1091945 - %IteratorPrototypeObject% should not be Iterator.prototype

Status: RESOLVED DUPLICATE of bug 1225396

(Core :: JavaScript Engine, defect)

Description

[Tom S. (please needinfo tschuster)]

Right now %IteratorPrototypeObject% is Iterator.prototype. I think it would be better for %IteratorPrototypeObject% to be a distinct object. "Iterator" is kind of legacy and shouldn't pollute %IteratorPrototypeObject%  with e.g. the legacy next method.

Comment 1

[Tom S. (please needinfo tschuster)]

Attachment: WIP

Comment 2

[Tom S. (please needinfo tschuster)]

Comment on attachment 8514692 [details] [diff] [review]
WIP

What do you think of this idea?

Comment 3

[Andy Wingo [:wingo]]

Comment on attachment 8514692 [details] [diff] [review]
WIP

Review of attachment 8514692 [details] [diff] [review]:
-----------------------------------------------------------------

I don't really know.  It could make sense to have an Iterator object in some future ES.  OTOH it probably shouldn't be the current Iterator object, as Iterator.prototype does the StopIteration thing, so you're probably right there.  feedback+ on my side then.

::: js/src/jsiter.cpp
@@ +1251,5 @@
>      nullptr                  /* construct   */
>  };
>  
>  /* static */ bool
> +GlobalObject::initLegayIteratorClass(JSContext *cx, Handle<GlobalObject *> global)

"legacy"

@@ +1311,5 @@
>  JSObject *
>  js_InitIteratorClasses(JSContext *cx, HandleObject obj)
>  {
>      Rooted<GlobalObject*> global(cx, &obj->as<GlobalObject>());
> +    if (!GlobalObject::initLegayIteratorClass(cx, global))

"legacy"

::: js/src/vm/GlobalObject.h
@@ +448,5 @@
>          return &self->getSlot(slot).toObject();
>      }
>  
>    public:
> +    // todo should be nativeobject

what does this mean?

Comment 5

[Jan de Mooij [:jandem]]

Patches in bug 1225396 address this.