[05:10] morning [05:17] Up early again? [05:24] yeah [05:24] couldn't sleep shit this night [05:49] Better than me. I've been unable to get to sleep. Was up until 4:30am last night. [05:50] yeah and i woke up at 4:30 am [05:50] what a satisfaction :/ [05:51] who feels in ia64 porting mode? [05:54] Depends. What's broken? [05:55] infinity: not sure.. that's why [05:55] infinity: if you want to look at it just ssh to halley [05:55] the source is in my ~ [05:56] rhcluster [05:56] just temporary drop the versione build-dep on gcc-3.4 [05:56] and try to build [05:56] it will fail with an interesting error [05:56] i blame glibc or kernel-headers === Seveas [~seveas@seveas.demon.nl] has joined #ubuntu-toolchain [06:00] [06:02] fabbione: I need to go pass out. If it's still an issue tomorrow, lemme know. [06:02] Oh, nice. [06:02] jbailey : It's all your fault, you horrible man, fix it now. [06:02] bite me [06:03] Yessir. [06:03] Biting, sir. [06:03] jbailey: i am pretty sure it's either glibc or kernel-headers [06:04] the only thing that will "break" is the new version of lvm2 that will b-d on rhcluster [06:04] do it [06:04] so it's not that bad [06:04] fabbione: Sure. It's been a while since I've infliced a glibc upload on breezy. [06:05] jbailey: ehehhe [06:06] I think I need one more day to get the EarlyUserspace to where mdz can finish his ltsp stuff. [06:07] oh neat [06:07] Since dash and busybox don't support posix shell arrays. =( [06:07] jbailey: next week you will need to tell me how to test that stuff [06:07] fabbione: Yup! I can basically handle 3 booting scenarios now: [06:07] 1) Booting off of a harddrive [06:07] 2) NFS Boot [06:07] 3) Booting off of software raid [06:08] I don't understand how lvm and evms interact enough to do those. [06:08] #3 what kind of software raid? [06:08] But they should be easy enough to do. [06:08] i can help you with lvm [06:08] fabbione: Whatever mdrun/mdadm can assemble. [06:08] ah neeeeeaaaat [06:08] My local test box is a raid1 setup. === Seveas [~seveas@seveas.demon.nl] has joined #ubuntu-toolchain [06:21] Definitely looks like l-k-h breakage. [06:21] But if that's jbailey's fault of upstream just sucking, I dunno. :) [06:21] eheh [06:23] Or... [06:23] Maybe it's RedHat's fault for being shithouse. [06:24] Google saves the day. [06:24] Check this patch against qemu: [06:24] http://lists.gnu.org/archive/html/qemu-devel/2005-02/msg00355.html [06:24] Search for "gettid" [06:24] Notice that _syscall10 construct, which is also failing for us, and the more elegant looking replacement. [06:27] I think all the _syscall## macros are gone on ia64. [06:27] Looks like it, yes. [06:27] IIRC, there's a proposal to get rid of them on all the other archs. [06:27] so why are they still there for other arches? [06:27] And good riddance to them. [06:27] ah ok [06:27] Too many magic numbers. [06:27] Mhm [06:29] infinity: this "elegant" patch is not exactly easy to understand [06:29] fabbione : I'll give you a patch, then. [06:29] thanks [06:30] fabbione: I'll try to do the sparc glibc stuff this weekend too. [06:30] At least for the 32-bit arch's. [06:31] jbailey: i will be out all weekend. do you want me to stop the buildd? so you can have full power on box? [06:31] Hmm. How caught up are you? [06:31] jbailey: it's building universe, since all the stuff in main is either banned or FTBFS [06:32] (or dep-wait) [06:32] jbailey: + you can stop the buildd yourself if you need to [06:32] jbailey: you just touch ~/EXIT-DAEMON-PLEASE [06:32] and it will stop at the end of the build it is doing [06:32] and i will restart it when i am back on monday [06:55] Kaboom. Works. Now to make the patch portable, so you can forward it upstream. [07:05] infinity: So it's officially Not My Problem(tm)? [07:07] Not anymore. [07:07] Not