At 12:49 PM 1/4/01 +0000, Simon Cozens wrote:
>On Wed, Jan 03, 2001 at 11:34:04PM -0500, Uri Guttman wrote:
> > >>>>> "DS" == Dan Sugalski <[EMAIL PROTECTED]> writes:
> > DS> A subsystem is, in this instance at least, a piece of perl that
> > DS> can more or less stand alone, but cross modules
>...
> > DS> Safe signals live pretty much entirely inside the interpreter
> > DS> piece, so don't really count as a separate subsystem.
> >
> > so you can't isolate safe signals easily.
> >
> > anyhow, that is my $.02. i do see this modular approach being a big win
> > but there are some things in perl which don't fall into one module. we
> > have to allow for that in special cases.
>
>Sounds like violent agreement to me. Incidentally, how about the dynamic
>scoping mechanism as another subsystem? You might think ENTER and LEAVE are
>ugly, but if we need dynamic scoping in C (which is highly likely) then we
>need some scoping mechanism to replace them.
Well, that's really part of the interpreter, or so I'd think. Though, come
to think of it, it might not be given the way it works now. (Which is
clever, but more than a little mind-bending on first glance)
I'm also thinking of separating the three scope effects--enter/leave
lexical scope, enter/leave dynamic scope, and enter/leave control
scope--into separate actions under the hood, though how much that gets
exposed is still up in the air, and mostly Larry's call. (Conditionally
closing a scope on exiting the block the scope is attached to is also
something I'm thinking about, but I'm not sure that we want to go there)
Dan
--------------------------------------"it's like this"-------------------
Dan Sugalski even samurai
[EMAIL PROTECTED] have teddy bears and even
teddy bears get drunk
