Irix6 long doubles implemented wrong? (27_io/ostream_inserter_arith)

[Alexandre Oliva]

On Jan  7, 2003, Richard Henderson <rth@redhat.com> wrote:

>> * real.h (struct real_format): Add denorm_p, remove has_denorm.
>> * real.c: Adjust all formats and references to has_denorm.
>> * c-common.c (builtin_define_float_constants): Use denorm_p to
>> define DENORM_MIN.

> This one's sticky.  Strictly speaking, the double-double format
> isn't LIA-1 compliant (too few denormal bits), and so libstdc++
> ought to be setting denorm_min to zero.  On the other hand, I
> can see that this value might still be useful for some people.

> Run it by the libstdc++ language lawyers.

I've just gone through the C++ Standard's numeric_limits definitions
and the C99's float.h specification, and I couldn't find any reason
to not use the definitions the way I've introduced them.  The fact
that they get the libstdc++-v3 testsuite to pass in cases it didn't
before seems to be a good indication that it is good enough.

Ok to check it all in?

-- 
Alexandre Oliva   Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer                 aoliva@{redhat.com, gcc.gnu.org}
CS PhD student at IC-Unicamp        oliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist                Professional serial bug killer