RDi Debug problem?

Recently, Barbara Morris submitted RFE 123323,

https://www.ibm.com/developerworks/rfe/execute?use_case=viewRfe&CR_ID=123323

regarding RDi debug. Out of curiosity I copied and modified the code sample as follows:

**Free
   dcl-s a_job char(3) inz('a!b');
   dcl-s a_37 char(3) ccsid(37) inz('a!b');
   dcl-s a_500 char(3) ccsid(500) inz('a!b');
   dcl-s a_utf8 char(3) ccsid(*utf8) inz('a!b');
   dump(a);          // Breakpoint set here (1)
   a_job = 'ABC';    // Breakpoint set here (2)
   *InLR = *On;
   return;

I found an odd inconsistency in RDi debug. I set two breakpoints (1) and (2) above. When the program was run under RDi (SEP) debug, it halted at (1). When Resume or Step Into were invoked, the program continued to run to termination.
The same program with the same breakpoints was run under the host debug facility, (STRDBG).  Execution halted at (1), but when Resumed/Stepped, the program to halted at (2) - as expected.

It would appear that the RPG IV DUMP opcode causes the RDi debug to terminate program execution prematurely, ignoring subsequent breakpoints or statement stepping.

Running RDi 9.6.0.4 - anyone else experience similar?

Brian.