2010/3/13 Jonathan S. Shapiro <[email protected]>:
> Bjarne Steensgaard recently asked me a question about BitC unboxed unions
> that revealed a design problem. I'ld appreciate thoughts about whether the
> fixes I see are acceptable, or if not, whether we should drop the language
> feature that raises the problem.

Nice catch. As both fixes are entirely acceptable, I see no reason to
get rid of unboxed unions. Of course, the first approach does have the
implication that the tag field must have space for a reserved bit,
perhaps making representing an external data structure as a defrepr
impossible in some very limited number of cases.

I favor the second approach.
_______________________________________________
bitc-dev mailing list
[email protected]
http://www.coyotos.org/mailman/listinfo/bitc-dev

Reply via email to