[00:05] cool [00:10] is it possible that the repository recorded my g-w permission bit when I first created it, and now is using g-w on all mkdirs because of that? [00:11] It doesn't record it, except in the perms of the dirs. [00:11] If you chmod -R xyz .bzr/*, it'll use xyz. [00:12] oh maybe its n04m's .bzr that has g-w bit... [00:12] he is the one making dirs with g-w :) === mc_ is now known as mc__ [00:16] Well, obviously. '04' doesn't include g+w ;p [00:17] ya I am not sure what the deal with the l33tsp34k is there :) [00:19] hi jelmer, I'm trying to fix some things up in olive, but it seems that any changes I make to /ui.py do nothing... any ideas? [00:19] jelmer: bugmfiling time on the cli [00:20] lifeless, which cli? [00:20] nekohayo: hi jeff [00:20] I want to alleviate those jumping progressbars, but no changes are visible... even if I change the window title from "Progress" to "Monkey", it does not work [00:20] branch registration [00:24] nekohayo: maybe it's using the one installed on the system? [00:25] jelmer: that's strange, locate "ui.py" |grep bzr does not return an exact match [00:26] nekohayo, not sure then what may be causing it [00:27] Peaker: Did you set the *remote* umask? [00:28] abentley: Its fixed now, the problem was the commiter's .bzr directory, not mine [00:28] abentley: when he created branches under my .bzr, permissions from his remote .bzr were used [00:28] That seems extremely unlikely. [00:29] abentley: he used: bzr get URL-TO-ME URL2-TO-ME and URL2-TO-ME inherited the .bzr permissions of hiw CWD, not of URL-TO-ME [00:29] When you use bzr+ssh, the permissions of the remote bzr are used, as if you were ssh-ing in, and creating the files by hand. [00:30] abentley: Looked a bit at the bzr+ssh code, it has an mkdir method that takes a mode and which means its likely that his local machine is telling the bzr+ssh url handler to set the modes inherited from his local .bzr, which in fact happened (as he fixed his .bzr permissions, the url2 started being created with g+w) [00:30] Oh, you're not talking about creating files over bzr+ssh. [00:30] jelmer: hmm. Now that makes me wonder, what is ui.py for inside the bzr-gtk trunk? [00:30] if it doesn't use it [00:30] nekohayo: it *should* use it [00:30] You're talking about retrieving data from bzr+ssh, and creating files locally. [00:32] jelmer: any way to *force* it to use it? [00:32] Owner and group can be inherited from CWD, but permissions cannot. [00:32] nekohayo: It should be using it in all cases [00:32] then wtf o_o [00:32] abentley: no, I am talking about creating a remote bzr+ssh branch of another bzr+ssh branch. For some reason it used his local unrelated CWD repository perms for that [00:32] Permissions, i.e. r/w/x, come only from the umask. [00:32] nekohayo: do this; python -c 'import bzrlib.plugins.gtk; print bzrlib.plugins.gtk.__file__' [00:33] /home/jeff/.bazaar/plugins/gtk/__init__.pyc [00:33] ah ha! wtf² :) /me removes that dir [00:34] Well, I consider it extremely unlikely that his local CWD permissions have anything to do with it. [00:34] abentley: Nope, the umask was and is 002. He created branches and they had g-w until he changed his local CWD .bzr to be g+w and then the branches he created remotely in my machine started having g+W [00:34] abentley: that's just what happened [00:35] Okay, if you say so. [00:35] yay, it works, thanks lifeless and jelmer... I can get back to trying to sanitize that progress dialog [00:39] holy cow [00:39] those progressbars are weird [00:39] they remove themselves dozens of times per second! [00:40] don't quite know what to do with that [00:40] Tell 'em to quit taking so many breaks or you'll fire them? [00:40] I should [00:40] nekohayo: I've been thinking about making it possible to set a "progress bar widget" [00:41] so windows can tell the progress bar to appear in a specific place before they start doing something [00:41] jelmer: ? [00:41] (like viz analysing history) [00:41] a big question I'd like to ask [00:42] wouldn't it be simpler to just have *one* master progressbar? [00:42] nekohayo: There is a top-level progressbar [00:42] that doesn't jump around and hold-up grandma [00:42] so you could choose to ignore all sub-progress