[01:02] <Dvalin> fabbione: around?
[04:47] <fabbione> Dvalin: i am now so to speak
[05:50] <Dvalin> still?
[05:51] <Dvalin> fabbione: what's the status of gcc 4.1.2 & kernel on sparc at your camp?
[05:51] <fabbione> Dvalin: gcc dunno.. what's in the archive...
[05:52] <fabbione> kernel we are going for .20 + some patches
[05:53] <Dvalin> okay, using 4.1 is no good for kernel it seems here
[05:53] <Dvalin> 4.0 works fine
[05:53] <Dvalin> :/
[05:53] <Dvalin> so I blame the compiler
[05:53] <fabbione> the one we are using is good
[05:53] <fabbione> i am booting those kernels
[05:58] <Dvalin> compiled with gcc 4.1.2?
[05:58] <Dvalin> mine will only boot without display
[05:59] <Dvalin> but then again
[05:59] <Dvalin> other hardware etc. too
[06:24] <fabbione> Dvalin: compiled with whatever we have in the archive now
[06:25] <Dvalin> okay
[06:25] <Dvalin> which kernel version was that again?
[06:26] <fabbione> .19 and .20-rc...
[06:26] <Dvalin> okay
[06:38] <Dvalin> I'll go ahead and try one of those then :)
[06:39] <Dvalin> btw. why is sparc not visible on packages.ubuntu.com?
[06:40] <fabbione> no idea
[06:40] <fabbione> i did never used packages.u.c
[06:40] <Dvalin> okay
[06:40] <fabbione> file a bug :)
[06:40] <Dvalin> nah
[06:41] <Dvalin> I'm just asking out of curiousity, couldn't care less for myself ;)
[06:42] <Dvalin> btw. there's one thing that puzzles me, why isn't different archs reported on sparc? eg. like sparcv8, sparcv9 etc. as i586, i686 etc. on intel?
[06:42] <fabbione> uhm?
[06:43] <Dvalin> eg. by 'arch'
[06:43] <fabbione> yeah but where..
[06:43] <Dvalin> uhm, in general? by kernel?`:)
[06:45] <fabbione> you make no sense
[06:45] <fabbione> where do you want to read that information?
[06:45] <Dvalin> hmhm
[06:45] <fabbione> proc/cpuinfo? uname -a ?
[06:46] <Dvalin> uname -m for an instance
[06:46] <fabbione> because you don't need it
[06:46] <fabbione> the kernel is either 32 or 64 bit. all optimizations are autodetected and turned on at boot
[06:47] <fabbione> so there is no point for userland to know
[06:48] <fabbione> and you can still see the hw cap via glibc
[06:48] <Dvalin> why is there less of a need for it than on x86?
[06:49] <Dvalin> I'd still say there is a use for userland to know, eg. when building stuff..?
[06:49] <fabbione> fabbione@rainy:~$ LD_SHOW_AUXV=1 /lib/libc.so.6 --version  | grep AT_HWCAP
[06:49] <fabbione> AT_HWCAP:    flush stbar swap muldiv v9 v9v
[06:49] <fabbione> no there is no need to know.. how build needs to know what's doing
[06:50] <fabbione> all these extra automatic features have been proven to be a real disaster
[06:50] <Dvalin> oh?
[06:58] <fabbione> when you autodetect something at build time and thenD you distribute a precompiled binary on a machine that does not have that feature it's BA
[06:58] <fabbione> BAD
[06:58] <Dvalin> yeah, but that's why you specify the arch of binary ;p
[07:06] <fabbione> Dvalin: that's also why stuff works for us but not for you
[08:15] <Dvalin> uh
[08:15] <Dvalin> I don't see the relevance of that
[08:17] <Dvalin> actually, AFAIK you build for sparcv8 by default without packages saying anything about it NOT being compatible with sparcv7 which just 'sparc' usually implies ;p