Sparge, and is Clearwin+ standard-conforming Fortran

I was wrong: it isn't. @ is not in the character set (except in character constants) for other compilers.

Whether or not WINIO@'s variable number of parameters is OK, WINIO@'s @ isn't OK. Some of the interfaces in the insert files may just not be standards-conforming, and certainly it is possible to use non-standard INCLUDE syntax in FTN95. None of the compiler directives, including WINAPP, will compile with lesser compilers.

So, Clearwin+ isn't (standard-conforming) Fortran - and you were right - but by heck it looks like it is!

Eddie

I suspect that is why we use '@' (and $) on the end of routine names -- so they are a) not going to clash and b) easy to spot.

Robert,

How does FTN95 know that @ is admissable sometimes? More to the point, why do we need WINAPP - since FTN95 will always know if it has seen WINIO@ or another of the @ routines that don't work other than with WINAPP, and it could presumably put a flag in the OBJ code, so that SLINK didn't need to look too hard afterwards ... just a thought.

Oh, and the RESOURCES section (if there is one) won't pass a generic compiler.

Eddie

Some but only a minority of @ subroutines are hard coded into the compiler.

WINAPP attaches the Windows (i.e. not Console) attribute to the executable and, optionally, allows the declaration of a resource script. It can be used without winio@, for example with standard WRITE/PRINT. At one time it also used to work with READ but I have not tried that for a while. In other words is converts any Fortran program from a console app to a Windows app.