Martin Buchholz <martin(a)xemacs.org> writes:
The biggest problem with lstreams is possible garbage-collection
overhead.
Why? If lstreams are properly managed in frob blocks, shouldn't their
"allocation" and "deallocation" be a very cheap operation?
Someone should study java and common lisp streams, and then
implement exposing lstreams to lisp code in a similar fashion.
Java and CL streams are a part of the interface, while lstreams are
more of an implementation helper, aimed at simplifying XEmacs C base.
It's quite reasonable that they differ from whatever Java and CL have.
--
Hrvoje Niksic <hniksic(a)srce.hr> | Student at FER Zagreb, Croatia
--------------------------------+--------------------------------
The meta-Turing test counts a thing as intelligent if it seeks to
devise and apply Turing tests to objects of its own creation.