Re: QPROCT and QPRODT

I'm not convinced that MI allows such behaviour - but, as always, I'll
be
happy to be proven wrong.

I cannot imagine any reason why it would care. And your empirical evidence
of what MI does or does not allow would seem to supersede either of our
opinions. :)

But beyond that I can only leave it to the authorities. I am not privy to
nor expert in the design of the machine interface.

Dennis Lovelady
http://www.linkedin.com/in/dennislovelady
--
"More than half (55 percent) of women undergraduates are female."
-- About Women on Campus, newsletter published by the National

I can't see the subtle difference you allude to - but that means very
little
:-)

I mean that those are the result flags that are set. You are assuming
that
MI limits you to test only the flags that are set by the most recent
instruction. I am suggesting that you can test for things that were
not set
by this instruction.

I used to do this in TRS80 assembler. I forget now the exact situation
(and
instructions), but I used to do something like:
ADDC RESULT,12 Add 12 to RESULT, with carry
CMP RESULT,16 Compare RESULT against 16
JMP MYTAG,EC Jump to MYTAG if equal or carry

Assume that carry was not set nor reset by CMP. I could still test for
it
after CMP.

Now I realize this is not TRS80 assembler (thankfully). I also realize
that
I'm not doing a branch on the same instruction line as compare (TRS
lacked
that). But I imagine similar rules apply.