Patch for HP-UX IA64 bootstrap failure

[Richard Henderson]

On Fri, Sep 10, 2004 at 10:27:04AM -0700, H. J. Lu wrote:
> On Fri, Sep 10, 2004 at 10:11:46AM -0700, Steve Ellcey wrote:
> > I am trying to figure out if ia64_scalar_mode_supported_p should return
> > TRUE or FALSE for TFmode on IA64 Linux.

False, at least for now.

> I do have a glibc patch to support __float128 on Linux:
> 
> http://sources.redhat.com/ml/libc-alpha/2004-05/msg00049.html

You know as well as I do, HJ, that that patch isn't going to be
applied.  Uli has told you that directly.  Do not be disingenuous.

That said, supporting IEEE QUAD for Fortran has very little to
do with supporting __float128 for C.  It's certainly possible for
someone to add the bits for 128-bit arithmetic to glibc without
stumbling over the other issues.  To-date no one has submitted a
patch to do that (and ONLY that).


r~