On Wed, 5 Aug 2026 20:26:27 +0200, Abe Kornelis wrote:
>Richard,
>
>I understand your disappointment.
>HLASM is a bit picky in that it will use lookahead
>in many cases, but not all.
>
Is the "but not all" clearly documented in the Ref. It should be
better than, "The developers tried their best. We hope you
find it useful."
What error message should be given for cases which
have no valid resolution, such as:
AIF (T'A EQ 'F').SKIP
A DC F'42'
.SKIP ANOP
This should result in an error message. Which?
With the expressive power of conditional HLASM it
might possible to write code that requires that the
solve logic puzzles that would please Lewis Carroll:
<https://en.wikipedia.org/wiki/Zebra_Puzzle>
Should HLASM incorporate AI to solve all resolvable
constructs? Perhaps it should fail on complexity
limits.
>This is one of those examples.
>
>I got into the habit of defining DSECTs,
>constants and the like at the top,
>rather than at the bottom of my pgm.
>
>By default neatly wrapped in PRINT OFF/ON
>to keep the listing clean of clutter.
>
>Some of my code relies heavily on
>inheritance. For some attributes I ran
>into comparable issues.
>
>I cannot speak on behalf of the team,
>but my guess is that fixing this issue
>might potentially break someone else's code.
>
>Kind regards,
>Abe
>===
>
>
>On 05/08/2026 16:22, Richard Zierdt wrote:
>> This assembles w/o error:
>> LENGTH EQU 12
>> MVC APPLES,=CL(LENGTH)' '
>> APPLES DS CL16
>>
>> This fails:
>> MVC APPLES,=CL(LENGTH2)' '
>> LENGTH2 EQU 12
>> APPLES DS CL16
>>
>> with this message:
>> ASMA151E Literal expression modifiers must be absolute and predefined -
>> =CL(LENGTH2)' '
--
gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
