Desugar bitwise negation into XOR and kill all UnaryOp stuff.

Desugar bitwise negation into XOR and kill all UnaryOp stuff.

R=mstarzinger@chromium.org, verwaest@chromium.org

Committed: http://code.google.com/p/v8/source/detail?r=16073

[Sven Panne, 2013-08-05 14:10:13]

New CL due to git cl hiccup

[Toon Verwaest, 2013-08-05 14:27:00]

drive-by lgtm

[Michael Starzinger, 2013-08-06 08:12:06]

LGTM (with comments).

https://codereview.chromium.org/22184004/diff/1/src/hydrogen-uint32-analysis.cc
File src/hydrogen-uint32-analysis.cc (right):

https://codereview.chromium.org/22184004/diff/1/src/hydrogen-uint32-analysis....
src/hydrogen-uint32-analysis.cc:36: if (use->IsBitwise() ||
nit: Condition should fit into one line now.

https://codereview.chromium.org/22184004/diff/1/src/objects.h
File src/objects.h (left):

https://codereview.chromium.org/22184004/diff/1/src/objects.h#oldcode4993
src/objects.h:4993: // [type-recording unary op type]: For kind UNARY_OP_IC.
It should be possible to remove the UNARY_OP_IC from the IC_KIND_LIST now.

https://codereview.chromium.org/22184004/diff/1/src/runtime.js
File src/runtime.js (left):

https://codereview.chromium.org/22184004/diff/1/src/runtime.js#oldcode300
src/runtime.js:300: return %NumberNot(x);
I would be tempted to leave this BIT_NOT runtime function in. Or, if we rip it
out, also rip out the Runtime_NumberNot supporting function. Either way, one
doesn't make sense without the other.

[Sven Panne, 2013-08-06 13:33:38]

Comments addressed, landing...

https://codereview.chromium.org/22184004/diff/1/src/hydrogen-uint32-analysis.cc
File src/hydrogen-uint32-analysis.cc (right):

https://codereview.chromium.org/22184004/diff/1/src/hydrogen-uint32-analysis....
src/hydrogen-uint32-analysis.cc:36: if (use->IsBitwise() ||
On 2013/08/06 08:12:06, Michael Starzinger wrote:
> nit: Condition should fit into one line now.

Done.

https://codereview.chromium.org/22184004/diff/1/src/objects.h
File src/objects.h (left):

https://codereview.chromium.org/22184004/diff/1/src/objects.h#oldcode4993
src/objects.h:4993: // [type-recording unary op type]: For kind UNARY_OP_IC.
On 2013/08/06 08:12:06, Michael Starzinger wrote:
> It should be possible to remove the UNARY_OP_IC from the IC_KIND_LIST now.

Done.

https://codereview.chromium.org/22184004/diff/1/src/runtime.js
File src/runtime.js (left):

https://codereview.chromium.org/22184004/diff/1/src/runtime.js#oldcode300
src/runtime.js:300: return %NumberNot(x);
On 2013/08/06 08:12:06, Michael Starzinger wrote:
> I would be tempted to leave this BIT_NOT runtime function in. Or, if we rip it
> out, also rip out the Runtime_NumberNot supporting function. Either way, one
> doesn't make sense without the other.

Runtime_Number not is dead, Jim...

[Sven Panne, 2013-08-06 13:35:15]

Committed patchset #2 manually as r16073 (presubmit successful).

[Jakob Kummerow, 2013-09-06 15:18:19]

https://codereview.chromium.org/22184004/diff/7001/src/x64/lithium-codegen-x6...
File src/x64/lithium-codegen-x64.cc (right):

https://codereview.chromium.org/22184004/diff/7001/src/x64/lithium-codegen-x6...
src/x64/lithium-codegen-x64.cc:1375: __ not_(ToRegister(left));
Leave the upper half of the register alone! :-(
crbug.com/285355