(j3.2006) Note 8.30 in F2008

[Malcolm Cohen]

>While we're looking at this section, another question about converting
><stop-code> values to process exit codes - what if some images execute
>STOP and others execute STOP 1  (or some other inconsistent
>combination)?  Right now we (Cray) tell people that they cannot rely on
>the exit status value if they have different stop codes.

Your advice is fine and good.

The standard says the stop code is per image, not per program execution.
The term "process exit status" is also clearly a per-image concept.

We could say something about the "whole program exit status", but it would
just be so many words with little use or effect, I think we should leave it
up to the vendor how this inherently processor-dependent thing is handled.

Cheers,
-- 
.....................Malcolm in Delft.