[MLton] Porting MLton... C99?

Wesley W. Terpstra terpstra@gkec.tu-darmstadt.de
Fri, 10 Dec 2004 13:05:23 +0100

On Thu, Dec 09, 2004 at 11:22:23PM -0600, Henry Cejtin wrote:
> I would say that by far the most useful case is the much simpler one of
> 	Int31.int option

That's a good point. =)

> Re C99, isn't it still true that with newer gcc's we have to turn off some of
> the aliasing analysis because our casts are not really C99-legal?

C99 lets you control the aliasing rules with the 'restrict' keyword.
That said, I have not yet given intense scrutiny to MLton's casts.

It is certainly true that a lot of the casts I saw have 'undefined'
behaviour, but most of the ones I saw were easily fixed. I will have
more to say on this after I've tried converting it all.

PS. Are there any (short) papers I could read about the type of GC MLton
implements? It would help if I understood how the code I was modifying
worked 'in the large'.

Wesley W. Terpstra <wesley@terpstra.ca>