Re: Unconstrained Arrays

[sjw <simon.j.wright@mac.com>]

On Mar 20, 11:40 am, Jean-Pierre Rosen <ro...@adalog.fr> wrote:
> sjw a écrit :> There is a deep language-lawyerly reason (which I don't understand)
> > why an array like your My_Array can't be aliased (at any rate in
> > Ada95); you have to use the initialize-with-aggregate approach.
> > Perhaps that's what leads to the initialize-with-aggregate style.
>
> Here is an example:
>
> procedure essai is
>    type Acc is access all String;
>    V : Acc;
>    S1 : aliased String := "abcd";
>    S2 : aliased String (1..4);
> begin
>    V := S1'Access;
>    V := S2'Access;
> end;
>
> And this is what GNAT says:
>
> essai.adb:5:04: warning: aliased object has explicit bounds
> essai.adb:5:04: warning: declare without bounds (and with explicit
> initialization)
> essai.adb:5:04: warning: for use with unconstrained access
> essai.adb:8:09: object subtype must statically match designated subtype
>
> The message is pretty explicit. The underlying reason has to do with
> dope vectors (i.e. how the bounds of an array are attached to the
> array). It might be different, depending on whether the array is
> /declared/ constrained or not (S1 is declared unconstrained, although
> constrained by initialization).
>
> For example, in the constrained case, the array might include a pointer
> to a structure that describes the subtype, while in the unconstrained
> case, the array might keep the bounds with the data. Therefore, a
> pointer to a constrained array is incompatible with a pointer to an
> unconstrained array.

That makes so much sense! I think my brain freezes when I see the
compiler quoting rules about statically matching, so I just remember
the first 2 lines of the message (as you say, pretty explicit) and
pass over the reason why in the last 2 lines. After all, GNAT is much
more likely to be right than I am.