Re: RE: Binary vs. Integer Parameters -- MIDRANGE-L

Brad,

By using IBM's supplied definitions, you are have a much better chance things 
will continue to work on future releases.  Worst case, just a recompile, 
instead 
of searching for a field that was changed, etc.  The names are not great, but 
they 
work, and overall, I think they save time.  Just getting a guaranteed match for 
each variables attributes is great.  It cost me about three days when I missed 
the 
little bit about optional parameters in the API programmers guide.  They must 
be 
coded as *OMIT or you will have very sporadic, seemingly random problems.  
After talking with tech support, I know we are not the only ones who made this 
error.  We were also billed several hours of research time by IBM, so it is not 
common knowledge.  Also, pointers need to be passed by VALUE to 
several of the APIs.  It is not spelled out why in any manual I have read.  It 
just 
works that way.  That is why I would like to see the prototypes included.

David Morris

>>> "Stone, Brad V (TC)" <bvstone@taylorcorp.com> 11/18 3:33 PM >>>
I thought these were available on the system somewhere already.  I know our
shop hasn't installed them, but I know I've heard people speak of them.  The
only problem is the variable names are IBM, so if it's something like file
name, the variable will be similar to QFJFLANAME.. ;)

What we do is simply only code the variables we want to use using FROM and
TO posistions, and name them whatever we want... like "FileName".  I think
it's easier in the long run.

Bradley V. Stone
Taylor Corporation - OASIS Programmer/Analyst   
bvstone@taylorcorp.com 


> -----Original Message-----
> From: Michael Calabro/commsoft [SMTP:mcalabro@commsoft.net] 
> Sent: Wednesday, November 18, 1998 3:04 PM
> To:   MIDRANGE-L@midrange.com 
> Subject:      Re: Binary vs. Integer Parameters
> 
> I'd like to second David's motion.  It'd be exquisite to have structures 
> for every API!
> Buck Calabro
> 
> 
> >>> <dmorris@plumcreek.com> 11/18/1998 14:22 >>>
> ..That sounds like a very good idea.  Currently we define a lot of space 
> that is never used, just to get a base definition.  It would also be nice 
> to find structures for every API/exit.  I would also like to see you 
> include 
> prototypes.
> 
> >>> <bvining@VNET.IBM.COM> 11/18 11:02 AM >>>
> 
> ..Due to all the wonderful enhancements to ILE
> RPG over the last several releases I think it would be nice to update
> the RPG QSYSINC includes though that would probably require using a new
> source file as incompatible changes would occur if the current ILE RPG
> source was simply replaced by more current techniques (Integers, Based
> structures rather than storage allocating structures, etc.).
> 
> Bruce Vining


+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---