Re: Why extra blanks in stat() IFS data structure?

Scott Klement wrote:
>
> On Mon, 15 Jul 2002 craigs@dekko.com wrote:
> >
> > Why the extra blanks in the stat() data structure for the stat() IFS API's?
> >
>
> Because of the way variables are aligned in C structures.   To make
> things more efficient, 16-bit integers are aligned on 2-byte boundaries.
> 32-bit integers are aligned on 4 byte boundaries, 64-bit integers are
> aligned on 8 byte boundaries, pointers are aligned on 16-byte boundaries,
> etc, etc.
>
> In RPG however, this type of alignment does not occur unless you use
> the ALIGN keyword in your D-spec.

When working with C structs, there should be either a _Packed keyword on
the C struct or an ALIGN keyword on the RPG DS.  Exactly one or the
other.

Aside:

Just to let people know - there's a new version of my "Converting from C
prototypes to RPG" (http://www.opensource400.org/callc.html) on the
opensource.org site.  It's been cleaned up and expanded, and it's also
been updated with V5R1 information (it's easier to call C from RPG now).

The opensource.org site was down for a while but it has a new owner.