with-input-from-file and co.

[R. Kent Dybvig <dyb@iuvax.cs.indiana.edu>]

I have never felt that "with-input-from-file" and "with-output-to-file"
were a good thing, and I would like to see them flushed from R4RS.
Whether or not this happens, I would certainly rather see them omitted
from the standard.  This goes for the proposed "with-input-from-port"
and "with-output-to-port" procedures as well.  I don't think that it is
a good idea to encourage rebinding of these ports, especially if we are
not going to specify what happens when an escape procedure is invoked.
It would mean that every independent subcomputation that may need to
read from or write to the original ports would have to stash a copy of
the port for itself to use explicitly.  Bugs resulting from this sort
of problem are difficult to track down ... especially for those in the
habit of adding "print" statements for debugging purposes.