Re: Uninitialized out parameters.

[Georg Bauhaus <bauhaus@futureapps.invalid>]

On 06/04/16 22:53, Stefan.Lucks@uni-weimar.de wrote:
> If people really think they need to declare something like the above procedure X, a revised Ada standard could allow declarations such as
>
>      procedure X(Formal: T) is raise;
>
> and
>
>       function Y return T is raise;
>
> where "is raise" stands for "raise Program_Error". That may sometimes come handy in the context of inheritance, or so.


Two prominent use cases will suggest this hammer won't fly ;-)
if used as a replacement:

1. Testing, or integration builds at various incomplete stages:
    one couldn't test part B of a program if part A of the program,
    calling X, always raises before it gets to B.

2. OPENSTEP style programming (a.k.a. Cocoa) as an example, meaning
    that, conceptually, whenever an object is not (yet) defined (nil),
    but then receives a message, nothing happens.

So, if a procedure does nothing and out comes nothing,
then it doesn't do harm either, at least if programmers know what
to expect. They fill in the blanks later.

I'd rather accept null functions in the set of oddities of
Ada than be forced to always have finished everything.
Note that raising disallows stubs like

   procedure X (Formal : out T) is
   begin
       pragma Warning ("TODO: compute a value for out Formal!")
       null;
   end X;

Treating warnings as errors should help with strict rejection
rules where the projects demand them.