of the DB schema, but that's easy. It's in the VCS anyway. [19:14] Yeah, that would be a useful first step :p [19:14] (of course, if it were properly secured, you COULD post it on boards. That's kinda a touchstone, actually...) [19:15] that gives me an idea: placing fake haxxor links on script kiddies boards to pages rimfilled with adds (or disguised subscription forms) [19:15] ads [19:18] but yes, i should probably rework my workflow [19:20] I have what seems a very simple question, but Google couldn't help me [19:20] I want to export my current revision, but without any history nor logs [19:20] just the current files, in the state they are [19:20] gcerquant: Something different from 'bzr export'? [19:21] I guess I could remove all .bzr files, but was looking for something less hacky [19:21] I tried bzr export, but it seems all the info is still there (commit logs, ...) [19:21] unless I miss something? [19:22] * fullermd blinks. [19:22] What's the _opposite_ of missing something? [19:22] export should just export the files... [19:24] I just tested again [19:24] Here is what I did: [19:24] In my repo, I did: bzr export /tmp/test [19:24] it exports all my files as expected into /tmp/test [19:25] but when in /tmp/test, if I do bzr logs, I get all the commit logs [19:25] I don't want them [19:25] Are you sure you're in /tmp/test? [19:25] export has never exported a .bzr/ dir at all, TTBOMK, much less one with a copy of the history. [19:26] And doesn't for me here in a quick test. [19:26] there is no .bzr directory in /tmp/test [19:26] but bzr revno gives me the revision number [19:26] and bzr logs all the commit logs [19:26] What does info tell you? [19:27] Checkout (format: pack-0.92) [19:27] Location: [19:27] checkout root: /private/tmp [19:27] checkout of branch: /directory_of_initial_branch [19:28] Is /private/tmp the same as /tmp? [19:28] yes [19:28] So, you're not seeing logs of /tmp/test. [19:28] You're seeing logs of /tmp. Because you're inside that branch (presumably, in a 'unknown' dir inside it, but that doesn't bother bzr) [19:28] ? [19:30] Ok, I get it [19:30] Branches extend down to infinity. If you mkdir /tmp/foo/bar/baz ; cd /tmp/foo/bzr/baz ; bzr log you'll get the same thing. [19:30] ok, I got it [19:30] at the beginning of my test, I did a bzr branch to /tmp [19:31] I had removed it, but the invisible .bzr directory was still there [19:31] and that's what I was seeing [19:31] thanks for the help [19:31] NP :) [19:31] good evening [19:59] does bzr-svn support something like partial clone/checkout? of a svn repo? [23:08] hi [23:10] if i create a bzr server, how do i actually write to the branch? [23:10] meoblast001: That question is rather nebulous to try and answer. Can you be a little more specific? [23:11] do i just do bzr push bzr+ssh://mysshaccount@myserver.com/foo/bar? [23:12] If (a) bzr is installed or otherwise accessible on the server, and (b) you can write to /foo/bar, yes. [23:12] if the ssh user can, not the bzr server user, right? [23:13] They're the same thing in this case. [23:14] bzr+ssh boils down to "ssh host -c 'bzr --talk-to-myself-on-stdin'" [23:21] fullermd: nothing is happening :( [23:21] if the bzr server user doesn't own the directory, does it fail? [23:24] It'll error if it can't write there, yes. [23:24] i just get a 0 transfer rate [23:26] fullermd: / 0kB @ 0kB/s [23:26] i did bzr push bzr+ssh://braden@192.168.1.100:9090/meo3d [23:29] Odd port to be running sshd on, but that's your business. Path seems a little more suspect; you really mean to put things right off the root? [23:29] oooh [23:29] the sshd port? [23:29] i thought it wanted the BZR port [23:29] which do i give it? [23:29] If you're talking to a "bzr serve" server, you want to use bzr://, not bzr+ssh://. [23:29] Very different things :) [23:30] what is the difference? [23:30] bzr: