[08:11] <slangasek> cjwatson: livecd-rootfs 1.137 - thanks and argh; set -u drives me batty
[08:32] <cjwatson> slangasek: likewise
[11:45] <ogra> lamont, do you see any issues (buildd wise) with that livecd-rootfs change ? http://paste.ubuntu.com/466935/
[11:53] <lamont> ogra: cleanup() needs to deal with making sure it's unmounted, if we pass that way
[11:53] <lamont> as in, think about when the script fails...
[11:54] <ogra> oh, yeah, indeed, do you see any probs beyond that ? i.e. loop mounting not being supported or some such ?
[11:54] <lamont> nothing else hits me in the face, no.
[11:55] <lamont> wut? you can do count=0 and have it do something useful?  COOL
[11:55] <ogra> yup :)
[11:55] <ogra> its extremely fast too
[11:55] <lamont> and here I'd been doing all this work to be able to say count=1
[11:56] <ogra> and doesnt really us space on the filesystem until you write to it
[11:56] <lamont> this is most cool
[11:56] <ogra> *use
[11:56] <lamont> yah
[11:56]  * ogra hopes to shove off 1-2h from the current build time with that change
[11:57] <lamont> genext2fs is that slow?
[11:57] <ogra> 3h per image is simply wrong
[11:57] <ogra> 1h for the chroot installation, 2h for genext2fs
[11:57] <lamont> that will be nice
[11:57] <lamont> bbiab
[11:57] <ogra> the prob is that it allocs the whole imagesize in ram before adding the image content
[11:58] <ogra> which means endless swapping
[11:58] <lamont> wtf? AAAAAAAAAAAAAAAAAAaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
[11:58] <ogra> its relatively fast on my 4G laptop ... but on the babbages it just sucks
[11:58] <lamont> big time
[13:00] <ogra> lamont, http://paste.ubuntu.com/466965/ ... better ?
[13:01] <ogra> (seems DEV was existing but not used by anything)
[14:03] <lamont> looks good
[14:11] <ogra> great, committing
[14:29] <lamont> ogra: if I wanted to steal acorn or sycamore for a bit, which one should I steal?
[14:30] <lamont> meh. nm - I'll just abuse kakadu
[14:31] <ogra> lamont, better talk to slangasek, i have nothing thats really urgent but he is rolling test images atm
[14:31] <lamont> that's what I'd be stealing it for :-D
[14:31] <ogra> ah
[14:31] <lamont> but kakadu will do fine
[14:31] <ogra> well, then any would be ok
[14:31] <lamont> is your change built yet?
[14:32] <ogra> not sure
[14:32]  * ogra checks
[14:32] <ogra> built isnt the issue though, mirrored to ports.u.c takes usually long
[14:32] <ogra> built
[14:32] <lamont> published?  (which rev is this?)
[14:33] <ogra> 1.139
[14:33] <ogra> LP says Accepted
[14:34] <lamont> ok.  actually going to use sycamore, to avoid the competition
[14:34] <ogra> http://ports.ubuntu.com/ubuntu-ports/pool/main/l/livecd-rootfs/
[14:34] <ogra> not mirrored yet
[14:36] <lamont> not published yet, more to the point
[14:36]  * lamont doesn't care about ports.u.c's copy
[14:36] <ogra> yeh
[14:37]  * lamont has manually taken the livecd builder lock on sycamore.
[14:37] <ogra> k
[14:38] <ogra> i'll restrict myself to acorn for test builds
[14:38]  * ogra just wants to do a timed build to see the gain
[14:39] <lamont> I believe your gain, and wish to reap its benefit
[14:40] <ogra> yeah, i just want to know if its faster than squashfs
[14:40] <ogra> since we only cp now
[14:40] <ogra> but its still 1.5G
[14:40] <lamont> I almost asked why you didn't rsync when I saw that cp...
[14:41] <ogra> heh, yeah, i could have done that too indeed
[14:41] <lamont> though cp is prolly slightly faster, given the virgin target
[14:41] <ogra> though i dont think it would speed up anything
[14:41] <lamont> wouldn't do a thing
[14:42] <ogra> yep