This is the mail archive of the gcc-bugs@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: g77: missing intrinsic atan2d


But they should be very easy to implement - 
------------------------------------------
# include <math.h>
double atan2d(double a, double b)
{
  return 180.0d0 * atan2 (a, b) / PI;
}

etc. 
------------------------------------------

When I write a patch, I 'll send it over :-) ; but it is not just the
c source code, , but also the info file, right? 

Is there a list of what these missing intrinsics are supposed to do
somewhere? (I didn't know atan2D was the degree version, etc); I also
notice that a whole lot of quadri-precision functionality are not yet
implemented; but gcc doesn't seem to have long doubles (at least not on
linux...).

> >I found a whole chuck of missing intrinsics of this form (atan2D, atand,
> >acosd, etc) in g77 - and got a error message porting some SunSparc fortran
> >codes involving this function; it is probably no big deal as I know a
> >similar function is in the libc or libm; but then, shouldn't something be
> >done about it? 
> 
> g77 doesn't yet support the "degree" forms of trig intrinsics that
> some other compilers (like Digital's, and, I guess, Sun's) do.  It'd
> be nice to support them someday -- patches to do so are welcome!
> 
> Note that these intrinsics are not standard AFAIK -- certainly they're
> not in F77 or F90, but I gather they're also not in F95, and don't
> recall seeing them on the list of things for F2K to support.
> 
> But I assume they can be quite useful.
> 
>         tq vm, (burley)



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]