🚀 go-pugleaf

In article <1993Jan11.215111.11291@qualcomm.com> greg@harvey.qualcomm.com (Greg Noel) writes:
>Recursion implies that the context of a pointer must be kept as part of
>the pointer, which leads to either expanding the size of a function pointer
>or generating trampoline code dynamically.  Expanding function pointers
>would be a cost even when nested functions are not used and there are
>some architectures where code cannot be generated on the fly.

I know that code can be generated on the fly on Sparc, 680x0, and
80286, 80386, 80486, PDP-11, VAX, 6502, and IBM 370.  What
architecture were you thinking of?

Given such an architecture, are you sure that its problems cannot be
treated by keeping a cache of generated code fragments (you keep a
per-thread cache of code+data pairs to avoid the cost of page
remapping and cache flushing).  Remember, most machines ARE able to
load programs and run them, so they must have some rudimentary ability
to "generate" code at run-time.

In truth, I don't high hopes for nested functions appearing in C++,
but arguments against them based on their "cost" or "difficulty of
implementation" are not very strong.

David Chase
Sun