Newsgroups: comp.org.acm
Path: utzoo!utgpu!watserv1!watmath!watmsg.waterloo.edu!gvreugdenhil
From: gvreugdenhil@watmsg.waterloo.edu (Gord Vreugdenhil)
Subject: Re: Turbo the standard?  I think not.
Message-ID: <1991May8.190631.14173@watmath.waterloo.edu>
Sender: news@watmath.waterloo.edu (News Owner)
Organization: University of Waterloo
References: <?.bGejk$1@cs.psu.edu> <1497@caslon.cs.arizona.edu> <a$2Gg.5#1@cs.psu.edu>
Date: Wed, 8 May 1991 19:06:31 GMT
Lines: 35

In article <a$2Gg.5#1@cs.psu.edu> hvozda@mirkwood.endor.cs.psu.edu (Eric S Hvozda) writes:
>Well I always had to use:
>
>IF condition_not_in_the_case THEN
>   do_what_the_else_would_have
>ELSE
>   case_here
>
>>String in the Turbo syntax is much simpler to use and easier to
>>read.  What's a "packed array"?  Why go to the bother of defining
>
>PACKED ARRAY is to tell the complier to nake the best use of the machines
>memory.  It certian cases like that of an array of BOOLEAN, it would normally
>put one bit per word of storage, but with PACKED, it tells it to use
>all the bits in each word.  However, a trade off of a greater execution time
>may result.
>

Hmm.  I don't have my copy of the Pascal standard handy, but I am pretty sure
that the compiler is allowed to interpret ``PACKED'' in any way it wants --
i.e. it *may* choose a more efficient storage method, it may not do anything.
Pascal doesn't define storage of any structures; any array may be 
"packed" by default and have ``PACKED'' be a nop, or never "pack" elements, 
etc.  In an undergrad/grad class we wrote (most of) an ISO Pascal compiler;
it stored arrays of chars in a packed format independent of the ``PACKED''
keyword.  All other types were the opposite (always unpacked -- including
boolean... bleck) due to time constraints. 

[...stuff deleted...]

>Ack!                             esh101@psuvm.psu.edu             ESH101@PSUVM
>hvozda@vivaldi.psu.edu           hvozda@endor.cs.psu.edu          DoD #0217

Gord Vreugdenhil
gvreugdenhil@watmsg.waterloo.edu
