Re: Fortran: Fix compile of caf_shmem on cygwin [PR124371]

On Fri, Mar 06, 2026 at 01:57:49AM -0800, Andrew Pinski wrote:
> On Fri, Mar 6, 2026 at 1:55 AM Andre Vehreschild <[email protected]> wrote:
> >
> > Sorry, I don't get you there.
> >
> > On cygwin WIN32 is set but it also has HAVE_FORK. Therefore checking for
> > HAVE_FORK is in my opinion sufficient. Or what is your comment about? "I am
> > standing on the pipe" as Germans might say.
> So your changelog was:
> `Use HAVE_FORK instead of WIN32 for preprocessor conditional.`
> What Jakub was mentioning was that the preprocessor check was on WIN32
> `NOT` being defined.
> So maybe:
> `Use HAVE_FORK instead of not WIN32 for preprocessor conditional.`

Yeah.  Even more precise would be Use defined(HAVE_FORK) instead of
!defined(WIN32) for preprocessor conditional.
because it isn't testing values but whether those are defined.

	Jakub