> No, it is not possible to set caller variables from a native method
> context. That entire concept is an obsolete holdover from Rexx's
> early days, and has not been extended to methods. And on top of that,
> you're making an unwarranted assumption that the caller of a method is
> going to be Rexx code. In many cases, this will not be true, so the
> concept of setting a caller's variable has no meaning.
Thanks for the explanation.
(Not to start a fruitless discussion - therefore this comment in
parenthesis -, nevertheless a few thoughts here:
- as the interpreter is able to do that, e.g. setting variables (e.g.
like RC or RESULT) in methods that call functions/procedures or routines
or exits, the API programmer should be able to that as well,
- if it does not make sense the API programmer would not employ it, nor
would she/he request it,
- if it does make sense and it was possible to find out whether the
calling method is native or not, one could code accordingly,
- last, but not least: it would be orthogonal with classic Rexx as it
allows API-programmers to interact directly with variable pools; there
would be no conceptual difference of that concept when executing code
called from a method, given that a method obviously has a variable pool
of its own.)