=== Snova_ is now known as Snova [00:38] Hi guys...I'm suddenly having doubts about whether I use bzr correctly....If I want to operate in a distributed fashion, every time I or a team mate make a change to our feature branch, we should merge it into the mirror, then push the mirror, correct? [00:39] what happens if I push, and then he pushes without first pulling? [00:43] wade: it sort of depends how you define "mirror" [00:43] if you are both collaborating on a single branch [00:43] you can both 'just push' to the mirror [00:43] you don't have to merge [00:43] if he pushes without pulling [00:43] bzr will tell him that the branches have diverged [00:43] and he needs to merge before he can push again, etc. [00:43] you could also use checkouts, which automate some of that book-keeping [00:43] though it means that "bzr commit" will block and tell you to update [00:44] Ok...I was just getting worried that I might be blowing away his changes when I pushed [00:44] wade: we're pretty good about not letting you "blow away" anything [00:44] as long as you don't use --force or --overwrite style flags [00:45] Ok, good to know - thanks [00:48] anyone able to provide insight on why I get "bzr: ERROR: unknown command "push-and-update" "? [00:48] syncrondi: because there isn't such a command? [00:48] Are you installing the "push-and-update" plugin? [00:49] I'm pretty sure it just piggy backs on regular "bzr push" anymore [00:49] I've used it before on another machine.. it's a plugin? [00:49] I don't remember installing a plugin to get thte capability [00:50] syncrondi: "lp:bzr-push-and-update" [00:50] when I do just a push, I get bzr: ERROR: No push location known or specified. [00:50] definitely a plugin [00:50] syncrondi: for "bzr push" if you haven't told it where to push before, how would it know? [00:50] "bzr push $URL" [00:50] and then future pushes will use $URL if not given [00:51] ah [00:51] Hmm, it looks like the plugin is only for linux, but I definately never used it on Linux before [00:51] syncrondi: not at all [00:51] It needs access to "ssh" [00:51] though it could be updated to not have that dependency [00:52] oh, ok [00:52] (and use the python 'paramiko' library) [00:52] there is an open bug on that [00:52] I've been putting off installing the paramiko library [00:52] syncrondi: well, on windows you either need paramiko or ssh anyway [00:52] plink is possible, but really doesn't work well [00:53] I've been using plink for a while without hitch on another system, but a friend helped me set it up [00:53] Now with my other box, I've had some tough luck getting everything working [00:55] bzr push seems to be working now that I've included the URL [00:55] But what's the difference between that and push-and-update? [00:55] syncrondi: so the main issue is auth support for paramiko [00:55] which is... poor [00:55] but as long as you only use pageant, it should be ok [00:55] 'bzr push' will push the history changes to a remote location [00:56] but won't update a remote working tree [00:56] 'push-and-update' goes out and after pushing the history, runs "bzr update" on the remote host [00:56] ok [00:56] So I can just run them separately [00:56] syncrondi: well, if you *need* to have a remote working tree, and update it, yse [00:56] yes [00:56] thanks jam [00:56] many use cases don't require a remote working tree [00:57] What do you mean by 'don't require' ? [00:57] syncrondi: if you are sharing your code between developers [00:57] you don't need a working tree on the shared server [00:58] just the *history* [00:58] It's a live site. :) [00:58] the cases for having a working tree tend to be stuff like "I want to maintain my http website on a remote server" [00:58] note that you *might* prefer the 'bzr-upload' plugin for tha [00:58] that [00:58] What are the advantages of the bzr-upload? [00:59] syncrondi: many people maintaining a live site *don't* want the history copied there [00:59] in case there is a security breach in the http requests [00:59] exposing the history of the site to 3rd parties [00:59] ah, ok jam. [00:59] also, bzr-upload works over things like ftp [01:00] which is insecure, no? [01:00] I'd rather stick with the ssh. [01:00] syncrondi: many hosting