[00:08] If I feel like trying again [00:11] Hy guys. Did anyone of you try the current windows build 1.9? I'm unable to get the svn plugin working. [00:12] It says Unable to load plugin 'svn' from 'C:\\Python25\\lib\\site-packages\\bzrlib\\plugins' . But if I understoof the website correctly svn should be included [00:30] wondering how to get bzr to recognize my authentication.conf file. [00:30] I'm using the bzr-svn plugin over https and it asks me for my pw multiple times [01:07] dD0T: can you please file a bug in https://bugs.launchpad.net/bzr-svn, including the contents of ~/.bzr.log [01:07] (see bzr --version for the location) [01:20] bzr: ERROR: These branches have diverged. Use the merge command to reconcile them. [01:20] is there a way I can tell _how_ they've diverged [01:20] bzr missing --mine-only, perhaps? [01:20] jml: yes === StoneToad_ is now known as stoneToad [01:20] hmm. either that interacts poorly with looms [01:21] or my loom is shagged. [01:21] (both are possible, I hit Ctrl-C during a combine-thread and since then it's been a bit wonky) [01:22] jml: combine-thread is basically atomic [01:23] lifeless: I ended up with a checkout/limbo dir and something else. [01:24] * jml resorts to export-loom again [01:24] jml: st;diff;revert should all fix it [01:26] yeah, so it looks like my bottom thread has all sorts of stuff that doesn't belong there :( [01:37] jml, poke :) [01:38] pygi: hi [01:47] :( :( :( [01:47] _every_ time I work with looms I end up being sad. [01:51] jml, :/ [02:30] jml: is your bottom thread upstream? === Mario__ is now known as pygi [03:00] Hi, I made some changes to a project some time ago and now I need to temporarily undo them and apply a change to an older version of a branch and publish it somewhere. [03:00] I did bzr branch project project-2 -r 4 and now have a copy tat is the way I want it [03:00] presumably I can make my change to this copy and get what i want, but how do i avoid bad stuff happening in the future? [03:01] what's the "right" thing to do I guess? [03:02] newz2000: what 'bad stuff' are you afraid of? [03:02] losing a revision or making it hard to put stuff back together down the road [03:03] let me give more details [03:03] I have a branch I share with sysadmins. They codereview and publish it to a website. [03:03] Temporarily we'll be making a change for the holidays [03:04] I want to share my branch with them so that they can push the changes out but later on (after the new year) they'll use my changes that are in progress in the other branch [03:04] I don't want to make their life hard === Mario__ is now known as pygi [03:06] i think you should be fine [03:07] just tell the admins to use push/pull --overwrite to move your changes onto the sever [03:07] ok, so just commit my change and push to the place I normally push to? [03:07] server [03:07] ok [03:07] you might want to push to somewhere else [03:07] ok, that makes sense [03:08] thanks for the help mwhudson [03:08] np [03:21] hi all [03:59] hi igc [04:00] hey lifeless! [06:07] I am trying bzr update command to get updates from bzr repository on launchpad. it says tree is upto date at revision 183 but when i see it via web interface it shows revision 184 [06:08] do I need to do anything before issuing bzr update for it to check the server for new revision [06:08] pulsar: if you have a branch, you need "bzr pull" to get the new revisions [06:09] "bzr update" updates a checkout to be current with the branch that it is a checkout of, "bzr pull" updates a branch from another branch. [06:09] spiv: ok trying that. my version control basics are really weak. [06:10] spiv: it to be at revision 185 now at both places. :) [06:12] thank you [07:07] hi all [08:12] night all [10:24] does anyone know why I cant add a subdirectory in my bzr branch which includes a .svn subdirectory?? I'm using bzr1.9 and have bzr-svn installed [10:39] ieei: bzr-svn allows you to use bzr to operate on SVN checkouts [10:40] ieei: so if you have a .svn directory in a bzr checkout, then bzr-svn will notice that and so bzr will open the checkout as a SVN checkout and not as a bzr one. [10:41] ieei: you could temporarily disable bzr-svn by passing --no-plugins to bzr [10:41] ieei: but by far the simplest thing to do is to not have the same directory on disk versioned by both svn and bzr [10:42] ieei: e.g. things can get very confusing if