67714 - Rhino needs to allow indirect eval, as SpiderMonkey now does

Status: VERIFIED WONTFIX

(Rhino Graveyard :: Core, defect)

Description

[Phil Schwartau]

Thanks to bug 38512, Spidermonkey now allows indirect calls to eval.
If the "strict" option is on, it will warn about it - but permit it:


--------------------  STRICT WARNING ON --------------------------
[/d/JS_TRUNK/mozilla/js/src/WINNT4.0_DBG.OBJ] ./js
js> var f = eval
js> f('1+1')
2: strict warning: function eval must be called directly, and not by way of a 
function of another name.
2


--------------------  STRICT WARNING OFF --------------------------
[/d/JS_TRUNK/mozilla/js/src/WINNT4.0_DBG.OBJ] ./js -W
js> var f = eval
js> f('1+1')
2



Rhino needs to do the same - currently we get this:

[/d/JS_TRUNK/mozilla/js/rhino] java org.mozilla.javascript.tools.shell.Main
js> var f=eval
js> f('1+1')
js: "<stdin>", line 2: uncaught JavaScript exception: EvalError: Function "eval" 
must be called directly, and not by way of a function of another name. 
(<stdin>; line 2)

Comment 1

[Norris Boyd]

I argue that Rhino can continue to throw an error for indirect eval as it is
permitted to do by ECMA. This allows us to perform the optimizations we 
currently perform when our optimizer is enabled.

Phil, could you make the test work for the case where we throw the error?

Comment 2

[Phil Schwartau]

OK, marking WONTFIX, and adding these testcases to the Rhino "skip list" 
(js/tests/rhino-n.tests):

                 js/tests/js1_4/Eval/eval-001.js
                 js/tests/js1_4/Eval/eval-002.js
                 js/tests/js1_4/Eval/eval-003.js

Comment 3

[Phil Schwartau]

Marking Verified WONTFIX -