sites only let you upload via ftp [01:00] so not an option [01:00] if you are in control over the server [01:00] *I* would probably use a bzr lightweight checkout [01:01] and "ssh host; cd website; bzr update" [01:01] I see [01:02] You've been very helpful. Thanks jam. I'm not the only one running the show, so I've been just going with the flow so far. [01:08] Hm. I'm seeing an actual update on the server side [01:08] not* [01:11] is it possible to retrieve a snapshot of a shared-repo atomically? [01:11] I wouldn't be sure how to do that nyu [01:14] I'm not even sure how to install the push-and-update plugin, jam [01:14] I feel like a frickin' noob at this.. because I am [01:20] nyu: not easily. If it is a requirement, I would suggest staging it to another location, and then using that. [01:20] 'bzr pull' etc know how to get a stable view of the data. [01:20] Which you could emulate, but requires copying things in a specific order [01:20] (read all the branches first, then read file X, then Y, etc.) [01:21] quite a bit easier to just do: [01:21] jam: could I then pull all branches automatically? I know multi-pull, but AFAICT it only pulls branches I already have locally [01:21] I branched the push_and_update plugin to my plugins folder, but I still get the error with unknown command. [01:21] nyu: for b in `bzr branches`; ... [01:21] syncrondi: make sure 'bzr help plugins' mentions it [01:21] oh [01:22] 'bzr branches' looks useful [01:22] and then it is probably just "bzr push" and it will check to see if the remote has a working tree, and if so spawn "ssh host bzr update" [01:22] jam: [01:22] bzr plugins doesn't mention [01:22] jam: I see, thanks [01:23] syncrondi: where did you actually put it, and what did you call it [01:23] on linux [01:23] cd ~/.bazaar/plugins [01:23] bzr branch lp:bzr-push-and-update push_and_update [01:23] on Windows [01:23] cd %APPDATA%/Bazaar/2.0/plugins [01:23] IIRC [01:23] Yeah, I just did a bzr branch https://launchpad.net/bzr-push-and-update push_and_update on C:\Program Files\Bazaar\plugins [01:24] I should put it in appdata, jam? [01:25] jam: working towards it [01:25] lifeless: sure. I'm also still seeing skips as failures [01:25] the subunit protocol branch is most of the way there, the details api seems nice, want to get a TestCase test-code support [01:25] jam: land vila's branch [01:26] if you like - bug, uhm, [01:26] https://code.launchpad.net/~vila/bzr/419776-subunit/+merge/14413 ? [01:26] doesn't seem the right one [01:26] bug #419776 [01:26] Launchpad bug 419776 in bzr "selftest --subunit output incompatible with --parallel=fork" [Medium,Fix committed] https://launchpad.net/bugs/419776 [01:27] I thought it was *your* branch that does it [01:27] ye sthat one [01:27] or are you saying a branch of subunit? [01:27] no [01:27] so my branch for bzr, which is landed, fixes things for the upstream API [01:28] released subunit though has a bug itself, in that its serialiser doesn't have addExpectedFailure [01:28] *and* the downgrade logic makes that a fail not a success [01:28] vilas branch just changes bzr to use the same logic the next release of subunit will, to degrade in that case to success [01:29] my protocol branch of subunit has addExpectedFailure on the subunit serialiser [01:29] I may backport that to subunit trunk when I get a minute [01:29] lifeless: well, I'll submit vincent's patch, since I *did* approve it :) [01:29] as did you it seems [01:30] http://feeds.laughingsquid.com/~r/laughingsquid/~3/iTCZZ6YbvfE/ <- wow [01:31] jam: shame to hear about the flight snafu [01:31] lifeless: for those who don't get enough rock band/ etc [01:31] lifeless: thanks for your sympathy [01:32] I suppose it means I got some quality hacking done with a real power supply :0 [01:33] :) [01:37] 3 minutes and I run out of free wifi apparently <> [01:37] you'd think, with a plane ticket they could throw in 5$ of bandwidth [01:39] lifeless: :) [01:39] you'd think the same thing about a $200/night hotel [01:39] lifeless: what is the subunit command that you can then pass to "bzr selftest --load-list" ? [01:40] ah, perhaps subunit-ls [03:37] Stupid question: How do you make --no-strict the default when pushing? An alias? [03:38] Another stupid