On Jul 14, 2011, at 2:14 PM, Mark S. Miller wrote: > Would we just be postponing all those same edge cases from 2**32 to 2**53?
No, because you lose precision adding 1 to an index to get length after that. No bignums. /be > > On Thu, Jul 14, 2011 at 1:54 PM, Brendan Eich <[email protected]> wrote: > On Jul 14, 2011, at 12:35 PM, Allen Wirfs-Brock wrote: > > > What about the possibility of simply eliminating the range error and the > > Uint32 restriction on the association between array indexed properties and > > the "length" property. Instead replace it with a ToInteger constraint. > > This is essentially how string operations are defined. Implementations > > could still optimize for lengths <2^32 and or any other size they deemed > > appropriate. > > I would prefer that. Jeff may agree. > > The uint32 business from ES1 days never paid off beyond allowing certain > storage optimizations, and even then many (most?) engines do not optimize > uint32 values. It was a flop on the optimization front, but it sure requires > extra code in Array's implementation. > > ToInteger matches string and makes full use of the integral domain in the > number type. > > The edge-case nature of this change suggests we could get away with it, even > though it's an incompatible change. But we need to be careful that we're not > turning errors into working code that could be exploited somehow. Again there > is no memory safety issue, rather "index safety". > > /be > > _______________________________________________ > es-discuss mailing list > [email protected] > https://mail.mozilla.org/listinfo/es-discuss > > > > -- > Cheers, > --MarkM
_______________________________________________ es-discuss mailing list [email protected] https://mail.mozilla.org/listinfo/es-discuss
