Ok, I've only had time to do a cursory check, but based upon quick inspection the ES6 spec. for unshift appears to to produce observably identical results when operating upon an an Array instance of max length. That's what supposed to happen, ES5 didn't specify any precondition checks, so to do so now is, in theory, a breaking change. In practice it might not matter, but nobody knows for sure.
if there are what appears to be unintentional differences between the behavior or the ES5 and ES6 spec, then bugs should be file. It's also reasonable to file bugs suggesting that a change to the ES5 behavior should be make but the submitter should try to justify why it is ok to make such a breaking change. Finally, at this stage of the release game, bugs that contain detailed "patches" to the currently specified algorithms are more likely to get attention then those that lack that level of detail. Allen On Oct 23, 2014, at 5:44 PM, Brendan Eich wrote: > Allen Wirfs-Brock wrote: >> I can only answer briefly right now: This is intentional. Array instances >> are still limited to 2^32-2 elements. Compat issues to change. But, the >> generic array methods aren't restricted to Array instances and support >> larger lengths in those cases. > > Hi Allen, appreciate brief reply. > > Still leaves mutation before error problem -- want a spec bug? > > Impl bugs to follow. > > /be > _______________________________________________ es-discuss mailing list [email protected] https://mail.mozilla.org/listinfo/es-discuss
