At 10:12 PM 12/3/2001 -0500, James Mastros wrote:
>On Mon, 3 Dec 2001, Dan Sugalski wrote:
> > I'm getting tempted to have some sort of multi-level ENV thing that, for
> > most single-interpreter cases, collapses down to a plain getenv/putenv.
>Umm, is everybody here somking crack?
>
>I thought the point of %ENV was to deal with the environment, the OS's
>poor excuse for a hash that was useful for taking persistant parameters
>from the shell.

Except when you embed the interpreter in a persistent server, another app, 
or start firing up safe interpreters.

The fact that the environment is such a poor excuse for a lot of things is 
a reason that something may want to get between a non-top-level interpreter 
and the real environment. If you invoke an interpreter in any way other 
than "perl foo.pl" there's a reasonable potential to want to step between 
the interpreter and the outside world.

>I didn't think that it was an advanced IPC mechinisim.  I didn't think
>anything that even resembled IPC was any of core.ops & pals domain.

It's not. The question is how do we make 'normal' %ENV access and 
potentially more complex 'environment' access work as a reasonably unified 
whole? If we can't, we punt, and that's fine.

                                        Dan

--------------------------------------"it's like this"-------------------
Dan Sugalski                          even samurai
[EMAIL PROTECTED]                         have teddy bears and even
                                      teddy bears get drunk

Reply via email to