Re: Environment Variables in RPGLE

Chuck,

I can't seem to locate the specific PTF, but before leaving IBM I found that CL was picking up residual values for uninitialized pointers. I reported the problem and CL was fixed so that a declared pointer, that is not set to a value, is initialized to NULL. So for ILE CL I simply declare a &NULL_PTR TYPE(*PTR) and use it for NULL comparisions.

Bruce


CRPence <crp@xxxxxxxxxxxxxxxxxxxx> wrote:
I did not find in the API doc for Qp0zGetEnv [could not find it by
search, only by API list] any sizes, but the CL commands suggest in the
help text that ADDENVVAR limits _environment variable name_ to a maximum
of 128 bytes in length and limits _value_ to a maximum of 1024 bytes in
length." Are the real maximums consistent with those program sources?

My CL version of RTVENVVAR using getenv() shares the same limits as
the above commands. The *PTR type was problematic for me [as Scott
alluded] because there is no documented manner [not that I could find]
to test for a *NULL, except the proper age-old MCH3601 monmsg\rcvmsg;
seems that if I was given a TYPE(*PTR), I really should also be able to
compare to %addr(*NULL)??

FWiW regarding the *PTR: In testing I did notice that the null
pointer was always getting set to an /AF/ [so called; another I am more
familiar with is the /FB/ null], so for laziness, and to see if it might
ever give a different result, I do not monitor a mch3601 so that the the
possibility of anything other than an /AF/ pointer is manifest as a *FC
so the failed assumption is obvious.

Regards, Chuck