| harmw | hm, isn't there something like a codestandard on readthedocs..? I know of some openstack projects that have just that | 08:08 |
|---|---|---|
| smoser | philips, it doesn't actually matter there. | 12:40 |
| smoser | the '-' get converted to '_'. | 12:40 |
| smoser | harmw, suck. | 12:41 |
| smoser | i don't know what to do about that really. i like that there is a version string in the image, but I really wish that | 12:42 |
| smoser | a.) i didn't screw that up | 12:42 |
| smoser | b.) i could properly "promote" a release | 12:42 |
| smoser | (ie, rename) | 12:42 |
| smoser | build-release doens't do anything that shouldn't be doable in a ci build | 12:43 |
| smoser | harmw, the '~pre1' is by design (versus '-pre1'). '~' in dpkg means "less than". | 12:47 |
| harmw | uhm, whats wrong with committing a change in bin/build-release since that's where you keep track of version numbering :) | 13:04 |
| harmw | and using the ~ annoyed me, I wasn't aware of a dpkg design thingy | 13:05 |
| harmw | but whats wrong with using - instead of ~? 0.3.2-pre1 would still be the same as 0.3.2~pre1, right? | 13:06 |
| smoser | harmw, 0.3.2-pre1 is > 0.3.2 | 13:30 |
| smoser | $ dpkg --compare-versions 0.3.2-pre1 gt 0.3.2 && echo yes, greater || echo no | 13:31 |
| smoser | yes, greater | 13:31 |
| harmw | ah like that | 13:31 |
| smoser | the issue with committing a change in bin/build-release is that then you cant feed it a tag as input. | 13:32 |
| smoser | the other issue is that then 0.3.2~pre4 can't just be "promoted" to 0.3.2 | 13:33 |
| smoser | as it has that yucky string in it. | 13:33 |
| smoser | i don tknow. its over thinking things i know. | 13:33 |
| harmw | can't we just go from 0.3.2rc1 to 0.3.2? | 13:37 |
| smoser | what do you mean? | 13:37 |
| harmw | nah, just thinking out loud | 13:38 |
| smoser | harmw, i'm ok to punt on that, and just change that one thing and call it 0.3.2. | 13:44 |
| smoser | assuming my kernel change didn't cause anything else | 13:44 |
| harmw | it builds just fine, I haven't tested the new image though | 13:45 |
| harmw | planning to do just that tonight | 13:45 |
| smoser | crap. | 13:53 |
| smoser | /dev/null has bad perms in the tarball. | 13:53 |
| smoser | and disk image | 13:55 |
| harmw | and you only updated the kernel :p stupid kernel | 14:04 |
| smoser | ? | 14:04 |
| smoser | oh. no. it was a bug before. | 14:04 |
| smoser | fix in hand. | 14:04 |
| smoser | and it explains why running 'cirros-status' as root would not get you network debug info | 14:05 |
| harmw | hm, so my images are broken as well? | 14:05 |
| smoser | right | 14:05 |
| harmw | (since I'm not using bin/bla) | 14:05 |
| smoser | fixed and pushed. | 14:06 |
| harmw | nice, now I only need to wait till midnight :p | 14:06 |
| smoser | you weren't using bin/bla ? | 14:06 |
| smoser | at all ? | 14:06 |
| smoser | you probably were. | 14:06 |
| harmw | haha, then I'm not going to get the fun out of this commit | 14:07 |
| harmw | I was using the steps you describe in README | 14:07 |
| smoser | you were. xmakedeves gets called from | 14:07 |
| smoser | * build disk images using bin/bundle | 14:07 |
| smoser | $ sudo ./bin/bundle -v output/$ARCH/rootfs.tar download/kernel-$ARCH.deb output/$ARCH/images | 14:07 |
| harmw | bin/bundle, yes | 14:08 |
| harmw | not bin/build-release | 14:08 |
| smoser | right. build-release just does what is in the README | 14:08 |
| smoser | its proably almost identical to yours | 14:08 |
| harmw | indeed it is :) | 14:08 |
| harmw | (though mine doesn't upload to LP ) | 14:08 |
| smoser | it doesn't upload | 14:09 |
| smoser | where did you think it did that ? | 14:09 |
| smoser | (actually, theres a bug that *i*haven't uploaded to launchpad :) | 14:09 |
| harmw | hm well I thought to have skimmed over some bits that pushed something to LP | 14:15 |
| * harmw should probably just stop thinking every know and then | 14:15 | |
