[00:19] hello, i have a problem when trying to pull a branch: http://ddorda.pastebin.com/d2e2d83e3 [00:24] Ddorda2: It shouldn't be an ugly error like that, but it means you need to get a newer version of bzr [00:25] does it have any PPA or something? [00:25] Ddorda2: Since you seem to be on Ubuntu, try https://launchpad.net/~bzr/+archive/ppa [00:25] wgrant: thank you for reading my mind [00:46] wgrant: it fixed the problem, thanks a lot :D [00:46] Ddorda2: Great. [01:05] jelmer: Just noticed you are working on bzr-hg. Are you close to doing another release? [01:05] I was just about to look at the debian package as it is blocking bzr from moving into testing === lamont` is now known as lamont [02:27] Do BzrEclipse works with bazaar 2.0? [02:29] ?? [02:31] Do BzrEclipse works with bazaar 2.0? === michaelchaffin is now known as poolfool [02:32] jonasfa: probably? try it and see? [02:59] is there a way to clean bzr --info infos? [03:01] you mean forget the related branches? [03:31] yes [03:32] you can use the --remember option to make it remember a new one [03:32] I know, but forget [03:32] how to forget [03:33] hmm, I don't know of a way [03:36] RenatoSilva: Try looking in .bzr/branch/branch.conf [03:40] wgrant: I`m already using that ?) I wondered if bzr had an option though [03:40] RenatoSilva: Not that I've seen. [03:41] s/?)/:) [03:41] ok thanks anyway [05:18] hello again [05:20] i just installed bazaar on my windows machine and i'm not quite sure how to share my project with other "scattered acoss the world" [05:20] as the page puts it [05:24] jackalborn: if you have server space anywhere you can push to that (using various protocols, the manual lists a few). If you don't have server space of your own you can use launchpad [05:27] i cant run a server off my spare machine? [05:28] jackalborn: sure. You may already be running enough of one though. For example: if you have a plain http server on the server, and ssh access to it off your windows system, you're set [05:28] Or FTP access. [05:29] or that, or being able to run bzr on the server in smart server mode, or one or two others I forgot. I'm currently fighting with the documentation, I thought it had an overview [05:29] i'm not really sure how i would tell if this spare computer has ssh or if it's running a http server [05:29] http://doc.bazaar-vcs.org/latest/en/user-reference/index.html#url-identifiers and/or http://doc.bazaar-vcs.org/latest/en/user-guide/branching_a_project.html but I think I'm not finding the right page [05:30] jackalborn: what OS does it have ? [05:30] windows xp [05:30] ok, then it's not running ssh and probably not running an ftp server [05:30] okay, my guess is "no" on both counts [05:30] who do you want to share with? [05:31] another computer on the network and another devolper outside of my network [05:31] idid find this [05:31] http://doc.bazaar-vcs.org/bzr-0.12/server.htm [05:31] for "outside of my network" you need an internet-facing server, which I'm guessing (hoping :P) this windows xp box isn't [05:31] if this is an opensource project using launchpad is an option [05:31] it probaly is not [05:31] it's not so much opensource [05:31] if it's not an open project you'd need your own server, or you'd need to communicate by sending each other bundles [05:32] indie game dev [05:32] http://doc.bazaar-vcs.org/latest/en/user-guide/sending_changes.html may be of use there [05:33] (haven't read the entire page, sorry) [05:41] no worries i'll read it and fill you in on what it says :) [05:44] Jackalborn: Are you new to revision control in general or just to bazaar [05:44] Jackalborn: ? [05:46] i use revision control at work, perforce. and i've actually set up an svn server on a windows machine and use tortoise cvs [05:46] Jackalborn: I use perforce at work as well, setup and maintain CVS and SVN at work as well (Yuk). [05:46] but i think that whole thing was just luck [05:47] yuk is a good word for it [05:47] i've also used google code [05:47] Jackalborn: I found the Scenarios page (http://bazaar-vcs.org/Scenarios) to be of hudge help ... [05:47] with tortoisecvs [05:47] Jackalborn : but maybe you can try each local hosting your own repo and then pass back and forth packs(?) [05:48] Not pretty but a solution [05:48] i'm going to try to set up a repo on my ftp [05:49] if that doesnt work then i will definatly be doing that [05:49] ugly as it may be [05:49] I have to say that bzr is a much nicer tool after you learn it ... but the perforce GUI can lead to some bad habits ... Oops [05:50] haha [05:50] perforce spoils me rotten [05:51] What's so good about Perforce? [05:51] it's simple and i didnt have to set up a thing [05:51] check in, check out [05:52] no worries [05:52] same with bzr [05:52] Isn't that rather because somebody already set it up? [05:52] although we did have a guy check out the entire depot once by accident [05:52] cd source ; bzr init ; bzr add . ; bzr ci -m "initial import" [05:52] wgrant: yes [05:53] I guess after you have been subjected to CVS hell and some of it's warts (non-atomic) well ... Perforce can be nice. [05:53] Heaven help you if your meta data gets corrupt ... and your snap shot is old ... Oops. [05:53] it took me like 2 weeks to set up that svn server and get it working [05:53] and then a month later that machine died [05:53] SVN is awkward like that. [05:54] With bzr you don't need any specialised software on the server. [05:54] I guess one of the 'nice' things about administrating a Perfoce server vs. CVS/SVN is it comes out of the 'box' and just runs on Linux /or/ Windows. [05:58] I also had to do a lot (less then a week) of hunting around the web to figure out how to setup CVS (and then SVN) on a Linux box with SSH and password authen back to active directory ... kind of crazy. [06:00] just reading that hurts my head [06:00] i cant even imagine [06:01] a matter of pointing pam_ldap/nss_ldap at the ad server? [06:02] Actually I tried the pam_winbindd (I think) ... but the pam_ldap would have been much nicer. [06:02] If you have SFU on the AD server, sure. [06:02] SFU ? [06:02] But most people use winbind or likewise, I believe. [06:03] Services for Unix. IIRC it adds some attributes to AD to make it less impossible to use nss_ldap with. [06:03] Ahhhh .... I tried that .... long time ago and well ... I hear it's much better in the last 5 years. [06:04] But again. Lets say I was setting up Bzr on an Linux box, I have an Active Directory server for my company. [06:04] Lets say SSH access, how would I tie access back to the AD ... key here is single password (or key) per user to cut down on their end? [06:48] so i set up an ftp server on my other computer and i'm able to connect to my repot and it's trunck from this computer [06:49] but now i cant seem to do anything with it all the buttons in bazzar give me an error [06:49] "Action not yet supported by current app_suite [06:53] Jackalborn: .... I am guessing you are using the new gui interface (gnome)? [06:58] windows gui interface [06:59] think i got it working though [07:00] Jackalborn: I use the command line mostly .. different from perforce but you can do a /lot/ from the command line. [07:07] davidstrauss: as answered; bzr uses a _heavily_ extended verrsion of python's unittest. If you want a easy-to-write-tests-in-environment, nose is good [07:08] its also, like bzr's tests, a single extension dimension - you only get to step away from the core unittest once. [07:08] davidstrauss: grab me Monday/Tuesday and I'll happily give you a quick rout of some other things you can use [07:09] testtools isn't a testing framework, its useful flue for folk maintaining test suites [07:09] subunit also isn't a framework, but its great interoperabolity glue between tests and reporting toolchains (e.g. things like Hudson) [07:43] thank you all for all your help [07:43] night [08:42] how do I keep bzr svn-import from consuming 10G of RAM when importing my svn repo? :) [08:43] is there a nice way to spit out a series of patches (one per rev) to separate files? "bzr log -p" is close, but I'd have to manually split it into files, or invoke it in a loop, or something [08:52] dash: do it incrementally, I guess. [08:53] dash: I assume you're running the latest, shiniest bzr and bzr-svn and subvertpy? [08:55] bzr 2.0, bzr-svn 1.0 [08:56] i am doing it incrementally [08:57] that just means it spends as much time in the spin-up phase as in the actual import phase [15:21] any idea how I can fix: bzr: ERROR: No such file: 'f2jvnez07njqaw00l791.pack': [Errno 2] No such file or directory: '/Applications/World of Warcraft Public Test/Interface/AddOns/.bzr/repository/upload/f2jvnez07njqaw00l791.pack'? [15:21] command is just 'bzr merge' [15:25] huh, createing an "upload" directory under .bzr/repository fixed it [18:37] Hi. I'm running fedora 10 64-bit. I installed bzr as an rpm using yum. I then downloaded the qbzr plugin and placed it in the plugins directory ~/.bazaar/plugins/. Now when I issue the command "bzr self test qbzr", I get the error message " [18:37] bzr: ERROR: No module named elementtree.ElementTree [18:38] You may need to install this Python library separately. [18:39] The version installed by yum is 1.16.1. [18:51] Just don't use qbzr [18:51] Command line is so much cooler [18:52] Sorry, terminal. "Windows" terms have gotten into my head. Eek. [18:53] I originally installed qbzr as a dependency for qbzr-explorer. [18:55] If I recall [18:55] ElementTree is extremely recommended for bazaar [18:55] but not needed [18:55] Like, i think it speeds up performance on XML by A LOT. [18:55] Let me find the page [18:57] I can't find the exact page [18:57] I'm not an expert on Fedora since I use Ubuntu, but this package looks like it should do it [18:57] https://admin.fedoraproject.org/pkgdb/packages/name/python-elementtree [19:05] it is needed by bzr [19:05] all the xml serializers/deserializers use it [19:06] but it's been built-in since python 2.4 or 2.5 [19:06] Anyway, it's no longer built in fedora 10 and upwards. [19:06] what version of python and bzr are you using? [19:06] Yes, I've python 2.5 installed system-wide(the one used by default_. [19:07] But the problem is that it still complains about it not existent. [19:07] can you try running the command with -Derror [19:07] I think I might have figured out the problem from bazaar log. [19:07] and pastebin the whole traceback? [19:07] something is probably not using the builtin module [19:08] Actually it was refering to elementtree.ElementTree inside bzrlib. [19:08] yes, I wanted to know where exactly [19:08] I don't have the log(; I've deleted it in favour of a fresh install of bzr 2.0). [19:08] I've installed 2.0, and going to try over. [19:09] I installed it using easy_install; hope it works this time. [19:10] Also, qbzr technically isn't a requirement for Bazaar Explorer - When I was looking into it, I think it said you can configure it to use bzr-gtk instead [19:11] I didn't have either, so I decided to try with qbzr. [19:13] mathepic: it is a requirement for bazaar explorer [19:13] it builds on qbzr [19:13] you can tell it to use bzr-gtk [19:14] "If you wish to use bzr-gtk dialogs with Explorer, select gtk as the application suite. gtk will only appear as an option if you have the bzr-gtk plugin installed." [19:14] Do a search for gtk at http://doc.bazaar-vcs.org/explorer/en/visual-tour-gnome.html [19:14] ok, we can excange quotes :) [19:14] "BzrExplorer provides most of its functionality by calling out to mini-applications provided by "application suites", namely the QBzr and bzr-gtk plugins. QBzr is required while bzr-gtk is optional." [19:15] Contradicting documentation [19:15] not really [19:15] Mine makes it sound like you can replace the qbzr applets with gtk applets [19:15] you can [19:15] but note that BE is a Qt application [19:15] Ahh., [19:15] it uses some of qbzr's functionality internally [19:16] you *can* use bzr-gtk for external dialogs where possible [19:16] K [19:16] Terminal is still the fastest way to go. :D [19:17] browsing history on terminal is definitely not the fast way to go [19:18] bzr log / / bzr diff -c / hm, no bzr st -c / back to bzr log / / bzr diff -c , and so on :) [19:18] that's now very efficient [19:18] Partially good news! "bzr selftest qbzr" doesn't complain now. [19:18] Cool. [19:18] It seems qbzr(at least the version I downloaded) was built against the latest version of qbzr. [19:20] Err, typo? [19:21] However, I still get some failing tests as illustrated here: http://pastebin.com/d7ff6f5fd [19:22] However, the self test of the explorer plugin itself is fine. [19:23] Personally, I wouldn't worry about it all that much. [19:27] Me, neither. [19:29] And now the big surprise, "bzr explorer" works. Thanks to all of you guys for your suggestions/hints. [20:21] I'd like to cleanup/fix the Bazaar repositories for a project. Is anyone available who might be able to assist me? :) [20:59] What do you mean "cleanup/fix the Bazaar repositories" [22:11] Suggestion for bzr: when committing, ask the user if he wants to add the unknown files that appeared since last commit. [22:12] Man. that would be super annoying [22:12] why [22:12] He said the *new* unknown files, so it wouldn't be *that* annoying. [22:13] yeah that would be less bad. [22:14] For example: bzr status = file 1; touch file2; bzr commit ---> The following new files has been added since your last commit: file2, do you want to add them before committing? [Y/n] [22:14] RenatoSilva: write an extension :)_ [22:15] That reminds me, I need to turn off the strict pushing thing. [22:15] Yes, please [22:16] "Yes, please" what? [22:16] I