[00:44] mkanat: the warning is harmless [00:44] mkanat: is that an imported branch on launchpad, or elsewhere? [00:44] poolie1: Yeah, I figured. [00:44] nekohayo: he's away this week probably === poolie1 is now known as poolie [00:45] poolie1: I imported it locally. The problem seemed to be that I was importing into a rich-root-pack repo. [00:45] poolie: When I used the default repo type instead, it worked. [00:45] poolie: It also could have been some old bug, since it was bzr 1.5. [00:45] ok [00:46] poolie: .... this "week"?! he leaves IRC on for weeks on end? :) [00:46] oh well too bad I guess === raof is now known as RAOF [01:14] asac, hi [01:14] asac, authentication should work out of the box with bzr-svn [01:14] asac, http or https authentication doesn't work with bzr itself, so you may have to specify "svn+http://" or "svn+https://" to get it to bypass the standard bzr http transport. === fta_ is now known as fta [01:19] jelmer: oh. cool. thank you! === mw is now known as mw|out [05:26] Hi. You guys had any luck running a bzr smart-server on Vista? [05:27] AFAIK the TCP server works fine. I think I recall hearing about someone using it over HTTP with NTLM auth at some point, I'm not sure which version of Windows though. [05:30] Thanks. Strangest thing, Vista seems to be blocking the smart server. It allows Perl echo servers to run, and I can access bzr locally, but it doesn't show up over the netwoek. It is Vista that is doing something weird. [05:37] grettke: try explicitly using the external IP address in the --port argument? e.g. --port 1.2.3.4:4155 [05:38] spiv: Will do. [05:44] spiv: That does it. Why is this necessary? There is something with which I am not familiar because it didn't occur to me to bind it in that manner. [05:50] (is that the IPv6 bug again?) [05:51] grettke: (we had to do that, but in our case it was a problem of `bzr serve` not starting (blowing up) and this forced it to think of the address as IPv4 only which made it work right) [05:51] grettke: good question, it might be to do with IPv6 [05:51] Or it might be that passing None as the host to Python's socket.listen method is choosing the wrong interface for some reason. [05:51] grettke: it might be related to https://bugs.edge.launchpad.net/bzr/+bug/293697 [05:51] Ubuntu bug 293697 in bzr "bzr serve shows error TypeError: not all arguments converted during string formatting" [High,Confirmed] [05:51] grettke: (this was on a Linux server, so even more tangential to what you're asking about) [05:52] Hmm, actually, I don't think that's the bug I was thinking of. [05:52] Ah, https://bugs.edge.launchpad.net/bugs/286871 is what I was thinking of. That's also what AfC is referring to. [05:52] Ubuntu bug 286871 in bzr "IndexError raising CannotBindAddress error starting bzr serve" [Medium,Confirmed] [05:53] spiv, AfC, ubottu: I had pretty much ruled out the router, the switch, the other computer, and even bzr since I can access from the server box itself, just not other computer; the last thing about which I wondered was whether or not I should disable IPv6 completely on the card. [05:54] * AfC {chortle}s at the addressee list in the above line [05:57] grettke: I imagine you can just specify a 4 byte IPv4 address as Andrew suggested and move on. I haven't the faintest whether Microsoft uses IPv6 internally in its operating system (if it did, then turning it off might make things go splat) or for the E.T. phone home "feature". [05:59] Thanks everyone, now I can sleep well again; this has been a thorn in my side I tell you! [05:59] grettke: you could try --port 0.0.0.0:4155 to further narrow down what's going wrong. If that works, it suggests that it's trying an IPv6 address. [06:00] grettke: if it doesn't, it suggests for some reason passing None as the host is causing it to bind to the wrong interface (or perhaps simply that it's not binding to all interfaces). [06:00] grettke: either way, a bug report with the details would be good I think. [06:00] spiv: That does work: bzr serve --port 0.0.0.0:4155 [06:01] spiv: Will do. [06:04] bye guys [06:16] help? "bzr: ERROR: [Errno 14] Bad address" === jamesh__ is now