bars [00:42] ignore? [00:43] nekohayo, not display [00:45] when I "bzr pull" into a bound checkout, does it pull into my working-dir/branch, or into the remote bound branch and syncs us both? [00:46] jelmer: hrmm. Do you have a branch of that or something? [00:46] Peaker: The latter. [00:46] a bit beyond my skillset i think, but if you want a UI reviewer I'd be glad to comment, etc [00:46] fullermd: cool, what I'd expect :) [00:47] [the coding being beyond my skillset :] [00:47] nekohayo: No, not yet. The progress bars need a lot of work, not just the way they look but also the way they're handled inside of bzr-gtk/olive [00:48] nekohayo: Comments the UI are more than welcome on the mailing list or the bug tracker [00:48] jelmer: do you have a bug/discussion thread about what you're planning to make them look like yet? [00:49] nekohayo: Ideally, they'll be integrated in the various windows [00:49] nekohayo: and no longer be standalone windows [00:49] nekohayo: they really need a lot of refactoring first before we can look at the specific ui [00:50] jelmer: Hm. Is the weird clipboard interaction of the post-tab viz intentional? [00:51] fullermd: clipboard interaction? [00:51] I'll take that as a no, then? ;) [00:52] jelmer: actually, I think the best way would be having an "info pane" like gedit's at the top below the toolbar [00:52] have you ever seen it? [00:52] fullermd: yes :-) [00:52] I pull up viz, flip over to the Relations tab. When I flip back to General, it highlights the revid and sticks it in my clipboard. [00:52] Odder still, when I flip back to Relations or exit viz, it restores the clipboard to what it was beforehand. [00:52] nekohayo: bzr-gtk is not just olive, the various dialogs can appear without olive as well [00:53] jelmer: but then how would you integrate them into the various windows? [00:53] fullermd: Ah, I see it now too [00:54] nekohayo: integrate them into the various windows themselve rather than in some top-level window [00:54] nekohayo: or did I misunderstand you earlier? [00:55] jelmer: well isn't olive just one top level window? [00:55] nekohayo: some of the windows that can appear with olive as top level window can also appear standalone [00:55] jelmer: that's what I was thinking about, with one and only master progressbar http://blogs.gnome.org/lucasr/files/2007/02/eog-error.png [00:57] nekohayo: the problem is, the master progress bar isn't always linear [00:57] we can't do linear [00:57] let's have an infamous mac OS beach ball :) [00:58] we can do linear so some parts of the operation [00:58] nekohayo: so you may want to see the subprogressbar in some cases [01:00] nekohayo: I think this would be an interesting thing to bring up on the list [01:00] but I'm getting some sleep first, g'night! [01:05] jelmer: I'll copy-paste snippets of this chat onto https://bugs.launchpad.net/bzr-gtk/+bug/127734 [01:05] Launchpad bug 127734 in bzr-gtk "Progress bars as widgets" [Medium,Triaged] [01:05] (I'm really not comfortable with mailing lists) [01:09] is it safe to delete ~/.bazaar/svn-cache/ ? bzr-svn is giving me weird trouble [01:12] nekohayo: sure, that'd be useful [01:12] Peaker: yes, it'll just be a bit slower trying to regenerate that data next time you use it [01:12] Peaker: how is it giving you trouble? [01:13] jelmer: depends on which svn branch scheme i use and which URL, I either get "invalid revision id", or divergence errors, or bad scheme/svn url [01:14] jelmer: at one point pull failed on the same URL but bind+updatee worked, but then commit failed claiming divergence [01:15] Peaker: which version are you using? [01:17] Bazaar (bzr) 0.90.0 - how do I get the bzr-svn version? [01:17] "bzr plugins" should print it [01:17] it just printed the directory. I got it from dpkg -l: ii bzr-svn 0.4.1-1 [01:18] btw: now it seems completely stuck at pulling from the svn... [01:18] Peaker: stuck how? [01:18] maybe its downloading, I don't know [01:18] I used "bzr pull https://eyal.lotem@enough.googlecode.com/svn" [01:19] entered the password, and its working for a looong while [01:19] 0.4.1 is quite old [01:19] and it worked fine on that url before? [01:20] Yeah, I think its decided to re-download the entire history, because my