Ada95 language design question (accessibility rules and type extensions)

Ada95 language design question (accessibility rules and type extensions)

[Mark Lundquist]

The invocation of accessibility rules in RM 3.9.1(3) is designed to prevent
dangling dispatch.  It seems to me that another way would have been to
disallow assignment of a tagged type to the classwide type of a statically
shallower parent.  Does anyone know why that would not have worked?

What got me started thinking on this was the common complaint about the
implementation of controlledness, and the ramification of this rule that
controlled types can't be declared deeper than library-level...

-- mark
--------------
Reply by email to: Mark dot Lundquist at ACM dot org
Consulting services: http://home.attbi.com/~mlundquist2/consulting

Re: Ada95 language design question (accessibility rules and type extensions)

[Nick Roberts]

I think the answer to your question, Mark, is that something like what you
suggest would have worked, but (the designers feared?) it would have ended
up with a lot of code with dynamic checks in it. I guess the designers felt
uncomfortable with this happening, and that the existing rules were not a
problem in practice (and they generally aren't).

But IANALL (I Am Not A Language Lawyer ;-) and a comment from one of the 9X
design team would be interesting.

--
Nick Roberts

"Mark Lundquist" <no.spam@getalife.com> wrote in message
news:dULT7.2153$NM4.98008@rwcrnsc53...
> The invocation of accessibility rules in RM 3.9.1(3) is designed to
prevent
> dangling dispatch.  It seems to me that another way would have been to
> disallow assignment of a tagged type to the classwide type of a statically
> shallower parent.  Does anyone know why that would not have worked?
>
> What got me started thinking on this was the common complaint about the
> implementation of controlledness, and the ramification of this rule that
> controlled types can't be declared deeper than library-level...