On 2015-12-29, at 06:59, Peter Relson wrote:

> FWIW,
> 
> Adding at the beginning of the module
> *PROCESS SECTALGN(16)
> 
> makes this work. I always think of needing SECTALGN(16) when there's a 
> quadword-aligned item (the default for a DSECT is doubleword aligned). I 
> don't know if my thought is fully correct.
> 
> Of course "Statement is unresolvable" is not a good diagnostic if that is 
> the situation.
>  
Oh, this is tricky.  If you're envisioning a new diagnostic to be issued
whenever the alignment requirement of any component exceeds the required
alignment of the section, Ed's two examples show the detection is
inconsistent:

On 2015-12-28, at 17:44, Ed Jaffe wrote:

> Check this out...
> 
> Try to assemble the following test program. Attempts to use AIALENTH as a 
> duplication factor fail with 'ASMA080E Statement is unresolvable' while 
> BIALENTH works just fine. Why?
> 
> AIADSECT DSECT ,
> AIAVRS   DS    0CL512
>         DS    32LQ
> AIALENTH EQU   *-AIADSECT
> 
> BIADSECT DSECT ,
> BIAVRS   DS    0CL512
>         DS    32XL16
> BIALENTH EQU   *-BIADSECT
> 
> TEST     CSECT
>         DC    (AIALENTH)X'00'
>         DC    (BIALENTH)X'00'
>         DC    (AIALENTH)X'00'
>         DC    (BIALENTH)X'00'
>         END

-- gil

Reply via email to