known as jamesh [06:21] mtaylor: that's a new one [06:21] mwhudson: well that's not what I want to hear . :( [06:21] mtaylor: what were you doing? [06:21] mtaylor: which version? [06:21] mtaylor: and which OS? [06:22] mwhudson: I get it doing just about anything - bzr diff , bzr commit [06:22] spiv: opensolaris [06:22] mtaylor: I think this may be a bug that was recently fixed in bzr.dev [06:22] and version is... [06:22] mtaylor: ah, yep, then it probably is that bug. [06:22] spiv: 1.9 [06:22] ok [06:22] phew. I'll just update then [06:22] oh, the chdir('') one? [06:22] that makes sense [06:23] mwhudson: right [06:23] mtaylor: https://bugs.edge.launchpad.net/bzr/+bug/297831 if you're curious [06:23] Ubuntu bug 297831 in bzr "EADDR inside pyrex readdir on x86 Solaris" [Medium,Fix released] [06:23] mtaylor: a workaround would be to delete the _readdir_pyx.so, although that will impact performance a bit [06:24] * spiv heads off [06:24] nah. I'm running a pre-release version of opensolaris anyway - I might as well upgrade bzr to bzr.dev :) [06:24] spiv: thanks! [06:24] mtaylor: :) [06:24] * mtaylor stabs solaris in the face [07:11] hi all [07:12] (fullermd:) What's the deal with the new repo format, 1.9? [07:17] What about it? It has a faster and more space-efficient index format. [07:19] How would I use it? [07:20] Is there documentation on it somewhere? [07:23] If you don't need your repo to be compatible with bzr 1.8 or earlier, you'd run "bzr upgrade --1.9". [07:26] Peng_: And how is 1.9 better? [07:41] GPHemsley: Like I said, it uses a smaller and faster index format. [07:41] GPHemsley: If your project is pretty small, it probably won't make a huge difference. [07:56] hello, is there any installer that has "everything included" to install bzr with everything needed to be able to access ssh+bzr repositories from windows? [08:06] Can't the Windows installer do that? [08:06] Though you might want to install putty anyway. [08:07] * Peng_ leaves. Sorry, but I don't know much about Windows anyway. [08:09] well, I installed bzr and it tells me that ssh+bzr is illegal protocol [08:12] sounds like a serious bug in the installer [08:14] is it possible the reason to be that I previously had python installed? [08:19] well, I uninstalled python and installed a "fresh" copy... and it still can't connect to ssh+bzr repositories [08:41] I installed the latest version from 1.7 series. Now it doesn't tell that it's illegal protocol, but the connection times out [08:43] and there is no networking problem, because I can connect via putty without any problems [08:52] same with 1.8 series :( [09:01] OK, it's possible I'm doing it wrong [09:01] is bzr co bzr+ssh://user@some.address.com/some/directory/ valid syntax? [09:31] luks, any idea? [09:31] loxs[]: yes, it's a valid syntax [09:31] no idea about the problem though [09:32] it would be probably useful to file a bug report [09:32] yes, now I'm searching for a similar thing [09:32] the problem is that I can't provide any useful information :) [09:35] and I am not sure that something is not wrong with my settings [09:35] my windows I mean (i have no windows knowledge) === loxs[] is now known as loxs [11:51] poolie1: lifeless: My ~bzr membership is about to expire, the emails keep telling me to poke one of you about it. :) [12:33] What it means: Contents conflict in ? [12:33] How can I resolve it ? [12:34] ah bzr mv ... [12:35] hi everyone [12:35] matkor: "bzr help conflicts" might be helpful [12:35] i have to "split" a bazaar repository into three parts: a big repository should be split into three smaller ones (due to some changes in our company). is this possible? [13:20] skam: If your big repositories contains 3 independents projects, yes. You can create one repository by project, and pull all project branches in their specific repo [13:20] s/big repositories/big repository/ [13:38] im using ftp (vsftpd) to push, however .bzr dir umask is shown as "drwx------", and i cannot clone because of "bzr: ERROR: Transport error: Server refuses to fulfill the request (403 Forbidden)" message. === abentley1 is now known as abentley [15:38] hiredman, any plan to update