[plt-scheme] Interacting w/ MzScheme

From: Robby Findler (robby at cs.uchicago.edu)
Date: Fri Dec 10 17:06:10 EST 2004

At Sat, 11 Dec 2004 04:43:17 +0900, Daniel Pinto de Mello e Silva wrote:
> On Fri, 10 Dec 2004 13:24:52 -0600, Robby Findler <robby at cs.uchicago.edu> wrote:
> > That gets you part of the way there but, for example, closing the inner
> > repl doesn't kill any threads or close any network ports it may have
> > started/opened. The paper I posted earlier is worth reading, if you're
> > interested in this.
> Thanks!  It reminds me of Matthew's LL2 presentation with MrEd popping
> out of his slides.
> By the way, speaking of 'require' re-executing modules in new
> namespaces: would it be faster to namespace-attach modules from the
> PLT standard library instead of re-requiring them in DrScheme when the
> user clicks Execute/Run a few times?

A little, probably. I'd guess you'd not notice if the files are
compiled to .zo files and the program is tens of thousands of lines or
less. There is a danger to doing it, tho. Values can leak out between
drscheme and the user's program, which can mean that, if things go
wrong, the Run button might not clear everything out.

Note that DrScheme does currently share (lib "mred.ss" "mred") and
everything it requires because it has the connection to the underlying
GUI toolkit and there can only be one of those connections per
application. So, we have to be careful to make sure there are no ways
that a user program can poke at that library to affect drscheme.


Posted on the users mailing list.