PROLOG-REQUEST@SUSHI.STANFORD.EDU (Chuck Restivo, The Moderator) (06/24/87)
PROLOG Digest Wednesday, 24 Jun 1987 Volume 5 : Issue 42
Today's Topics:
Query - Stack Problems & Catch
----------------------------------------------------------------------
Date: 22 Jun 87 20:35:25 GMT
From: Lan See Chen <dalcs!chen@seismo.css.gov>
Subject: Out of global stack problems
I am implementing an Information Retrieval system in CProlog on
BSD 4.3 UNIX. Could anyone tell me how to handle an "OUT OF
GLOBAL/LOCAL STACK" error message while I'm running this IR program
with a large number of recursions. Besides, I would also like to know
if there is any limitations on the number of recursions of a procedure
as well as the length of the list.
Thank you.
------------------------------
Date: Tue, 23 Jun 87 11:53:34 MDT
From: ted%nmsu.csnet@RELAY.CS.NET
Subject: new predicate
I have recently had a need to handle signals from other processes in
Prolog. It seems that this need will become pervasive as people try
to write interface programs of all sorts since it becomes very
important to allow the user to interrupt the current process and
return control to the top-level user interface (which will often not
be the user interface of the underlying Prolog).
The predicate that I would like to see would allow code of the
following form:
top :- catch(kill), loop.
top :- write(escaped).
loop :- loop.
In this code, loop would run until the `kill' signal is received, at
which time, the catch would fail, and the second clause of top would
be executed.
My wish list for the catch predicate includes the following:
a) signals should be named, preferably in a fashion
similar to naming convention for signals in the base
operating system.
b) catches should be nestable, so that when catch either
returns, or fails due to the receipt of the signal, then
the previous handler is made current again. This would
allow the code in loop to catch signals if need be.
What is the opinion of the Prolog community about this predicate?
What similar features currently exist in available Prologs? What
low level facilities are required to implement this capability?
------------------------------
End of PROLOG Digest
********************