[GCC-XML] fails to parse __float128 on linux 32-bits, OK in 64-bits

[Brad King]

On 12/21/2011 12:26 AM, Steve M. Robbins wrote:
> OK.  So if I cannot get anything sensible, I may as well hack around
> this by using "gccxml -D__float128=double ...".

This may be necessary.

> I am curious, however, why the 64-bit version of gccxml does not seem to
> have a problem, while 32-bits does.  Any ideas?

When gccxml builds its builtin GCC parser gets configured as a compiler
targeting the host platform.  Look at

   GCC/gcc/config/i386/i386.c
   GCC/gcc/config/ia64/ia64.c

to see how each target treats __float128.  Note that x86_64 targets will
use the i386.c source (ia64 is for Itanium).  That source contains the
condition "if (TARGET_64BIT)" around its support for __float128.

-Brad