On 6-Jun-09, at 6:24 PM, Mark Renouf wrote: > Wow, I like it! This isn't as crazy as it sounds. After just watching > the V8 talk from I/O, I've learned the JavaScript library is > implemented in JavaScript (preloaded in a heap snapshot). > > The efficiency level they are hitting now makes this seem like a very > sensible approach, especially for Chrome, etc. If the WebSocket > standard ever materializes, it could be even better (and standards > based), and act as a last-resort fallback on all platforms. > > I'm curious if it performs well enough for general use on Firefox? I > suspect it's quite a bit slower right now.
The XPCOM version never made it past the prototype stage (enough to connect to the server, handle LoadJSNI calls and just enough support for JS types to step through a couple of calls). I couldn't tell you what performance would be like, but with "javascript.options.jit.chrome" set to true, it should be running some fairly fast native code. BTW, one advantage of this XPCOM version is that it would work on both FF 3.0 and 3.5. The current OOPHM XPI fails on 3.5, as the Moz devs changed some of the JS type constants. JSVAL_VOID (the effective internal signature of the "undefined" value) is unfortunately an entirely different constant value in 3.5. Matt. --~--~---------~--~----~------------~-------~--~----~ http://groups.google.com/group/Google-Web-Toolkit-Contributors -~----------~----~----~----~------~----~------~--~---