the .bzr directory says the checkout is at one revision of the SVN repo and the .svn directory says it's at another [10:45] I've tried to keep dual bzr/cvs checkouts before... It's doable, but it takes a lot of careful management [10:45] Easy to mess it up, and end up with conflicting histories in each side [10:45] LeoNerd: dual bzr/cvs checkouts is easier because there's no bzr-cvs [10:46] How does that help? [10:46] LeoNerd: so you don't have the problem that doing "bzr ci" will try to operate on the cvs data rather than the bzr data. [10:46] spiv: thanks, I know that it is a bit stupid and confusing, but what I'm doing here is really stupid in the first place. I want to version the svn metadata in my bzr branch.. so anyway, thanks, i'll just temp. disable the bzr-svn [10:55] For extra points, version the bzr metadata in svn at the same time! [10:58] hehe :P [11:26] Where are the Windows installers ;-) ? [13:56] lifeless, ping [14:03] pwong [14:27] morning all [14:27] hei jam [15:14] hi [15:14] bzr revert does not revert the last commit. How come? [15:15] Because revert doesn't revert commits. [15:15] bzr revert reverts the working tree [15:15] bzr uncommit then [15:15] I see [15:16] thank you [15:28] jelmer: yo [15:38] *headdesk* Someone at work has decided that half of the tree I am working on needs to be in another, separate SVN repository. [15:44] NfNitLoop: _nice_ [16:58] lifeless: isn't it? This place so fails at sane source control. [18:27] lifeless: ping === Mario__ is now known as pygi [18:45] jam: yo [18:46] how's MV? [18:46] intense :) [18:46] some very good discussion [18:46] chatted with jerry from samba today [18:47] how was that? [18:48] good [18:48] I was wondering if you had any thoughts on my "chunked" patch. [18:49] saw it go by last night, haven't had a chance to review; I was feeling quite ill last night [18:49] I'll poke at it asap [18:49] it isn't critically urgent, but I was hoping for "design" thoughts [18:49] more than code-level [18:49] well [18:49] we've discussed previously yes? seems fine to me [18:49] outputting thefragments a compressor has as they arrive is ok [18:49] it does seem a little special cased [18:50] So the big thing I wanted to get away from was needing to ''.join() and split_lines() [18:50] when some implementation thinks in fulltexts [18:50] and others think in line [18:50] lines [18:50] a fulltext => chunks is just (fulltext,) [18:51] which has a nice effect that ''.join([text]) is text [18:51] (if the list is a single string, it just returns the string) === BasicPRO is now known as BasicOSX [18:55] mm-mysql: hey. Any updates on a Lunch schedule? [18:56] lifeless: btw, the 'chunked' work shows a decrease of about 300MB for the conversion. So we still use 700MB, but that is down from 1GB [18:56] And it seems to start at 400MB and then grow to 700. I'm not entirely sure why [19:16] jam: I have a possibly-stupid question: what's the point of running chunks_to_lines() over the result of file.readlines()? Won't it be identical? [19:17] Peng_: the previous code did "split_lines(''.join(readlines())" [19:17] so there was obviously some doubt [19:17] jam: Ah, good point. [19:17] I believe abentley was concerned about platform differences for file.readlines() [19:17] file.readlines() can be an arbitrary object [19:18] Like on Mac 9 it would split on '\r' [19:18] though we don't support OS 9 [19:18] worth a comment I think [19:18] lifeless: it *can* but that code just did "open(filename)" so we know it is a real File object [19:18] Peng_: Also if you don't open the file in "binary" mode on Windows [19:18] jam: ah, heh. Certainly things like commit() don't know that [19:18] but we should always be doing that anyway [19:19] because they ask the tree and the tree is virtual [19:19] lifeless: yeah, that is TreeTransform code [19:19] where we do now === bac_uds is now known as bac === Mario__ is now known as pygi [19:35] I changed svn-revision-scheme and now I get check_file_version_parents errors when checking and when trying to commit :( I've tried to remove the svn properties in the repo, but no luck [19:40] jelmer: ping [19:46] vila: ping [20:10] mwhudson: pong [20:11] vila: so we're vaguely thinking of not using apache to serve bazaar branches from bazaar.launchpad.net [20:11] (and using something more dynamic instead) [20:12] and were wondering what sort of features