Re: Garbage Collection in Ada

[eachus@spectre.mitre.org (Robert I. Eachus)]

In article <dewar.845470908@merv> dewar@merv.cs.nyu.edu (Robert Dewar) writes:

  > Well I would not call it ignorance, since this is certainly not a
  > recognized technical term. In fact I can't even remember if GNAT
  > invented it or not, I suspect we did.

   You may have, but every PL/I compiler implementor I know has come
up with the same term or a similar one.  In PL/I you have to support
pointers to subprograms, which have to designate the code, the
environment/stack and the location of associated static local data.
Some compilers use three word pointers, at Stratus we used two by
having a static pointer and code pointer pair for each subprogram in
global memory, so a fat pointer designated the stack frame and this
vector.

   > We use it simply to mean, as you say, a pointer with other
   > information.  For example, in GNAT, the default (but overridable)
   > form of pointers to unconstrained arrays is a structure with two
   > pointers, one to the data (which could be, but is not yet, a
   > virtual origin), and one a pointer to a record containing the
   > array bounds.

   Just out of curiousity, how do you implement access to subprogram
types in GNAT?  Do you use a pair of addresses or stick the code
pointer in the stack frame?  Or does this vary from target to target?






--

					Robert I. Eachus

with Standard_Disclaimer;
use  Standard_Disclaimer;
function Message (Text: in Clever_Ideas) return Better_Ideas is...