question: How can I unlock a file that bzr locked, preventing me from doing a commit? [03:38] Apparently, I'm HM [03:39] xD [03:41] "unlock"? In what way? [03:42] :: Unable to obtain lock file:///C:/Documents%20and%20Setting.... [03:42] Oh god, Windows. :P [03:43] Pastebin the full error. If you're sure bzr isn't running, "bzr break-lock" will probably fix it, depending on the problem. [03:43] held by syncrondi@host on host host[process #1712] locked 21 minutes, 27 seconds ago Will continue to try until 21:42:20, unless you press Ctrl-C If you're sure that it's not being mdified, use bzr break-lock [03:44] but the break-lock doesn't work [03:44] it complains that it isn't a branch [03:44] wait [03:45] I don't need all that jibberish after the command [03:45] it was saying I needed to have some arguments, but it worked without arguments [03:46] Yeah, "bzr break-lock" defaults to the current directory if you don't specify a path/URL. That trips people up a lot. [03:47] Thanks Peng! [03:49] :) [08:33] * jml considers dinner plans [08:41] jml: I suggest eating. [10:35] *yawn* [10:44] morning [10:44] i've read that bzr's interoperability with hg is not as good as with git & svn repos...what can be done atm? [10:46] i need to keep hg repo of the project in sync, but have lot of problems using forest extension with my flakey internet (mobile 3g) connection and wonder if i could convert repo to bzr locally and somehow pull and commit from bzr? [10:48] gour: it needs more work :) [10:48] LarstiQ: anything works now? [10:48] gour: yeah, jelmer has been improving it lately I gather [10:48] LarstiQ: is it in 2.1 trunk? [10:49] gour: I'd assume it's in lp:bzr-hg [10:49] here i've 2.0.0 [10:49] ahh, let check it at lp [10:50] hg's subrepos is not quite ready, hg does not have, afaik, working fast-export/import...:-/ [10:51] hmm, web says: " [10:51] A plugin for bzr which provides support for Mercurial file formats. It currently allows pulling from (and eventually pushing to) hg repositories." [10:51] the description might be old, I'd look at the recent revisions / NEWS [10:53] ok, let me pull it [10:59] i've installed plugin, created new bzr repo and tried to pull from it, but get: bzr: ERROR: No module named hg.ui [11:04] i cannot see anything similar in bugs [11:06] jelmer: shall i open a ticket for the above problem? [11:06] gour: have you got all dependencies installed? [11:06] LarstiQ: hmm, i just pulled the plugin [11:06] gour: not sure if hg.ui is part of standard mercurial, or maybe version differences [11:07] LarstiQ: hg works here [11:07] gour: can you `python -c 'import hg.ui'`? [11:08] LarstiQ: ImportError: No module named hg.ui [11:09] i'm still not sure about hg.ui extension/module [11:10] back [11:11] LarstiQ: i.e.i believe it's due to [ui] section in my ~/.hgrc [11:11] gour: maybe, what does google say? And what happens if comment the ui section out [11:13] LarstiQ: commenting ui section does not help [11:16] otoh, "from mercurial import hg, ui" works [11:32] * gour submitted bug-report [11:38] hmm, 'make check' for bzr-hg fails as well [11:45] hmm, this line is the problem: from bzrlib.plugins.hg.ui import ui [11:48] bze cannot find the plugin which is installed in ~/.bazaar/plugins [11:48] *bzr [11:49] but i wonder how is it that 'bzr plugins' lists the plugin? [11:50] hello, i want to create a bzr store, but without publishers, i want the contributors commits to be publish automatically [11:50] how to do that? [11:59] LarstiQ: the problem was that i did not run: python setup.py build_ext -i in plugin dir [12:03] gour: aha [12:18] hi garyvdm [13:05] is fast-export atomic? [13:10] in which sense? [13:10] hi bialix [13:10] hi nyu [13:10] i.e. if I fast-export from a remote repo while someone is pushing [13:10] I assume fast-export will always obtain a consistent snapshot? [13:11] hi phcoder [13:11] IIUC it should keep read lock and will use only latest branch tip as of operation start [13:11] jam said yesterday "'bzr pull' etc know how to get a stable view of the data." but didn't mention fast-export [13:11] but there is possible quirks with repo re-packing [13:12] nyu: without looking at the fast-export internals I'd say most likely it uses the same code as