Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!floyd!harpo!seismo!hao!hplabs!sri-unix!arbab%USC-CSE@USC-ECL.ARPA
From: arbab%USC-CSE@USC-ECL.ARPA
Newsgroups: net.micro.pc
Subject: IBM Pascal Problems
Message-ID: <17591@sri-arpa.UUCP>
Date: Mon, 12-Mar-84 19:43:00 EST
Article-I.D.: sri-arpa.17591
Posted: Mon Mar 12 19:43:00 1984
Date-Received: Mon, 19-Mar-84 07:43:18 EST
Lines: 46


I am having problems with the IBM Pascal 1.0 and would appreciate
any hints, pointers, etc.
	
     1. I have a program that goes through pass 1 without
	any errors or warnings, but gives a "compiler internal
	error" in pass 2.
	The only things in the program that may be considered
	"funny" are use of units and a bunch of string manipulation
	procedures (standard), which should not matter at compile
	time anyway.

	I tried juggling the program around, commenting out segments
	of source code, systematically deleting source code, replacing
	procedure bodies with single writeln statements, and never
	got through the second pass. If I am lucky, I get one of these
	error numbers: 100, 101, 116, 180, 186, etc., otherwise all
	I get is "Error: Compiler Internal Error, Contact Technical
	Support".

	I finally took out all unit, extern, public, etc., reduced it
	to a very simple program, and then cut it down to the point
	that I got it to compile by deleting one of my variables and
	the couple of statements referring to it. Deleting the same
	variable from the original program does not work, though.

   	I do not believe it is the length of the program (about 10
	pages of source listing) as I have successfully compiled
	larger ones. I think I am somehow pushing the compiler
	beyond one of its internal size limits (too many variables
	with birthdays in October, perhaps...) but after wasting
	a perfect weekend, I could not see any pattern to what
	bothered the compiler. WHY ISN'T THERE ANY MENTION OF SUCH
	INTERNAL ERRORS AND THEIR CAUSES IN THE MANUAL?

     2. It seems like both passes of the compiler and the linker
	are "hard-wired" to run from the A diskette. This is very
	annoying when you have a hard-disk or extra memory to use
	as RAM-disk and you cannot transfer Pascal to run from
	either. Is there any way to make the compiler run from
	another disk? Maybe there have been discussions on this
	issue before, but, again, WHY ISN'T THERE ANY EXPLICIT
	MENTION OF THIS IN THE MANUAL?


Farhad Arbab