Re: MATAGPAT strange behavior?

The person who owns this instruction implementation confirms that you have
detected a bug, apparently introduced in v5r1. Please feel free to open an
APAR. In this case the problem would get fixed in the next release anyway.
However, a report on a general mailing list doesn't guarantee an official
response, as I'm sure you can understand.

Thanks for asking about this problem.

Paul Godtland
plg@us.ibm.com; t/l 553-7103; Rochester, MN; 030-2, A412




                      Gene_Gaunt@Review
                      Works.com                To:       mi400@midrange.com
                      Sent by:                 cc:
                      mi400-admin@midra        Subject:  Re: [MI400] MATAGPAT 
strange behavior?
                      nge.com


                      04/04/02 12:11 PM
                      Please respond to
                      mi400






AUTO or STAT, same bug.

This seems a bug to me: the MATAGPAT option X"02" on V5R1 doesn't honor the
template value I specify in position 1-4 (the "number of bytes provided for
materialization").  Instead MATAGPAT returns the entire list of program
activations, clobbering my storage after my template declaration.

dcl dd     Template    auto char(8192);
dcl dd     *           def(Template) pos(1) bin(4) init(64);
dcl dd     Test        def(Template) pos(65) char(16) init("THIS IS A
TEST");
dcl spcptr ?Template   auto init(Template);

matagpat ?Template, GroupMark, X"02";

/* this instruction returns more than 64 bytes                   */
/* variable "Test" is now clobbered with activation list entries */