pull to get revisions from repo [13:12] I don't understand read lock. do you mean write lock? [13:12] but I can't be 100% sure [13:12] bzr has both read and write lock on internal structures [13:12] hello all. I fast-imported my old repo but some branches are missing because they were merged in other branches. Can I somehow create a branch from commit-id? [13:13] to get stable reads either you use atomic test-and-set operations or you lock writes, right? I don't understand what locking reads means [13:13] if somebody pushes then dest repo will be write locked, therefore attempt to read lock it will fail [13:13] ah, ok [13:13] write lock excludes readers [13:14] phcoder: just branch from desired revision [13:14] shit, I think I was lectured on this less than 6 months ago. I forgot everything [13:14] thanks bialix, I remember now :-) [13:14] :-) [13:15] phcoder@debian:~/grub2/bzr-new$ bzr branch -r phcoder@gmail.com-20091106215044-ehljkps46fa20o1s mips-all mips [13:15] bzr: ERROR: No namespace registered for string: u'phcoder@gmail.com-20091106215044-ehljkps46fa20o1s' [13:17] bzr branch -r revid:phcoder@gmail.com-20091106215044-ehljkps46fa20o1s mips-all mips [13:17] phcoder: ^ [13:18] `bzr help revisionspec` may help [13:35] Is there any file in the .bzr directory that gives you just the current revision? (Trying to use current revision as a dependency in the Makefile, since one of my scripts depends on the revision) [13:37] mathepic: Use `bzr revno` or `bzr version-info` [13:38] Yeah, but I need to know when `bzr revno` in the makefile changes. [13:38] oh, hmm [13:39] is that an "oh, hmm, I figured it out"? :) [13:39] No, its "hmm, I think I see your point" [13:39] ok [13:40] Would make be able to find the change in the shell output from the call `bzr revno` [13:40] you could cache it for one, and use a .PHONY target [13:41] mathepic: but maybe it's good to take a step back and look at what task you're trying to accomplish with this mechanism? [13:41] Generation of a header file for use in the main program [13:42] It uses bzr version-info to get it, but then the output of course would change when the revision changes, so it needs to be run again [13:43] * LarstiQ waits for mathepic to come back [13:51] bialix: thanks. I'll look into it after restoring our experimental branch [13:52] mathepic: is it your question on stackoverflow about Makefile and revnos? [13:53] phcoder: looking into it? into what? [13:53] phcoder: into restoring my branches which were merged [13:56] LarstiQ: hi, when and if mathepic will come back say him he had answers on his SO question [13:57] bialix: I will remember to tell him that [13:59] here is http://stackoverflow.com/questions/1694264/how-to-make-dependency-in-makefile-so-that-target-is-built-when-bazaar-revision-c [14:57] how can I see which revision two branches diverged at? [15:00] cody-somerville: I'd use `bzr missing` [15:15] bialix: Hi - I think my network connection sorted now [15:15] Hi GaryvdM ! [15:16] bialix Any other regressions other than Bug 478239? [15:16] Launchpad bug 478239 in qbzr "treewidget missed checkboxes for dialogs that require them" [Critical,Fix released] https://launchpad.net/bugs/478239 [15:20] bialix: Otherwise I would like to release 0.16 now. [15:20] Oh - Just noticed Bug 47827 [15:20] Launchpad bug 47827 in vmware-player "vmware-player lintian warnings" [Medium,Invalid] https://launchpad.net/bugs/47827 [15:20] I mean Bug 478277 [15:20] Launchpad bug 478277 in qbzr "Test test_model_working_tree failed: list index out of range" [Undecided,New] https://launchpad.net/bugs/478277 [15:22] GaryvdM: I'll update by qbzr copy right now, have no more bugs off hands [15:27] GaryvdM: what about https://bugs.launchpad.net/qbzr/+bug/475286 [15:27] Launchpad bug 475286 in qbzr "Tools/Plugins not working - bzr: ERROR: exceptions.IndexError: tuple index out of range" [High,Confirmed] [15:28] bialix: I don't think it is a regression, but it should be easy to fix. [15:29] it seems it's already fixed [15:29] lemme check [15:30] no, it's not [15:30] I'll fix it, wait 15 mins [15:32] Cool - 478277 is going to take me a while, because I'm having a failer in that test, due to a pyqt 4.6.0 bug, so I'm going to have to upgrade