[05:48] so, the version of bazaar in hardy is 1.3.0 [05:48] how do I make it use 2.0.x [05:48] ? [05:49] that or how do I cause the windows shell shortcuts use the the protocol 1.3 requires instead of the 2.0 version [05:49] ? [06:12] add the bzr ppa on hardy [06:13] bzr ppa on hardy contains 1.3.x [06:14] no it doesn't [06:14] see the bzr downloads page for instructions [06:15] oh, weak. [06:15] I forgot to apt-get update :P [07:44] happy easter [07:44] anyone here? :D [08:02] so, can anyone familiar with TortoiseBZR tell me how to make it login to a remote respository as a different user than the one that is logged in on windows? I would like to log in as a user on the linux box, but not as the user on the windows box. [08:02] using bzr+ssh:// [08:15] Infl8ableSoulm8: Tried bzr+ssh://user@host/path? [08:34] yeah [08:35] I mean that by default it is logging in as windows-user@boomheadshot.org instead of matt@boomheasdhot.org [08:36] but it doesn't seem to really matter at this point. [08:36] every time I commit something as one user, I can't update as a different user [08:36] the files are created as being owned by user1:usr1 so I have to chown -R the whole repository again [08:44] is it possible to set things up so all committed files are set 660 *:bzr ? [08:49] oh, I see... [08:49] user@ [08:49] hrm. [08:50] that would work, but it would still require that the permissions on new files be set correctly. [09:12] hrm... looks like sticky chmod is the winner :/ [09:23] Does Bazaar or ancilliary tools have support for keeping track of un-cherry picks (the revisions you do not want)? Last I knew, it did not. Does any DVCS? Vague neural impressions says, "Darcs might." [09:24] As I recall, you have to respecify the cherries or uncherries every time you merge with bzr, which is manually tedious. [09:30] No, it's not sticky. It's setgid. [09:30] * fullermd does the whole rant again. [09:31] I called it 'sticky' because I found an email that referred to it that way [09:31] edakiri: Well, you can't have a revision without having all of its ancestors. So, it's always THERE in the history. But you can commit a reversal of the changes, and that will propogate forward fine. [09:31] (of course, if you then merge that branch into a branch where you WANT those changes, things get uglier) [09:31] I have no actual idea what it was talking about. [09:31] Well, find the author of that email and kill them. [09:32] Well, OK, maybe that's a LITTLE over the top. Take off at least 3 limbs, anyway. [09:33] well, it actually was a chmod command. [09:33] something like chmod 6774 [09:34] I am only used to using the three digit commands. [09:34] 0___4 on a dir would be bizaare. Presumably it's more like 02775 or 02770. [09:34] You certainly don't need 04000 in there. [09:35] well, if I had any clue what I was doing, I'm sure I would do it correctly. [09:35] and it was a recursive command [09:36] On filesystems with SysV semantics, setgid on a directory causes files within that directory to inherit the dir's group. [09:36] (under BSD semantics, that's the default behavior) [09:36] setuid doesn't do anything in general. [09:37] You certainly do NOT want setuid/setgid on any of the _files_. [09:39] well, it's letting me commit and update as different users without throwing errors, now. [09:54] fullermd: The case which turns ugly is exactly what the main thing desired for the D in DVCS: maintaining multiple branches which desire to keep different code variations. So it is this case i am seeking a way to solve. [09:57] darcs is the only thing that might work for that, since it treats each change as an independent entity floating in space. [10:01] vila: File "/usr/lib/python2.6/dist-packages/bzrlib/transport/http/__init__.py", line 132, in get [10:01] return StringIO(response_file.read()) [10:03] vila: thats urllib, and its buffering [10:23] lifeless: grr, yeah, someone put a FIXME there, in anger I presume, that means you need to override get(). You don't need to seek backwards right ? [10:27] vila: no, I have a 4GB+ stream to deal with ;) [10:27] vila: I'm calling _get for now. [10:27] vila: this will be breaking http smart server initial clones [10:27] vila: by breaking I mean 'this is why we are running out of memory' [10:28] lifeless: bzr is supposed to use readv() for packs so should not be affected [10:29] lifeless: or are you saying we *do* use get() ?? [10:30] lifeless: I have to go *now*, be back in a couple of hours [10:30] lifeless: as hinted in the FIXME this *can* be fixed, bundle reader just needs to keep track of whatever it reads and give it back when needed instead of seeking backwards [10:31] vila: not packs [10:31] bzr+http [10:31] vila: actually, I think that case is clean [10:32] bye [12:12] vila: back yet ? [12:14] lifeless: wow, back ~3 minutes ago :) [12:15] are you aware of any 32 bit limitations in python socket/urllib on 64bit Ubuntu ? [12:16] no, and babune, as well as most its slave run with 64 bits OSes. I don't know how python is built there though [12:17] lifeless: any specific case ? [12:19] http://pastebin.ca/1856482 [12:20] I hit ctrl-C when I was seeing no network traffic, and no error, on my test lmirror server [12:21] thats the traceback ont he client [12:22] I'm not saying that bzrlib.transport is necessarily buggy... but something strange happened *somewhere* [12:26] lifeless: weird, self._rbuf seems to be a StringIO() buffer, I fail to see how that could block, could it be that you were *very* unlucky on yout ctrl-C ? [12:26] vila: well,i ts more that I'm wondering why my server would suddenly halt [12:26] a 32 bit issue would explain it, if something somwhere has a counter and it wraps at 2G [12:27] or even