Elemantary Ada question

Elemantary Ada question

[Reinert Korsnes]

Hi,

Given the following simple test program:

with Text_IO;
use  Text_IO;
procedure Atest1 is
   type E1 is (a, b, c);
   package E1_Io is new Text_IO.Enumeration_Io (E1);
   use E1_Io;

   A : Float;
   begin
    Put(E1'First);
end Atest1;

Can anybody explain me why "A : Float" conflicts
with the declaration:  "type E1 is (a, b, c);" ?

reinert

-- 
Norwegian Polar Institute
Polar Environment Center
N-9296 Tromso
Norway
Fax: +47 77750501

http://geophys.npolar.no/~reinert/personal.html

Re: Elemantary Ada question

[Lutz Donnerhacke]

* Reinert Korsnes wrote:
>Can anybody explain me why "A : Float" conflicts
>with the declaration:  "type E1 is (a, b, c);" ?

Ada is case-in-sensitive.

-- 
	      http://www.tm.oneiros.de/calendar/2001/index.html

Re: Elemantary Ada question

[Reinert Korsnes]

In article <slrn927ai8.tq.lutz@taranis.iks-jena.de>,
 lutz@iks-jena.de (Lutz Donnerhacke) writes:
>* Reinert Korsnes wrote:
>>Can anybody explain me why "A : Float" conflicts
>>with the declaration:  "type E1 is (a, b, c);" ?
>
>Ada is case-in-sensitive.

Ok, let A be a possible value of a variable of type E1
(let say "type E1 is (A, B, C)").
And A is a variable name.  

Where is the conflict ?


-----------------the test program is as follows (repeating):

with Text_IO;
use  Text_IO;
procedure Atest1 is
   type E1 is (a, b, c);
   package E1_Io is new Text_IO.Enumeration_Io (E1);
   use E1_Io;

   A : Float;
   begin
    Put(E1'First);
end Atest1;

reinert





>
>-- 
>	      http://www.tm.oneiros.de/calendar/2001/index.html

-- 
Norwegian Polar Institute
Polar Environment Center
N-9296 Tromso
Norway
Fax: +47 77750501

http://geophys.npolar.no/~reinert/personal.html

Re: Elemantary Ada question

[Ted Dennison]

In article <900bgt$i6t$1@news.uit.no>,
  Reinert.Korsnes@npolar.no wrote:
> In article <slrn927ai8.tq.lutz@taranis.iks-jena.de>,
>  lutz@iks-jena.de (Lutz Donnerhacke) writes:
> >* Reinert Korsnes wrote:
> >>Can anybody explain me why "A : Float" conflicts
> >>with the declaration:  "type E1 is (a, b, c);" ?
> >
> >Ada is case-in-sensitive.
>
> Ok, let A be a possible value of a variable of type E1
> (let say "type E1 is (A, B, C)").
> And A is a variable name.
>
> Where is the conflict ?

If you are saying that in many cases where you would write "A", its type
could theoreticly be figured out by the context of the statement, then
you are right. However, its pretty easy to construct examples where it
couldn't. What matters here to you isn't what's theoreticly possible for
a compiler to do, but what Ada allows you to do.

In Ada, reusing an Identifier for multiple objects can cause one of two
basic situations; hiding or overloading (in truth, there's a few more,
but we'll ignore those situations for now). "Overloading" is where both
objects exist simultaniously and the compiler will figure out which was
intended from context at compile time. Overloading is only allowed for
subprograms (functions and procedures) with different parameter
profiles, and enumeration literals.

Any other time an indentifier is reused, it should cause hiding; the
later declaration "hides" the earlier one for the entirety of its scope.
This implies that they can't be declared at the same scope (otherwise,
there's no point to the earlier declaration).

Now in your case, then enumeration literal "A" could be overloaded, if
your other "A" were a subprogram or enumeration literal. But since your
other "A" is just a variable declation, this must be a hiding situation.
But it can't be a hiding situation because they are both declared at the
same scope. Therefore its illegal.

Now this was a very informal (read: wrong in many details) treatment of
the subject. If you want to go over the actual legaleese rules involved,
see 8.3 of the LRM (available online at
http://www.adapower.com/rm95/arm95_128.html#SEC128 )


--
T.E.D.

http://www.telepath.com/~dennison/Ted/TED.html


Sent via Deja.com http://www.deja.com/
Before you buy.

Re: Elemantary Ada question

[Mats Weber]

If you want to declare an identifier twice in the same declarative
region, _both_ declarations must be either an enumeration literal, a
function or a procedure (i.e. overloadable entities). If one of them is
something else, then it's illegal. The actual rule is a little more
complicated than that, you can have a look at LRM 8.3.

Re: Elemantary Ada question

[Lutz Donnerhacke]

* Reinert Korsnes wrote:
>In article <slrn927ai8.tq.lutz@taranis.iks-jena.de>,
> lutz@iks-jena.de (Lutz Donnerhacke) writes:
>>* Reinert Korsnes wrote:
>>>Can anybody explain me why "A : Float" conflicts
>>>with the declaration:  "type E1 is (a, b, c);" ?
>>
>>Ada is case-in-sensitive.
>
>Ok, let A be a possible value of a variable of type E1
>(let say "type E1 is (A, B, C)").
>And A is a variable name.  
>
>Where is the conflict ?

There are two identical terms in the same namespace.

What should 'if A = A then ...' mean? How to determine the Float value and
the enumeration constant?

  type E1 is (A, B, C);
can be assumed as
  type E1 is range 1 .. 3;
  A : constant E1 := 1;
  B : constant E1 := 2;
  C : constant E1 := 3;

-- 
	      http://www.tm.oneiros.de/calendar/2001/index.html

Re: Elemantary Ada question

[Robert A Duff]

reinert@ola.npolar.no (Reinert Korsnes) writes:

> Can anybody explain me why "A : Float" conflicts
> with the declaration:  "type E1 is (a, b, c);" ?

Enumeration literals are overloadable, but variables are not
overloadable.  You are not allowed to have a non-overloadable A if there
is some other A in the same scope (even if the other one is
overloadable).

The exact rules are in chapter 8 of the RM.  It says that two
declarations with the same name are "homographs" if one or both are
non-overloadable.  Also, if both are overloadable, and have the same
parameter and result types (where the enumeration literal A is like a
parameterless function returning type E1).  It is illegal to have two
homographs in the same "declarative region".

Or perhaps you are asking why was the language defined that way.
Well, variable names could be overloadable, and the language rules
would work just fine.  The above would be legal, and things like

    X: E1 := A;

would resolve to the right A.  Things like "if A = A..." would be
ambiguous, and therefore illegal.

- Bob

Re: Elemantary Ada question

[Scott Ingram]

Reinert Korsnes wrote:

> 
> Can anybody explain me why "A : Float" conflicts
> with the declaration:  "type E1 is (a, b, c);" ?
> 
> reinert
>

Because identifiers in Ada are case insensitive, so
'a : Float' is a redeclaration of an already declared
enumeration type ('type E1 is (A, B, C);')
-- 
Scott Ingram
Vice-Chair, Baltimore SIGAda
Sonar Processing and Analysis Laboratory
Johns Hopkins University Applied Physics Laboratory