[03:13] how would I check out an svn repository into a subdirectory of my bazaar branch without any ill effects? I want to still pull changes from the svn if needed, and also have the svn be managed by bzr. [03:22] cellofellow: Why not use a bzr-svn checkout? [03:22] cellofellow: That way you have a bzr branch that can cooperate with the svn repo. [03:24] I just did that, and it seems to work ok. [05:39] phoenixz: Hey. Did you figure out what you needed? [05:40] Hi, yeah, more or less.. Im still toying a bit with it before I will transer all repos.. I need the different ideas of BZR to settle a bit.. [05:42] Absolutely. Much fiddling preceeds real work :) [05:43] Its not so diffuicult I suppose, its like... switching querty to dvorak.. :) it sucks! [05:47] Did the mainline stuff make sense? That's probably a pretty big mental departure. [05:49] It did actually, yes, its just something.. needs to sink in [05:49] I have 2-3 years of SVN / SVK stuff to squeeze out of my brains [05:51] Well, don't squeeze it all out. Otherwise we can't pressgang you into helping run the SVN User Reeducation Camps. [05:52] Heheheh, I doubt it will all sip away === lionel_ is now known as lionel [10:43] morning [10:44] $ bzr pull . [10:44] bzr: ERROR: Cannot lock LockDir(http://bazaar.launchpad.net/~exaile-devel/exaile/exaile-0.3.0/.bzr/branch/lock): Transport operation not possible: http does not support mkdir() [10:44] can anyone elaborate on this? [10:46] you probably have a checkout [10:47] over http [10:47] and you can't do mutating operations - push/pull/uncommit/commit to a checkout if you can't write to the master [10:49] thanks lifeless [10:49] so how can update my checkout? [10:49] bzr update [10:49] (very non-intuitively named command :P) [10:49] * BUGabundo tries [10:53] Updated to revision 2567. [10:53] seems to work :) [10:53] thanks lifeless luks [10:53] BUGabundo: the rule to remember is: [10:53] if you use it like svn, *use it like svn* [10:54] I don't :) [10:54] bzr is my 1st cvs [10:54] you do if you use 'checkout' [10:54] its centralised-style workflow, which is all that svn does [10:54] I mean, I don't use it like svn [10:54] BUGabundo: I think you mean 'vcs', ''cvs'' is the name of a tool [10:54] which one is better for single offline keeping? [10:54] if it's your first vcs, you should read the manual [10:54] bzr branch or checkout ? [10:55] luks been using it for a while, read the man long ago. much has changed since bazaar started :) [10:56] but not this [10:56] http://doc.bazaar-vcs.org/bzr.2.0/en/user-reference/bzr_man.html#concepts [10:58] which one is better for single offline keeping? bzr branch or checkout ? [10:59] seems to be CO [10:59] since I don't need the history :) [10:59] specially lightweight CO [10:59] lightweight CO is very slow over the network [11:00] :( [11:00] * BUGabundo goes reading more [11:02] Hi Luks [11:03] luks: I generated quite a bit of bug mail. So you may have missed my response to bug 421039 [11:03] Launchpad bug 421039 in qbzr "Diff windows launched from qci are always in the background" [High,Confirmed] https://launchpad.net/bugs/421039 [11:03] luks: Is that still a problem for you? [11:04] GaryvdM: let me check [11:06] GaryvdM: yes, the problem still exists in lp:qbzr [11:06] the commit window is always above the diff window [11:07] Luks: ok - I'll get hold of that other laptop, and try debug it. [11:08] As I said, it dose not do it on my computer - And I have an almost identical setup - so it is very weird. [11:08] luks: What version of qt and pyqt do you have installed? [11:10] GaryvdM: PyQt 4.4.3 and Qt 4.4.0 [11:12] luks - Ah - I'm running Qt 4.5 - I'll try it with 4.4 [11:20] luks: Asking you this cause you did qconflicts - do you think bug bug 174509 is important? [11:20] Launchpad bug 174509 in qbzr "qcommit/qbrowse: Show conflict status in file list and an option to launch a merge tool to resolve them" [High,Confirmed] https://launchpad.net/bugs/174509 [11:27] GaryvdM: I wrote qconflicts, but not for myself :) [11:27] I don't really have an opinion on the bug [11:29] I guess it's not 'important' [11:29] maybe just a way to launch qconflicts from qcommit would do it === davidstrauss_ is now known as davidstrauss === Adys_ is now known as Adys [12:53] lifeless: you're the author of cia_baz.sh ? this script runs on the server, right? Savannah admins seem a bit confused about it, and I don't know what to answer them === davidstrauss_ is now known as davidstrauss === davidstrauss_ is now known as davidstrauss === abentley1 is now known as abentley === abentley1 is now known as abentley === abentley1 is now known as abentley === abentley1 is now known as abentley === abentley1 is now known as abentley === abentley1 is now known as abentley === asac__ is now known as asac [21:05] Hmm. I am getting an interesting error [21:06] 'bzr info' in my branch says 'bzr: ERROR: No repository present: ""' [21:06] pretty sure it never was in a repo, I branched it from a standalone branch on another machine [21:07] Every branch always has a repo, otherwise there's no place to store revisions. [21:07] Can you access that branch locally, wherever it is? [21:07] oh, of course [21:07] nope, bzr info says that locally, it's a file:/// url [21:08] What's in .bzr/ ? [21:08] Hmmmm [21:08] there is indeed no repository directory [21:08] what a curious turn of events! [21:09] oh, there's checkout/ [21:09] anyway all i have is: [21:09] 'bout the only way you'd ever get bzr to produce that would be if it were in a shared repo, and then moved out somehow. [21:09] branch branch-format branch-lock checkout [21:10] fullermd: fwiw, bzr-git now supports dwim revspecs :-) [21:11] jelmer: So now all we need is bzr-nongit to do to ;> [21:23] nyu: I'm not the author [21:24] nyu: but I'll do what I can to help [21:24] lifeless: thanks! [21:25] lifeless: the savannah admin said he thinks this script would have to be run in client side [21:25] is this correct? [21:26] I don't know, I haven't seen it [21:27] uhm... "Robert Collins contributed a client for Bazaar, using email delivery: cia_baz.sh." [21:27] wow [21:27] that's you? [21:27] must have been a long long time ago [21:27] in a galaxy far far away ;) [21:28] http://cia.vc/doc/clients/ distinguishes between bazaar and bazaar-ng [21:28] I'm using bazaar-ng, right? [21:28] yes [21:28] this page looks ancient [21:28] cia_baz.sh is for 'baz' which was a fork of tla [21:29] cia_bzr.py instructions say "Copy this file to ~/.bazaar/plugins [21:29] " [21:29] looks like client side === smerrill-away is now known as smerrill [21:31] I've just dropped micah an email [21:31] to update the page [21:31] jelmer write the cia plugin for bzr [21:31] cia-clients - clients scripts for CIA commit notification on IRC [21:31] heh [21:31] it's even packaged [21:32] as for client/server - all bzr plugins get installed in the same place; whether they work on servers or not depends on whether they need all history or only the most recent commit [21:33] I'd expect the CIA plugin to work on a server - but savannah are not running a server at the moment [21:33] (sftp:// isn't a server,its a file system :)) [21:35] lifeless: ssh+bzr:// ? [21:35] yes [21:35] or bzr:// [21:35] they run that too [21:35] or http:// with a server configured [21:35] nyu: then you should use bzr+ssh! [21:35] its much faster than sftp [21:35] https://savannah.gnu.org/support/index.php?107077 [21:36] ok good [21:47] the script now lives in /usr/share/pyshared/bzrlib/plugins/cia/__init__.py [21:47] as provided by cia-clients package [21:48] do I still have to copy it to ~/.bzr ? this one looks like a system-wide plugin repository [21:49] config = branch.get_config() [21:49] project = config.get_user_option('cia_project') [21:49] uhm [21:50] my modest python skills tell me this is a setting I need to feed into bazaar somehow [21:52] systemwide is fine [21:52] and yes you need to get the cia_project option [21:52] either in .bzr/branch/branch.conf [21:52] or ~/.bazaar/locations.conf [21:52] or ~/.bazaar/bazaar.conf [21:52] bzr help configuration for an overview [21:54] thanks [22:48] Hey all - quick question= Does anyone know what the BZR_SSH environment variable should be set to on win ? [22:48] uhm any idea what I'm doing wrong? [22:48] rmh@thorin:/tmp$ bzr co bzr+ssh://robertmh@bzr.savannah.gnu.org/grub/branches/experimental/ [22:48] No handlers could be found for logger "bzr" [22:48] bzr: ERROR: Repository KnitPackRepository('file:///tmp/experimental/.bzr/repository/') is not compatible with repository