Re: 'SIZE attribute of a type

[elroy.jpl.nasa.gov!swrinde!mips!samsung!caen!umeecs!dip.eecs.umich.edu!wa rack@ames.arc.nasa.gov (Christopher Warack)]

In article <5898@email.sp.unisys.com> parkhill@email.sp.unisys.com (parkhill) w
rites:
>>parkhill@email.sp.unisys.com (parkhill) writes: 
>>>Image a compiler was smart enough to understand baised number ranges. 
>>
>>>   type b is range 9 .. 10; 
>>>        or 
>>>   type b is range 2**31 - 2 .. 2**31 - 1;
>>
>>
>>>If the compiler can generate code that only uses 1 bit then b'Size 
>>>should return 1.
>> 
>>Oh, that compilers were so cleverly written! 
>>
>
>> Deleted text.
>
>My point on the compiler having the capability to use one bit is that I
>find the utility of type_name'size is now questionable.  Lets say the
>compiler is capable of using 1 bit but won't do it in most situations.
>How can any programmer use type_name'size to any practical purpose?  It 
>seems less useful than Entry_Name'Count.  However, I am sure that the 
>justification for the change to type_name'size is rock solid.

Seems to me that the only reason you'd want the minimum number of bits
used to represent a type is if you wanted to implement that type with the
minimum number of bits allowed, eg, in a maximally packed record using
a "somewhat" portable rep clause.  (Not the most beautiful piece of code I
can imagine).

If you wanted to know the number of bits in a certain type of object (e.g,
a variable of a type) use object'size -- that's what it's for.

-- Chris

-- 
Christopher A. Warack                   warack@eecs.umich.edu
Graduate Dept, EECS			(313) 665-4789
University of Michigan