> Why is CODE32 defined on the x86-64 architecture?  This architecture
> has 64 bit pointers, and it seems like it's only by luck that the
> bytecode interpreter normally works.

No, it's not by luck: the x86-64 ABI defines several "code models",
one of which -- the "small code model" -- guarantees that code and
static data is in the lower 2G of the address space, and is the
default for gcc (and I suspect other compilers as well).

I would have expected the dynamic loader to implement the small model
like the static linker does, but from what you say it's not the case.

- Xavier Leroy