[00:22] poolie: quick sanitty check [00:22] /usr/lib/python2.6/dist-packages/bzrlib/transport/local.py(484)stat() [00:22] that does a 'stat()' [00:22] not an 'lstat()' [00:22] is it just me or is that nuts ? [00:27] hi [00:28] this was discussed a while ago [00:28] during the review that added it [00:28] i can see arguments either way [00:28] consistency vs probably wanting to know about symlinks explicitly [00:30] oh, actually. [00:30] I think my package version is just a little old [00:30] trunk does an lstat [00:35] i just did a checkout lp:django and realised i should have really used a shared repo [00:35] is there an easy way to copy a stand alone repo into a shared one? [00:35] sure [00:35] make a shared repo and move your branch under it [00:35] then do bzr reconfigure --use-shared (It hink thats the option - --help will tell you) [00:35] cool [00:37] lifeless: thanks [00:37] lifeless: and you want it to do an lstat? [00:37] i think that's what i recommended [00:37] poolie: trunk does an lstat, and yes thats what I want. [00:37] generally we want to handle symlinks explicitly/consciously [00:37] 2.1.1 does a stat [00:43] that's slightly surprising, i thought it all landed together [00:45] transport.stat is very old [00:45] IIRC [01:02] Good morning. [01:04] Saw an O'Reilly book on Git. It was very confusing :) [02:10] spiv: did you sleep on https://code.edge.launchpad.net/~lifeless/bzr/commands/+merge/23068 [02:10] ? [02:13] lifeless: oh right, thanks [02:20] lifeless: reviewed. (Spoiler: I approved it) [02:20] changes needed ? [02:21] Suggested a docstring change. [02:21] Your call on what to do with that. [02:24] I have a question... [02:24] if I call b.iter_merge_sorted_revisions(some_rev_id), i.next() is the revision I specify [02:24] but if I call b.iter_merge_sorted_revisions(some_rev_id, direction='forward'), i.next() is the first revision [02:25] as in revno 1 [02:25] why?/ [02:25] is it an error [02:25] or expected behaviour [02:26] thumper: The docstring does say "forward returns tuples in the opposite order to reverse." [02:26] And direction='reverse' is the default. [02:26] spiv: thanks [02:32] spiv: ah [02:45] * igc out for a few hours [02:48] thumper: hey, can I chat with you for a minute (voice) ? [02:48] lifeless: sure [02:48] lifeless: pots or skype? [03:02] skype :P [03:04] poolie: subunit : http://pqm.bazaar-vcs.org/ [03:05] sounds intriguing... [03:05] meaning... progress bars? or it now mails back subunit spam? [03:10] both [03:11] the display is pqm parsing the subunit stream [04:14] lifeless: do you know of an object that acts like a list, but is fed by a generator and reusable? [04:14] uhm [04:15] I've written things that are likish that [04:15] storm might have one [04:16] lifeless: it would make my logic much more easy to follow [04:16] lifeless: and still benefit from not calling a whole history operation [04:18] thumper: "acts like a list" meaning "indexable"? [04:18] spiv: no, I mean iterable again [04:19] Oh, reiterable. [04:19] caching iterator [04:19] yeah [04:19] http://code.activestate.com/recipes/576941-caching-iterable-wrapper/ [04:25] thumper: ^ [04:25] lifeless: yeah, saw it, looks overly complicated for what I want [04:25] making something simpler [04:25] thanks though [04:25] de nada [04:36] is there anyway to get bazaar to push changes to a host filesystem? i've tried using bzr+ssh://domain.com but keep getting the error "bzr: command not found" and then ERROR: Connection closed: Unexpected end of message. [04:37] wolfgang00: Just drop the “bzr+” bit of that url and you'll be able to push the changes. [04:38] RAOF: bzr doesn't support ssh urls [04:38] wolfgang00: you can relpace bzr+ssh with sftp [04:38] Gah, of course. [04:40] when i just use ssh: bzr: ERROR: Unsupported protocol for url "ssh://yardbutl@yardbutlerstore.com": bzr