bazaar needs from an http server for efficient operation [20:12] Something more dynamic? [20:12] range: header support is an obvious one [20:12] mwhudson: have a look at bzrlib/tests/http_server.py ? :-) [20:13] Peng: branches are accessed via ~user/project/name but stored by database id [20:13] Peng: this ends up being a pain [20:13] vila: good idea! [20:14] mwhudson: from the top of my head, missing parts will be authentication, but even there I think there are some basis [20:14] vila: we really don't care about authentication; yet, anyway :) [20:15] and of course you may want https too which I'm still working on [20:16] mwhudson: so if you aren't thinking to enable 'smart' support [20:16] then the #1 thing is Range: operation [20:16] mwhudson: there are also a couple of bugs pending regarding xmlrpc from behind proxies (which is used by the launchpad plugin) [20:16] vila: that shouldn't effect the server end, though [20:16] mwhudson: we can have some fairly large Range: requests [20:16] both number of bytes [20:16] and complexity of request [20:17] jam: yes, i've noticed :) [20:17] (they break the squid my isp lumbers me with) [20:17] I don't know how that would effect a potential Squid proxy [20:17] mwhudson: well that specifically is just because we start at 0 [20:17] and then make a request >1MB away [20:17] we could break that into 2 [20:17] but it adds a round-trip [20:17] and, unfortunately, we don't know we are going to have the problem [20:18] until after we have waiting for a *very* long time [20:18] I think we should consider a config-ish entry for controlling that, for people who know they are behind a broken proxy [20:19] mwhudson: Anyway, other than maybe 5-10 GET requests for a couple files, everything else is a readv() [20:19] we use it on both indexes and raw content [20:19] ok, that's what i thought [20:19] jam: the squid bug doesn't occur so often AFAICT and is fixed in recent versions [20:19] I would *like* to see launchpad implementing bzr+http [20:20] jam: +1 [20:20] squid bug is fixed in all supported versions; just some folk are bad about upgrading [20:20] jam: that would be nice, yes [20:20] vila: what are you still doing around ? :) [20:20] i get faintly terrified by resource consumption [20:20] but that's step 2, anyway [20:20] jam: ponging mwhudson :-) [20:21] vila: when *I* work until after 10pm, its because I'm home all alone. I didn't think you had the same excuse [20:22] I don't, just back from restaurant and showing blinking icon to gf: look ! They're pinping again ! [20:22] pinging even [20:22] Seems like you respond a bit too much to the little blinking icon :) [20:22] I'll keep it in mind [20:23] vila: sorry for disturbing your evening [20:23] mwhudson: don't worry, I won't stay there :-) [20:23] vila, vila, vila: please pay attention to me :) [20:24] jam: I already voted tweak to your chunked patch :) [20:24] Sorry, but my wife and son are away for a month... [20:27] jam: Ah, hi. [20:27] Sorry. [20:27] hi mm-mysql [20:28] I was going to get back to you about *this* week, but unfortunately, I had a death in the family (grandma, not unexpected, totally, but still busy because of it), and a release to get out the door [20:28] What does next week look like for you? [20:28] * mm-mysql pulls up the calendar [20:28] mm-mysql: I'm very sorry to hear that. [20:28] Next week I'm probably going to be out of town. [20:29] Ah. [20:29] Okay, so.... [20:29] It's the week of xmas, what about the 22nd? [20:30] should be fine by my calendar (it is pretty much wide open :) [20:31] Let's pencil that in? [20:31] Where/when would you like to meet? [20:31] sure [20:31] Anywhere/anytime :) [20:31] heh. [20:31] Panera or the Brewery are both fun places [20:32] lunch or dinner is fine [20:32] Let's do the brewery, lunch? [20:32] High noon, or? [20:32] sounds good. [20:32] around noon [20:32] Okay. You're on the calendar. [20:33] I'll see you at the Brewery, around 12 noon, on Dec 22nd. [20:39] mwhudson: a word of caution about bzrlib/tests/http_server.py, look at it for features, not implementation, it's nowhere near to be optimized (quite the opposite even as it's the base for all sort of torture test servers, so the emphasis is on easy extension/redefinition rather than performance) [20:39] vila: sure, i think i'm probably going to start with something from twisted [20:42] mwhudson: have you considered