network is working [01:21] which is weird because I am pulling into a branch inside a repository that already has the entire history [01:21] I hope it wont do damage to my repo :P [01:22] whether it pulls doesn't depend on whether your network is up [01:22] Peaker: it says "fetching revision info" ? or "copying revision" ? [01:23] jelmer: my network is "working", as in something's downloading. I dont have progress report, so its my only way of knowing [01:23] it doesn't say anything beyond "https://enough.googlecode.com/svn is permanently redirected to http://enough.googlecode.com/svn/" after the password prompt [01:23] its just waiting [01:25] Peaker: are you sure it worked earlier? because I wouldn't know what could be asking for a password... [01:26] jelmer: it worked days ago, since then I pulled/pushed and did lots of things to that branch [01:26] nm, its not that important I gave up on pushing back to svn [01:26] how do I get a diff out of a "bzr missing"? [01:27] Peaker: I'm very curious why it broke, because it really shouldn't [01:28] jelmer: Well, I used trial&error and messed around with it a lot, so it will be hard to separate the variables here [01:28] jelmer: maybe I did something wrong.. [01:28] Peaker: bzr diff prints diffs, missing can't [01:29] Peaker: if you've only used the bzr command-line, it shouldn't break [01:29] 0.4.1 is quite old though.. :-/ [01:29] jelmer: I edited and then deleted ~/.bazaar/subversion.conf too [01:30] it had "trunk0" I tried changing to "trunk" to "none" and then to delete it [01:30] (in branching-scheme) [01:31] that shouldn't be a problem [01:32] well, I think debugging a non-important problem on a non-recent version of a plugin in a not-documented state of branches is probably not a priority :) [01:32] If I want to review the work done on a branch, relatively to its base branch, do I have to merge back, look at the diff, and then revert? [01:33] wow I'm getting 4KB/sec of ARP/DHCP background broadcast noise from my ISP :) [01:34] was wondering where that is coming from [01:34] Peaker: diff -r ancestor:../base [01:35] ancestor is a special rev token? [01:37] -> help revisionspec [01:37] ah thanks. "bzr vis" seems like a nice way of reviewing branches too [01:37] I wish it showed the whole file in meld/kompare tho [01:42] a diff is not showing me enough context. meld doesn't seem to take input from stdin, and kompare hangs :( How do you view/review your diffs? [01:44] moin [01:45] moin? [01:46] Yeah, niom spelt backward. [01:47] is that a way to view diffs? :) [01:52] ups, I just was saying Hi! [01:53] Peaker: try with, bzr cat -r revno file_name > file_name_revno, and then use kompare/vimdiff/kdiff3 [01:54] Verterok: I want the diff of a branch, not a specific file [01:55] ok [01:57] I guess I can just do the merge, and then use emacs's dvc [01:59] Peaker: I just executed, bzr diff -r branch:../trunk | kompare - [01:59] and (for me) worked just fine [02:00] Peaker: bzr difftools can run meld for you IIRC [02:00] Verterok: it should, it just hangs here :-( [02:00] night all [02:01] lifeless: what command would that be? [02:03] maybe this http://bazaar.launchpad.net/~sward-dev/bzr-difftools/trunk or http://mysite.verizon.net/sward.dev/projects/bzr_difftools.tar.gz ? [02:07] lifeless: thanks, night [02:11] difftools is giving me a diff on the .bzr dir itself :P [02:22] ok, kompare eats a -p1 diff just fine! whew [02:22] it probably doesn't like that the src/dest are identical in the default diff [02:24] or kompare doesn't like actually finding the involved files [02:25] Peaker: try passing -n switch to kompare [02:26] to avoid search for the files [02:27] nope it hangs anyway [02:27] If I use -p1 and replace old/ and new/ with parent-branch/ and this-branch/ it works tho [02:27] i killed bzr while commiting,now i cant comit cause there still is the lock.what shall i do? [02:28] I'm using kompare 3.5.8, maybe it's fixed in this version [02:28] mc__: try with 'bzr break-lock' [02:29] Verterok: you saved my day! [02:30] Verterok: 3.5.8 here too [02:30] Verterok: you can bzr diff | kompare -o - ? [02:31] yes, it's work fine here, my branches are small ~50 revisions [02:32] and I'm making a diff of 16 revisions [02:33] Wow. 2592.1.25.2.7.1.28.1.6.1.3.1.9.2.1.3.74.1.31.3.18.1.9.1.2.1.12.1.8.1.46.1.18.1.1.2.21.2.10.3.8.2.9.1.9.1.1 is an AWESOME revno. [02:34] hehehe [02:34] Verterok: well i did bzr break-lock,and it worked great. but now it hangs while commiting first it goes super fast but then it hangs with "[====================== ] Uploading data to master branch - Stage 4/6" [02:35] the command i use is "bzr ci -m "now we've got working scaffolds(really)"" [02:36] why does difftools show the .bzr in the diff? [02:36] [02:36] bzr(python) uses 0% cpu [02:36] Sure, but what's it using on the network? [02:36] Uploading data doesn't take much suds. [02:37] i dont how i can find out its bandwith usage [02:37] I guess I can use difftools and ignore the .bzr dir [02:37] but it shouldnt take that long,the files are only several kb [02:40] trying to write a plugin to create a command that combines two commands. How do you invoke a clean-tree --missing via bzrlib? and other commands? === mw is now known as mw|out [02:41] Isn't clean-tree bzrtools, not bzrlib? [02:41] oops it is yeah [02:41] Yes it is. [02:42] ok but to run remove_tree does it make sense to instantiate its cmd_remove_tree and run that? [02:42] hmpf,now i got "Read from remote host austriangeekforce.net: Operation timed out [02:42] " today is not my luckiest day [02:42] or use bzrdir.BzrDir.open(location) that's behind the cmd_remove_tree? [02:43] That, I have no idea. I suspect reimplementing the command isn't the best course, though. [02:45] I am not sure how to instantiate a cmd_* object [02:52] ok cool I have a new remove-full-tree that kills ignored files and then runs remove-tree :) [02:53] "find . -exec 'rm -rf {}' ';'"? [02:53] Well, excluding .bzr. :P [02:54] Yeah, that sure would save some space... [02:54] As a bonus, it not only cleans the tree, but removes unreferenced revisions from the repository :p [02:54] fullermd: heh its also to avoid a mess where by my emacs is editing files from multiple branches [02:54] I prefer to have only the active branch's files in place [02:55] I think ignored files should be considered disposable, and thus remove-tree should remove them... since it doesn't I made this plugin [03:08] now this looks like a bug,commiting small changes work,but if there are several files added bzr hangs for several minutes and then aborts [03:09] where can I publish a branch of the plugin I made? [03:09] Peaker: launchpad.net ? [03:10] open a whole project for it? overkill, its 10 lines of python [03:10] is there some wiki somewhere where I can paste a branch address? [03:10] if the project is on launchpad you can list your branch there i think [03:11] yeah I don't want to open a whole project for it [03:12] contact the original author,maybe he will adopt your changes [03:14] editing the site wiki table is insane [03:27] jelmer: about? [04:09] Peaker: You can push a branch to your own personal space without creating a project. [08:19] bzr currently 'crashes' when it tries to make a symlink on Windows, instead of just printing a warning and doing nothing or making a text file like in svn [08:23] s/currently/Bazaar (bzr) 0.90.0 [08:25] when making a light weight checkout, does the download not include the history either? [08:27] cause a friend is making a light weight checkout from my machine and its taking ages, as if its downloading the entire history [08:29] or can that happen because I'm using a dumb server here? [08:35] Peaker: Lightweight checkouts don't save any of the history, but if it's a dumb server, it probably needs to download quite a lot to get the latest version of all files. [08:35] Peng: Yeah that's what must have happened [08:35] Peng: I was wondering if there's a silly download-history-to-discard-it thing [08:35] but I guess that's only because of the dumb server [08:35] anyhow, thanks, gotta go [09:05] Good day everybody [09:05] I've got a, probably easy, question [09:07] I have a file that contains (default) configuration options. Users edit this file with for example their username. [09:07] I've want to keep track of the default. But want to ignore the edits made by users. [09:07] Is there a way to ignore edits to this file. Except when explicit commited? [09:10] I