supports bzr+ssh to operate over ssh, use "bzr+ssh://user@domain.com". [06:10] our network fetching really is choppy :( [06:15] :( [06:16] hi spiv [06:16] welcome back to full time [06:18] Thanks :) [06:35] poolie: theres another hydrazine branch up for you, fwiw [06:36] heh, that's more cheering [06:37] I've tested status='Queued' and it works ok on staging [06:37] not flawless, but the web pages don't blow up [06:38] igc: ping [07:14] hi bialix [07:36] hi bialix [07:38] poolie: I have a fairly big hydrazine change [07:39] lp:~lifeless/hydrazine/cron - I hope you'll like it [07:39] k [07:39] hi igc [07:39] i was just merging some of them now actually [07:39] hi poolie [07:40] poolie: I'm calling it a day [07:41] poolie: if you want to talk this one through with me before I sign off, I'd be delighted to do so. [07:41] poolie: btw tomorrow mornign I'll be awolish - popping up to the doctor at sparrows fart [07:46] wow, interesting [07:46] let's chat tomorrow afternoon? [07:47] spiv: do you have any ideas on how to test the ssh rekeying thing locally? [07:48] it would be good to test both paramiko against an openssh server and openssh client against conch i guess [07:49] mwhudson: manually hack REKEY_BYTES in paramiko/packet.py, maybe? [07:51] poolie: ok [07:51] spiv: ah, that might work [07:51] poolie: I have some personal stuff to organise with you too, this week before the weekend if at al possible [07:52] mwhudson: it's possible there's something in conch for that too [07:52] mwhudson: I don't know if it's something either side can trigger [07:52] spiv: ah, i was going to say [07:52] i /thought/ it was a client side thing [07:52] maybe not though [07:52] mwhudson: maybe Conch gets confused if both sides trigger at about the same time? ;) [07:53] spiv: oh god, if it's that complicated someone else can fix it :) [07:53] although noone will :/ [07:54] mwhudson: heh [07:54] mwhudson: thinking of concurrency, I happened to see your lp code review about CachingIterator, and you need to be more devious :) [07:54] mwhudson: I sent mail explaining how :) [07:54] spiv: "don't use it with threading" [07:55] mwhudson: pfft [07:55] ah i see [07:55] mwhudson: I don't need no stinking threads to break stuff ;) [07:55] spiv: does thumper's update work better with your code? [07:55] Haven't seen it, so don't know. [07:56] If it does better than giving three different answers, all wrong, then that's a start :) [07:58] hi all [07:58] vila: hi! [07:58] vila: hey, lost+found merges; is that in progress, or just something we've been discussing [08:01] lifeless: still not in progress but my next target once I come back to conflicts [08:01] * mwhudson becomes a little stabby trying to get paramiko to connect to launchpad.dev [08:07] hi vila, spiv [08:07] hi mwhudson [08:07] hi Ian [08:07] hi vila [08:07] hi poolie [08:08] igc: hello [08:12] spiv: it's conch [08:12] mwhudson: \o/ [08:13] mwhudson: I guess the rekeying logic in conch is not carefully unit tested... [08:13] mwhudson: btw [08:13] mwhudson: thumper's updated code is differently broken :) [08:13] spiv: what an outrageous accusation [08:14] (both of those!) [08:15] spiv: what are the chances that the hardest part of fixing this upstream will be figuring out how to test it [08:16] mwhudson: depressingly high :( [08:23] i wonder if it's something as dumb as the server continuing to send the data from the subprocess while key exchange is happening [08:24] mwhudson: that sounds quite plausible. [08:24] spiv: what update code? [08:25] what also sounds plausible is that it's time for dinner [08:26] thumper: your CachingIterator has bugs, see my post to the code review [08:26] spiv: do you have a solution too? [08:27] Has Command.pligin_name() been deprecated or something? Doesn't seem to work on my bzr 2.1.1 any more, though it used