No one would be obligated to use 64-bit set symbols. There could be an 
assembler option to produce "old" or "new" ASMADATA/disallow or allow 64-bit 
set symbols. ASMADATA has "version" indicators in every record. This is a 
solvable problem.

Charles

-----Original Message-----
From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf 
Of Paul Gilmartin
Sent: Friday, July 04, 2014 9:27 AM
To: [email protected]
Subject: HLASM doubleword set-symbol arithmetic

On Fri, 4 Jul 2014 11:18:12 -0400, John Gilmore wrote:
>
>..., but z/OS is still full of gaps in
>the facilities it makes available for work above the bar, through many 
>of which one could drive the proverbial tank.
> 
Such as executing code above the bar.  But I've wondered before, and seen no 
comment, whether 31-bit address space constraint is onerous, particularly for 
LPA.  I.e. how often is address space constraint a consideration when deciding 
what should go in LPA for performance reasons otherwise?

>When, for example, the HLASM does not yet support doubleword set-symbol 
>arithmetic . . .
> 
This has been discussed on ASSEMBLER-LIST; one of the infrequent topics on 
which you and I agree.  At that time an IBM employe defended HLASM; I agree 
with half his technical arguments.

o 64-bit expressions would impel incompatible changes in ASMADATA
  format.  Yes; and I'd be not much satisfied with a compatibliity option.

o There would be quiet changes in the results of experession evaluation,
  possibly incompatible with existing art.  Here, I disagree.  HLASM
  responsibly reports as errors overflows in expression evaluation, so any
  (adverse) effect on assembler behavior would be limited to programers
  accustomed to ignoring those errors.  I don't sympathize with them.

  (But I do know one instance in base-displacement resolution where
  HLASM produces results which would be incorrect in AMODE 64.  It's
  so bizarre a case that I doubt any programmers are relying on the
  behavior.  HLASM appears to ignore overflows in base-displacement
  resolution.  I consider this a bug that should be corrected.  But I
  haven't reported it.)

-- gil

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions, send email to 
[email protected] with the message: INFO IBM-MAIN

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to