Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!hellgate.utah.edu!helios.ee.lbl.gov!ucsd!ucsdhub!hp-sdd!hplabsy!hp-ses!hpcuhb!hpcllla!hpclisp!hpcljms!brians
From: brians@hpcljms.HP.COM (Brian Sullivan)
Newsgroups: comp.lang.pascal
Subject: Re: Pascal sets
Message-ID: <950032@hpcljms.HP.COM>
Date: 22 Jan 90 19:49:00 GMT
References: <11866@goofy.megatest.UUCP>
Organization: Hewlett-Packard Calif. Language Lab
Lines: 28


  Most of the Pascal compiliers that I know which implement the above set
constructs use one of these two methods.

  Method 1.  Used by UCSD Pascal is to require that sets by limited to 256
elements or less.  They choose 256 because it allow them to use set of char
in the compiler.  I belive that it would reject the use of a set constant
that contained an element outside the range 0..255.  In this implementation
all integer based sets are represent by a bit array of 256 bits. 


  Method 2. This method rejects any overloaded set constant.  Instead each
set constant must be qualified by the type it is a member of.  For example
your statement:

>     if([i..j] <> [j] ) then
>        writeln('How big are constructed sets?');

 would have to be written


>     if(t35([i..j]) <> t35([j]) ) then
>        writeln('How big are constructed sets?');

  For references on this method see Oregon Pascal, Microsoft Pascal and 
Meridian Pascal.

                -- Brian --