My Script:function DEBUG_NOTE( msg ) {...} // stuffs a message into our
database logvar obj = new Object(); DEBUG_NOTE( "obj.foo = " + obj.foo ); DEBUG_NOTE( "obj.foo == null " + (obj.foo == null) ); DEBUG_NOTE( "!obj.foo " + !obj.foo ); obj.foo = null; DEBUG_NOTE( "obj.foo = " + obj.foo ); DEBUG_NOTE( "!obj.foo " + !obj.foo ); My Output: obj.foo = undefined // as it should be obj.foo == null false // THATS not good !obj.foo true // okay obj.foo = null // okay !obj.foo true // okay "undef == null" is required by the spec (11.9.3) to be true, yet my script's test of such returns false. A pile of our script expecting undef==null is now failing horribly. This seems like a Really Basic feature of the language. Is it even theoretically possible to break this on my end? OTOH, I would expect a Vast Wailing and Gnashing of Teeth if it wasn't just me. I'm using a context that inherits a sealed parent scope containing "initStandardObjects" (and I'm kinda surprised this isn't the default behavior). This particular scope also has executed a collection of files compiled into a script, and there are various exposed java objects & classes floating around. I'm using the standard jsFunction_blah means of exposing things, not the expermental annotations stuff (though it was tempting). Recent changes include: 1) Upgrade from Rhino 1.5r1 2) Added the sealed parent (or is it a prototype... meh) scope with "initStandardObjects" instead of creating a new one for each context 3) That collection of files compiled and executed. I could have sworn all this stuff was working properly after the above changes, and I haven't upgraded my 1.6 JVM with the new embedded rhino thing (that i'm aware of... damn auto-installers). JS.jar is definitely in the class path, dated 3/22/09. Rewriting our scripts to work around a deviation from The Standard is pretty vile. Expecting our customers to work around it is unacceptable. Any suggestions? --Mark Storer Senior Software Engineer Cardiff.com #include <disclaimer> typedef std::Disclaimer<Cardiff> DisCard; _______________________________________________ dev-tech-js-engine-rhino mailing list [email protected] https://lists.mozilla.org/listinfo/dev-tech-js-engine-rhino