RemoteRepository(bzr+ssh://robertmh@bzr.savannah.gnu.org/grub/.bzr/) [22:52] syncrondi: You probably just need to Pageant from Putty [22:52] nyu: KnitPackRepository is a very old format. [22:53] GaryvdM: I've got that .. but I think I accidentally pointed my env var at plink instead [22:53] syncrondi: then just set BZR_SSH= [22:53] GaryvdM: why would bzr want to use it? [22:53] I didn't tell it to [22:54] nyu: what version of bzr do you have? [22:54] should I create a shared-repo locally? [22:54] 1.5 [22:54] nyu: That won't help (shared-repo) [22:55] nyu: 1.5 should support packs - so if you run bzr upgrade on the branch it should work. [22:56] hang on [22:56] nyu: don't do that [22:56] GaryvdM: _always_ gather more data before suggesting that [22:56] nyu: bzr 2.0 is release - I would recommend using that. [22:56] lifeless: sorry [22:57] GaryvdM: you need to check that a) you won't break interop with the users community and b) that you're not going to be accidentically causing a rich-root transition [22:57] nyu: bzr info -v nosmart+bzr+ssh://robertmh@bzr.savannah.gnu.org/grub/branches/experimental/ [22:57] nyu: that will tell use the actual format in use by grub [22:58] lifeless: I meant that he should upgrade his local branch. [22:59] GaryvdM: right, but its a shared repo [22:59] GaryvdM: and if he has other projects in that repo [22:59] GaryvdM: what will happen to them ? :) [22:59] I c [22:59] exactly, they will become incompatible with their upstreams [22:59] its the downside of the waterfall transition that is 2a [23:00] lifeless: it's odd, I created everything myself using this version :-) [23:00] nyu: did you run 'bzr init-repo' at some point ? [23:00] nyu: is there anything in your local shared repo? [23:01] http://pastebin.com/d7c182b58 [23:01] lifeless: am I supposed to have a local shared repo? [23:01] nyu: you have one, or you would not have got that error [23:01] I don't think I have any, just a bunch of checkouts and branches [23:01] lifeless: but I'm in /tmp [23:01] run 'bzr info' please [23:02] in /tmp? [23:02] whereever your cwd was when you got that error [23:02] $ bzr info [23:02] bzr: ERROR: Not a branch: "/tmp/". [23:02] ok [23:02] and /tmp/experimental was an empty dir ? [23:03] I removed it, then tried a checkout and got the error [23:03] ok [23:03] what bzr version? this could be a bug.. [23:03] 1.5 (from debian lenny) [23:03] GaryvdM: I cleared BZR_SSH, but I now I get the error saying that it isn't set. I set it to pageant.exe but it didn't like that either [23:04] nyu: so, I /really/ recommend you upgrade and use bzr 2a [23:04] it will be a lot faster, and I suspect that you have indeed found a bug in checkout [23:05] I see [23:05] syncrondi: sorry - I checked the docs - You need set BZR_SSH=paramiko [23:05] ah [23:05] nyu: sorry, the version is 2.0.0, the format name is 2a [23:06] syncrondi: see bzr help env-variables [23:06] nyu: you don't need to use the 2a format [23:06] but the 2.0.0 client is more than a year newer [23:06] GaryvdM: In my frustration with the command line verison, I installed the full package for win because I've had that running before.. I'll take a look at that doc [23:07] lifeless: will that switch to a new format? I don't want to block access to 1.x users [23:07] only if you tell it to [23:07] ok [23:07] [or run 'bzr init' or 'bzr init-repo' somewhere - those commands create new projects] [23:08] GaryvdM: So it appears that I can use plink? [23:08] to create a local shared repo, I should use "bzr init-repo --rich-root-pack" just like i did on the server? [23:08] syncrondi: Ok - I've never used it myself. [23:09] nyu: yes [23:09] GaryvdM: oh, ok. I just saw that it has plink listed in the doc [23:09] ok [23:13] oh, in the shared repo the problem doesn't happen anymore [23:13] Well, instead of doing C:\plink.exe I set it to just 'plink' and its progressed a bit ( I think ). Now it says bzr: ERROR: [Error 2] The system cannot find the file specified [23:13] Any ideas on that? [23:14] I'm using bzr branch bzr+ssh://syncrondi@myhost.com:port/path/ [23:14] I have a local tree which is used for the bazaar mirror (pulls from subversion and pushes to bzr). should I move this into the shared repo as well? [23:15] nyu: sure [23:15] do I need to re-create