RE: Compute Date Duration

Simon Coulter wroute:

> > The MI reference specifies that the Data Definitional Attribute 
> > Template is 112-N bytes long,
> > but if I add the lengths of all fields I get 100 bytes as a minimum 
> > length.
> 
> The 100 byte length does not include the end calendar element. I get 
> 116 bytes as the minimum length of a DDAT structure.
> 
>       Fixed portion            32
>       Era table                        50     presuming the SAA era 
> table--different for China or 
> Japan.
>       Calendar table
>               start                    18
>               end                              16
>                                               ---
>                                               116
>                                               ===

Do you think that you need the end element if you specify 0 for the number or 
elements?

> 
> I can't get any valid combination to equal 112 bytes. However the DDAT 
> length is ignored by the instructions so it doesn't matter what value 
> is specified. The parts of the DDAT are found via the offset fields.
> 
> > If I specify 100 bytes for dat 1 _CDD fails because the format of dat 
> > 2 is wrong.

I had forgotten to adjust the offsets. I too get template value not valid now:

 Message . . . . :   Template value not valid for instruction.                 
 Cause . . . . . :   The location of the value is template  with an offset to  
   field in bytes X'0004', an offset in field in bits X'0000', a length of     
   field of 2, and an instruction operand number of 4.  The reason code is     
   X'0000'. If the reason code is X'0000', a reason code may not be available.