[00:00] Peng_: pageant has to kick in once; latency kicks in hundreds of times [00:01] Peng_: if the CPU is so slow that starting pageant and having it hand a key out once is slower than some hundred * latency, bzr will be having many other issues ;) [00:01] maybe you want for any reason use http, a shortcut would be nice [00:01] lph: or so [00:01] or lpa: for ignoring -lp-login [00:12] RenatoSilva: the link you asked about in #launchpad is stable, its got the GUIDs embedded [00:12] RenatoSilva: if you want to point to the most recent version of a file, you can use head: or -1, for the revision component. [00:14] http://bazaar.launchpad.net/%7Erenatosilva/%2Bjunk/moin.macro.childpages/download/renato_silva-20090620220448-vcluh5vueefsrn2i/childpages.py-20090620220344-3uvqg24v5ninpo1x-1/ChildPages.py [00:14] where do I put head: ? [00:16] RenatoSilva: You'd replace the revid, renato_silva-20090620220448-vcluh5vueefsrn2i [00:16] RenatoSilva: Of course, in a future revision, that file could be deleted, so the URL isn't stable. [00:16] RenatoSilva: Or it could just be modified into something different. [00:16] Peng_: well, it would be stable, but not reliable :) [00:17] and only in the case that the file gets deleted [00:20] It would be nice to have something simpler like http://bazaar.launchpad.net/~renatosilva/+junk/moin.macro.childpages/download/head:/ChildPages.py [00:30] how do I cancel a bzr add? [00:30] bzr remove is not working [00:33] bzr revert [00:39] ok thanks [00:48] mtaylor: dropping by ? [00:48] hey lifeless [00:48] lifeless, dropping by what? [00:48] bzr ;P [00:49] I'm _always_ in #bzr :) [00:49] :> [00:49] something is unhappy with my laptop [00:50] I shut it down and started it again and now nothing seems to be able to find any of my settings [00:50] ouch [00:50] even though the files are in the home dir [00:50] like, .mozilla-thunderbird contains an intact profile, but thunderbird can't find it - xchat didn't remember that I want : for nick completion instead of , [00:50] check $HOME? [00:51] yup: /home/mtaylor [00:52] I'm going to blame xfs [00:52] I'm pretty sure stewart is wrong about it being any good [00:53] this is really going to wind up pissing me off [00:53] reboot again? :< [00:53] already did [00:54] ah [00:54] so, its not temporary buffer issues [00:54] what about perms? xacls? [00:54] lifeless: I don't even know how those work [00:55] well, I know how perms work [00:55] thanks everybody [00:55] I'd find the smallest app that fails; and strace it [00:57] well... I'm going to say that the thunderbird problem is that my pref.js file is truncated [00:57] ah yes [00:57] thats the 'posix compliant but useless' behaviour we expect from fast file systems. [00:57] [you've seen the ext4 threads/bugs about this right?] [00:58] lifeless: no. ... but I'm running xfs [00:58] is this what I get for listening to stew? a fast but useless filesystem? [00:58] on a laptop, fast and dangerous [00:58] * mtaylor throws something in the direction of melbourne [00:58] no kidding [00:58] so here's the basic problem [00:59] writes are slow [00:59] yup [00:59] apps routinely *don't* fsync or fdsync because it makes ext3 machines die [01:00] bleh [01:00] so, when you do the following: write, close, rename [01:00] to atomically replace a file foo [01:01] its possible for the directory link for foo to be flushed to disk without the data for foo [01:02] so - when I go to shutdown my system now and it sits there with a blinking cursor for forever, I'm guessing I should _not_ force power it off, because perhaps it's syncing data to disk? [01:02] yes [01:02] ass [01:02] * mtaylor is going to re-install without xfs [01:03] now to be 'correct' you can do [01:03] fdsync(tmpfile) [01:03] rename [01:03] but noone is doing that, are they? [01:03] sorry, fsync(tmpfile), rename, fsync(dirfd) [01:04] people that do it, like FF, make linux desktops incredibly slow [01:06] its appropriate for server environments do say 'force X to disk' but its not appropriate for most software [01:06] posix doesn't let you