at 4G [12:27] lifeless: it's an extrenal server right, not one in the same process as your http client ? [12:27] the stream is - let me just wget it [12:27] 4126895989 bytes long [12:28] du -s /tmp/t/music/ [12:28] 2099976 /tmp/t/music/ [12:28] IIRC bt.http.response will read by 500K chunks [12:29] is the amount writen to the output dir ont he machine I'm streamin to [12:34] lifeless: 2099976 is not a multiple of 512 * 1024, weird [12:34] its stuck, same place [12:35] client side si looping: [12:35] recvfrom(4, "", 524288, 0, NULL, NULL) = 0 [12:35] 100% CPU use :( [12:35] lunch time here, [12:35] you mean it loops *trying* to read but getting nothing ? [12:36] right [12:36] or [12:36] its looping with a zero length buffer [12:36] I'm going to check the params now [12:37] at least 524288 is 512 * 1024, so we are indeed respecting that [12:38] why deoes 512*1024 matter ? [12:38] oh, so I *know* that the server can send a 4G stream [12:39] because wget can get it all. [12:39] man recvfrom [12:39] 524288 would appear to be the length we're willing to receive [12:39] so that call is ok [12:41] yes, that bt.http.response imposing that 512K to avoid memory errors [12:41] s/that/this/ [12:46] I'm running it a gain with an assert in osutils.pumpfile [12:46] just incase its gotten a zero-length read or something silly organised [12:55] vila: http://mail.python.org/pipermail/python-bugs-list/2005-January/027258.html is int he sapce [12:58] hah! I think I know. [12:58] 32-bit squid build, I think [12:59] yeah [13:00] yup [13:04] gnight [13:05] lifeless: gnight, so the problem is solved for you ? [13:05] very sure it will be [13:05] ok [13:06] the client was using a 320bit squid build without -DLARGEFILE wor wahtever [13:06] I'll look into it more tomorrowish [13:12] http://tinypic.zapto.org/2kn4m8.png?t=1270382681 do my breasts look to big? [13:13] It must be hard to have a 320 bit build without LARGEFILEs... [13:13] Annaa: dunno about your breats, but your brain, definitely no [13:13] fullermd: naah, he emant really LARGE, like Annaa's [13:16] * fullermd keeps waiting for you to type 'Anaaa' [13:16] bananaaaaa [13:18] Is that just a typo, or are you a sheep? 'cuz I've been thinking about making a new sweater... [13:25] I was thinking about some string search algorithm that mentioned banana as a useful test case :) [15:15] banana-rama mo-mana fee-fi fo fana, banaaana [15:15] morning vila and fullermd [16:46] so... random question, using bzr-svn, now trying to merge the work back (quite a few moves, bunch of modifications)... bzr-svn however seems to just be attempting to do a remove/add instead of a move [16:47] bit non optimal since history will be lost [16:47] any ideas? === radoe_ is now known as radoe [19:28] curious, how would one go about pushing revs/history of a bzr branch into svn when there is no common ancestor? [20:10] ferringb: uh, carefully ;-) [20:10] you may wish to try it with a local toy SVN repo first [20:19] SamB_XP: or toy branches [20:19] * ferringb figured it out [20:19] basically push to an empty location in svn [20:20] trickier was how to get revs from a bzr branch directly into svn w/ the history fully merged in (hacked that one, but it worked) [23:07] Hi [23:07] Guys, do you know a place where I can have a private bazaar repository for free ? [23:07] (I need this temporary actually :)) [23:11] any server you can sftp or ssh into [23:16] (or access as a filesystem -- say, a flash drive?) [23:22] I will give you a free, private bzr repository for $100! [23:22] oh. [23:23] I think he meant money-free rather than free-of-restrictions [23:23] right. [23:24] besides the fact that I'm a newb and no one would want to have their stuff hosted on my box to begin with :D [23:24] a Linux newb ? [23:24] not so much a linux newb. [23:24] I am pretty good with linux for desktop purposes. [23:25] goundy: what organizations are you affiliated with? [23:25] any? [23:25] but all this security and permissions stuff that comes with running a server is kinda over my head at the moment. [23:25] SamB_XP, nope... not yet actually that's why I'm looking for a quicky :°) [23:25] I thought maybe you had a school or something [23:25] I will give you a free, private bzr repository for $100!> Ohooo generous :p [23:26] SamB_XP, I do have one, but my code isn't targeted to it [23:26] well, they often have CS departments, and those often have some way of getting a shell account... [23:28] another option would be to use one of those computers your mother wishes you'd just throw away as a *nix server [23:35] See, I thought about trying that once. But getting *nix running on an Apple ][ is a bit of a challenge. [23:36] SamB_XP, yeah, unfortunately the only machine my mum wishes I throw away is yet down :p [23:36] fullermd: lol [23:37] fullermd: I'm thinking college was a few years back for you, yes? [23:38] It wasn't THAT long ago! It was just... umm... OK, shaddup. [23:38] it could just be that their CS dept is that far back. [23:38] She does still wish I'd throw away the Apples, though! [23:42] it's only that I don't think people young enough to still be in college were old enough to have any control when any Apple IIs the family may have had would probably have been thrown out ;-) [23:42] Pah. Slackers. [23:43] whu? [23:43] are you implying that more people your age should be getting PhDs or something ? [23:43] Actually, I acquired most of them from a school that was throwing them out (many, MANY years after they should've) [23:43] ah [23:44] Most of the people I went to school with who intended to get PhD's have them by now. [23:44] okay, that may throw off my reckoning ? [23:44] (and the remainder probably won't get them anytime soon)