Bill Baxter Wrote:
> 
> Is it possible to do the kind of "goroutine" scheduling they do purely
> as a library?

I think so.  It sounds like they basically just have a thread pool that 
executes fibers, and D 2.0 already has fibers.

> That wasn't really clear to me how their "segmented stacks" thing
> works.  Sounds like it would need low-level runtime system support,
> though.

In the description, it sounds like they're just talking about the stacks 
resizing on demand.  Maybe the "segmented" bit comes from the stack not being 
contiguous in memory, but that sounds a bit weird.  Either way, the 
underpinnings are already in place in D's fibers for auto stack growth (thanks 
to Mikola Lysenko), so I don't see this as a compiler-dependent feature or 
anything like that.

Reply via email to