Primitive subprogram for a subtype?

Primitive subprogram for a subtype?

[Victor Porton]

Is it possible to define a primitive subprogram for a subtype rather than 
the base type?

-- 
Victor Porton - http://portonvictor.org

Re: Primitive subprogram for a subtype?

[Dmitry A. Kazakov]

On 21/08/2017 00:45, Victor Porton wrote:
> Is it possible to define a primitive subprogram for a subtype rather than
> the base type?

No, a subtype (Ada subtype) would have same tag.

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de

Re: Primitive subprogram for a subtype?

[mockturtle]

On Monday, August 21, 2017 at 12:45:44 AM UTC+2, Victor Porton wrote:
> Is it possible to define a primitive subprogram for a subtype rather than 
> the base type?
> 
> -- 
> Victor Porton - http://portonvictor.org

As I understand (but I am no language lawyer) a subtype is actually the base type, but just with a restriction on the set of possible values.  Therefore, it seems to me that it does not make sense to have type and subtype with different primitive subprograms.

Re: Primitive subprogram for a subtype?

[Victor Porton]

Dmitry A. Kazakov wrote:

> On 21/08/2017 00:45, Victor Porton wrote:
>> Is it possible to define a primitive subprogram for a subtype rather than
>> the base type?
> 
> No, a subtype (Ada subtype) would have same tag.

How is it related to having the same tag? I do NOT speak about defining a 
primitive subprogram for both the type and its subtype but for the subtype 
only.

-- 
Victor Porton - http://portonvictor.org

Re: Primitive subprogram for a subtype?

[Egil H H]

On Monday, August 21, 2017 at 2:32:29 PM UTC+2, Victor Porton wrote:
> 
> How is it related to having the same tag?

It's not. Types can have primitive subprograms without being tagged.

-- 
~egilhh

Re: Primitive subprogram for a subtype?

[Dmitry A. Kazakov]

On 21/08/2017 14:32, Victor Porton wrote:
> Dmitry A. Kazakov wrote:
> 
>> On 21/08/2017 00:45, Victor Porton wrote:
>>> Is it possible to define a primitive subprogram for a subtype rather than
>>> the base type?
>>
>> No, a subtype (Ada subtype) would have same tag.
> 
> How is it related to having the same tag?

Because a primitive subprogram must have a unique slot in the 
dispatching table. If tags are same dispatch cannot lead to different 
bodies.

> I do NOT speak about defining a
> primitive subprogram for both the type and its subtype but for the subtype
> only.

A subtype is a type that inherits subprograms = considered substitutable 
for its supertype. If it does not, then it is not a subtype in any 
sense, including Ada's, Liskov's, whatever flavor of.

P.S. Considering non-primitive subprograms and non-tagged types. Ada 
subtype is an "equivalent" type. It is both sub- and supertype of its 
base type and all other subtypes of whatever subtypes and their subtypes 
(a closure set). Thus anything you define on it is exported to the base 
type and to all subtypes from this type equivalence class.

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de

Re: Primitive subprogram for a subtype?

[J-P. Rosen]

Le 21/08/2017 à 14:32, Victor Porton a écrit :
> How is it related to having the same tag? I do NOT speak about defining a 
> primitive subprogram for both the type and its subtype but for the subtype 
> only.
A value (as opposed to a variable) has no subtype, just a type. Suppose
P is a primitive subprogram of some integer type, if you define another
P for some subtype of it, which one would you call for P(1) ?

Remember that the bounds of a subtype can be computed at runtime, but
that overload resolution happens at compile-time!
-- 
J-P. Rosen
Adalog
2 rue du Docteur Lombard, 92441 Issy-les-Moulineaux CEDEX
Tel: +33 1 45 29 21 52, Fax: +33 1 45 29 25 00
http://www.adalog.fr