Number: 298

Date: 26-Mar-84 15':05':59

Submitter: Sannella.PA


Subject: POSSIBILITIES screwing up stack

Lisp Version: 

Description: '
Received': from USC-ECL.ARPA by PARC-MAXC.ARPA; 26 MAR 84 08':48':11 PST'
Return-path': <@USC-ECL.ARPA':FIKES@ECLD>'
Received': from ECLD by ECLA with ECLnet; Mon 26 Mar 84 08':48':39-PST'
Date': Mon, 26 Mar 84 08':48':03 PST'
From': Richard Fikes <FIKES@ECLD.#ECLnet.ARPA>'
Subject': Problems With POSSIBILITIES'
To': vanmelle.PA'
cc': FIKES@ECLD.#ECLnet.ARPA, lispsupport.PA'
	I am using the possibilities list interface to generators and'
am getting frequent display panel stops (i.e., 9xxx) that usually'
cause loss of my virtual memory.  For example, yesterday I could cause'
a 9016 each time TRYNEXT was called in one of my functions.  We have'
very litle information as to what the display panel codes mean, so I'
don''t know what to make of that.  I suspect the possibilities list'
functions may be screwing up the stack somehow, perhaps in an'
interaction with the process mechanism.  I would greatly appreciate it'
if you could take a quick look at POSSIBILITIES, TRYNEXT, AU-REVOIR,'
and ADIEU to see if you see any problems.  I will try to get a simple'
reproducable case of failure, but that may be difficult and there may'
be something going on that would be relatively obvious to your eye.'
Received': from USC-ECL.ARPA by PARC-MAXC.ARPA; 29 MAR 84 17':27':22 PST'
Return-path': <@USC-ECL.ARPA':FIKES@ECLD>'
Received': from ECLD by ECLA with ECLnet; Thu 29 Mar 84 17':27':43-PST'
Date': Thu, 29 Mar 84 17':26':49 PST'
From': Richard Fikes <FIKES%ECLD@ECLA>   '
Subject': Stacks and Generators'
To': lispsupport.PA'
cc': FIKES@ECLD.#ECLnet.ARPA, Masinter.PA, VanMelle.PA'
This is a followup to my earlier messages regarding POSSIBILITIES.  I'
have found bugs in my program that apparently were causing me to lose'
my virtual memory due to stack problems.  Hence, I have no reason to'
believe there are maicrocode problems regarding generators.  However,'
because generators tie up so much stack space, they are of very'
limited use.  So, I would like to reiterate my request for'
modifications to POSSIBILITIES, GENERATOR, and COROUTINE to allow'
creation of generators and coroutines with an empty stack (as with'
processes).  That option would make them a much more useful part of'
the programming environment.'
Bill and Larry, thanks for the help in analyzing the situation.'


Test Case: 

Edit-By: Sannella.PA

Edit-Date: 30-Mar-84 15':44':52


Assigned To: 



System: Language Support

Subsystem: Stack and Interpreter



Microcode Version: 

Memory Size: 

File Server: 

Server Software Version: 


Frequency: Intermittent

Impact: Moderate

Priority: Hopefully

Status: Open

Problem Type: Bug

Source Files: