Andy Wingo scripsit:

> It is useful to distinguish exceptional exits from general nonlocal
> control flow.  R7RS should specify that the port is closed on
> exceptional exits, but not on other exits.  It is possible to implement
> this with exception handlers.

An exceptional exit is simply a non-local exit from an exception handler.
It's not particularly obvious when this is happening.

> How does this square with `guard', whose cond clauses are specified as
> being evaluated in the dynamic environment of the guard expression?  I
> still think this is an error.  Giving intervening dynamic-wind
> expressions a workout is not something I want to happen when exceptions
> are raised.

`Guard` works by capturing its continuation and setting a handler that
captures the continuation at the point of raise.  The guard continuation
is then restored, which gives us the outer dynamic environment, and the
cond-clauses are executed.  If we fall off the end of the cond-clauses,
the point-of-raise continuation is restored (giving us the inner dynamic
environment) and the condition object is reraised.

--
That you can cover for the plentiful            John Cowan
and often gaping errors, misconstruals,         http://www.ccil.org/~cowan
and disinformation in your posts                cowan@ccil.org
through sheer volume -- that is another misconception.  --Mike to Peter

_______________________________________________
Scheme-reports mailing list
Scheme-reports@scheme-reports.org
http://lists.scheme-reports.org/cgi-bin/mailman/listinfo/scheme-reports