usually version myconfig.conf.default, which everybody has to copy to myconfig.conf [09:11] Ok... that's a solution [09:14] Thanks until a better solution is found this one will do [09:14] I haven't found anything better yet :/ [09:15] Ok... to bad .bzrignore ignores only those file not yet in the repository [09:30] Someone else asked for that feature a couple months ago. It would be pretty useful. === weigon_ is now known as weigon [09:36] Yes it would be very useful was there a feature request or something created? [09:36] Yeah, I think so. [09:36] A spec. [09:36] Any idea where I can find that one? [09:37] Honestly, I didn't like it much. It was kind of complex. [09:37] Bloged: Somewhere on http://bazaar-vcs.org/ :P [09:38] Yep that is something I thought so... but I've searched for this feature and couldn't find anything remotely like it... Feature/Specification or already implemented :| [09:38] Maybe it was deleted. [09:42] Shouldn't it be on launchpad? [10:40] i'm trying bzr for the first time. i want to push something using bzr+ssh but i run ssh on a different port. is it still possible? [10:40] I've never tried it but: : isn't working? [10:41] doesn't seem to be [10:41] ahh no [10:41] i'm lieing it does work! [10:41] i was putting a slash in the wrong place [10:45] ok i'm getting somewhere, it's saying something about the langauge is incorrect [10:46] i have to set $LANG somewhere for the python interpretter [10:46] how do i do that? [10:46] New bug: #158972 in bzr "http server use both settimeout and makefile" [Undecided,New] https://launchpad.net/bugs/158972 [10:52] bzr: warning: unsupported locale setting [10:52] Could not determine what text encoding to use. [10:52] This error usually means your Python interpreter [10:52] doesn't support the locale set by $LANG (en_GB.UTF-8) [10:52] Continuing with ascii encoding. [10:52] bzr: ERROR: exceptions.AssertionError: end of file reading from server. [10:52] ^^^ that's what i'm getting [10:53] fluidics_ss2, the remote server is probably missing the en_GB.UTF-8 locale [10:53] is that in python? [10:54] * fluidics_ss2 knows very little about python [10:54] bzr+ssh just calls bzr on the server [10:54] oh ok [10:54] you just need to install the locale [10:54] nothing python related [10:56] what's a locale? [10:57] i mean, what sort of package name am i looking for to install it? [10:57] locale stands for Language Collection [10:57] What kind of server? [10:57] ubuntu [10:57] sudo apt-get install language-support-en [10:58] cool [10:58] thanks [10:58] to search for packages via command line use "apt-cache search [10:59] Minus the " [10:59] apt-cache search [10:59] yeah, i trying that [10:59] well that just installed a lot of stuff [11:00] stuff to do with fonts and open-office [11:00] i'm not sure that was particularly good [11:00] If it works now you could always autoremove this metapackage [11:02] well i'm still getting the same error about locale but i don't think that's the real issue [11:02] sudo locale-gen en_GB.UTF-8 [11:02] sudo apt-get autoremove language-support-en [11:03] That will undo your install of the metapackage and dependencies if you want [11:03] do you mean autoclean? [11:03] yeah that didn't do anything [11:03] nevermind [11:03] No autoremove is remove and autoremove dependencies [11:05] ok [11:05] autoremove only works on edgy and newer it would seem [11:05] don't worry [11:05] Ok... [11:05] i think the bzr push is completing [11:05] i'm trying to push a branch like this bzr push bzr+ssh://ip:port/test [11:06] and it's creating the test dir on the remote machine [11:06] but it's empty [11:06] No it's not, it's got .bzr/ in it [11:06] ls -a reveals: . .. .bzr [11:06] where's my stuff then? :P [11:07] I'm still very new to Bazaar to... but that seems to be correct [11:07] In .bzr :) [11:07] .bzr really is empty [11:08] ls -Ra /test [11:08] . .. .bzr [11:08] . .. [11:08] Well, then push didn't complete. [11:08] ls [11:08] branch branch-format branch-lock checkout README repository [11:09] yeah, i think this error "bzr: ERROR: exceptions.AssertionError: end of file reading from server." might have something to do with it [11:09] btw thanks for all the help so far guys [11:09] If you're getting