On Oct 9, 1:38 pm, Hannes Wallnoefer <[email protected]> wrote:
> On Oct 9, 10:06 am, sleepnova <[email protected]> wrote:
>
>
>
> > On 10月9日, 下午3時00分, Hannes Wallnoefer <[email protected]> wrote:
>
> > > On Oct 9, 8:25 am, sleepnova <[email protected]> wrote:
>
> > > > Is there any document about this?
>
> > > Not really. This was introduced with a small patch to make
> > > JavaAdapters implementing org.mozilla.javascript.Scriptable be treated
> > > as Scriptables (native JS objects) by Rhino.
>
> > >https://bugzilla.mozilla.org/show_bug.cgi?id=462827
>
> > > This is a neat hack, but it has its perils, so while it's great for
> > > prototyping, I recommend implementing host objects in Java (which is
> > > what I do now in Helma NG).
>
> > So, you suggest to implement host object in Java instead of js.
> > I am wondering if there is any plan to define an official way to
> > implement host object in js.
> > I think it would be great to have such extension so people can
> > implement functionalities through uniform js abstraction.
>
> I'm not saying you shouldn't use this feature. It's just that there's
> a price you pay for the simlicity you get. If you implement a host
> object in JS using JavaAdapter, each object access goes through
> Rhino's LiveConnect JS->Java method dispatch mechanism, which means

Sorry, it's not LiveConnect JS->Java, but JavaAdapter's Java->JS of
course. It's all too confusing. ;)

Hannes

> far more overhead and less control (because of automatic argument and
> return value conversion) than native host objects.
>
> But if it works for you there's no reason you shouldn't use it.
>
> Hannes

_______________________________________________
dev-tech-js-engine-rhino mailing list
[email protected]
https://lists.mozilla.org/listinfo/dev-tech-js-engine-rhino

Reply via email to