Not wanting to beat a dead horse too much, but I ran an experiment using
the code in my original post:
Macro
&Lab MVC2 &Target,&Source
&Lab MVC &Target.(L'&Source),&Source
MEND
vs. John's provided macro and they generate identical code. I read John's
response about the symbol table so I made sure the test using my macro used
a unique literal for source (I also have a second test using a designated
field) so it was a "first reference" (I don't think we're doing more than
the standard 2 pass assembly). The only reason I bring this up is that it
seems this could be helpful in a few situations.
As I said, I don't want to beat this horse any more, but even though I've
been writing assembler for 30 years (including z80, 808x, and 68000
assembler), I still learn a lot from this list.
Changing tack slightly, has anyone used MVST in a case where there was the
possibility of a missing terminator (i.e. something other than a x'00' as
the terminator). It's easily addressed (stick the expected terminator in
the last character of the sending field), but just something I encountered
today.
Thanks again,
Mark Hammack
On Wed, May 16, 2012 at 4:14 PM, Paul Gilmartin <[email protected]>wrote:
> On 2012-05-16 13:10, Ray Mansell wrote:
>
>>
>> Of course, if the assembler were to support the L' construct for
>> literals, this would not be a problem.
>>
>> At least during Pass 2.
>
> Lookahead is confusing.
>
> -- gil
>