that locale error, then yeah, that would probably be stopping it at that point. [11:10] hmm ok [11:10] well the apt-get install lang.... stuff didn't seem to fix that [11:10] i ran it on both servers [11:10] The problem there is likely to be not a lack of a package on the server side, but that your LANG is set wrong. [11:10] set | grep LANG outputs nothing [11:11] Try 'locale' [11:11] ooooo [11:12] well yes, now we have identified a discrepency :) [11:12] Maybe it's set via LC_ALL or something. I'm not entirely clear on the intentional difference between the ways of setting it. [11:14] Anyway, the problem is likely to be that you've set your locale to ".UTF-8" and your system-defined locales are ".utf8" or something like that. [11:14] on one of the servers the locale is blank [11:15] which can't be right :) [11:15] Blank would mean C, which should be fine (may be _incorrect_, but it's not broken) [11:16] ok, i've run export LANG="en_GB.UTF-8" [11:16] and now locale output is the same on both servers, yay! [11:17] ok lets give this push thing another go [11:17] Shouldn't matter if they're different, as long as both are valid. [11:18] made no difference [11:18] i think it's python saying "i don't support en/gb utf-8" [11:19] Well, yeah. That's my point :p [11:19] Look at what locale tells you it's set to, and look at what locale -a says are valid choices. You'll probably find a discrepancy. [11:19] ok [11:20] And it's probably a .UTF-8 vs .utf8 thing (at least, that's the most common one I've seen) [11:20] one server says: [11:20] C [11:20] en_GB.utf8 [11:20] en_US.utf8 [11:20] POSIX [11:20] and the other says: [11:20] locale: Cannot set LC_CTYPE to default locale: No such file or directory [11:20] locale: Cannot set LC_MESSAGES to default locale: No such file or directory [11:20] locale: Cannot set LC_COLLATE to default locale: No such file or directory [11:20] C [11:20] POSIX [11:21] so i guess the remote server is missing some files [11:21] The second, I presume is the one that was C (or blank), and you set to en_GB[...] [11:21] yeah [11:22] And the first... well, there it is. [11:23] yep i pasted both in up there [11:24] i have no idea how to fix this [11:24] Eh? Set the locale to the valid choice. [11:25] Find where you're getting en_GB.UTF-8 set, and change it to en_GB.utf8. Computer don't guess words that 'look close' when you tell 'em something :p [11:26] In your shell rc file, most likely. [11:26] ah [11:26] ok [11:28] ~/.bashrc mentions nothing about locales [11:28] Mmm. I'm not sure what sequence of files bash reads. I think there's a .profile or .bash_profile somewhere in the lineup too. [11:28] It may be set in the system-wide ones in /etc. [11:29] yeah ok i'll have a bit of a dig around [11:35] i can't find anything [11:35] i've grepped * on /etc and ~ for en, gb and utf8 and the only mentions seem to be irrelevent [11:39] what's the difference between declare and export? [12:06] * fullermd mutters. [12:07] Yeah, so who needs power anyway? [12:08] this locale thing is a nightmare [12:08] i give up (for now) [12:08] How's that? [12:09] i think i need to do something with localedef [12:09] * fullermd blinks. [12:09] Huh? [12:09] Does just overriding the env not do the job? [12:09] nope [12:09] i get errors in locale -a if i do that [12:09] locale: Cannot set LC_CTYPE to default locale: No such file or directory [12:09] locale: Cannot set LC_MESSAGES to default locale: No such file or directory [12:09] locale: Cannot set LC_COLLATE to default locale: No such file or directory [12:09] C [12:09] POSIX [12:10] Well, that's the server that did have it all unset, right? [12:10] yeah [12:10] It should be left unset, since it doesn't have any other locales anyway. Your problem was on the other one. [12:11] but seeing as i'm in the UK shouldn't it be set as so [12:11] having it blank doesn't seem right to me [12:11] C is a perfectly valid locale. [12:11] what does the locale affect anyway? [12:12] It may not give you everything you want, but that's not relevant to your _problem_. Your problem is on the other server, where you have it set to an invalid value. [12:12] A number of things, but the most important in this connection is character set. [12:13]