Re: autoconf-2.59 and AC_C_LONG_DOUBLE: comments and a proposal

[Paul Eggert]

"Nelson H. F. Beebe" <address@hidden> writes:

> AC_C_LONG_DOUBLE therefore needs to be rewritten with a test program
> that is linked and run to report the result of the expression.

I'd rather not do that, since that makes cross-compilation a pain.
But I see your problem.

> it should instead specify
>
>  - Macro: AC_C_LONG_DOUBLE
>      If the C compiler supports a working `long double' type with more
>      range and precision than the `double' type, define
>            ^^^ critical change here
>      `HAVE_LONG_DOUBLE'.

Why is this change advisable?  If long double merely has more range,
or merely has more precision, that should be good enough.  Or are
you thinking it might have _less_ range or precision?  OK, yes, that
should be checked for.

> In numerical code, when a "long double" type is being exploited to
> achieve higher accuracy (I'm working on just such code right now), one
> needs BOTH increased range and increased precision.

Surely this depends on the application.  Some applications might just
need extra range, for example.

Perhaps we need to have multiple tests, for different application
needs?  The original goal for AC_C_LONG_DOUBLE was fairly simple:
"long double won't hurt".  It sounds like you need something more
elaborate?