Re: Q: access to subprogram

[ncohen@watson.ibm.com (Norman H. Cohen)]

In article <Dtyt4K.Jvx@world.std.com>, bobduff@world.std.com (Robert A Duff)
writes: 

|>                                                                 With
|> displays, it is somewhat difficult to implement passing nested
|> procedures as parameters, because you have to pass the display, which is
|> not of compile-time-known size.

The size of a display for a given procedure IS known at compile time.
It is equal to the number of procedures that lexically surround that
procedure (give or take one, depending on the details of a given
implementation).

In essence, a display is an array of pointers to the stack frames for the
current activations of each subprogram surrounding the current
subprogram.  The array is typically small, fast access to the values of
the array elements is crucial for efficient access to nonlocal variables,
and the array indices needed for addressing nonlocal variables are always
known at compile time, so the elements of this array are typically stored
in registers.

(By the way, I share Bob Duff's amazement at the noninclusion of downward
closures--we called them downward at the time--in Ada 95, and I say this
as one of those Bob described as urging the removal rather than the
addition of proposed Ada-95 features.  I viewed the inclusion of downward
closures as the REMOVAL of an arbitrary restriction.  The decision was
not made in ignorance.  Bill Taylor had made what I considered an
irrefutable case for downward closures, showing how much easier it would
be to write iterators if downward closures were allowed.  It came down to
a conflict between the interests of Ada programmers and the interests of
a minority of Ada implementors, and in this case the interests of the few
implementors using displays prevailed.)

--
Norman H. Cohen    ncohen@watson.ibm.com