think I've done 3 commits today where immediately after committing I realized I didn't add the files yet [22:17] mathepic: I committed and pushed once without remembering to add the files, that's the main reason for my suggestion :) [22:17] * wgrant wonders why people don't 'bzr st' more often. [22:18] wgrant: Cuz we think we remember the status. [22:19] wgrant: also the unknown files list is at the bottom [22:20] pushing the interesting part off the top of the terminal :) [22:20] wgrant: in my case, status just does not have my attention all the time. When I added the files, that was not pretty clear in my mind, so that "oh mush bzr add", it was more like "oh these are my changes" [22:20] also there's like a zillion unversioned files [22:21] dash: doesn't bzr have extensions lists for ignoring files? [22:21] Why? They aren't ignored? [22:21] RenatoSilva: Yeah but they don't all have the same extension [22:21] RenatoSilva: also some files I'd want to add have the same extensions [22:23] dash: if they have the same extensions but you want to ignore just some of them, then you must not add the extension to ignore list right? [22:24] RenatoSilva: Yeah, I did that three times today. Luckily it was to my own project... [22:26] RenatoSilva: yes! I did not. [22:28] dash: by extension you mean a bzr plugin? [23:16] bug 448310 :) [23:20] where's the bug bot ha [23:22] https://bugs.launchpad.net/bzr/+bug/448310 [23:27] That would be better as a blueprint, if bazaar supports them [23:27] And Bazaar does, so yeah... [23:27] Its not really a bug [23:28] bugs also are feature requests [23:28] Blueprints are very, very rarely a better idea. [23:28] That is more appropriate as a bug. [23:28] afaik blueprints are for BIG feature requests [23:29] is it possible for files and revisions to have multiple ids? [23:29] Oh [23:29] Wow [23:29] One of my most hated grammar mistakes [23:30] "Remember him to do this" [23:30] And its in the bug title [23:31] mathepic: sorry I'm not English native haha [23:31] mathepic: what's the right form? [23:31] RenatoSilva: Remind [23:31] ... [23:40] :/ [23:40] so I have 3 bazaar trees [23:40] README in all 3 refers to the same file [23:40] but they have different file-ids [23:40] how can I make the 3 trees cross-mergable? [23:44] luke-jr: /win 2 [23:44] Argh. [23:45] luke-jr: There's probably something in bzr-rebase to help with that. [23:45] luke-jr: But you'd have to modify the history of two of the trees to be compatible with the third. [23:48] Hm, no, rebase probably won't work. [23:49] Bug #231674 is relevant. [23:49] Launchpad bug 231674 in bzr-rewrite "can't replay, need maptree support in rebase" [Wishlist,Triaged] https://launchpad.net/bugs/231674 [23:50] wgrant: we have numerous branches in the wild. I don't want to break their mergability [23:50] wgrant: I really need a repository which can understand that files have multiple file-ids [23:50] and tolerate any of them on merging trees [23:51] does such a format exist, and if not, how hard would it be to make a custom format for it? [23:51] jelmer: you about? [23:51] luke-jr: There was a discussion recently (on the bazaar mailing list, maybe?) about having revision ID aliases, but there's nothing done about that yet. [23:52] :/ [23:52] luke-jr: How did these multiple sets of history come about? [23:52] wgrant: bzr-svn's shortfalls [23:52] or rather, that's part of the cause [23:52] wgrant: our project is very old [23:52] we in fact have 6 data sources for the repositories [23:53] original CVS [23:53] forked CVS #1 [23:53] forked CVS #2 [23:53] Subversion migration [23:53] bzr-svn branch of Subversion trunk [23:53] luke-jr: Heh, the alias thing came up in exactly the bzr-svn case... [23:53] bzr-svn branch of Subversion 0.2.8 [23:53] the two bzr-svn branches are incompatible at the moment [23:53] Sounds like you imported it badly. [23:54] bzr-svn can do branches/tags fine if you tell it to. [23:55] If I were in your situation, I think I would declare a flag day and declare one history as canonical. [23:56] no, it can't. [23:57] wgrant: both branches have active development. we can't do that. [23:58] luke-jr: You would prefer to maintain a custom repository format forever rather than writing a bit of conversion code? [23:58] wgrant: hmm, that might be an option [23:58] is it possible to change file-ids easily? [23:59] luke-jr: See the conversation transcript in the bug I referenced. It suggests how to do it with rebase. [23:59] * luke-jr doesn't see what rebase has to do with this O.o