On 7/31/06, Ken Moffat <[EMAIL PROTECTED]> wrote:
On Mon, Jul 31, 2006 at 03:11:11PM -0700, Dan Nicholson wrote:
> Looks like this could be the simple fix, but I'm not on a system to test.
>
> $ sed -i '[EMAIL PROTECTED] \(/dev/fd/.\)@-e \1@' configure
>
As it stands, not acceptable - we can't assume sed supports -i
until we boot or chroot (unlike LFS). That's a minor reservation.
I'm building on ppc64 at the moment to see if the
bash_cv_dev_fd=standard fix does indeed fix the problem with
c++-types.out (don't hold your breath, I'll be in bed before both
sets of glibc tests have finished, even with the hours I keep ;)
Yeah, Joe is also right that for CLFS you can't depend on the host.
Might be acceptable for native builds, though (after reverting -i from
sed).
Greg just brought up an interesting point to me, though. Why hasn't
this affected LFS/DIY? I just checked my build log from a couple weeks
ago, and my temp bash has
checking whether /dev/fd is available... absent
Except that was glibc-2.3.6. Greg's doing native glibc-2.4 for DIY
with bash-3.1, and I'm pretty sure he builds the temp stage
unprivelaged. Something goofy going on.
--
Dan
_______________________________________________
Clfs-dev mailing list
[email protected]
http://lists.cross-lfs.org/cgi-bin/mailman/listinfo/clfs-dev
