Re: Initialization and Finalization of limited object "returned" by a function

["Hibou57 (Yannick Duchêne)" <yannick_duchene@yahoo.fr>]

Le Sat, 13 Feb 2010 16:59:44 +0100, Robert A Duff  
<bobduff@shell01.theworld.com> a écrit:
>> What's the new Professor Sir Tucker Taft's project ?
>
> The ARG is inventing some syntactic sugar to make user-defined
> container code easier to read, and Tucker is doing most of
> the work on that.  One example is that there will probably be some
> mechanism to say:
>
>     A(X) := A(X) + 1;
>
> where A is (say) a hash table mapping strings to integers,
> and X is a string.  User-defined array-indexing notation.
>
> Also, some new kinds of for loops.  For example, looping
> through the components of an array without any
> horsing around with index values:
>
>     for C : Character of My_String loop
>         Do_Something(C);
>     end loop;
>
> The ": Character" above is optional, according to the latest proposal.
That's high-level language capability.

So that's to be added with the other project I've heard about here at  
comp.lang.ada, about first-class-functions (it was during the thread about  
making constant declarations overloadable if my mind is right) and the  
pre-post-condition discussed here and on fr.c.l.a.

As Santa Claus seems to be actually checking orders, do you know if there  
will be something like co-routines ? We already have tasks, but tasks are  
a bit too much heavy as co-routines. I was to open a thread to investigate  
this topic two weeks ago (I finally did not start it). I'm thinking about  
it, beside another topic about unknown discriminants.

-- 
No-no, this isn't an oops ...or I hope (TM) - Don't blame me... I'm just  
not lucky