Enhance es5shim.js to adapt to presence of bugs to be worked around.

For each bug handled by es5shim, test and report for its presence, and
only work around a bug on those platforms where that bug is detected.

Now that we work around
https://bugzilla.mozilla.org/show_bug.cgi?id=637994 , we restore the
original getter logic to WeakMap.js, so the WeakMap emulation can
prevent identity stealing even on platforms suffering from that bug.

[kpreid2, 2011-05-31 18:15:40]

I see nothing wrong code-style-wise, or possibly affecting my work, with this
but I am not qualified to evaluate its correctness. LG(qualified)TM.

Please address code reviews to kpreid.switchb.org@gmail.com rather than
kpreid@switchb.org (unfortunately, Rietveld doesn't understand secondary
addresses).

[ihab.awad, 2011-06-02 04:42:25]

lgtm++

http://codereview.appspot.com/4547070/diff/6003/src/com/google/caja/ses/es5sh...
File src/com/google/caja/ses/es5shim.js (right):

http://codereview.appspot.com/4547070/diff/6003/src/com/google/caja/ses/es5sh...
src/com/google/caja/ses/es5shim.js:21: * >secureable ES5</a>.
Add comment noting what you said verbally, that you intend this to be useable
outside Caja as a transparent way to "repair" an almost-ES5 platform.

http://codereview.appspot.com/4547070/diff/6003/src/com/google/caja/ses/es5sh...
src/com/google/caja/ses/es5shim.js:67: * this platform.
Isn't this TODO already to-done by the test_* function below?

http://codereview.appspot.com/4547070/diff/6003/src/com/google/caja/ses/es5sh...
src/com/google/caja/ses/es5shim.js:80: //var TOLERATE_MISSING_CALLEE_DESCRIPTOR
= false;
Why not just delete these commented-out lines now?

http://codereview.appspot.com/4547070/diff/6003/src/com/google/caja/ses/es5sh...
src/com/google/caja/ses/es5shim.js:299: * TODO(erights): isolate and report the
V8 bug mentioned below.
I wasn't sure what the V8 bug actually is ... no reference to bug. :)

http://codereview.appspot.com/4547070/diff/6003/src/com/google/caja/ses/es5sh...
src/com/google/caja/ses/es5shim.js:310: * Work arond for
https://bugzilla.mozilla.org/show_bug.cgi?id=637994
... around ...

http://codereview.appspot.com/4547070/diff/6003/src/com/google/caja/ses/es5sh...
src/com/google/caja/ses/es5shim.js:328: * <p>A little known fast about
JavaScript is that
... fact ...

http://codereview.appspot.com/4547070/diff/6003/src/com/google/caja/ses/es5sh...
src/com/google/caja/ses/es5shim.js:598: return defProp(base, name, fullDesc);
I'm a bit confused that dummySetter never calls 'desc.set()'.

[MarkM, 2011-06-05 17:38:18]

http://codereview.appspot.com/4547070/diff/6003/src/com/google/caja/ses/es5sh...
File src/com/google/caja/ses/es5shim.js (right):

http://codereview.appspot.com/4547070/diff/6003/src/com/google/caja/ses/es5sh...
src/com/google/caja/ses/es5shim.js:21: * >secureable ES5</a>.
On 2011/06/02 04:42:26, ihab.awad wrote:
> Add comment noting what you said verbally, that you intend this to be useable
> outside Caja as a transparent way to "repair" an almost-ES5 platform.

Done.

http://codereview.appspot.com/4547070/diff/6003/src/com/google/caja/ses/es5sh...
src/com/google/caja/ses/es5shim.js:67: * this platform.
On 2011/06/02 04:42:26, ihab.awad wrote:
> Isn't this TODO already to-done by the test_* function below?

Done.

http://codereview.appspot.com/4547070/diff/6003/src/com/google/caja/ses/es5sh...
src/com/google/caja/ses/es5shim.js:80: //var TOLERATE_MISSING_CALLEE_DESCRIPTOR
= false;
On 2011/06/02 04:42:26, ihab.awad wrote:
> Why not just delete these commented-out lines now?

Done.

http://codereview.appspot.com/4547070/diff/6003/src/com/google/caja/ses/es5sh...
src/com/google/caja/ses/es5shim.js:299: * TODO(erights): isolate and report the
V8 bug mentioned below.
On 2011/06/02 04:42:26, ihab.awad wrote:
> I wasn't sure what the V8 bug actually is ... no reference to bug. :)

Done.

http://codereview.appspot.com/4547070/diff/6003/src/com/google/caja/ses/es5sh...
src/com/google/caja/ses/es5shim.js:310: * Work arond for
https://bugzilla.mozilla.org/show_bug.cgi?id=637994
On 2011/06/02 04:42:26, ihab.awad wrote:
> ... around ...

Done.

http://codereview.appspot.com/4547070/diff/6003/src/com/google/caja/ses/es5sh...
src/com/google/caja/ses/es5shim.js:328: * <p>A little known fast about
JavaScript is that
On 2011/06/02 04:42:26, ihab.awad wrote:
> ... fact ...

Done.

http://codereview.appspot.com/4547070/diff/6003/src/com/google/caja/ses/es5sh...
src/com/google/caja/ses/es5shim.js:598: return defProp(base, name, fullDesc);
On 2011/06/02 04:42:26, ihab.awad wrote:
> I'm a bit confused that dummySetter never calls 'desc.set()'.

Because we only install the dummySetter if fullDesc.set === undefined. This
should only be true if desc.set is absent or undefined. If desc.set contains a
setter, then we use that one.