- From: Robert Carr <carrr rpi edu>
- To: johan gnome org
- Cc: desktop-devel-list gnome org
- Date: Tue, 06 Jan 2009 16:07:18 -0500
JSCore is open to all these extensions, and they are on the slate to be
implemented, just not a priority. It's somewhat likely I could end up
implementing some of them, as a few would be nice to have in Seed.
Long term though I think there are considerations that have to be
considered beyond "Which one supports yield/let today!". Notably, the
general push towards WebKit migration (as in, migrating all GNOME modules
to WebKit), speed/memory differences, and potential speed/memory
differences in the future.
Once these JS extensions are implemented in JSCore, it would essentially
be a faster spider/tracemonkey, with a cleaner API, so it would be a
shame to have decided on Mozilla prematurely.
==============Original message text===============
On Tue, 06 Jan 2009 12:30:25 EST Johan Dahlin wrote:
Jason D. Clinton wrote:
> On Tue, Jan 6, 2009 at 11:14 AM, Johan Dahlin <johan gnome org> wrote:
>> The language is pretty different, SpiderMonkey supports quite a few
>> /language/ extensions which JSCore doesn't.
> s/doesn't./doesn't yet./g
I don't think JSCore is going to implement all features present in
Spidermonkey since they have a different use case. WebKit/JSCore uses
different as it's in addition to that is used to develop native applications
(firefox, thunderbird etc). In other words, JSCore (V8 & JScript etc) will
reasonably only support what web pages actually uses. So sure, the language
extensions present in Spidermonkey might eventually be supported in other
engines, but /only/ if web pages actually start to use them.
Anyway, what subset of JS that's going to be used in the newest and fanciest
web pages seems like a less than ideal criteria to use to select a
which will help developers to write great applications *today*.
===========End of original message text===========
] [Thread Prev