[01:17] hi al [01:17] all [01:18] I'm trying to serve bzr trhu http... but instead of using Apache of FastCGI I would like to use cherrypy... but I tried some code and don't have a clue... does anyone made it work with cherrypy? [01:24] bzr will happily work over plain http [01:25] bob2: what I want to do is to use a standalone app to serve trhu http without the need of an apache or so... [01:25] bob2: maybe i'm missing something [01:26] If you're going to run standalone anyway, why not just use bzr:// ? [01:26] that's fine, just serve it as a regular directory [01:26] bob2: trhu http? [01:27] if you wnat to serve it via http (rather than bzr://) yes [01:28] bob2, sorry dont follow... i just use bzr server PATH ? and that will serve over http? === Bambi_BOFH is now known as Kamping_Kaiser [04:09] lifeless: Well, once you stop swapping, could you check the mailing list queue for me? [04:31] Peng_: if its held you geta mail saying so [04:41] Peng_: have you had one ? [05:08] lifeless: Oh, really? No, I didn't. It just vanished and I was guessing that might be why. [05:08] * Peng_ looks at his mail host. [05:18] * Peng_ sends again. :D [05:22] Hmm, I can send the email to myself, but maybe not my Gmail account. [05:25] Yay, Gmail got it. [05:30] lifeless: Do you know how large messages have to be before they need to be approved? [05:32] I wish I was sending something *worth* all of this inconvenience. :\ It's just the ConfigObj thing. [05:45] lifeless: Could you check the queue anyway? Because if it's not that, I have no idea where the messages are going. :\ === BasicPRO is now known as BasicOSX [14:27] hey is anyone home [14:34] Maybe. [14:36] perfect [14:36] i am using it for checkin and chekcout in a way to modify my code. that my issue is when i commit it it says for notes, fixed bugs, and author. when i put stuff in fixed bugs i get unrecognized bug ...... commite refused. it has to be in 'tag:id' which i am confused about [14:36] i am wanting to push everything to launchpad [14:36] but i am just committing my code right now [14:37] i have entered this in 'LP:363538' and recieved:> bzr: ERROR: Unrecognized bug 'LP:363538'. Commit refused. [14:39] lp:363538 [14:39] so like this: lp:363538 lp:363537 lp:363534 lp:363520? [14:39] perfect tahnk you [14:39] --fixes before each [14:40] ? [14:40] "--fixes lp:363538 --fixes lp:363537 --fixes lp:363534 --fixes lp:363520" [14:40] what does the --fixes do [14:40] are you using a GUI tool? [14:41] yes and no [14:41] yes for that [14:41] for updating hte branch i am using command line [14:41] hi guys. Wondering if this is possible. In my bzr project, I've got a sub-folder. I now want to make this a standalone bzr project. Is it possible to do this whilst carrying the history? [14:42] thewrath:try your line then [14:42] its `/myapp/plugins/myplugin` . I want to maintain 'myplugin' as a seperate project now [14:43] when i do that james_w is it supposed to auto update the status to fix released? [14:43] nope [14:43] not yet at least [14:44] ok [14:44] CaMason: You can use "bzr split" (probably). It really will keep the *full* history though, not just the history from /myapp/plugins/myplugin. [14:44] Peng_ that's no problem really. I just want to have it as standalone [14:45] is it possible to have a bzr project within a bzr project tree? i.e. if I added `myapp/plugins/myplugin` to bzrignore, and then had a branch in the plugins folder [14:46] CaMason: I think it'll automatically be ignored. You don't have to use .bzrignore. [14:46] ok. I've just read about split/join. Are those new features? [14:46] CaMason: see http://bazaar-vcs.org/NestedTreeProgress [14:48] right - so it works? [14:50] I see I need to bzr upgrade. Any idea which type I need? [14:50] CaMason: Note that you won't be able to go back after upgrading. And everybody else who has a copy of your branch will have to as well. [14:50] CaMason: Which format are you using now? [14:51] Pack 0.92 [14:52] james_w: is that something that they are looking at [14:52] yep [14:52] CaMason: Then you should probably upgrade to rich-root-pack. [14:52] CaMason: alternatively, you could use lp:bzr-scmproj, or version symlinks [14:52] CaMason: You might need pack-0.92-subtree, which is still considered kinda experimental. [14:52] whoa info overkill [14:53] what's lp:bzr-scmproj ? [14:56] CaMason: