[webkit-dev] interested in js speed-up

[ToolmakerSteve]

Maciej Stachowiak wrote:
> 
> Also, while the Tamarin project is an interesting piece of technology,  
> let's be careful not to oversell it. Tamarin only provides half of a  
> JavaScript engine (the execution engine but not a parser/compiler to  
> generate the bytecodes), it cannot yet handle arbitrary JavaScript  
> code from the web, and at least the classic version is not  
> particularly fast on normal JS, only on typed code (which no one uses  
> yet and which is syntactically incompatible with current JavaScript).  
> It's true that the tracing JIT branch promises to do reasonably well  
> on normal untyped JS as well, but now we're talking about a rewrite,  
> not a proven technology.
> 

I appreciate hearing your insights into the Tamarin code; that helps me
understand it, and where it stands in the bigger picture. Good to hear that
you are continuing your efforts with WebKit.

Now as to the larger question I intended to raise:
is it worth doing much to the current JS code, or will that all soon be
moot, as people's focus shifts to the next generation of JS? Perhaps that
next generation will require a total re-think anyway?

-- 
View this message in context: http://www.nabble.com/interested-in-js-speed-up-tp15822510p16045463.html
Sent from the Webkit mailing list archive at Nabble.com.