to 4.6.1 [15:33] GaryvdM: bug 475286 is not regression per se, but traceback in qbzr is bad [15:33] Launchpad bug 475286 in qbzr "Tools/Plugins not working - bzr: ERROR: exceptions.IndexError: tuple index out of range" [High,Confirmed] https://launchpad.net/bugs/475286 [15:35] GaryvdM: done, pushed [15:38] cody-somerville: qlog is your friend ;-) [16:05] how can I incorporate a file from one unrelated (no common ancestor) branch to another without losing history? [16:06] hello, i want to create a bzr store, but without publishers, i want the contributors commits to be publish automatically, how to do that? [16:09] nyu: Dose the other branch have file other than the one you want to import? [16:10] elroboto: I'm not sure I understand you correctly, but if you create a public place that people can push to, other people will be able to pull stright away? [16:11] elroboto: I'm not sure what you mean by "without publishers" [16:12] nyu: fast-filter the file first, then merge filtered branch [16:14] but you will lose liason to original history [16:14] bialix: uhm this discards original history? [16:14] GaryvdM: they're totally unrelated branches, with different files [16:14] filtered history will have new revision-ids and file-ids [16:15] GaryvdM: i.e. one project borrowing files from another [16:15] bialix: oh, that's ok [16:15] bialix: but it'll have whole history of the imported file? [16:15] most likely [16:16] ok thanks [16:16] fast-filter can't support moves of files between directories [16:16] otherwise it's ok [16:16] try it [16:17] you now almost expert in it ;-) [16:17] hehe [16:18] Hi. [16:18] I might be having a case of the stupid, but I'm having trouble installing bzr 2.0 from the PPA on Hardy [16:19] because of the following error: "bzr: Depends: python-central (>= 0.6.7) but 0.6.5ubuntu1 is installed." [16:19] Does anyone have any advice for eliminating my stupid? [16:24] gmb: Acording to https://edge.launchpad.net/ubuntu/+source/python-central/ , hardy-updates has python-central 0.6.7 [16:25] GaryvdM: Aaargh. I have the dumb. Not only is the hardy-updates line commented out in sources.list, it actually points to feisty, not hardy. God knows what I was tripping on when I upgraded. [16:25] Thanks for that! [16:27] Yep, that fixed that and a number of other things. Awesome. [16:27] GaryvdM: Thanks again :) [16:27] gmb: It's a pleasure. [16:34] Hi there! [16:34] Anyone available for a bit of help? [16:35] Well, let's try anyway :s [16:35] Abstract: checkout permanently out of date [16:36] symptoms: last part of http://osdir.com/ml/bazaar/2009-08/msg00479.html [16:36] in short: [16:36] inside a working repo, a checkout from another on an ftp [16:36] bzr up > "tree is up to date" [16:37] bzr ci --local … [16:37] bzr ci --local … [16:37] bzr ci > "bound branch is out of date with master branch" [16:38] of course, this is not true [16:38] done this twice already, no changes [16:38] any ideas? [16:39] but it is [16:39] Alcmene: `bzr up` after the --local should pivot them away, but you might prefer pushing those revisions to master first [16:42] I don't get it :s [16:42] I don't mean to update my working tree [16:42] but the distant repo only [16:43] plus, if I _do_ update, it of course reverts to the version on the server, which is the one before all my local commits [16:44] Alcmene: then just `bzr push` [16:44] (to the master) [16:45] yes, ok, but why does it tell me this?? [16:45] Alcmene: you're trying to commit, but your local branch is not at the same revision as the master [16:45] hence, it tells you to sort that first [16:47] mmh… I get it, but it still sounds like a very strange way of telling it IMO [16:48] and I'm almost sure I already did this without a problem [16:49] Alcmene: the message itself tells you to run `bzr update` and then commit, did you try that? [16:49] yes! [16:49] GaryvdM: ping [16:50] but updating _reverts_ : it updates to the tip of the distatnt repo, which is the same as reverting to the last no-local commit [16:50] Alcmene: no [16:51] Alcmene: it pivots your local commits as pending merges [16:51] well, how come a bzr stat after the update didn't show me _anything_ then? [16:51] * LarstiQ tried it locally and it works exactly as