Re: ILE Exporting Variables and *DUPVAR

[Rory Hewitt <rory.hewitt@xxxxxxxxx>]

Yeah, this seems like an obvious solution to me. I know most people hate
the idea of data exports, but I use them quite a bit, but that's probably
because I'm used the PL/I where you can define variables as EXTERNAL, which
is effectively the same thing.

In fact, I have one RPGLE module which allocates memory for a
data-structure and exports that pointer to that data-structure. Rather like
various IBM API's, the first 10 characters in the data-structure specify
the format of the contents of the data-structure. Other procedures import
that pointer and have their own data-structure based over the pointer,
depending on the format. It all works very well.

That being said, I would have preferred the option that if a variable is
defined as IMPORT, then it is read-only. I think that would have made it a
much more 'accepted' mechanism...

On Fri, Mar 16, 2012 at 8:22 AM, Buck <kc2hiz@xxxxxxxxx> wrote:

> Aside from aesthetics, does it really matter if you randomly choose the
> one procedure to own the storage for this DS?  That is, pick one
> procedure to be labelled EXPORT and make the rest IMPORT.  The compiler
> will let any IMPORT procedure update/read the DS, won't it?
>   --buck