| === harlowja_away is now known as harlowja | ||
| harmw | smoser: I see 0.3.2 is tagged as such? | 17:50 |
| harmw | I'm auto-building in about 10 minutes | 17:50 |
| smoser | no tag | 17:51 |
| harmw | just ned to think of something to only build a new buildroot when it is really -really- neded | 17:51 |
| smoser | i'mo building trunk right now. | 17:52 |
| smoser | i'll put it on download and just not call it | 17:52 |
| smoser | but can point you at it | 17:52 |
| harmw | fair enough :) | 17:53 |
| smoser | its built i386, amd64 and is 10 minutes into arm | 17:53 |
| smoser | i suspect 20 minutes it'll finish. | 17:54 |
| harmw | ah yes, the other archs - I'm still only building x86_64 | 17:54 |
| harmw | btw why is make br-source done witj i386 as arch? | 17:54 |
| smoser | just because | 17:54 |
| harmw | hmk, and its wrong to have it build with $ARCH ? | 17:55 |
| harmw | i386/arm/x86_64 ? | 17:55 |
| smoser | probably, yeah. they do have different packages though i think. | 17:56 |
| smoser | i forget. | 17:57 |
| harmw | well having it build against x86_64 went great, haven't tried it on arm/i386 though | 17:57 |
| smoser | oh. right. it iwll work. | 18:00 |
| smoser | and it i think does it anyway. | 18:00 |
| smoser | the explicit step is just to have the download done | 18:00 |
| smoser | so that it hits cache | 18:00 |
| harmw | ah lol | 18:01 |
| harmw | well its caching those downloads anyway | 18:01 |
| harmw | any thoughts on this btw https://bugs.launchpad.net/cirros/+bug/1292973 ? | 18:01 |
| harmw | harlowja: news from the bsd front? | 18:07 |
| harlowja | harmw ah, let me see if i can get sean in, didn't chat last week | 18:07 |
| harlowja | pinging now | 18:07 |
| harmw | hm, sucks my server hasn't got enough firepower left to host a jenkins instance for c-i nightlies | 18:08 |
| smoser | harmw, is that busted in current buildroot ? | 18:09 |
| smoser | we should forward there if it is | 18:09 |
| harmw | yea, it regards the 2012.05 buildroot | 18:09 |
| harmw | or just upgrde buildroot | 18:09 |
| harmw | which i've already done in some branch | 18:10 |
| harlowja | harmw no response ping from sean, i'll let u know | 18:15 |
| harlowja | will get him in here or bug him about whats happening/status | 18:15 |
| smoser | http://paste.ubuntu.com/7109590/ | 18:16 |
| smoser | harmw, ^ . its odd. i built on an azure instance to a tmpfs, thkning it would go crazy faster. | 18:17 |
| smoser | but it did not. | 18:17 |
| smoser | harmw, http://download.cirros-cloud.net/pre-release/0.3.2/ | 18:19 |
| harmw | tmpfs, nice, I was thinking of going that way aswell | 18:29 |
| harmw | my buildroot is 2G, thats gonna be a close call... | 18:30 |
| smoser | well, 7G was within a cuple hundred meg. | 18:31 |
| smoser | but that included everything. | 18:31 |
| harmw | 2.1GB, without the downloads folder | 18:32 |
| harmw | oh and its still compiling | 18:32 |
| harmw | ah no, just finishing up | 18:33 |
| harmw | took 33 minutes | 18:33 |
| harmw | http://buildservice.weites.com/cirros/trunk/ | 18:34 |
| harmw | nice :> | 18:34 |
| harmw | smoser: do you have reallife results from ARM-based cirros boots? | 18:35 |
| smoser | no. other than lxc. | 18:35 |
| harmw | cool, I'm expecting a cubietruck which has hardware virt stuff | 18:36 |
| smoser | oh nice. | 18:36 |
| smoser | canonical has some such hardware, and its possible people here have used it. | 18:37 |
| smoser | but i'm not sure actually. i really did it for lxc. but i'd like to have it known functional on som eactual hardware | 18:37 |
| smoser | and at some point would want arm64 stuff too | 18:37 |
| harmw | it should arrive next week or so | 18:38 |
| harmw | and then Ill load it up as nova-compute node | 18:38 |
| === harlowja is now known as harlowja_away | ||
| === harlowja_away is now known as harlowja | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!