Replace BIT_NOT by XOR with ~0

Replace BIT_NOT by XOR with ~0

Remove the whole UnaryOp stuff as https://codereview.chromium.org/21782002

[Sven Panne, 2013-08-05 07:15:14]

This CL is missing something: In HOptimizedGraphBuilder::BuildBinaryOperation we
want to transform an BIT_XOR with a ~0 back into an HBitNot, otherwise we lose
some optimizations, e.g. ~~x => x.

https://codereview.chromium.org/21813010/diff/1/src/parser.cc
File src/parser.cc (right):

https://codereview.chromium.org/21813010/diff/1/src/parser.cc#newcode3211
src/parser.cc:3211: factory()->NewNumberLiteral(-1),
Replace the -1 by ~0 here and in the comment, it's a bit clearer.

[Sven Panne, 2013-08-05 07:17:04]

... and one more organizational remark I forgot in me previous reply: It would
be good to coordinate efforts a bit more, because I have basically the same CL
sitting on my disk, which should not be very surprising given the comments in
the previous CLs, so this is basically duplicated work. :-/

[Sven Panne, 2013-08-05 07:30:23]

Hmmm, the CL is not even compiling. I'll open a clean new CL...

[weiliang.lin2, 2013-08-05 07:54:38]

On 2013/08/05 07:30:23, Sven Panne wrote:
> Hmmm, the CL is not even compiling. I'll open a clean new CL...

Hi Sven, 
Sorry for that duplicated work.in future i will give a claim before I start to
do it. 

So, if you have your CL ready, please submit it and just close mine. It seems my
uploaded code is not incomplete. Weird~

[weiliang.lin2, 2013-08-05 08:00:24]

On 2013/08/05 07:54:38, weiliang.lin2 wrote:
> On 2013/08/05 07:30:23, Sven Panne wrote:
> > Hmmm, the CL is not even compiling. I'll open a clean new CL...
> 
> Hi Sven, 
> Sorry for that duplicated work.in future i will give a claim before I start to
> do it. 
> 
> So, if you have your CL ready, please submit it and just close mine. It seems
my
> uploaded code is not incomplete. Weird~

The patch set 3 is OK