it? [23:15] * lifeless shrugs [23:16] I can live with that, it's just 30min ;-) [23:19] nyu: you can bzr branch form the old bzr standalone branch into the shared-repo, insted of branching from svn [23:19] nyu - That should be quicker. [23:19] *from [23:20] syncrondi: check My Documents\.bzr.log - That should give you more info. [23:21] GaryvdM: well, in fact it's a checkout rather than a standalone branch [23:21] which now that I think, seems a bit odd [23:21] A heavywight checkout, or a lightweight checkout? [23:22] default is heavy, right? [23:22] yes [23:22] then heavyweight. sounds better now [23:22] If heavy - it will be the faster from the local bzr branch [23:22] * checkout [23:23] but if I branch that checkout, it's not a checkout anymore, right? [23:23] nyu: I think - a heavyweight checkout == bound branch [23:23] so you can bzr branch [23:24] and then bzr bind svn+.... [23:25] GaryvdM: Thanks for the tip. I'm looking there now, but nothing appears obvious to me. I see this as the last line before the error: [23:25] File "C:\Python26\lib\subprocess.py", line 830, in _execute_child [23:25] startupinfo) [23:25] WindowsError: [Error 2] The system cannot find the file specified [23:26] syncrondi: please put the last section of the file in http://pastebin.org/ [23:26] GaryvdM: ah, ok [23:28] nyu: if you branch the checkout,t he new branch is a branch [23:28] nyu: it doesn't alter the checkout to branch *from* it [23:29] lifeless: but I want a checkout as result [23:29] it seems that "bzr bind" turned it into one, as GaryvdM said [23:29] but it still remembers the local dir it originated from as "parent branch" [23:30] can I make it forget that? [23:30] nyu: if you do bzr pull svn+... --remember, it will change the parent branch [23:30] nyu: or edit branch.conf [23:31] nyu: then just make a checkout [23:31] nyu: I suggest just doing what you wanted [23:32] GaryvdM: http://pastie.org/668389.txt?key=cqmpumyv5bm1lglmhidj7w [23:32] ignore all the hints about speed, they add confusion and complexity [23:33] ok [23:34] so I rm -rf trunk, and check it out again [23:34] when removing stuff from a shared-repo, do I need to tell bzr to remove it, or may I do that by hand? [23:34] just delete it [23:34] the backing database will keep the revisions [23:34] we don't currently have a gc command [23:35] syncrondi: I'm sorry - I don't know how to solve that . Maybe someone else can help you. [23:35] ok here we go [23:35] checkout in 1s!! [23:35] Thanks GaryvdM [23:36] Would you recommmend perhaps uninstalling everything and reinstalling the standalone? [23:36] no [23:36] its trying to run a process [23:36] run with BZR_PDB=1 [23:37] when the error happens you will be put into a debugger and be able to confirm the program its trying to find [23:40] lifeless: Ok, I'm not familiar with how to run with BZR_PDB=1 [23:41] Did a google, but nothing substantial came up; any hints? :) [23:41] BZR_PDB=1 bzr .... [23:41] oh, you're on windows [23:41] set BZR_PDB= [23:41] set BZR_PDB=1 [23:41] bzr ... [23:43] > c:\python26\lib\subprocess.py(830)_execute_child() [23:43] -> startupinfo) [23:43] (Pdb) [23:43] locals() [23:44] locals() ? [23:44] type that in [23:44] hit enter [23:44] show the result [23:46] http://pastie.org/668398 [23:46] its trying to run plink [23:46] is plink on your path ? [23:47] if I commit in "trunk" (whose parent branch is in svn), is my commit pushed to svn as well? [23:48] or should I use svn for commits as usual [23:48] nyu: In a hw checkout - it will push to trunk when you commit [23:49] sorry - will push to svn when you commit [23:49] ah, great [23:49] nyu: with just a branch - it will push only when you do bzr push [23:49] and I take it that if I branch trunk foo, then commit in foo, then push in foo, it also goes to svn? [23:50] (trunk is a bzr checkout of svn) [23:50] push foo trunk [23:50] yes, that will go to svn as well [23:50] however! [23:50] because of svn limitations I would not do that [23:50] cd trunk [23:50] bzr merge ../foo [23:51] bzr commit [23:51] will behave better [23:51] you wouldn't use bzr to commit to svn, or you wouldn't use branches? [23:51] ah I see [23:51] then I just do it in that tree [23:55] lifeless: Wow, I think that fixed it! [23:55] I figured C: was in my path but I had to add it