[fortran patch] Don't use TREE_LISTs for storing arguments to intrinsic functions

[FX Coudert]

> Do you happen to have a testcase that highlighted this problem?

Not that I know of. I think it was generating code that is valid, but  
inefficient. Instead of translating MAX(A,B) into
     if (A > B)
       MAX = A;
     else
       MAX = B;
your patch was generating
     if (A > A)
       MAX = A;
     else
       MAX = A;
     if (B > MAX)
       MAX = B;

This yields the right result, but is inefficient. Thus, I don't have  
a testcase.

FX