RE: Using API QTWCHKSP

[Scott Klement <klemscot@xxxxxxxxxxxx>]

I'm not sure that IBM would change the data structures in QSYSINC to use
a pointer... I rather doubt they would, especially since it would break
backwards compatability...

But... there's a lot of things that could/should be done to QSYSINC:

   --  Use less cryptic names.  I mean, I try to make my programs
         easy to read.  When someone goes to change my program, and
         sees "EZHQEP" they don't automatically think "data queue exit
         program!"  or "QMHMML00" doesn't exactly scream "maximum
         message length".

       I never use these header files because they make my programs
       into an incomprehensible mess that nobody wants to work on.

   --  Give prototypes, constants and structure defs for the UNIX-type
         APIs... and give them the same names as their C counterparts.

I strongly agree about the "B" data type.  You should get a compiler
warning every time you use it that says "The B data type is deprecated!
Consider using the I or U data type."


On Tue, 5 Feb 2002, David Morris wrote:

> Buck,
>
> It sure would be more useful if IBM fixed those definitions. I would
> also like to see the structures in qsysinc be based on a pointer.
> It wouldn't take much of a program to generate them in place
> with conditional compile directives to select a version.
>
> David Morris
>
> >>> Buck.Calabro@commsoft.net 02/05/02 02:52PM >>>
>
> I couldn't agree more.  Alas, the system includes all use "B"!
> (wrkmbrpdm
> qsysinc/qrpglesrc)  Anybody looking there for examples will have to do
> the
> conversion.  Also, some manuals still use "B" instead of "I" for their
> examples.
>   --buck