unless you want to readd all those _syscall macros to ia64. [07:07] And I know how much you'd love that. [07:10] Yeah, that'd be a case of tell the porter to cope. =) [07:13] S'ok, I have a deep hatred of magic numbers, so this patch is more sane anyway IMO. [07:18] fabbione : halley:~adconrad/rhcluster_1.20050527.debdiff [07:18] fabbione : There's also a build.log and a bunch of debs in that directory as proof that it builds. :) [07:20] is it portable? :) [07:21] infinity: http://www.inspot.org/assets/tell_card4_pop.jpg [07:21] Gotta love links on Savage Love. [07:22] fabbione : More than the old one, that's for sure. :) [07:22] jbailey : Hahah. [07:24] fabbione : It's portable in the sense that any system that defines __NR_gettid will now behave properly, and the rest get -ENOSYS. That's the best I'm willing to do. :) [07:24] fabbione : That's lightyears ahead of the magic number code I was replacing, so I say "good enough". [07:25] If gettid isn't available, getpid is usually the replacement. [07:25] jbailey : Well, want to revise my patch further?... I somehow doubt upstream cares about non-Linux platforms at the moment, but making it even better doesn't hurt. [07:25] The current patch should work for all the platforms we care about currently. [07:26] Not at the moment, I'm just flossing before bed. [07:26] You wear a g-string to bed? [07:26] well in any case upstream doesn't really care about portability [07:26] they support only i386/amd64/ppc64 afaik [07:26] fabbione : They should care about portability to Linux arches. I'd forward this patch if I were you. [07:26] but we do care :) [07:26] i am going to anyway [07:27] (Besides, if what jbailey says is true, this patch will be needed for future portability on the "major" arches too) [07:27] Also, halley sucks. No devscripts, no patchutils... [07:28] Sleep time. [07:28] jbailey : G'night, you sick freak. [07:28] night jeff === chmj [~d3vic3@196.11.241.45] has joined #ubuntu-toolchain [09:15] fabbione : So, other than the ia64 porting, are those packages ready for upload, SUPERSTAR?! :) [09:15] infinity: yes [09:15] i am testing the last 2 init scripts right now [09:16] Rock. [09:16] I assume my shell-fu from the other day was also for these. Didn't bother looking. :) [09:16] yeah :) [09:16] it's for the gnbd-*.init [09:17] i need to remember to add the copyrights and do the anal changelog :) === infinity sees that construct in two different init scripts, and decides this shell-fu has become fabbione's new hammer for every nail. === daniels [~daniels@amnesiac.heapspace.net] has joined #ubuntu-toolchain [09:19] daniels : Bring me dinner. I'm hungry. [09:19] daniels: bring us females.. we are hungry :) [09:19] infinity: what happened to your steak? [09:19] fabbione: eh dude, you're married. i don't even have any prospects. :P [09:19] daniels : I ate it all. [09:20] infinity: well, there you go [09:20] fabbione : I ate them all. [09:20] let this be a lesson to you [09:21] hehe [09:21] ehehe === doko [~doko___@dsl-082-082-194-124.arcor-ip.net] has joined #ubuntu-toolchain [01:46] Yar [01:48] arh [02:04] yo [03:17] jbailey: ping [03:21] pong [03:21] Sorry, I've been working something out with pen and paper. =) [04:54] In file included from tracepath.c:21: [04:54] /usr/include/resolv.h:110: error: array type has incomplete element type === lamont-away whaps jbailey [04:54] (from iputils) [04:55] Eh? Are you trying to blame another set of third party bad code on my beautiful glibc and lkh? =) [05:04] l-k-h doesn't have bugs. === lamont [~lamont@15.238.6.115] has joined #ubuntu-toolchain === katoc [~katoc@201.145.208.167] has joined #ubuntu-toolchain [10:17] hi!! i need to add packages to installation cd === katoc [~katoc@201.145.208.167] has left #ubuntu-toolchain ["Konversation]