Untyped "constant"?

Untyped "constant"?

[cis.ohio-state.edu!zaphod.mps.ohio-state.edu!uakari.primate.wisc.edu!aplc]

I have a question about Ada.  It has to do with "named numbers" and 
their subsequent use.

a) Suppose I have a compiler/RTE in which integer'last is 32_767.
b) And suppose the system has a long_integer type in which
   long_integer'last = 2_147_483_647.
c) And suppose I have the program:
          procedure Named1 is
             X : constant := integer'last+5;
          begin
              for j in 1..X loop
                  null;
              end loop;
          end Named1;

d) Finally, suppose the "loop" variables are really used. E.g., by
   disabling optimization when I invoke the compiler.  


Then:  is this a legal program or not?


Thanks,
Mars Gralia



DISCUSSION

Some compilers, admittedly archaic or for very small computers, 
will raise a Constraint_Error, thereby forcing me to say:
             X : constant long_integer := integer'last+5;

I would prefer the program given above, because it would appear to be 
more portable.
     

Re: Untyped "constant"?

[agate!spool.mu.edu!caen!uakari.primate.wisc.edu!aplcen!ddsdx2.jhuapl.edu!]

In <1991Sep23.164446.13412@aplcen.apl.jhu.edu> gralia@ddsdx2.jhuapl.edu (Mars J
. Gralia) writes:


>I have a question about Ada.  It has to do with "named numbers" and 
>their subsequent use.

>a) Suppose I have a compiler/RTE in which integer'last is 32_767.
>b) And suppose the system has a long_integer type in which
>   long_integer'last = 2_147_483_647.
>c) And suppose I have the program:
>          procedure Named1 is
>             X : constant := integer'last+5;
>          begin
>              for j in 1..X loop
>                  null;
>              end loop;
>          end Named1;

>d) Finally, suppose the "loop" variables are really used. E.g., by
>   disabling optimization when I invoke the compiler.  


>Then:  is this a legal program or not?
Nope.  The declaration
  X : constant := integer'last+5;
is illegal.   The expression given for the initialization of a number 
declaration must always be an expression of universal type.  The attribute 
integer'last returns an integer, not a universal_integer.  Therefore, the
result of the addition will resolve to integer.  Universal types must be made
up entirely of expressions involving only other universal types (literals and
other named numbers).

--Thor


>DISCUSSION

>Some compilers, admittedly archaic or for very small computers, 
>will raise a Constraint_Error, thereby forcing me to say:
>             X : constant long_integer := integer'last+5;

This, being a constant declaration and not a number declaration should
be fine if you convert the result of 'last to long_integer:

  X : constant long_integer := long_integer(integer'last) + 5.

>I would prefer the program given above, because it would appear to be 
>more portable.
>     

--Thor
collard@capsrv.jhuapl.edu

Re: Untyped "constant"?

[John Goodenough]

In article Untyped "constant"? of 23 Sep 91 16:44:46 GMT
gralia@ddsdx2.jhuapl.edu (Mars J. Gralia) writes: 

>a) Suppose I have a compiler/RTE in which integer'last is 32_767.
>b) And suppose the system has a long_integer type in which
>   long_integer'last = 2_147_483_647.
>c) And suppose I have the program:
>          procedure Named1 is
>             X : constant := integer'last+5;
>          begin
>              for j in 1..X loop
>                  null;
>              end loop;
>          end Named1;

>Then:  is this a legal program or not?

The named number declaration is illegal, because INTEGER'LAST is of type
INTEGER (see 3.5(9)) and the initialization expression for a named number must
have a universal type (see 3.2.2(1)).  So suppose you fix this problem by
writing:

	X : constant := INTEGER'POS(INTEGER'LAST) + 5;

Note that INTEGER'POS returns a value of type universal integer (see 3.5.5(6)),
INTEGER'LAST is a static expression (4.9(8)), and INTEGER'POS(INTEGER'LAST) is
static (4.9(8)).

But this doesn't solve the problem, since

	for J in 1..X loop

is now equivalent to

	for J in 1..32_767+5 loop

The upper and lower bounds of this range will be converted to the predefined
INTEGER type (3.6.1(2)), and this conversion will raise CONSTRAINT_ERROR.



John B. Goodenough					Goodenough@sei.cmu.edu
Software Engineering Institute				412-268-6391

Re: Untyped "constant"?