bzrtools in jaunty ? bzr is stuck to 1.6.1 instead of 1.9 because of this. [15:38] hiredman, nm [15:38] i meant "hi," (damn xchat) [15:39] fta2: iirc James had requested a sync for bzrtools as well [15:40] nope, I forgot [15:40] looks like we need bzrtools and bzr-gtk, anything else? [15:41] bzr-builddeb is stuck too [15:41] jam: ping [15:41] fta: stuck? [15:42] well, i mean, upgrading bzr removes bzr-builddeb and bzrtools [15:42] ah, ok [15:42] that will take an upload, need to check that it works first though [15:44] thanks [16:03] is there a more recommended protocol than sftp these days? [16:04] bzr+ssh [16:06] does that work better over slightly higher latency links? [16:06] fresh branch checkouts are a bit painful over adsl. [16:06] bzr+ssh has 1/3rd the roundtrips of sftp [16:07] quicksilver: so you'll notice with high latency links [16:08] excellent. [16:19] Hmm, well bzr+ssh is making much more substantial use of my network connection, but it's not saying much while it does it. [16:19] no funky progress bar [16:19] yeah, progress reporting sucks a bit atm :-/ [16:20] sftp was 14 minutes to check out 604 revision, 70 megabyte repo [16:21] bzr+ssh was 2 minutes 40 seconds. [16:21] woot :) [16:23] LarstiQ: hi ! Where did you get that 1/3rd ? [16:23] vila: here from what lifeless has said, I suppose [16:25] ok, it's for bzr branch without shared repos I presume then [16:25] I think so, yes. [16:26] For a split second I thought you had implemented some effort tests to compare sftp vs bzr+ssh without telling us :-) [16:27] no no, sorry :) [16:27] * LarstiQ goes to his mom for dinner, ciao [16:34] How do I set up bzr for launchpad? [16:36] set up? [16:41] mmm [16:42] so what's the preferred repo layout [16:42] i was trying to do like, repo/project/trunk repo/project/branches, etc [16:43] what would the /project/ be... making it a branch doesn't seem to work [16:43] ie: branch within a branch [16:44] Kobaz: if you want then it can just be a plain directory [16:44] k [16:44] I would advise you make repo/ a plain directory [16:44] How do I set up SSH keys with Bazaar? It seems like it isn't detecting my key... [16:44] each /project/ a shared repository [16:45] yeah [16:45] then branches underneath that [16:50] Dvyjones: What OS? [16:50] Linux [16:50] Dvyjones: There is absolutely no magic involved. It just calls out to OpenSSH. If "ssh foo@bar" works, it will work with bzr. [16:51] So it's probably me specifying the wrong SSH key in launchpad, or it hasn't updated yet... [16:53] Probably, yeah. Or your SSH key suffers from the OpenSSL vulnerability. [16:53] Or you made some sort of mistake when pasting it into the box on LP. [16:53] It was generated just a couple of days ago on Ubuntu 8.120, so I guess not :P [16:55] heh, ok [16:55] I don't know what's wrong, then. Sorry. [16:56] * Dvyjones copied it again, and it worked [17:06] er [17:06] whoops [17:06] how do i do an unpush === beuno_ is now known as beuno [17:56] can someone explain me why chmod values are different in bzr push sftp:// and bzr push ftp:// ? [18:20] How do I configure who it says made the commit? [18:20] Right now it is using my hostname [18:20] cody-somerville, bzr whoami [18:21] thanks [18:25] I accidentally did some commits without having my whoami configured. is there a way to change the committer's names afterwards? [18:26] danser, only by redoing the commits [18:26] jelmer: ok, thanks [18:29] Unless your default username is really embarrassing, it's probably best to just live with the mistake. [18:30] it's only about two recent commits, and I'm still the only author, so I don't mind redoing these :-) [19:00] is it possible to encode a password in a location? or noninteractively at the commandline at all? [19:00] Tak, yes, "bzr://username:password@hostname/path" [19:02] hmm - I get: ERROR: Invalid url supplied to transport: "invalid port number (part of password) in url [19:02] Maybe supply a port too? [19:02] I am, actually [19:05] the entire url looks like: sftp://user:pass@host:port/path/ [19:49] do I need to urlescape the password? (the url is already in single-quotes for the shell) [20:43] i'm thankful