https://gcc.gnu.org/bugzilla/show_bug.cgi?id=126305

--- Comment #9 from dave.anglin at bell dot net ---
On 2026-07-17 3:14 p.m., kargl at gcc dot gnu.org wrote:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=126305
> 
> Steve Kargl <kargl at gcc dot gnu.org> changed:
> 
>            What    |Removed                     |Added
> ----------------------------------------------------------------------------
>                  CC|                            |kargl at gcc dot gnu.org
> 
> --- Comment #8 from Steve Kargl <kargl at gcc dot gnu.org> ---
> (In reply to dave.anglin from comment #7)
> 
>> %rename lib liborig
>> *lib: %{static-libquadmath:-aarchive_shared} -lquadmath
>> %{static-libquadmath:-adefault} -lm %(liborig)
> 
> Can you confirm what floating point HPUX supports?
> gfortran assumes at least a C99 libc with float,
> double, and long double math functions.

I wouldn't say HP-UX libc is C99 compliant.  It has many C99 functions but
lacks
long double math support and some other features.

In the past, what was missing was provided by c99_functions.c.  I just added
frexpl
and scalbnl.
> Does it have an IEEE 128-bit long double?

The PA_RISC architecture specifies the IEEE 128-bit long double format.

> Is it a hardware or software implementation?

Software.  There is basic arithmetic, comparisons and printf support in libc
for long double.  But there are no long double math functions in libm.
> If it is software, is libquadmath the backing 
> implementation?

No.
> Does HPUX have prototypes for the *l(long double)
> functions in math.h where 'long double' maps to
> libquadmath's __float128_t.

No.  The HP-UX system doesn't know about libquadmath. Any mapping would need to
be
done by libgfortran.
> Can gfortran be configured on HPUX without REAL(16)
> (or REAL(17)) support.  Perhaps, that should be the
> default?

I don't think that's currently possible.

Reply via email to