On Tue, Jul 31, 2007 at 06:07:44PM -0400, Bryan Kadzban wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: RIPEMD160
> 
> Ken Moffat wrote:
> >  First, 64-bit flex failed when my script tried to 'touch
> > doc/flex.1' - there was a '21933 Killed' message.
> 
> The OOM killer, perhaps?  Anything in the kernel logs at this point that
> may point to that?
> 
 Brian, thanks for that thought.  I've just powered that box up
again to check the logs [ normally, I use a laptop to catch up on
mail while watching the TV at this hour - '65 Dylan at the moment ].
OMG, how embarrassing - the builds on Monday (gcc-4.2.1) and
Tuesday (gcc-4.2.0) seem to have been using a 2.6.22-rc1 kernel, but
chroot was built against 2.6.22 (release) headers, which might
explain why each of them show an oops for 'touch'.

 This is unlikely to explain why coreutils/glibc were giving
problems using 'echo' to create headers with 4.2.1, but it
definitely could explain the failure to compile bash natively.

 So, my next step for x86_64 has to be to compile 2.6.22.1.

 I suppose maybe I've got too many boxes/systems - for LFS-6.3-rc1 I
made sure I was using at least 2.6.22, but somehow missed that on
multilib (I think I started on another box which *was* running
2.6.22 with pure64, but after the problems with temp-system bash I
moved to the other box to try multilib and missed this.  Excuses.)

 The strange thing was that running touch from the command line in
chroot was fine.

 Oh well, at least I've got something else to try.  Thanks.

ĸen
-- 
das eine Mal als Tragödie, das andere Mal als Farce
_______________________________________________
Clfs-dev mailing list
[email protected]
http://lists.cross-lfs.org/cgi-bin/mailman/listinfo/clfs-dev

Reply via email to