https://launchpad.net/bzr-scmproj [14:57] CaMason: a tool to combine branches [14:57] all I'm really trying to do is have a facility like svn:externals [14:57] the split/join method seemed like common sense [14:58] LarstiQ, that SCM Project looks like it may be overkill for me [14:59] perhap I just symlink and add the link to .bzrignore? [15:02] CaMason: that's possible [15:02] CaMason: join --reference would be the thing to do, yes. But it's not finished. [15:02] heya phanatic [15:02] hey LarstiQ [15:02] I just tried `bzr split myapp/plugins/myplugin` and it's thrown up `bzr: ERROR: No WorkingTree exists for /myapp/plugins/myplugin/.bzr/checkout/` [15:06] CaMason: hmm [15:06] I've removed the created .bzr folder, and commited some uncommited changes (oops) [15:06] and re-done the split [15:07] both 'sides' of the split now show the other half as being removed, when using `bzr st` [15:09] ok well, that seems to have worked :o === `6og is now known as Kamping_Kaiser [16:53] Hi; I backed up in a hurry only the .bzr/ of a repository I wanted. Can I still check out the branches from there? [16:53] clemente: Yes. It's slightly complicated, though. [16:55] with help from `bzr heads` [16:55] clemente: Install the bzrtools plugin. Run "bzr heads" (or maybe "bzr heads --tips"?) to get the tip revision IDs of all of your branches. "bzr init" an empty branch, cd to it, "bzr pull . -r revid:", then do the same for the other branches, or (probably) "bzr branch abranch -r revid:". [16:56] some branches might not be on tips in the dag, that would then be harder [16:56] Good point. [16:56] clemente: but basically, if you know the revision (bzr heads helps to find those), you can restore a branch [16:57] Peng_: heads --all, it is. I'm trying what you suggested [16:58] I don't remember, what's a dead head? *checks* [16:58] Ok, it doesn't say. :P [17:01] This worked except for branches which don't show in „bzr heads --all“ [17:03] But I suppose I can find it somewhere inside .bzr/ [17:18] Hi, how do I undo a push? [17:21] joh: by using push --overwrite in combination with a revision you prefer [17:22] joh: say, `bzr push --overwrite -r -2` or `bzr uncommit; bzr push --overwrite` [17:22] LarstiQ: Ah, thanks! [18:05] * SamB wonders why bzr-svn's cache for http://svn.webkit.org/repository/webkit is so big for him ... [18:06] Isn't WebKit really gigantic? [18:06] I only checked out a little piece though! [18:07] just http://svn.webkit.org/repository/webkit/trunk/JavaScriptCore/pcre/ [18:07] SamB: my uneducated guess would be that it needs more than just that to determine the branch identity. [18:07] SamB: but maybe it does more than it needs to [18:10] It does it again after I delete the cache, too, apparantly ... [18:11] SamB: that's what I'd expect :) [18:18] jelmer: branching just http://svn.webkit.org/repository/webkit/trunk/JavaScriptCore/pcre/ ate about 183 MB! [18:19] in ~/.bazaar/svn-cache alone [18:33] hello [18:33] is it acceptable to create a new branch just by copying an existing one? [18:34] Stavros, yes [18:34] Stavros: cp -a wise? Yes. [18:34] ah, thank you [18:35] LarstiQ: what's -a? [18:35] Stavros: archive, implies -dpR [18:35] ah, the os x copy doesn't have that [18:36] Stavros: if the source branch uses a shared repositroy, it's best to have the destination remain in the shared repository (otherwise it can't find the revisions it depends on without handwork) [18:37] and some other differences in execution, but copying is a valid way of branching [18:41] what's a shared repository? [18:43] Stavros: the result of `bzr init-repo` [18:44] hmm [18:44] Stavros: if `bzr info` says something like: shared repository: /home/larstiq/src/bzr [18:44] Stavros: then the branch is using one [18:44] heya pygi [18:44] hi hi LarstiQ :) [18:44] how goes it? [18:44] pygi: busy busy busy [18:45] * LarstiQ relaxes on sunday with #bzr :) [18:45] LarstiQ, ha, I know the feeling! [18:45] heh, no relaxing here :( [18:45] I have to design a new system ... deadline in two days [18:45] and I can't really learn about it from anywhere, cause its a rather new thing :( [18:46] what kind of system? [19:05] are there any Python IDE's that integrate well with bzr? [19:06] lampliter1: http://bazaar-vcs.org/IDEIntegration [19:06] lampliter1: I personally use vim [19:07] lampliter1: other than that, I'd suggest looking