MATCTX (was Detecting changing objects...)

I think I see the confusion now.  Earlier you had refered to the "C
function matctx" which made me believe you were using the C matctx
builtins (the _MATCTX1 and _MATCTX2 in my previous note); but what you
are actually using is the system API QusMaterializeContext which is not
a C function (from my point of view) but rather a system API documented
in the System API Reference.  As such this API can also be called
by other ILE languages (that can define an Open or System pointer
anyway).

This is not a breach of security.  There are many system APIs that
provide a function that is not directly available to user state
application programs.  This is simply one of them.  In this particular
case IBM knew of developers using the MI MATCTX, knew that it was in
the best interest of everyone to allow these developers to work on
level 40 and 50 systems, evaluated the specific inputs and outputs of
MATCTX, considered the development cost to make developers change to a
more formal API (Receiver variable, length, format, error code, etc.)
and decided that a simple wrapper of the MATCTX MI instruction would
be the most appropriate solution in this specific case.

Bruce

>
>[Leif Svalgaard]  Bruce, I've left the entire thread in this posting
>to have
>the complete context.
>Here is first the C-program MATCTX:
>
>#include <qusmiapi.h>
>
>void main(int argc, char *argv[]) {
>typedef struct {
>  _SYSPTR SystemPtr;
>} s; s* SysPtr;
>SysPtr = (s*) argv[2];
>QusMaterializeContext((_SPCPTR) argv[1],
>                        SysPtr->SystemPtr,
>                      ( void *) argv[3]);
>}
>>
>> Interestingly enough, the C function matctx does not have
>> that problem. ...
>>
>

+---
| This is the MI Programmers Mailing List!
| To submit a new message, send your mail to MI400@midrange.com.
| To subscribe to this list send email to MI400-SUB@midrange.com.
| To unsubscribe from this list send email to MI400-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: dr2@cssas400.com
+---