to in the past. [08:27] spiv: is there any way i can change what options bzr passes to ssh? [08:27] thumper: no, although simply raising an exception if someone tries to use it re-entrantly would probably be adequate. [08:27] spiv: other than "hack the source dummy" [08:28] mwhudson: write a plugin to override the ssh vendor implementation, dummy? :) [08:28] mwhudson: (no) [08:28] luckily i can use .ssh/config for this [08:32] ah "cool" can reproduce with openssh as well as paramiko [08:33] mwhudson: That's... "good" :) [08:34] yeah [08:35] * mwhudson files a twisted bug [08:47] spm, can you answer the mail thread about bzr-email and the hpss chroot? [08:48] poolie: which list? [08:48] poolie: you me, perhaps? [08:48] oops, i did mean spiv [08:48] "you mean me", rather. [08:48] Obviously it's too late in the day to expect accurate typing :) [08:48] spiv: which should so do a nick change and really confuse people - som april fools ideas just come too late.... [08:49] poolie: could you land your 2.2b1 branch into lp:bzr/2.0 ? That would make it easier to track what is in the corresponding 2.2b1 tarball [08:49] do you mean into bzr/2.2? [08:49] poolie: what ? You still can't parse my typos ? :-/ Yeah I meant lp:bzr/2.2 of course === radoe_ is now known as radoe [09:23] vila, sure, i'll try to do that soon [09:24] poolie: thanks ! [09:24] hi all [09:27] hi bialix, spm [09:27] hi igc [09:29] igc: I think I also need CRT90 redistributable? [09:30] igc: I don't trust the tools that I don't understand [09:33] bialix: yes, you'll need that [09:33] bialix: do you have VS 2008? [09:33] bialix: http://www.py2exe.org/index.cgi/Tutorial#Step52 [09:33] igc: no, I haven't [09:34] bialix: thei nstaller code tries to implement that process [09:34] installer [09:35] * bialix reads [09:35] bialix: http://www.microsoft.com/downloads/details.aspx?FamilyID=9b2da534-3e03-4391-8a4d-074b9f2bc1bf&displaylang=en [09:35] * bialix starts to think using buildout is not the best idea [09:37] bialix: my feeling is that buildout is not the best fit here - wget + bzr would be less complex and more reliable IMO [09:38] igc: maybe I should install VC2008 Express? [09:38] http://www.microsoft.com/express/Downloads/#2008-Visual-CPP [09:39] igc: I think my traceback caused by damaged svn-1.6.6.zip file. [09:39] bialix: can you remember why we need pywin32? Was it needed for tbzr packaging? [09:39] *svn-win32-1.6.6.zip [09:39] bialix: got to run [09:39] delete that corrupted zip and try again [09:39] * igc dinner [09:39] pywin32 is must have for windows and bzr [09:39] igc: bon appetit and gnight [09:43] igc: deletinbg file does not help: I have the same problem again [10:47] hi [10:48] just wondering... what if i use colocated tree model and happen to delete the whole directory? [11:26] millun: how do you mean? [11:56] jelmer: if i delete the php project (not intentionally) including the .bzr directory [11:56] i am screwed right? i use bzr explorer. i obviously should PUSH quite often? [11:57] to some remote repository [11:57] millun: yes [11:57] unless you're using a shared repository that you haven't deleted? [11:57] no, i am not using a shared repository (not quite sure what you mean by that). i come from SVN, u know [11:58] i use "colocated branches" [11:59] because that's what i've learned in the tutorial [11:59] millun: colocated branches aren't an officially supported feature [12:00] millun: what tutorial was this? [12:00] on the official pages [12:00] lemme look it up [12:01] http://doc.bazaar.canonical.com/explorer/en/guide/processes/starting_a_project.html [12:01] at the very end [12:02] In most cases, the best model comes down to personal preference. In some case though, the colocated branches model is the only choice, e.g. if you want to version control files in a special location (like /etc or a set