barmar@Think.COM (Barry Margolin) (11/27/89)
In a previous posting, I think I made a slight misstatement when I
described function linkage on Symbolics workstations. I said that the
caller has a pointer to the callee function object. Repenning Alexander
asked how redefinition and forward reference works in such a scheme.
Actually, I think it has a pointer to the function cell of the callee, not
the function object itself. In the case of functions named by symbols,
this is the actual function cell in the symbol structure; for functions
named by lists it's a pointer to a cons cell in the function specifier hash
table. And in the case of forward references it's a pointer into a fixup
table that contains back pointers to be used when the function is loaded.
Barry Margolin, Thinking Machines Corp.
barmar@think.com
{uunet,harvard}!think!barmar