at pida [19:07] I generally use Emacs but making it work with speech recognition is getting increasingly tedious. Additionally my development environment is so messed up, I'm trying to restore it to sanity [19:07] LarstiQ, a security one :) [19:07] lampliter1: possibly off topic, but are you aware of http://code.google.com/p/dragonfly/ ? [19:07] LarstiQ, sorry, you need to say my name so xchat would light up :P [19:08] pygi: did you say something? I only get highlights when people don't use my name ;P [19:08] LarstiQ, :P [19:09] LarstiQ, do I shoot you now or what? :D [19:09] dragonfly annoys me. We have already three of those frameworks. It's worse than Python web frameworks for replication of effort. Of course my Python web framework is aimed at handicapped people (Universal design principles) so that's okay. :-) [19:10] but yet, when you're disabled, you go through some really nasty hoops to do work. [19:10] bloody speech recognition... [19:11] I've been speech recognition on a Windows guest virtual machine. My host OS is linux and acts as a file server two all my virtual machines. [19:12] And I do work on remote machines. [19:12] The fundamental question is, how can I access all of these environments through various security hoops and barriers while still using speech recognition to drive whatever IDE I use [19:13] the best I've come up with is to run the IDE on Linux, drag it into Windows via X11, use fam to trigger rsync to update the remote machine [19:14] and if I can keep my bzr branch local, it solves a bunch of problems [19:15] either I bored you or I scared you. Sorry about either case [19:16] no, just occupied [19:17] lampliter1: I'm not active in the same area, but I'm not aware of competing libraries for Dragonfly? [19:19] there is the raw natlink, then on top of that, we have vocola and unimacro [19:21] hmm [19:21] I find that natlink is the most flexible since you implement your grammar actionsin Python but it is a lot of work. vocola make simple things easy, medium things complicated, and complicated things almost impossible. unimacro is something that takes quite a while to cross the learning curve and may be worth it but, it still doesn't make simple things easy to release the doesn't make difficult things hard [19:23] just noticed one thing about dragonfly. Looking through the code, the is a lot of the misspelled words with caps and concatenation of said fragments. [19:23] Do you have any idea how bloody hard it is to dictate those symbols? I mean, wouldn't you think that someone writing a macro package for speech recognition users write a package that could be dictated by speech recognition users [19:25] symbols should be nothing but lowercase words concatenated or joined by underscores [19:25] the stupid capitalization for classes etc. sets up a nasty situation where if you use the code, you are at greater risk to ruining your vocal track on top of your damaged hands [19:26] I'll pass it on. [19:27] sorry but it's incredibly frustrating for me. I have more ideas in my head than I can write code for. Then add to that dictating stuff by voice, debugging hard to see errors introduced by dictation and general frustration with computers that aren't meant to be used by disabled people [19:27] I get really cranky sometimes [19:27] lampliter1: I get that. [19:27] I've got one project that's almost 10 years old that I haven't been able to release because I can't write enough code to get it done. Damn frustrating [19:28] LarstiQ: thanks. Not a lot of people do [19:28] what I need is a smart coding monkey that will listen to what I say, ignore me, and do the right things to get what I want done done. :-) [19:29] unfortunately the MIT programmers assistant Project hasn't moved forward [19:29] lampliter1: that, would be useful to more people than just you ;) [19:30] Hell, a good program and by voice IDE would save so many people so much pain and return a lot of disabled programmers to productive life [19:30] Hmm, Google CADIE could do it, I bet. :D [19:30] * Peng_ hides [19:30] * lampliter1 tosses corned beef sandwich with sauerkraut and nice mustard in Peng_ direction [19:32] but seriously, the biggest impediment we have is editing. The features needed aren't very complex (ha) but the problem is, we can't get anyone interested in helping in doing the work that is