Major CVS LFS bug - binutils-184.108.40.206
ca9mbu at eos.sunderland.ac.uk
Mon Jan 27 15:19:34 PST 2003
"Chris Lingard" <chris at stockwith.co.uk> wrote in message
news:b14ck2$fn5$1 at quasar.highos.com...
> I suggest that this is very urgent. You have at least three options
> Revert to binutils-2.13.2. gcc works with triplets
> Put the HAVE_GAS_HIDDEN in chapter 5 and chapter 6
> Think of another method to fix this
3) Patch gcc's configure script to recognise binutils version strings >
triplets. This should avoid the problem in the future and I don't think
any Makefiles or such-like would need regenerating as a result of this,
we just need to make the test return "yes", right?
4) Patch binutils-220.127.116.11 to force it's version number back to a
triplet. Yuck, yuck & more yuck. This would be the king of all hacks
Out of the 4 methods so far, the HAVE_GAS_HIDDEN instruction is probably
the easiest to do, followed closely by reverting to the previous version
of binutils. However, IMO the patch to gcc's configure script would be
the more technically superior. The regex on line 7668 (current CVS) is
clearly looking at only 3 parts to the version string so it just needs
to be extended right? I'm assuming here that the current regex caters
for versions like 2.12, 2.11, etc.
Oh, and once this is sorted we'll have to remind someone to bash the
binutils guys over the head for doing this to us, I mean what was wrong
with 2.13.3 for pete's sake!
Unsubscribe: send email to listar at linuxfromscratch.org
and put 'unsubscribe lfs-dev' in the subject header of the message
More information about the lfs-dev