a j-i-t mapping lookup, feeding the result into the mod-rewrite core? [20:42] * emmajane submits her first-ever bzr bug. [20:43] lifeless: you mean the "prg" rewritemap thing, or however it's spelt? [20:43] yeah [20:43] it's another option, no decisions as yes [20:43] yet [20:49] Is there a way to convert a Mercurial repository to a Bazaar repository? [20:51] Devourer: you can you hg fast-export and bzr ast-import quite easily [20:51] *fast-import* [21:56] hi abentley, poolie [21:59] jelmer: Hi. [22:14] hi [22:24] does pushing a large revision often say 0/0 for a long time? [22:25] how do i drop a revision [22:25] i dont want to change any files [22:25] i just want to delete revision 2 [22:25] i really messed it up [22:25] you can't, really [22:25] you could branch from r1 and merge r3... into it [22:26] ok i'll just delete the files and recommit [22:26] that won't remove them from the history, though [22:27] revision 2 has too many files and it hangs trying to upload them [22:27] you might just need to wait longer [22:27] i put wxGTK and wxMozilla in it [22:27] im not supposed to do that right? [22:27] no [22:27] yeah [22:28] exactly [22:30] bob2: i broke bazaar [22:30] bzr: ERROR: Could not acquire lock "(remote lock)" [22:30] what do i do? [22:32] bob2: im about ready to hit myself in anger [22:34] meoblast001: If you're absolutely sure the branch/repo isn't actually in use, you can use "bzr break-lock $the_url" to unlock it. [22:34] ok [22:34] thanx [22:34] i know its not [22:34] no one ever uses it [22:35] meoblast001: You do. :) [22:36] i dont want to push revision 2 [22:36] only revision 3 [22:37] revision 2 has too much for my internet connection to handle [22:38] Peng_, is this imposible? if it is im gonna hit a wall or something [22:39] meoblast001: More or less, no. It would be easiest to uncommit revisions 2 and 3, delete the stuff from 2 and recommit 3. [22:39] ok [22:40] how do i uncommit? [22:40] i hate timelimits [22:40] meoblast001: bzr uncommit -r1 [22:40] "bzr uncommit" :D [22:40] to commit *to 1* [22:40] you'll probably want to follow that with "bzr revert" to stop versioning mozilla and wx [22:40] either revert everything, or just parts of the tree [22:41] thanx [22:42] i always forget this stuff [22:42] i gtg though [22:42] bye [23:50] spiv, heh. "This space left intentionally blank." [23:54] emmajane: pretty much :) [23:54] I like it. :) [23:55] I'd been getting random errors on .bzr/repository/obsolete_packs/': as well. But it wasn't anything I could reproduce reliably so I hadn't reported it yet. [23:55] It sounds like that might be the same problem though? [23:55] Right. [23:55] sometimes it's a problem, and then 2 seconds later I run the same commit again, and it's fine. [23:56] It's an often empty directory, but whenever packs are repacked (which happens automatically every so often) then you'll get that error. [23:56] ah [23:56] yes, "every so often" [23:56] "every so often" == every 10 commits, if commits are all you do (rather than pushing/pulling into the repo, say) [23:57] "random" problems are a little bit infuriating. :) [23:57] lifeless: is it a bug that for looms, 'switch' will change threads when there are uncommitted changes, but 'down-thread' won't? [23:57] lifeless: (hint: yes) [23:57] spiv, commits and uploads. [23:57] jml: down-thread won't? [23:57] spiv, I'm in the middle of tweaking CSS files... [23:57] jml: no [23:57] jml: I thought it did... [23:58] (but I guess I haven't tried doing that) [23:58] jml: I thought about it hard when merging abentley's branch, and its correct [23:58] lifeless: it's also causing me lots of problems. [23:58] jml: now is not a great time to discuss; but think about the issues of going down-thread with a pending merge - down-thread is recommended for various workflows [23:59] jml: alternative safety nets are good [23:59] as in, I'd like alternatives [23:59] lifeless: I want switch to stop me from switching unless I pass a --yes-i-know-what-im-doing flag [23:59] that would give me options [23:59] but as you say, we can discuss later. [23:59] lifeless: I'll review your branches now, btw. [23:59] It took me a while (and a lot of face on desk smashing) to figure out that "uploads" the directory is in no way related to the plugin uploads. [23:59] and we won't talk about the .bzr folders that I just plain old deleted assuming it was corrupt.