Re: Advice passing a large parameter

On 2019-04-15 10:56 AM, Jon Paris wrote:

Chris is correct but I would also suggest you consider getting rid of the Char(15) length on the DS - it serves no purpose - other than possibly to confuse people and give you one more thing to change should the DS get larger.

As to the "best" place. Personally I would place the DS template in one of two places depending on where it is used.

If it is widely used, even when not being used as a parm, then include it in a copy source for "standard data types" and include that in all programs that need it. Since a template takes up no memory there is no harm in having hundreds of templates even if you only use one.

If it is always used as a parm then I would include it in the same source member as all the prototypes.

As a rule of thumb, any definitions that the prototype depends on should also be in the same copy file.

I would define anything in a copy file as a template, and I would use some naming convention such as ending with _t or starting with template_, to indicate that it is a template. And I would define all template data structures as qualified.

So your copy file would look like this:

dcl-ds plist1_t extname('PLIST1DS') qualified template inz;
end-ds;

dcl-pr VS_RateShop char(15) ExtProc('VS_RateShop');
inPlist1 likeds(plist1_t) const;
end-pr;

And ideally, you would define any other PLIST1's in my code as being LIKEDS(plist1_t). But it would be non-trivial to change any existing definitions of PLIST1, since defining them as LIKEDS would make them qualified, which would mean you'd have to change all the code using the subfields.