[12:36] <mbiebl> Keybuk: do you know,  in what version of glibc, the WEXITED flag for waitid was introduced?
[12:36] <mbiebl> Tried to build upstart on sarge, and it failed miserably, because of that ;-)
[12:36] <Keybuk> mbiebl: oh, almost certainly very recently
[12:37] <Keybuk> waitid itself was only introduced in 2.6.9
[12:37] <Keybuk> (and is broken on compat arches in all kernel releases until the next one <g>)
[12:38] <mbiebl> until next one means 2.6.19?
[12:38] <Keybuk> 2.6.20 maybe
[12:38] <mbiebl> oh
[12:38] <Keybuk> (upstart works around it though
[12:39] <yankees26> how?
[12:39] <Keybuk> how works around, or how broken?
[12:39] <yankees26> how it works around it?
[12:39] <Keybuk> it clears the siginfo_t structure before passing it to waitid
[12:39] <Keybuk> the bug is that compat_sys_waitid() in the kernel doesn't copy zero values
[12:39] <yankees26> ah, so shouldn't it build for mbiebl in that case?
[12:40] <Keybuk> yes
[12:40] <Keybuk> dunno what the WEXITED thing might be, just looking that up atm
[12:40] <yankees26> ah
[12:40] <Keybuk> it's entirely possible that it's simply missing on really old libc
[12:40] <Keybuk> seeing as the syscall itself is recent
[12:41] <yankees26> mbiebl: what version of glibc do you have?
[12:42] <Keybuk> sarge doesn't have 2.6 as default, no?
[12:43] <Keybuk> and only has 2.6.8 as an option
[12:43] <Keybuk> so upstart won't work with that
[12:43] <yankees26> ah well im in luck 'cause i dont use debian
[12:44] <yankees26> of course i first have to get through the majority of LFS :P
[12:44] <mbiebl> yankees26: sarge has 2.3.2, current unstable (which I run) 2.3.6
[12:45] <yankees26> ah
[12:45] <yankees26> well by the time i get to LFS glibc 2.5 will probably be out of testing in arch (currently on 2.4) so
[12:45] <yankees26> hopefully i wont run into your problem
[12:47] <Keybuk> upstart fully supports glibc 2.4
[12:47] <Keybuk> glibc2.4 and kernel 2.6.17 were the recommended targets
[12:47] <yankees26> i run both of those :D
[12:48] <mbiebl> Well, it never hurts to support a broader audience
[12:48] <Keybuk> mbiebl: actually, it hurts a little bit
[12:49] <mbiebl> Sure, for developing.
[12:49] <mbiebl> But I rather meant acceptance.
[12:50] <mbiebl> Keybuk: you said you work around the waitid issue, does that mean upstart works on a 2.4 kernel?
[12:50] <Keybuk> no
[12:50] <Keybuk> upstart only works around the mis-behaviour of compat_sys_waitid
[12:50] <Keybuk> while I sit on Kyle's lap until he fixes it
[12:51] <Keybuk> it doesn't have any alternative for waitid if that syscall doesn't exist
[12:51] <yankees26> Kyle is?
[12:52] <_ion> keybuk: Are new conclusions related to upstart that were reached at uds-mtv documented somewhere?
[12:52] <mbiebl> Keybuk: what exact feature of waitid do you need, which couldn't be implemented otherwise?
[12:52] <Keybuk> mbiebl: the fact it can obtain information about a dead process without actually reaping it
[12:52] <Keybuk> so we can find out things like it's binary name, cpu time, etc.
[12:53] <mbiebl> ah, ok
[12:53] <Keybuk> _ion: not yet, got to finish writing them up
[12:53] <Keybuk> mbiebl: that will be necessary for supporting supervision of processes that fork
[12:53] <yankees26> does uds-mtv stand for ubuntu developer summit mountain view?
[12:53] <Keybuk> yankees26: yes
[12:53] <yankees26> ok, just making sure (im not intune with ubuntu development stuff)
[12:54] <mbiebl> Keybuk: maybe these requirements should be documented somewhere.
[12:55] <Keybuk> mbiebl: they'll be in the README
[12:56] <mbiebl> good
[01:08] <yankees26> hey guys (slightly off-topic): to have a project and bazaar in launchpad, does the project need to be affiliated with Ubuntu?
[01:14] <Keybuk> yankees26: nope
[01:14] <Keybuk> free hosting for all
[01:14] <yankees26> ok
[01:15] <yankees26> crimsun on #launchpad beat you to it ;) (thought you all had died here so asked there :P)
[01:15] <yankees26> thanks though
[01:16] <yankees26> i have added my project :D
[01:16] <yankees26> (now to setup bazaar)
[01:17] <Keybuk> I did die
[01:17] <Keybuk> updated to latest feisty
[01:17] <Keybuk> and rebooted
[01:17] <yankees26> hehe
[01:17] <yankees26> how is it?
[01:17] <Keybuk> unstable and buggy :p
[01:17] <yankees26> sounds like fun :D
[01:23] <yankees26> ugh, ~/.ssh/ has no id_dsa.pub or id_rsa.pub!?!?!
[01:25] <yankees26> ah, had to create one
[01:32] <yankees26> could someone with gpg experience help me?
[01:34] <_ion> PGP and SSH keys are a different thing. See ssh-keygen(1), ssh-add(1) and apt-cache show keychain
[01:34] <yankees26> i figured out the ssh key
[01:34] <yankees26> for gpg i did gpg --gen-key and that finished
[01:35] <yankees26> but im not sure what "number" is the actual key for gpg --send-key
[01:35] <_ion> There are quite good GnuPG tutorials in the web.
[01:35] <yankees26> ok
[01:35] <yankees26> TO THE INTERNET!
[01:36] <_ion> To the tubes
[01:36] <yankees26> that too P
[01:37] <yankees26> is it the thing after pub from gpg --list-keys?
[01:38] <_ion> It's the 2EE24E0B in 'pub   1024D/2EE24E0B 2005-07-03'
[01:38] <yankees26> ah
[01:38] <yankees26> thanks _ion
[03:27] <myren> is there a runtime daemon?
[03:27] <myren> what is it coded in?
[03:27] <myren> how portable is it?
[03:28] <myren> it uses dbus, right?  what dbus api does it make use of (low level or glibc)
[03:28] <Keybuk> runtime daemon?
[03:28] <Keybuk> upstart is an init replacement, so it does have a /sbin/init, yes
[03:28] <Keybuk> it's written in C
[03:28] <Keybuk> not portable at all; aimed squarely at current Linux and glibc
[03:28] <Keybuk> the currently released version does not use dbus, there's a branch which uses low level api
[03:36] <myren> i meant portable between architectures
[03:36] <myren> there's a lot more variants of linux than ubuntu alone supports
[03:36] <myren> ooh, glad there is dbus integration, thats swell.
[03:36] <Keybuk> it compiles on most, yeah
[03:36] <Keybuk> there's an open bug about hppa which is committed in the bzr tree
[03:36] <myren> the daemon keeps running right?
[03:36] <myren> beyond just startup
[03:36] <Keybuk> yes
[03:36] <Keybuk> it's pid 1
[03:37] <myren> hppa?  hp power risc?
[03:37] <myren> just a guess
[03:40] <Keybuk> yeah
[03:41] <myren> i know he's gone, but, thanks mate
[04:00] <Amaranth> if the daemon stopped running your computer would break :P
[04:40] <myren> duly noted Amon
[04:40] <myren> err, Amaranth
[05:38] <myren> as in, Seveasoft?
[10:37] <myren> is there any info on dbus integration besides the source?