[Yhc] some initial questions

Robert Dockins robdockins at fastmail.fm
Thu Mar 2 18:00:36 EST 2006


On Tuesday 28 February 2006 07:36 am, you wrote:
> Hi Rob,
>
> Just to add further comment to what Neil said:
>
> 1) Indeed, there is no need for backward jumps so they were deliberately
> not included. (This makes several of the compiler analyses easier too).

That's more or less what I expected.  The absence of explicit loops makes 
things considerably simpler.

> Compared to verification of Java verification of YHC should be quite
> easy for the simple reason that YHC does not include support for unboxed
> integers, so all variables on the stack must always be pointers to heap
> nodes (or stack frames).

That's an interesting point.  I'll think about this some more.

> 2) I plan to document my implementation of concurrency on the wiki soon.
> Since it changes some of the docs. Since the effect of concurrency on
> non-concurrent programs is fairly small support for concurrency will
> become part of the standard build soon :-)

That is very good to hear.  It seems to me that the two biggest issues are:

-- supporting multiple execution stacks
-- blocking I/O and FFI calls

How did you decide to handle these problems (mostly for my curiousity)?

Thanks,
Rob Dockins


More information about the Yhc mailing list