ERROR: Server sent an unexpected error: ('error', 'Permission denied: "/meo3d": [Errno 13] Permission denied: \'/meo3d\'') [23:30] who needs permission to that folder? [23:30] ssh or bzr? [23:30] i'm really running short on time and wasn't expecting all these problems [23:30] The difference is that bzr:// talks to a server you manually started (as whatever user you ran it is). [23:31] what if i want to talk to both? [23:31] bzr+ssh:// ssh's into a server (as whatever user you specify), spawns off bzr on the server side, and talks to it across the ssh channel. [23:31] oh, ok [23:31] now i get it [23:31] It's the difference between pserver and :ext: in CVS. [23:32] i chowned that directory to braden [23:32] how can i check that it is? [23:32] (also, not that bzr:// servers by default won't allow any writes at all, since they don't do any AAA. [23:33] fullermd: how do i get a listing of who owns a particular directory? [23:33] ls? [23:34] man ls told me nothing about how to get the permisions [23:34] at least not while skimming the manual [23:35] -l [23:35] drwxr-xr-x 2 braden braden 4096 2009-08-29 18:16 meo3d [23:35] ssh is ran by root, and i log in as braden [23:35] and bzr server is ran by user bzrserver [23:38] why do i choose to do things when i don't have time :( [23:38] meoblast001: when you say bzr server is run by [23:38] the user [23:38] if you use bzr+ssh, the bzr server is run by *the user that logs in via ssh*. [23:38] sudo su bzrserver, password, bzr server --blah blah blah [23:38] no [23:39] meoblast001: not with bzr+ssh [23:39] how do i make this not so [23:39] i want to run 1 bzr server for all users [23:39] if you're short on time, put it back the way it was ;) [23:40] meoblast001: I mean, I'm delighted to help you out, just thinking it might be less stressful for you to do it another day [23:40] with no vc and transferring a freaking tarball to my laptop [23:40] i'm stressed all the time [23:40] might as well get it out now than later [23:40] ok. so why do you want 1 server (this is a serious question, its not obvious to me) [23:41] yes [23:41] 1 server, and then i can use ssh for any user on the server, given he/she has access to the directory we are pushing to [23:41] you don't need 1 server for that [23:41] bzr+ssh does that anyway [23:42] so i just don't run a bzr server at all? [23:42] thats right [23:42] ok, so what about that error i'm getting? [23:42] bzr+ssh creates a server process when the user logs in [23:42] the permission denied one? [23:42] yes [23:42] wait [23:42] wtf [23:42] i'm confused [23:42] the user that is logging in doesn't have access to the folder [23:42] how does it know the root directory of the repo? [23:43] and yes i do [23:43] i checked [23:43] drwxr-xr-x 2 braden braden 4096 2009-08-29 18:16 meo3d [23:43] bzr push bzr+ssh://braden@192.168.1.100:8090/meo3d [23:43] when you type 'bzr push bzr+ssh://host/foo/bar/baz', it starts a server at /, and looks at foo/bar/baz for the repo [23:44] meoblast001: that will be trying to access /meo3d [23:44] what if i don't want the repo at the root of the server? [23:44] then put it in a subdirectory [23:44] e.g. /srv/bzr/meo3d [23:44] bzr push bzr+ssh://braden@192.168.1.100:8090/srv/bzr/meo3d [23:45] oh [23:48] lifeless: something weird (but not fatal) happened [23:48] i pushed on my desktop to my server [23:48] then branched on my laptop [23:48] and an OpenSSH message came up on my desktop [23:48] heh [23:49] is that normal? [23:49] I don't know. What was the message? [23:50] prompt for password [23:52] thanks for the help [23:52] i should be going though [23:52] this is only a temporary personal BZR server so that shouldn't matter [23:52] i use launchpad for my actual projects [23:52] bye [23:57] jelmer: welcome back [23:57] https://code.launchpad.net/~lifeless/subunit/next/+merge/10871