Re: The joy of FORTRAN

On 02/03/2025 13:41, Richard Kettlewell wrote:

The Natural Philosopher <tnp@invalid.invalid> writes:

And in fact these days with inline functions
>
if(x>y)
       fubar(x,y,z);
else
       rabuf (z,y,x);
>
Is even shorter and allows the program flow rather than the
implementation to be seen clearly.
>
I wonder if compilers actually inline functions that are only called
once.

 Modern compilers will often inline functions that aren’t declared
inline, yes; I don’t know if “only called once” is a consideration. It’s
not implausible (possibly in a more general form, e.g. “will inlining
all calls take more or less space overall?”)
 

My complier writing friend suggested that if the function were declared static, and the function only called once. it certainly could be inlined, but he was not sure at what optimisation level of gcc it would be.
--
Climate is what you expect but weather is what you get.
Mark Twain