On Nov 8,  3:34pm, John Dunlop wrote:
> Subject: Re: Cool "getenv" feature on MPE/iX

There are certain reserved names that are treated as JCWs in an expression
or command line substitution, but are not truely variables, in that they
consume no storage in our variable table:
   system
   fatal
   warn
   ok

These reserved names can have a number concatenated also, like:
   ok444

SHOWJCW shows them, CALC can evaluate them, echo !reserved_name works,
![reserved_name] works, the FINDJCW intrinsic returns them, but SHOWVAR
doesn't show them cuz they aren't really there!

All of the other ways of seeing these reserved names have been
kludged up to work -- to be compatible with MPE V/E back in the
very early releases of MPE XL.

Now, should these reserved names be PRIVATE ? ... :)

Jeff Vance, CSY

> Walter Murray wrote:
> >
> > Richard Trapp RAT ([log in to unmask]) wrote:
> > : Greetings folks,
> > :    I just wanted to pass on a interesting "feature" of the "getenv"
function
> > : which we ran into recently:
> > :   Doing a getenv on a non-existent variable named "SYSTEM" returned a
value
> > : instead of a null string.
> >
> > <snip>
> >
> > Hi, Rich!  Glad you enjoyed this!  Actually, it looks like a "feature"
> > of the underlying intrinsic, HPCIGETVAR.  It surprises me, too.
> >
> > Walter Murray
> > Hewlett-Packard
> > Support Technology Lab
>
> Funny that....also this :
>
> Dev:showvar system
>               ^
> Variable not found in variable table. (CIERR 8106)
> Dev:echo !system
> 49152   <----   I've no idea what this is but I assume it is similar to
> the above
>
> Cheers,
> John
> --
> Work : [log in to unmask]
> Home : [log in to unmask]
> Web  : http://www.tcp.co.uk/~jdunlop/index.html
>
> [STD disclaimer: All thoughts and opinions expressed are my own and do
> not reflect the opinions of any current (or former) employers]
>-- End of excerpt from John Dunlop



--