[Not important] the “not null” notation in Ada 2005

[Not important] the “not null” notation in Ada 2005

[Yannick Duchêne (Hibou57)]

I remember some peoples says here, this is a pity to have “not null”  
annotations, while the case which should be clearly marked, is the one  
where a reference may be null. I'm OK with that principle. But concretely  
what would have been the notation ? “may be null” ? Two reserved words  
added for that ? A simple “null” would not have been really expressive  
(just a though I get right a few minutes ago).


-- 
Si les chats miaulent et font autant de vocalises bizarres, c’est pas pour  
les chiens.
“ c++; /* this makes c bigger but returns the old value */ ” [Anonymous]

Re: [Not important] the “not null” notation in Ada 2005

[Robert A Duff]

"Yannick DuchÔøΩne (Hibou57)" <yannick_duchene@yahoo.fr> writes:

> I remember some peoples says here, this is a pity to have ÔøΩnot nullÔøΩ
> annotations, while the case which should be clearly marked, is the one
> where a reference may be null. I'm OK with that principle. But
> concretely  what would have been the notation ? ÔøΩmay be nullÔøΩ ? Two
> reserved words  added for that ? A simple ÔøΩnullÔøΩ would not have been
> really expressive  (just a though I get right a few minutes ago).

Take a look at OCaml, for example.

- Bob

Re: [Not important] the “not null” notation in Ada 2005

[Yannick Duchêne (Hibou57)]

Le Fri, 25 Mar 2011 16:51:24 +0100, Robert A Duff  
<bobduff@shell01.theworld.com> a écrit:
>
> Take a look at OCaml, for example.
>
> - Bob
Thanks Bob; but I know SML, not OCaml. Do I really have to learn all of  
OCaml to get an idea or can you give some tracks ?

-- 
Si les chats miaulent et font autant de vocalises bizarres, c’est pas pour  
les chiens.
“ c++; /* this makes c bigger but returns the old value */ ” [Anonymous]

Re: [Not important] the “not null” notation in Ada 2005

[Robert A Duff]

"Yannick Duch�ne (Hibou57)" <yannick_duchene@yahoo.fr> writes:

> Thanks Bob; but I know SML, not OCaml. Do I really have to learn all of
> OCaml to get an idea or can you give some tracks ?

The two languages are pretty similar, actually.
If you want a thing, you declare type thing.
If you want a thing-or-null you declare "Thing | Null".
Or maybe something like "List = Pair | Empty_List".
There's a Maybe type -- Maybe(T) means "a T or nothing".

You don't get any sort of null-like value unless you ask for it,
and you don't have to call it "null", and you can have several
such special values.

- Bob

Re: [Not important] the “not null” notation in Ada 2005

[Yannick Duchêne (Hibou57)]

Le Fri, 25 Mar 2011 18:13:45 +0100, Robert A Duff  
<bobduff@shell01.theworld.com> a écrit:
> The two languages are pretty similar, actually.
> If you want a thing, you declare type thing.
> If you want a thing-or-null you declare "Thing | Null".
Yes, the type constructors. Would have been a good idea, I agree.

-- 
Si les chats miaulent et font autant de vocalises bizarres, c’est pas pour  
les chiens.
“ c++; /* this makes c bigger but returns the old value */ ” [Anonymous]

Re: [Not important] the “not null” notation in Ada 2005

[Dmitry A. Kazakov]

On Fri, 25 Mar 2011 16:45:28 +0100, Yannick Duchêne (Hibou57) wrote:

> I remember some peoples says here, this is a pity to have “not null”  
> annotations, while the case which should be clearly marked, is the one  
> where a reference may be null. I'm OK with that principle. But concretely  
> what would have been the notation ? “may be null” ? Two reserved words  
> added for that ? A simple “null” would not have been really expressive  
> (just a though I get right a few minutes ago).

No notation needed. Before Ada 2005 broke it, "access T" was not null, if
anonymous type is what you mean.

Named type is a different beast. When you declare a named access type "not
null" is a constraint. You cannot have a constrained subtype before the
type.

Yet another story is an access type, which does not have null as a value.
It is not much different from being constrained. Compare it with integer
types. When you declare:

   type P is range 1..2;

You do not eliminate 0, as you might think. P has a "parent type," which
still has 0 (P'Base).

So which case you meant? (:-))

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