On 06/01/2013 07:29 AM, Alex Rønne Petersen wrote:
Hi,
I'm sure this has been brought up before, but I feel I need to bring it
up again (because I'm going to be writing a threaded-code interpreter):
http://gcc.gnu.org/onlinedocs/gcc/Labels-as-Values.html
This is an incredibly important extension. The final switch statement is
not a replacement because it doesn't allow the programmer to store a
label address directly into a code stream, which is what's essential to
write a threaded-code interpreter.
I'd also like to see this.
The Erlang folks went through hell just to use this feature; see the 5th
Q at:
http://www.erlang.org/doc/installation_guide/INSTALL-WIN32.html#Frequently-Asked-Questions
The idea is to be able to write code like this:
----
import std.algorithm;
enum Op : ubyte
{
imm,
add,
sub,
// ...
ret,
}
final class Insn
{
Op op;
size_t[] args;
void* lbl;
Insn next;
}
final class State
{
Insn pc;
size_t[64] regs;
}
size_t interp(Insn[] code)
{
// Set up the instruction stream with label addresses
// the first time that it is executed. Label addresses
// are stable, so we only do this once.
foreach (insn; code.filter!(x => !x.lbl)())
{
void* lbl;
with (Op)
{
final switch (insn.op)
{
case imm: lbl = &&handle_imm; break;
case add: lbl = &&handle_add; break;
case sub: lbl = &&handle_sub; break;
// ...
case ret: lbl = &&handle_ret; break;
}
}
insn.lbl = lbl;
}
...
(This must be supported at compile time!)
