Iterators for D

[Sean Kelly]

Daniel Keep wrote:
> 
> I suppose the bottleneck (if it is one) is how fast you can do
> user-space context switches.  Which is basically a function call, moving
> the base of the stack, and switching over stuff like the exception
> handling registers.  Or somesuch... Mikola Lysenko wrote StackThreads;
> my attempt only mostly worked :P

That's pretty much it.  Saving the current register data, loading the 
other register data, and jumping to a new instruction address.  There's 
no way this can be as efficient as an integer increment instruction (for 
loop iteration) but it's orders of magnitude faster than a kernel thread 
context switch.


Sean