Re: Sticking diagnostic information (up to 16k) where calling program can get it?

[Vernon Hamberg <vhamberg@xxxxxxxxxxxxxxx>]

James - you can use puts with a user space, too, I believe - just use 
the full IFS path name. It behaves like a stream file, after all.

Not sure if that helps, however!!

Vern

On 11/13/2014 12:17 PM, James H. H. Lampert wrote:
> I've currently got a situation where program A calls program B (both 
> ILE RPG).
>
> If program B gets an error, it generates up to 16k of diagnostics 
> before returning a "failed" return code to program A. It currently 
> uses a call to "puts()" from the ILE C runtime to send those 
> diagnostics to STDOUT (either the "red function key" screen for 
> interactive jobs, or a job-global spool file for server jobs.
>
> If Program A gets a "failed" return code from program B, it then adds 
> a few lines of its own to STDOUT, again with a call to the c "puts()," 
> which then seemlessly integrate with the diagnostics from program B. 
> Looking at either the "red function key" screen or the spool file, you 
> can't tell that the contents were written by two different programs.
>
> BUT NOW, I've had a request to put that diagnostic information in 
> other places, including possibly emailing them (without the need to 
> have MAILTOOL, MMAIL, or anything else installed).
>
> Dispatching the complete diagnostics, when program A has them, isn't 
> the problem. I'm currently trying to figure out a nice, 
> easy-to-implement, place for program B to stick its diagnostics (up to 
> 16k) where program A can get at them (without adding another parameter 
> to the call), and my sleep-deprived brain (Academy of Ancient Music at 
> Disney Hall, last night) doesn't want to cooperate.
>
> I can of course fall back on creating a USRSPC, but I was hoping for 
> something simpler.
>
> --
> JHHL