[PATCH] PR fortran/33229 -- Fix intrinsic subroutine as function call

[Tobias Burnus]

Hi Steve,

Steve Kargl wrote:
> I did not realize you use "dg-do compile".  You're right.
> One needs to execute the test.  The compile test is useless
> in that the code would compile before my patch and give
> a wrong result.

Here, I always got an ICE (segmentation fault) thus I have not
realized that there is a wrong-code problem. (For me it was
merely an ICE-on-valid-code PR.) In any case, even if the ICE
is not present everywhere, I would not call the test case
useless.  If you have an idea how to improve the test, I'm
interested. (But I do not intent to use the 8-GB-allocation
version of the patch ;-)

Tobias