for the vcs-imports team [20:44] but i have a question [20:44] i'm trying to branch this: https://code.launchpad.net/~vcs-imports/emc/trunk [20:45] if i branch it into its own, new repo (ie, i run bzr branch in a directory which is not initialized) it works nice & fast [20:45] is there anything special i have to do to push files from olive to a website through ftp? [20:45] but if i try to branch into an existing shared repo, with some other locally-created branches, it's awfully slow [20:46] it says "Transferring 0/4", then stops & bzr pegs the CPU, and it just sits there [20:46] seb_kuzminsky: your repository is probably in a different format and it has to convert it [20:46] ah [20:46] try bzr info on both [20:46] new branch of lp: pack-0.92 [20:47] branch into existing repo: rich-root-pack [20:47] thanks luks! [20:48] i'm a bit confused by the different repo formats [20:48] you are not the only one :) [20:48] i guess i should stay on the default pack-0.92, but it kept bugging me to "upgrade my repo" [20:49] :-/ [20:49] hi. i just started using bzr through olive - i am trying to push some test files to my server through ftp, i dont get an error, but nothing happens === james_w` is now known as james_w [21:00] I grabbed bzr-git from launch bad, I see "TypeError: 'Tree' object is not iterable" when I try to branch from a git repo [21:02] * mwhudson points at jelmer [21:02] hiredman: what version of bazaar? [21:03] 1.9 [21:04] hmm [21:04] python 2.5 [21:04] hiredman: are there bug reports for bzr-git? [21:10] http://gist.github.com/29575 is the output === mthaddon_ is now known as mthaddon [21:10] hi [21:10] I just got bit by a UI bug in loom. [21:10] but it's a pretty subtle one. [21:11] hiredman: hm, fairly mysterious [21:12] I was submitting threads to a PQM via exported branches, I ended up having one of the threads rejected because of a text conflict caused by one of my lower threads. [21:12] :( [21:13] I feel that loom should have warned me about this. [21:13] * james_w waves to jml [21:13] james_w: hello. [21:13] mwhudson: thanks again for loggerhead, just set it up on a site and it was painless and works beautifully [21:13] james_w: glad it worked for you! [21:14] jml: presumably they didn't conflict before you started exporting? === thunderstruck is now known as gnomefreak [21:14] james_w: they did! [21:14] james_w: the interesting thing here is that the lower thread had commits that were not yet merged into the upper threads. [21:14] ah, that'll do it [21:15] so you're able to export threads from above your current position in the stack? [21:15] thing is, the only way to find that out with loom is to go to the bottom thread and merge all the way up. [21:15] james_w: I think export-loom just does the whole stack -- not sure though. [21:16] james_w: I might have switched to a higher thread also, again not sure. [21:16] ah, I haven't used export-loom, I see that would be a problem [21:16] what warning would you have liked? [21:16] "You are exporting a loom that hasn't been fully merged. The resulting branches may conflict"? [21:17] james_w: I guess what I would like is something akin to "bzr status" that shows me if any threads have revisions that haven't been merged into the higher threads [21:17] and also what's changed since the last record (hah!) [21:18] it sounds like they are things that bzr-loom should provide [21:18] representing it obviously won't be easy, as you indicate [21:18] yeah. [21:38] Hello people [21:38] * jml shakes his fist at combine-thread [21:38] I was wondering if there was an semantical equivalent of "svn mv" something out of a branch into a new one. [21:39] That is, I have a project that has grown way too large, and I'd like to refactor a sub-component of it as a sub-project, with its own set of branches. [21:39] Is there a way to grab a sub-tree of a branch and turn that into a branch, or branch a sub-tree of a branch. Or is the infamous "move the file manually into a new branch" the way to go. [21:40] anyone know how I can get info from the last record? [21:57] beuno, hi [21:58] mwhudson, hi [21:58] jelmer: some bzr-git queries from hiredman [21:59] hiredman, pong [21:59] mwhudson, we (and by that I mean Lo-Lan-Do) are working on also getting loggerhead up and running on alioth [21:59] jelmer: nice [22:03] jelmer: cool [22:06] jelmer: does the version of loggerhead you are using include the serve-branches script? Is that what you are using? [22:07] james_w, no, we're the standard loggerhead package - which uses start-loggerhead [22:07] using serve-branches can display the hierarchy nicely [22:08] going by one of the bug reports it sounds like that is what would be preferred [22:09] ah, wasn't aware they're actually doing different things these days [22:10] serve-branches -- new, perhaps a little dwimmish, liked\ [22:10] start-loggerhead -- old, crappy, i personally hate it [22:11] I'm not sure where the extra port is coming from, do you have the public URL setting in loggerhead.conf? [22:22] mwhudson, but serve-branches doesn't use loggerhead.conf, does it? [22:22] jelmer: no, you have to use --style options [22:23] it would be nice if there was a way to just get the stylistic improvements by setting some variable in lgogerhead.conf [22:24] you mean the serving from directories stuff? [22:24] because there isn't any real difference in appearance, i thought [22:25] I'd like to get the hierarchy that serve-branches has [22:25] what else can't you get with serve-branches? [22:26] That's the only bit I'm aware of that's differnet (but I may not be very well informed) [22:26] then why not use serve-branches? [22:27] i realize you can't use loggerhead.conf so neatly, but a little shell script with --port this and --host that doesn't seem so hard [22:28] mwhudson, this is for the debian package, which already uses start-loggerhead [22:29] jelmer: now you've lost me [22:29] which part of the debian package? [22:29] the init script [22:30] oh right\ [22:30] did you see that trunk loggerhead has one of them too now? [22:30] the alternative is for me to allow the user to configure a choice between start-loggerhead and serve-branches [22:31] and to confuse users since their configuration for address/port to listen on is in different places [22:31] ah [22:31] the loggerhead.conf suits a package really well, and so having an option to turn on the hierarchical browsing would be great IMO [22:32] i'm fine with the idea of configuring serve-branches from a config file [22:32] i'm not really fine with it being the same as loggerhead.conf currently is though [22:34] or [22:34] nuke loggerhead.conf completely [22:35] and let a /etc/defaults/loggerhead set variables [22:35] (in the packaging) [22:38] mwhudson, serve-branches doesn't necessarily have to support a config file - it can just be a tool to quickly get a loggerhead up [22:38] the important thing (for me at least) is that the same functionality is in start-loggerhead [22:41] Looks like Dragonfly finished their git conversion. [22:54] jelmer: i don't see why you so want start-loggerhead [22:57] mwhudson, it means I don't have to worry about the configuration (since except for enabling it, it is all outside of /etc/default), because it can fork for me as well as log etc, and because it's already there (otherwise I'd have to make sure there's a smooth transition from start-loggerhead to serve-branches) [22:58] jelmer: well if you'd _asked_ me when you were packaging loggerhead, i'd have told you to stay away from it ;-p [22:58] jelmer: when serve-branches is feature compatible, we can do a compatibility shim somehow [22:59] mwhudson: or, as some of the features are insane [22:59] mwhudson: at that time, serve-branches wasn't even installed by setup.py IIRC.. [22:59] mwhudson: you could just deprecate it [22:59] there's always going to be a spot of friction though, because the way the branches are specified [22:59] +of === jfroy|work is now known as jfroy === jfroy is now known as jfroy|work [23:43] jelmer: ping [23:43] hiredman, pong [23:43] Yo [23:47] hiredman, you were hitting issues in bzr-git [23:48] Yes. [23:49] I was not able to branch from a git tree, or whatever you call it [23:50] I am not even sure how complete bzr-git is suposed to be [23:50] http://gist.github.com/29575 was the output