Re: The QSYGETPH API

[Barbara Morris <bmorris@xxxxxxxxxx>]

Pete Helgren wrote:
> 
> Barbara,
> 
> This is the whole CLLE tamale and it is running without a hitch...
> ...
>    
> (Don't confuse &Error and &Err....I did at first)

I'm not.  When you code a literal parameter of 37 on a CALL in a CL
program, it passes x'000000003700000f'.  If the API is only looking at
the first four bytes, it sees x'00000000'.

> 
> I can't explain why it works then (but I am glad it does because I look like
> a genius!), but what changes would you recommend to prevent a future
> surprise when this "feature" is fixed? (Its V5R3M0 - TL04153  Temporarily
> applied)
> 

It crashes with CPF3C1D on my v5r3 system.  To pass the last two
parameters the way the API wants them, try this:

   DCL        &PwdLen   *INT Len(4) VALUE(10)                    
   DCL        &PwdCcsid *INT Len(4) VALUE(0)  /* job CCSID */   

   CALL       PGM(QSYGETPH) PARM(&USR &PWD &PH &ERR &PwdLen &PwdCcsid)

or if you also need to compile it prior to V5R3:

   DCL        &PwdLen   *CHAR Len(4) 
   DCL        &PwdCcsid *CHAR Len(4) 

   CHGVAR  %BIN(&PwdLen) 10
   CHGVAR  %BIN(&PwdCcsid) 37
   CALL       PGM(QSYGETPH) PARM(&USR &PWD &PH &ERR &PwdLen &PwdCcsid)