Bug 102276 – Object.getOwnPropertyDescriptor() on a non-strict function exposes strict "caller" property

• Status: RESOLVED
• Resolution: CONFIGURATION CHANGED
• Component: JavaScriptCore

[Adam Klein]

The following code appropriately throws a TypeError as required by ECMAScript:

function nonStrict() { return nonStrict.caller; }
function strict() { "use strict"; return nonStrict() === strict; }
strict()

But the following code does not throw, and returns true:

function nonStrict() { return Object.getOwnPropertyDescriptor(nonStrict, 'caller').value; }
function strict() { "use strict"; return nonStrict() === strict; }
strict()

Though this is underspecified in the spec (it says that only [[Get]] should throw when asked for caller), it seems likely that [[GetOwnProperty]] should do the same, and that's what both Chrome and Firefox do.

[Adam Klein]

See https://bugs.ecmascript.org/show_bug.cgi?id=310 for further discussion of this issue.

[Alexey Shvayka]

(In reply to Adam Klein from comment #0)
> Though this is underspecified in the spec (it says that only [[Get]] should
> throw when asked for caller), it seems likely that [[GetOwnProperty]] should
> do the same, and that's what both Chrome and Firefox do.

As of r277026, [[Get]] is aligned with [[GetOwnProperty]] not to return strict callers.
In modern browsers both test cases (adjusted for Firefox having "caller" on Function.prototype) evaluate as `false`.
I am adding test coverage for [[GetOwnProperty]] as part of https://bugs.webkit.org/show_bug.cgi?id=225277.