advertised [16:52] Alcmene: that I don't know, but if you can give instructions on how to reproduce that I'd like to try [16:52] * Alcmene backs up his local tree before trying it, and will tell the result in 2 minutes [16:53] luckily most of the commits were merges, 'cause I would have lost a lot [16:54] all your local commits are still in your local repository [16:54] not too hard to retrieve with `bzr heads` [16:55] :| wtf [16:55] well, ok, I know I'll sound like an asshole right now [16:55] bzr up did do the trick [16:55] why would that make you sound like an asshole? [16:55] I swear it did the opposite yesterday [16:56] because I have wasted your time for no reason! [16:56] hah :) [16:56] such is triaging [16:56] * LarstiQ doesn't mind [16:58] well, I must have made a mistake yesterday evening, was in a rush [16:58] sorry again, thanks a lot LartstiQ [16:59] Alcmene: np :) [17:01] GaryvdM: I've made one more last-minute-fixes commit. Now I think road for 0.16 is free [17:15] bialix: pong [17:15] was waiting for pyqt 4.6.1 to finish building [17:15] GaryvdM: I've made couple of minor fixes for new cmd logging feature [17:16] GaryvdM: I understand [17:16] bzr qlog [17:31] bialix: I got my tests passing now. But I don't see Bug 478277 [17:31] Launchpad bug 478277 in qbzr "Test test_model_working_tree failed: list index out of range" [Undecided,New] https://launchpad.net/bugs/478277 [17:31] bailix: fix is simple, but will you please test it before I push to trunk. [17:32] where is your fix/patch/branch? [17:32] after updating to 1.18 push is pretty fast [17:33] bialix: lp:~garyvdm/qbzr/Bug478277 [17:34] * bialix branching [17:34] hsn: you may want to thanks lifeless and spiv I guess [17:36] GaryvdM: tests passed; please land it [17:36] cool === tchan1 is now known as tchan [20:49] abentley: awake? [20:49] lifeless: yep. Good morning! [20:53] I'm on a demo plan from a nearby wifi hotspot - do you know details about how we;re meant to get wifi? [20:53] and / or breakfast? [20:53] I checked in late last night - semi automated, so have no clue :( [20:54] lifeless: No wifi here. DSL in every room. We're on our own for breakfast. [20:54] lifeless: I can drop in on you in, say 10 minutes if you like. [20:55] lifeless: Or you can call me in room 50 [21:30] Hi guys...is there any way I can troubleshoot what appears to be a SSH hang with bzr? (using a sftp URL)...I got one the other day, and now my programming colleague is getting one when trying to do a merge - he issues the merge command, it asks for the password and then it just hangs [21:32] wade2: ~/.bzr.log, possibly with some debugging flags (bzr help debug-flags) [21:33] Ok, thanks === mtrudel_ is now known as cyphermox [21:54] does anyone has had problems using bzr in ubuntu with ntfs partitions [21:54] NTFS is usually mounted read-only. have you mounted the partition is question rw? [21:54] s/is/in/ [21:57] yes rw [21:57] the problem is that bzr detect permission changes [21:59] becxause of this I don't want to use it, I'm afraid of confusion [22:00] I mean, I would like to use bzr in ubuntu, but can't becasue of this problem [22:00] RenatoSilva: I tried that the other day. It kept on telling me the x bit had change. [22:01] RenatoSilva: I ended up just branching on to a ext4 drive [22:01] GaryvdM: I think that's it, don't want to test it now :) [22:02] RenatoSilva: Pulling and Pushing should work fine. It's just working with a Working Tree that is problematic. [22:02] GaryvdM: I don't understand [22:02] GaryvdM: nothing has changed, so it should not tell the contrary [22:03] GaryvdM: I don't understand why does it consider permission changes. Do bzr store them in the metadata, the permissions? [22:04] RenatoSilva: x (execution) does not get stored by ntfs, so the linux ntfs driver just says every thing is executable. [22:04] RenatoSilva: bzr stores this because it is important for unix dev [22:04] * RenatoSilva is going to windows to work in bzr, brb [22:19] Moring jam [22:19] morning GaryvdM [22:19] just got into Sydney [22:20] *morNing [22:20] still about 4-5 hours out from making it to the sprint [22:20] how's things with you GaryvdM? [22:21] Good [22:21] About to announce qbzr 0.16 [22:21] Hi vila [22:21]