Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ames!elroy!orion.cf.uci.edu!uci-ics!ics.uci.edu!milne
From: milne@ics.uci.edu (Alastair Milne)
Newsgroups: comp.lang.pascal
Subject: Re: Enumerated types in TP40
Message-ID: <7250@paris.ics.uci.edu>
Date: 11 Feb 89 12:10:38 GMT
References: <363@lafcol.UUCP> <950016@hpcllmr.HP.COM>
Sender: news@paris.ics.uci.edu
Reply-To: Alastair Milne <milne%ics.uci.edu@orion.cf.uci.edu>
Distribution: usa
Organization: University of California, Irvine - Dept of ICS
Lines: 28

In article <950016@hpcllmr.HP.COM> mark@hpcllmr.HP.COM (Mark Rozhin) writes:
>>I believe something like sizeof(Enum1) div sizeof(Enum1(ord(One))) will do the
>>trick, but I'm unable to test this right now...
>
>what does sizeof( <constant> ) mean? 

    The same as "sizeof(anything-else)": the amount of memory, typically in
    bytes, required to store that value; or if the argument is a type, then
    the amount needed for a value of that type.

    I'm not quite following the proposed use of "sizeof" here.  On the one
    hand, it looks as if it's being used to measure the ordinality (right
    term?) of an enumeration -- which it won't do.  And the expression 
    "sizeof(Enum1) div sizeof(Enum1(ord(One)))" looks as if it should yield
    one, since the arguments being passed to "sizeof" are equivalent.

    I have yet to encounter a way dynamically to determine the number of
    values in an enumeration, short of having a dialect which provides a 
    MAX function.  A fix that's not too hard to use is to have a constant
    "lastValue" declared as the final value in the enumeration.  Besides
    providing an upper limit for scans, its ordinal is the number of values in
    the type, less 1.  Of course, this takes advantage of TP4/5's relaxed
    ordering and counting of CONST, TYPE, and VAR sections.

    If anybody has a better way, I'd be interested to hear.


    Alastair Milne