Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!husc6!rutgers!ucla-cs!zen!ucbvax!ANDREW.CMU.EDU!mw22+
From: mw22+@ANDREW.CMU.EDU (Michael Alan Wertheim)
Newsgroups: comp.sys.apple
Subject: Re: Kyan Pascal and Programming Toolkits
Message-ID: <oVG2Exy00Uo5gaE0YD@andrew.cmu.edu>
Date: Fri, 11-Sep-87 12:26:37 EDT
Article-I.D.: andrew.oVG2Exy00Uo5gaE0YD
Posted: Fri Sep 11 12:26:37 1987
Date-Received: Sat, 12-Sep-87 22:28:44 EDT
References: <3791@ihlpg.ATT.COM>
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The ARPA Internet
Lines: 60

I use an old, old (pre-2.0) version of Kyan Pascal.  I haven't seen any of
the newer versions, so some of these bugs may have been fixed  (I would
hope).

1.)  The NEW procedure doesn't work.  No matter what kind of structure you
try to allocate memory for, it always gives you 400 bytes.  For example, the
following programs do exactly the same thing:


PROGRAM P1;

TYPE PTR:^NODE;
     NODE:RECORD
	    A:BOOLEAN
	  END;

VAR P:PTR;
    I:INTEGER;

BEGIN
    FOR I:= 1 TO 100 DO NEW (P)
END.



PROGRAM P2;

TYPE PTR:^NODE;
     NODE:RECORD
	    A:ARRAY [1..100] OF ARRAY [1..100] OF REAL
	  END;

VAR P:PTR;
    I:INTEGER;

BEGIN
    FOR I:= 1 TO 100 DO NEW (P)
END.


Both programs crash at I = 100 or so.  Even if a boolean value takes up a
whole word of memory, I should be up in the thousands before Program1
crashes.  A single node form Program2 is too big to fit into memory, but it
seems like you can create 100 of them before you run out of memory.

2.) Very long string literals (including comments) will crash the compiler.
i.e. you try to compile a program and the compiler crashes into the monitor.


As I said before, these bugs have probably been fixed.  No flames form
current Kyan Pascal users please.


Michael Wertheim
Carnegie Mellon University
Pittsburgh, PA

Arpa: mw22@andrew.cmu.edu
Bitnet: mw22@cmuccvma
UUCP: ...!{seismo, ucbvax, harvard}!andrew.cmu.edu!mw22#