explicitly describe 'file A or file B' except in the roundabout way of the write+rename idiom [01:07] so when ext4 started routinely hosing peoples configs in the same way [01:07] there was a huge furor [01:07] ext4 has been fixed [01:16] hrm. [01:16] should I perhaps use ext4 instead of xfs? [01:16] although - for that matter... why is my _prefs.js_ file _ever_ being written to except when I edit my config? [01:17] give you one guess [01:17] I'd think that after editing a config file, doing a force X to disk would be acceptable [01:17] and other times, for things like "last email opened" those could be stored in, oh, I don't know, a separate slightly more acceptably volatile file [01:17] sure [01:17] * mtaylor punches firefox people in the throat [01:17] but really [01:18] the main thing is 'all of old, or all of new never just the existence of new' [01:18] if you need more performance than ext3, I hear ext4 has been shaping up very well [01:18] and like I say, this pattern is supported on it now [01:18] ted tso's rants about it not being required by posix aside. [01:18] well... it's tough to say - it's a new laptop with better disks anyway [01:19] but stewart just keeps carping about how superior xfs is to everything [01:19] http://mjg59.livejournal.com/108257.html [01:19] this is worth reading [01:19] stewart is right *for servers with battery backed up raid and hardware with no driver quirks and none of this 'suspend' bullshit [01:20] mmm [01:20] I don't have that on my laptop [01:20] actualy - with the new video card, suspend works nicely, thank [01:20] thanks [01:21] on the other hand - I can't seem to get 1600x1050 working, and now there is actually nothing in xorg.conf anymore [01:21] so I can't actually fix it [01:21] is it ubuntu? [01:21] when did we stop having useful config files? [01:21] yes [01:21] several releases ago [01:21] sigh [01:21] dpkg-reconfigure xorg-xserver IIRC [01:21] yup. doesn't do squat [01:23] hrm. that time I got it to add Option "UseFBDev" "true" [01:23] lets see if that helps... [01:27] lifeless: nope [01:27] ctrl-alt-backspace seems to not work anymore too, and neither does ctrl-alt-+ or ctrl-alt-- [01:28] cab is disabled by default in recent ubuntu's [01:28] + and - depend on the virtual size + modes [01:28] if they are the same, nothing will happen [01:28] m. fair enough [01:29] what driver are you using? [01:29] so, where are the available screen resolutions actually configured these days [01:29] intel? [01:29] xrandr [01:29] you can persist them in Xorg.conf [01:29] ah [01:29] well, it tells me: Screen 0: minimum 320 x 200, current 1280 x 800, maximum 1280 x 1280 [01:29] but its all geared around dynamic changes === Kissaki is now known as Kissaki^0ff [02:51] LarstiQ: you about? [02:54] poolie: btw I have a slug membership card for you for this year [03:05] rockstar: mwhudson: http://werkzeug.pocoo.org/ [03:08] lifeless: anything in particular? [03:09] ran into thelink [03:09] seemed relevant to loggerhead [03:09] the only part of paste i'd like to replace is the server, perhaps, and werkzeug doesn't have anything special in that department [03:09] debugging aspects [03:09] (we should look at spawning) [03:09] if you already know of it, cool [03:10] yup [03:10] thanks though [03:42] hey there, I can't seem to get bazaar to acknowledge that I manually resolve my conflicts [03:42] "0 conflict(s) auto-resolved." [03:42] what to do at times like this? [03:44] oh, seems like doing a bzr resolve the-file instead of just bzr resolve ... solved the problem [07:51] hi [07:51] how can i get a diff from revision 8 to revision 9? [07:51] diff -c 9, or diff -r 8..9 [07:51] bzr diff -r8..9 [07:54] lifeless, bialix, thanks [09:31] igc: hi [09:52] hi bialix [10:06] Can I create a "pointer" project? A project which only makes reference to the original one? === Kissaki^0ff is now known as Kissaki [10:21] Hi igc [10:51] hi garyvdm [12:15] nooo, garyvdm is back :p [12:15] * pygi hides [12:16] Hi pygi [15:15] Hi, is there any good user-friendly UI for ubuntu? [15:21] pisecx: