Gil again:
>Do I remember correctly that it used to?  Assembler VS?
>Assembler H?

Correct; all predecessors of HLASM, and HLASM before R3.

>I was told back then that (part of) the
>nominal value was used if padding was necessary for
>alignment?

The nominal value has no effect on padding.

>What does HLASM do in this case: Leave the
>padding bytes undefined, as DC would do?  Pad with
>an assembler-selected pattern?  Other (but what)?

The rule is that if a DC needs padding and any nonzero bytes were generated
prior to the DC, pad with binary zeros. If a DS needs padding, or if the DC
was not preceded by nonzero bytes, leave a gap.  (This always starts a new
object module record.)

>If the programmer specifies the (optional) nominal value,
>will that be used as the padding pattern?

No. Padding within constants depends on programmer coding; but padding of
skipped bytes does not.

John Ehrman

Reply via email to