[committed] Replace quadmath_flt128tostr (gdtoa based) with quadmath_snprintf (glibc based) and printf hooks (PR fortran/47642)

[Steve Kargl]

On Mon, Feb 14, 2011 at 04:51:51PM +0100, Jakub Jelinek wrote:
> Hi!
> 
> This patch removes the buggy gdtoa based __float128 -> string
> printing code and adds instead floating point printing code from
> glibc, with various configury to make it more portable (sometimes
> at the expense of dumbing it down).

Well, I finally had time to test the patch in the pr
on FreeBSD.  

ul.lo printf/printf_fphex.lo printf/printf_fp.lo printf/quadmath-printf.lo printf/rshift.lo printf/submul_1.lo printf/sub_n.lo  
libtool: link: `printf/addmul_1.lo' is not a valid libtool object
gmake[3]: *** [libquadmath.la] Error 1
gmake[3]: Leaving directory `/usr/home/sgk/gcc/obj4x/x86_64-unknown-freebsd9.0/libquadmath'
gmake[2]: *** [all] Error 2
gmake[2]: Leaving directory `/usr/home/sgk/gcc/obj4x/x86_64-unknown-freebsd9.0/libquadmath'
gmake[1]: *** [all-target-libquadmath] Error 2
gmake[1]: Leaving directory `/usr/home/sgk/gcc/obj4x'
gmake: *** [all] Error 2

I'll revert my tree, svn update to your commit, and try again.

-- 
steve