Re: Q: access to subprogram

[dewar@cs.nyu.edu (Robert Dewar)]

Bob Duff said

"Now, GNAT has gone and provided the same capability, using an
implementation-defined attribute (Unrestricted_Access), but
unfortunately, this attribute allows dangling pointers to subprograms,
unlike either of the above options."

But unfortunately, virtually all the useful uses of Unrestricted_Access,
at least in my experience are cases where your limited mechanisms would
have been insufficient, since we definitely need to be able to assign
these values.

So we would probably have put the attribute in even if the language had
the safe cases. Note that the two safe cases that Bob mentions are in fact
cases where the work arounds required in the current language are pretty
straightforward, so the limited form of the feature proposed in Bob's
note is not very convincing since it provides rather limited capabilities
and no significant increase in expressive power.