New: Support ES6 classes (refs #10)

[nzakas]

Three of the original Esprima tests are failing due to a strict mode violation. For some reason, my changes introduced a problem where things like this aren't flagged:

function f() {"use strict"; x={ set s(eval){} }}

It should throw an error on eval, since that's now allowed in strict mode, but doesn't. The three test failures all have to do with this situation (same with getters and methods as well).

I've been staring at this for too long and can't figure out how to fix it. Any insights greatly appreciated. :)

[btmills]

Reading through from the start trying to grok this. If classes are disabled, shouldn't this go to the default case as with block bindings below?

[btmills]

Ah, I see. The default case and breaking out of the switch both end up calling parseStatement(). Doesn't that make the default case redundant?

[nzakas]

Yeah, probably. I mostly just copy-pasted. :)

[btmills]

espree.js line 2613 in the set portion of parseObjectProperty() is param = [ parseVariableIdentifier() ];. Nowhere does it call validateParam(), which gets called by parseParam() in normal methods. Could that be the issue?

[nzakas]

Ah could be! That method was completely refactored in Esprima, and I really didn't want to fight through that refactoring with this commit.

[nzakas]

@btmills That was it! All tests are passing! We have classes!