The only iterating mechanism that lets you access the “`done` value” is `yield*`. All other constructs (`for-of`, spread, destructuring, …) completely ignore it. Its main purpose is to let `yield*` make recursive generator calls with results, without disrupting normal yielding. That is, for normal iteration it is an out-of-band value.
Axel > On 06 Apr 2015, at 00:43, Mark Volkmann <[email protected]> wrote: > > Thanks, that helps. I had been thinking that for-of was a model for how > iterators should be consumed. > > When implementing any kind of iterator (generator or not), it seems important > to recognize that if you choose to return a value with done set to true, you > are precluding the use of for-of with that iterator (assuming you want all > the values). > > On Sun, Apr 5, 2015 at 11:40 AM, Allen Wirfs-Brock <[email protected] > <mailto:[email protected]>> wrote: > > On Apr 5, 2015, at 6:04 AM, Mark Volkmann wrote: > >> I thought that when an iterator returns done: true, the value should not be >> used. > > Why do you think that? The definition of the IteratorResult interface in the > the ES6 spec. doesn't say that. > >> However, if a generator function ends by returning a value, done will be >> true when that value is returned and the value should be used. > > Could be used. Whether a generator or any Iterator provides a meaningful > value when it reach the the 'done' state depends upon its specific definition. > >> Given this, how can a consumer know the correct way to handle the value when >> done is true? Clearly consumers shouldn't have to be aware of whether the >> iterator is actually a generator and whether it ends by returning a value. > > An Iterator does not have to be a generator to provide a 'done' state value. > In general, If you are going to do anything other than basic iteration, such > as is performed by for-of, you need to know about the specific Iterator you > are using. > > Allen > > > > > > -- > R. Mark Volkmann > Object Computing, Inc. > _______________________________________________ > es-discuss mailing list > [email protected] > https://mail.mozilla.org/listinfo/es-discuss -- Dr. Axel Rauschmayer [email protected] rauschma.de
_______________________________________________ es-discuss mailing list [email protected] https://mail.mozilla.org/listinfo/es-discuss