of web-server configuration files). [12:04] igc: hi [12:04] igc: The starting_a_project page seems to use the term colocated branches for something different than we do elsewhere [12:19] gnight === mrevell is now known as mrevell-lunch === mrevell-lunch is now known as mrevell [14:59] how can I create a branch which only contains code until a specific revision and not the subsequent revisions? the problem is that I forgot to create a milestone at some point for my project and I now need to go back in time to create a bugfix for an earlier milestone [15:01] cr3: it's been a while since I've done it, but you can do the branch command with -r rev# it should work [15:01] like this bzr branch -r 2.3.1 ./ ./new [15:01] IIRC that is. you can do bzr help branch to make sure I got the syntax right [15:03] rubbs: seems like you got it right, thanks! [15:04] np === deryck is now known as deryck[lunch] === IslandUsurper is now known as IslandUsurperAFK [17:11] Hi all [17:12] I hacked a bzr-gc.sh script (http://paste.debian.net/68592/), does it make you run away screaming? [17:16] * fullermd runs and screams. [17:23] <__monty__> How would I go about fixing this bug: https://bugs.launchpad.net/bzr/+bug/45599 ? [17:23] Launchpad bug 45599 in bzr ""no repository present" error should suggest using bzr branch" [Wishlist,In progress] === IslandUsurperAFK is now known as IslandUsurper === radoe_ is now known as radoe [19:13] is it possible to push files to a remote server? ive tried bzr push sftp:// but i get the error This transport does not update the working tree [19:17] wolfgang00: use the bzr-upload plugin [19:17] push will push only the branch-related info, not the real files === mordred_ is now known as mtaylor [19:30] hmmm. when i use bzr upload i get the message "bzr: ERROR: Cannot upload to a bzr managed working tree:" [19:42] wolfgang00: bzr upload is only for non-bzr managed trees [19:43] wolfgang00: I mean, it can only update a remote tree without a .bzr directory [19:44] jelmer: so is there a way to upload to a bzr managed tree? im not entirely opposed to just removing the .bzr folder and not having it managed but its been a nice feature up till now. [19:47] wolfgang00, there's the push-and-update plugin [19:52] IslandUsurper: do i have to then use checkouts? [19:59] wolfgang00, no, you just push from one branch to the other, and push-and-update runs `bzr update` over ssh [20:22] vila: if you are still around, /wave [20:50] with push and update installed i get "This transport does not update the working tree of: ..." still [20:51] wolfgang00, it lies [20:51] a bug has been filed [20:51] but you should log in to the other server to make sure [20:52] you can see it say "running ssh bzr update" right? [20:56] i do not get it running bzr update. [20:57] it pushes the changes but i have to do update manually [21:42] when you do `bzr hooks`, do you have update-after-push under "BranchHooks: post_push:" ? [21:42] wolfgang00 ^^ [21:43] How do I delete a branch? [21:46] dougx, delete the folder [21:47] awilkins: thanks [21:47] dougx, if it's in a shared repo, the revisions will remain but they will have no accessible tip [21:47] ok, I see [21:47] done:-) [21:48] ok figured out the push and update plugin was not installed properly. thats fixed but when i run bzr push i get the error that ssh cannot find the file specified but it looks like its dropping the ~ out of the location [21:53] Hi. I'm trying to upgrade an Ubuntu Jaunty server to Bazaar 2.x. I've added the PPA source to apt, but it's still nto upgrading it. Am I missing something? [23:33] when i try to push to a remote server i get the error "ssh yardbutl@yardbutlerstore.com bzr update "public_html/testing.yardbutlerstore/" its dropping the ~/ from public_html [23:35] and i do not seem to have update-after-push under branchhooks