Chrome canary is doing the right thing, our test is breaking

A test to check whether a frozen object's __proto__ property was mutable
was returning a symptom in the case of correct behavior, namely
silently ignoring the assignment.  This corrects the symptom report to
a simple "false".

[MarkM, 2012-11-30 07:48:14]

This is wrong. For the first, the assigning code is strict. For the second, it
is not an assignment but a call to defineProprty, which is independent of the
caller's strictness anyway.

[metaweta, 2012-12-03 19:19:33]

A test to check whether a frozen object's __proto__ property was mutable
was returning a symptom in the case of correct behavior, namely
silently ignoring the assignment.  This corrects the symptom report to
a simple "false".

[metaweta, 2012-12-03 19:20:22]

A test to check whether a frozen object's __proto__ property was mutable
was returning a symptom in the case of correct behavior, namely
silently ignoring the assignment.  This corrects the symptom report to
a simple "false".

[metaweta, 2012-12-03 19:25:52]

A test to check whether a frozen object's __proto__ property was mutable
was returning a symptom in the case of correct behavior, namely
silently ignoring the assignment.  This corrects the symptom report to
a simple "false".

[metaweta, 2012-12-03 19:26:39]

A test to check whether a frozen object's __proto__ property was mutable
was returning a symptom in the case of correct behavior, namely
silently ignoring the assignment.  This corrects the symptom report to
a simple "false".

[metaweta, 2012-12-03 19:28:36]

A test to check whether a frozen object's __proto__ property was mutable
was returning a symptom in the case of different but less harmful incorrect
behavior, namely silently ignoring the assignment.  This adds a second
test and a repair.

[metaweta, 2012-12-03 19:29:19]

A test to check whether a frozen object's __proto__ property was mutable
was returning a symptom in the case of different but less harmful incorrect
behavior, namely silently ignoring the assignment.  This adds a second
test and a repair.

[metaweta, 2012-12-03 19:31:25]

OK, please have another look.

[kpreid2, 2012-12-03 20:14:54]

LGTM modulo comments

https://codereview.appspot.com/6851128/diff/12002/src/com/google/caja/ses/rep...
File src/com/google/caja/ses/repairES5.js (right):

https://codereview.appspot.com/6851128/diff/12002/src/com/google/caja/ses/rep...
src/com/google/caja/ses/repairES5.js:1925: var isExtensible =
Object.isExtensible;
Shouldn't these be grouped above with hop, slice, etc. around line 1890?

https://codereview.appspot.com/6851128/diff/12002/src/com/google/caja/ses/rep...
src/com/google/caja/ses/repairES5.js:1926: function
repair_FUNCTION_PROTOTYPE_DESCRIPTOR_LIES() {
If this is now a repair for two different bugs, please note this in the function
name and a comment.

[MarkM, 2012-12-03 20:55:46]

LGTM