burning our hands [19:33] * LarstiQ nods at lampliter1 [19:33] anyway, I should go eat lunch then walk the dog. It's a gorgeous day out here near Boston and dog walking helps restore my sanity after my hands go poof [19:33] lampliter1: issues not shared by the majority often don't even register in their heads. [19:34] lampliter1: ok, have a good walk. [19:34] I know. I know. Here's an ironic that a friend of mine who is the head of development at Dragon Systems pooh-poohed the need for a speech recognition driven environment. Guess what, his hands are toast and now he can't build the environment.... [19:34] not enough market share... [19:35] that should be "who was the head of development" [19:35] no longer because of his hands now? [19:35] is/was/can/can't/... are all suspect. :-) [19:35] yes, his hands are bad. Maybe worse than mine [19:36] what we both need is a way to edit code by voice. Code creation is not wonderful but we can do it [19:36] we need to be able to identify things like predicates arguments array indices and then be able to modify them simply by voice [19:37] for example, "change the second argument" should allow me to do more detailed editing on the second argument found on the current line. [19:38] The basic principle is known as disambiguation vocal commands through reduction of scope. Most the time disambiguation is by increasing the specificity of the command but if you're smart, as you dictate you can reduce scope dynamically by using what you dictated to identify what you're going to work on [19:39] I should probably stop now because otherwise I will give you a two credit seminar in speech user interfaces [19:40] lunch calls... oh, and by the way, thanks for the advice on the IDE. [19:41] * LarstiQ tries to parse the sentence on scope again [19:41] and unfortunately still don't see concretely how that works [19:41] I get increasing the specifity of the command [19:44] sorry. [19:45] First principle with speech recognition commands as you try to reduce ambiguity in order to increase accuracy and recognition [19:45] there should be accuracy in recognition [19:46] kind of proves my point. :-) [19:46] most of the time you reduce ambiguity by making the command more specific. For example, set font color green [19:46] "move to end of line" [19:47] but you can also reduce ambiguity to reducing the scope of what you're working on. [19:47] You can also reduce ambiguity by reducing the scope of what you're working on [19:48] for example, editing a line of code (let me get one) [19:48] grammar = Grammar("notepad_example", context=grammar_context) [19:49] editing this is a bit tricky especially with select and say [19:49] so let's try to change grammar_context to grandpa_context [19:50] you'd couldn't just say change grandma to grandpa because there are multiple occurrences on this line and, in fact, there probably are multiple occurrences through the body of the code (method, class, file) [19:50] so if you said change grandma to grandpa, you have an ambiguous command because you don't know the scope of what you're operating over [19:51] if you said "edit second argument" you could reduce your scope to just that one argument [19:51] right [19:51] all your environment would see is: context = grandma_context [19:51] now obviously it is a further ambiguity here but I'm going to dodge that for the moment [19:52] at this point you could say replace drama grandma with grandpa and everything should work okay [19:52] but if you want to change context, a more specific command could reduce ambiguity by saying something like change first context or change its second context [19:53] there's no need to do anything fancy when you're done because you really do a specific command or one that refers to the greater context of lines or methods [19:53] it gets interesting when you have a partially formed expression. But, if you do things right with template driven code creation, you never have a partially formed expression. [19:54] The beauty of this technique is you can also use a variety of visual cues such as colors with high transparency highlighting what it thinks its operating on and indicators for where it sees arguments etc.. But that's really hard [19:56] like I said, a two credit seminar :-) [19:57] lunchtime [22:38] Peng, Peng_: hello? === abentley1 is now known as abentley