[Norman H . Cohen]

In article <1991Sep23.164446.13412@aplcen.apl.jhu.edu>,
gralia@ddsdx2.jhuapl.edu (Mars J. Gralia) writes:

|> a) Suppose I have a compiler/RTE in which integer'last is 32_767.
|> b) And suppose the system has a long_integer type in which
|>    long_integer'last = 2_147_483_647.
|> c) And suppose I have the program:
|>           procedure Named1 is
|>              X : constant := integer'last+5;
|>           begin
|>               for j in 1..X loop
|>                   null;
|>               end loop;
|>           end Named1;
|>
|> d) Finally, suppose the "loop" variables are really used. E.g., by
|>    disabling optimization when I invoke the compiler.
|>
|>
|> Then:  is this a legal program or not?

As Dave Collard points out, Integer'Last+5 is of type Integer, and the
initial-value expression for a named number must be of type
universal_integer, so this is not legal.  However, we can get around this
technicality by rewriting the initial-value expression as

   Integer'Pos(Integer'Last)+5

since the 'Pos attribute yields a universal_integer result.

However, RM 3.6.1(2), when deciphered, states that a discrete range in
which each bound is either a named number, an integer literal, or an
attribute with a universal_integer value has each bound implicitly
converted to type Integer.  Thus 1 .. X is equivalent to
Integer(1) .. Integer(X).  Since X has a value greater than Integer'Last,
the conversion in the upper bound raises Constraint_Error.

|> Some compilers, admittedly archaic or for very small computers,
|> will raise a Constraint_Error, thereby forcing me to say:
|>              X : constant long_integer := integer'last+5;
|>
|> I would prefer the program given above, because it would appear to be
|> more portable.

Well, it's portable in the sense that it is guaranteed to raise
Constraint_Error (except when the loop is optimized away, in which case
the implicit conversion that would have raised Constraint_Error can be
eliminated pursuant to RM 11.6(7)).

A more useful form of portability would come from defining your own
integer type with the required range:

     procedure Named1 is
         type Loop_Index_Type is range 1 .. Integer'Last+5;
     begin
         for J in Loop_Index_Type loop
             null;
         end loop;
     end Named1;

Implementations for which Integer'Last = System.Max_Int will reject this
at compile time.  Implementations for which System.Max_Int >=
Integer'Last+5 are guaranteed to execute without an exception.

Re: Untyped "constant"?

[cis.ohio-state.edu!zaphod.mps.ohio-state.edu!rpi!bu.edu!inmet!stt]

Re: Untyped "constant"? by gralia@ddsdx2.jhuapl.edu

> I have a question about Ada.  It has to do with "named numbers" and 
> their subsequent use.
> 
> a) Suppose I have a compiler/RTE in which integer'last is 32_767.
> b) And suppose the system has a long_integer type in which
>    long_integer'last = 2_147_483_647.
> c) And suppose I have the program:
>           procedure Named1 is
>              X : constant := integer'last+5;

This is not legal Ada, since the type of integer'last+5 is integer,
not universal-integer.  This can be made legal as follows:

               X : constant := Integer'POS(Integer'LAST) + 5;

This works because int-type'POS returns the value of its operand
converted to univ-integer type.

>           begin
>               for j in 1..X loop

This will not work in general, because when both bounds can
be interpreted as universal integer (as above), the 
type of J will be STANDARD.INTEGER.  Therefore, you should
get CONSTRAINT_ERROR on this line.

>                   null;
>               end loop;
>           end Named1;
> 
> d) Finally, suppose the "loop" variables are really used. E.g., by
>    disabling optimization when I invoke the compiler.  
> 
> 
> Then:  is this a legal program or not?

No, it is not legal as written (see above for correction).

> Thanks,
> Mars Gralia
> 
> 
> 
> DISCUSSION
> 
> Some compilers, admittedly archaic or for very small computers, 
> will raise a Constraint_Error, thereby forcing me to say:
>              X : constant long_integer := integer'last+5;

This is also illegal, because, again, the type on the right
is Integer, not Long_Integer.  It could be made legal as follows:

              X : constant Long_Integer := Long_Integer(Integer'LAST) + 5;

In any case, *all* compilers should raise CONSTRAINT_ERROR
if you say "for J in 1..X" when X is a universal-integer > Integer'LAST.

By the way, "X : constant := ..." must *never* raise constraint error
on the assigment, though some subexpression on the right hand side
might raise constraint error, presuming it is inside a 'POS (as above).
All static univ-integer calculations must be performed to arbitrary
range (no overflow allowed).

> I would prefer the program given above, because it would appear to be 
> more portable.

By using 'POS, you can make the declaration of X work without
constraint error, but you will still get a constraint error
on the loop unless you specify a type for J that includes
the low and high bound.  Furthermore, there exist compilers where 
Integer'LAST is equal to SYSTEM.MAX_INT, and in that case
you are out of luck.

S. Tucker Taft    stt@inmet.inmet.com
Intermetrics, Inc.
Cambridge, MA  01238