Olive? [15:21] its included in bzr-gtk package [15:21] installed, but can not start it [15:21] ops. started now.. [16:59] <_gpg_> hello [17:01] <_gpg_> we were using cmsynergy and switched few months a go to subversion. Actually i"m very interested by bazaar [17:01] <_gpg_> The "Decentralized with automatic gatekeeper" is what i need with my team actually, i cant find any documentation about how to use/set PQM [17:25] Is there a way to get a diff between two different branches? [19:39] "bzr rebase ../opencog" errors with NoSuchRevision: KnitPackRepository('file:///home/kizzo/python-bindings/.bzr/repository/') has no revision ('kizzo@crashtest-20090619010527-8dhvulrdodnszrtj',) [19:41] If anyone would want to try, they can do "bzr branch lp:opencog; bzr branch lp:~kizzobot/opencog/python-bindings; cd python-bindings; bzr rebase ../opencog". [20:25] hello folks. I want to push my bzr repo to my web domain using the ftp account. Buz bzr fails. I do not understand where is the error. [20:25] here is the error message : http://pastebin.com/m11ce7fe4 [20:26] I use bzr on my Ubuntu 8.10. [20:30] try bzr --no-plugins push aftp://web613f1@theadib.net/TDD/trunk [20:31] if it works file a bug against bzr-svn [20:40] bialix: thx for the hint. Now I come closer. Still some errors: [20:40] FTP temporary error: 451 /TDD/trunk/.bzr/repository/upload/nax1mcnkw7hpb8bffyso.fetch: Append/Restart not permitted, try again. Retrying. [20:40] bzr: ERROR: Transport error: Error setting up connection: 530 Sorry, we dont allow more than 3 connections per host! 530 Sorry, we dont allow more than 3 connections per host! [20:41] I tried using aftp// and ftp// [20:42] it seems like your ftp server can't append [20:43] I'm not sure if can workaround this [20:46] Here is the error I am getting: http://pastebin.com/d4bddf102 [20:58] try to pull lp:opencog first [22:32] it would be nice to bzr push lp:project/* [22:32] * pull [22:33] grrr, * branch === mario_ is now known as pygi [22:33] RenatoSilva: at a project level you need to make a series to do that [22:34] RenatoSilva: lp:project/series works. as does lp:~user/project/branch [22:34] I tried lp: ~renatosilva/* but doesn't work [22:34] oh [22:34] no, you can't wildcard [22:41] maybe in the future :) [22:46] pull pulls from a single branch [22:47] I'm not sure what it would /mean/ to tell it to pull many branches [22:47] what do you have in mind? [22:50] actually I just thought about bzr branch [22:51] bzr branch lp:/~user/project/* for example would fetch all branches under ~user/project [22:53] well, we have a planned review of the way bzr presents branches and repos that may make it easier to do things like that [22:53] not sure that it would end up being spelt identically ;P [23:00] hi [23:00] if i have a diff, how do i apply it? [23:00] have some uncommited code on my laptop [23:00] is it a plain old diff, or a bundle? [23:00] wanted to move it over to my desktop [23:00] just a plain diff [23:01] patch -p0 < diff [23:01] ok... thanks [23:01] myself, I'd usually commit and push it [23:01] then uncommit on the far end [23:02] i didn't finish the feature though [23:02] doing this can handle more situations than plain old diffs [23:02] its ok that its not finished - see where I say 'and uncommit on the far end' [23:04] lifeless: this way should be easiest as it's just a few changes in a few files [23:08] lifeless: you mean bzr patch -p0 < diff? [23:08] RenatoSilva: no [23:09] ok [23:09] RenatoSilva: you can do that too, but its generally not needed, and for this case, thigns that plain patch won't handle would still be handled wrong by bzr patch - e.g. file adds would show up as parallel imports [23:10] meoblast001: a small note if I may - you don't need to finish features to do commits in bzr; because you can have your own branch, you can make as many commits as you want while developing the feature. [23:10] yeah i know [23:10] just harder to make my own branch [23:11] than to pastebin a diff and patch it [23:11] meoblast001: not really [23:11] you could pastebin the