On 3/15/2013 10:43 AM, John Ehrman wrote:
One reason for the Assembler's not having implemented a fuller set of
extended branch mnemonics is that once the original ("incomplete") set
wasprovided, many users implemented their own macro-based extensions.

Given the decades-long lack of a mechanism to choose between a set of
built-in assembler mnemonics and a wide variety of user-implemented
mnemonics (not all of which were the same) it seemed safest to stay with
the original set.  It could be very risky for the assembler to provide
built-in extended mnemonics that had slightly different branch masks from
those a customer was already using, because programs could behave very
differently with no warning or other indication.

It seems to me the most obvious choice for "logical" extended mnemonics
would have been to insert the letter 'L' as in Jxxx -> JLxxx. Had I
written these macros years ago, I might have created (for example) JLO
to branch on overflow (i.e., on 'carry' aka BC 5,xxx) after an ADD
LOGICAL. Many developers might have already done this only to have it
all stop working when the assembler added JLO to mean "jump long on
overflow". So, you might have already silently broken existing macros.

--
Edward E Jaffe
Phoenix Software International, Inc
831 Parkview Drive North
El Segundo, CA 90245
http://www.phoenixsoftware.com/

Reply via email to