Unify the Count Operation assignment with other assignments.

Unify the Count Operation assignment with other assignments.
This does not enable inlining of setters (yet).

R=danno@chromium.org

Committed: https://code.google.com/p/v8/source/detail?r=15578

[Toon Verwaest, 2013-06-20 11:37:13]

PTAL

[Sven Panne, 2013-06-24 06:58:13]

https://codereview.chromium.org/17432004/diff/1/src/hydrogen.cc
File src/hydrogen.cc (right):

https://codereview.chromium.org/17432004/diff/1/src/hydrogen.cc#newcode6416
src/hydrogen.cc:6416: if (result != value) Push(result);
This comparison (and similar ones elsewhere) are quite confusing, and admittedly
I don't fully understand them. This should be refactored somehow, perhaps using
an enum describing the intention instead of the HValue*.

We should strive to make things understandable locally, without having to have
hydrogen.cc in one's head... :-)

[danno, 2013-06-27 08:34:10]

https://codereview.chromium.org/17432004/diff/1/src/hydrogen.cc
File src/hydrogen.cc (right):

https://codereview.chromium.org/17432004/diff/1/src/hydrogen.cc#newcode6416
src/hydrogen.cc:6416: if (result != value) Push(result);
On 2013/06/24 06:58:13, Sven Panne wrote:
> This comparison (and similar ones elsewhere) are quite confusing, and
admittedly
> I don't fully understand them. This should be refactored somehow, perhaps
using
> an enum describing the intention instead of the HValue*.
> 
> We should strive to make things understandable locally, without having to have
> hydrogen.cc in one's head... :-)

+1 to sven's comment. It's really difficult to understand what the comparison
between result and value is supposed to check or verify. The enum sounds like a
good idea.

... and some comments would help, too. Why is it OK, for example, to put the
result higher up the environment stack the the value when the simulate is made?
will this really do the right thing? Don't you want to push the result after the
value, and then after the simulate pop the result, pop the value and re-push the
value?

[Toon Verwaest, 2013-07-08 09:26:30]

Addressed comments. PTAL again.

[Toon Verwaest, 2013-07-09 08:52:17]

Addressed comments, PTAL again.

[danno, 2013-07-09 16:22:04]

lgtm

https://codereview.chromium.org/17432004/diff/1/src/hydrogen.cc
File src/hydrogen.cc (right):

https://codereview.chromium.org/17432004/diff/1/src/hydrogen.cc#newcode6515
src/hydrogen.cc:6515: if (result != value) Drop(1);
This is a bug

[Toon Verwaest, 2013-07-09 17:08:52]

Committed patchset #3 manually as r15578 (presubmit successful).