output of 'bzr send' [23:12] presumably if you can get the output of 'bzr diff', you already have a branch, it already has changes in it ;) [23:12] actually [23:12] glyph: he may be working in a checkout [23:12] lifeless: still [23:12] glyph: or there may be other factors [23:13] oh I didn't even see the earlier scrollback [23:13] meoblast001: one of the reasons I love bzr is exactly that case [23:13] moving code from my laptop to my desktop [23:13] I do it by doing 'bzr push bzr+ssh://my-desktop/home/me/mybranch' [23:15] hi all... just trying to get the GUI working but i dont know where to "After this, you can start the GUI by typing olive-gtk" :/ [23:15] Sheepherd: OS? [23:15] win xp 32bit [23:15] sp 3 [23:16] Sheepherd: start > run [23:16] thats all? [23:16] ssh was the easiest way by far [23:17] Sheepherd: I'm not sure what's olive-gtk tough, but that's how I'd read the message. I just tried and didn't work here :P [23:18] Sheepherd: what are you doing, installing bazaar? [23:18] RenatoSilva: followed every step here correctly: http://bazaar-vcs.org/WindowsInstall [23:18] RenatoSilva: yea exactly [23:19] RenatoSilva: olive-gtk comes from the 2nd last line of the site [23:19] Sheepherd: oh I had a bit of trouble setting up bzr on my win xp to. [23:19] Sheepherd: just forget all of this, undo everything. [23:20] Sheepherd: then use http://edge.launchpad.net/bzr/1.16/1.16/+download/bzr-setup-1.16-1.exe [23:20] RenatoSilva: :/ [23:20] Sheepherd: next, nextm finish, and you're [almost] done [23:21] Sheepherd: this installer has an embedded python [23:21] ya thats why i didnt use it [23:21] cuz i already have python [23:23] well, I have python too, but I had some trouble setting it up separately [23:24] RenatoSilva: hope this proceeds smoothly now ... [23:24] I read that page too, which is confusing imho, but I'm fine now after just installing the all-in-one [23:26] Sheepherd: what would you like bazaar for? launchpad by any chance? [23:27] RenatoSilva: first time hearing about that. but someone from #python recommened some svcs or how those tools are called to get a good overview over changes made to my app [23:28] Sheepherd: never heard of vcs or dvcs? [23:28] RenatoSilva: ok installed the whole thing... how do i access the gui now? [23:29] RenatoSilva: yea dvcs... thats what i meant :) and i never heard about that until like 30mins [23:29] did you get the idea of dvcs? [23:30] RenatoSilva: something to trace down changes made to a file? [23:31] Sheepherd: useful links... [23:31] http://en.wikipedia.org/wiki/Revision_control [23:31] http://en.wikipedia.org/wiki/Distributed_revision_control [23:32] http://ianclatworthy.files.wordpress.com/2007/10/dvcs-why-and-how3.pdf [23:33] When you're done with the overview of vcs and dvcs, you can read the 5 minutes tutorial for Bazaar: http://doc.bazaar-vcs.org/latest/en/mini-tutorial/index.html [23:34] and to learn more, the user guide: http://doc.bazaar-vcs.org/latest/en/mini-tutorial/index.html [23:34] jelmer: does bzr-git 0.4.0 require a new dulwich? [23:36] RenatoSilva: thx :) and where should all those lead me to? [23:36] RenatoSilva: was bazaar the wrong choice for me? [23:37] Sheepherd: bazaar is a fine choice for you; but you've just started to learn a new tool. VCS tools are not about inspecting your current code, they are about controlling it [and reporting] over time [23:39] lifeless: k... hope ill get used to this new system soon [23:39] Sheepherd: have you even used any vcs tool before? [23:40] RenatoSilva: nope, today is even the first day i heard about that [23:40] first time* [23:43] RenatoSilva: why are u asking? [23:43] Sheepherd: it helps us understand the things you already know, so we don't tell you too little or too much [23:43] Sheepherd: humm, try having some time to read the links, in the order I mentioned. I think this can avoid many trouble [23:43] k thx both of you :) [23:44] gonna read them carefully tomorrow [23:44] :) [23:44] should be sleeping already xD [23:45] im off then... thx again and cya tomorrow maybe [23:46] bye