> What is at issue is changing a datum that is not an instruction. Such > data typically have more references than instructions do, and the > references are more likely to be unrelated to each other. > > >Similarly overlooking a reference to a source code instruction > >about to be reassembled may also cause adverse affects. > > Only rarely will changing a literal reference in a source line cause > problems that could have been avoided by reading the code that > referred to that line.
All of that is true. My primary concern with a zap to a literal value is that there may be multiple references to that literal from "hidden" places like inside macro definitions. Casual visual inspection of source wouldn't typically find them. With a full assembly xref listing and no PRINT NOGEN or PRINT OFF obfuscations, you can probably determine all of the references to each literal that would be altered by the proposed zap. So it is true that one CAN safely make the change -once-, even if in some cases you had to alter instructions to refer to a new "literal" zapped into a patch area. If you ever needed to do it again, the second zap would be a b1tch... but (as Shmuel says) "its not my dog" CC ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html