=== Logan_ is now known as Guest85841 [13:13] is it possible to pull in bzr without updating the working tree? [13:35] yes [13:35] but in practice you want a shared repo, then [13:45] bob2, what's the command? [13:45] as above [13:45] you want a shared repo and bzr pull [13:45] and yeah, ot's just a hypothetical question [13:45] :) [13:45] oh [13:45] bob2, so with a single working directory & branch it's not possible eh? [13:48] didn't say that [13:48] but it hardly seems worth explaining for a hypothetical === wedgwood_away is now known as wedgwood [13:57] bob2, is it complicated heh? [13:57] ok! === wedgwood is now known as wedgwood_away [14:56] bob2, hmm i think you misread one or more of my messages [15:44] Noldorin: Repository.fetch() [15:45] Noldorin: not sure if we expose that on the command-line directly anywhere [15:45] Noldorin: (or why we should) [15:59] HI, I am getting this error: can anybody help?? [15:59] Connection error: Couldn't resolve host 'xmlrpc.launchpad.net' [Errno -2] Name or service not known [16:23] jelmer, ah right, i see [16:23] fair enough then === mmrazik is now known as mmrazik|afk [18:05] Noldorin: you can also do `bzr remove-tree` first ;) [19:41] LarstiQ: 'first'? === deryck_ is now known as deryck [21:42] bzr is taking hours to branch a ~1GB repo on a machine with 512MB ram :( [21:44] it appears to be swapping constantly (physical memory is full) [22:35] tuvwx: if you're cloning over http from LP, don't :) [22:37] lifeless: i'm using the launchpad-suggested command at the top of the 'code' tab: bzr branch lp:... [22:38] tuvwx: are you logged in? that makes a significant difference since that will make it switch from http to ssh [22:42] no, i'm not logged in. does bzr use ssh differently from http? [22:46] tuvwx: yes, significantly. bzr uses a custom (bzr-specific/optimized) protocol over ssh. over http it just does regular file access [22:47] but why would "regular file access" consume that much memory? [23:00] tuvwx: what version of bzr are you running? [23:00] jelmer: 2.6.0dev2 from debian wheezy [23:04] okay, that should have all the performance improvements [23:54] tuvwx: because if you thrash the internal page cache, it regets content over http [23:55] tuvwx: and each retrieval has approx 2x overhead in RAM, and unlike the os page cache isn't shared. [23:55] tuvwx: I suspect you're branching something like mysql [23:55] tuvwx: or java, and some of these have very large btrees which result in inventory cache thrashing [23:55] tuvwx: there is a bug on it I believe.