Implement experimental exponentiation operator via desugaring

Implement experimental exponentiation operator via desugaring

BUG=v8:3915
LOG=N
R=

[caitp (gmail), 2015-02-25 15:35:34]

This version doesn't touch the compiler at all. Note that the changed code is
not very different, because the changes in the compiler are mostly single line
changes.

There is an observable bug which traceur works around by allocating a temporary
variable, but maybe for an experimental feature this doesn't matter.

[arv (Not doing code reviews), 2015-02-25 16:58:25]

arv@chromium.org changed reviewers:
+ arv@chromium.org

[arv (Not doing code reviews), 2015-02-25 16:58:25]

Maybe skip **= until it works as expected?

https://codereview.chromium.org/952243002/diff/1/src/parser.cc
File src/parser.cc (right):

https://codereview.chromium.org/952243002/diff/1/src/parser.cc#newcode524
src/parser.cc:524: AstValueFactory* ast_value_factory =
parser_->ast_value_factory();
just inline this

https://codereview.chromium.org/952243002/diff/1/src/scanner.cc
File src/scanner.cc (right):

https://codereview.chromium.org/952243002/diff/1/src/scanner.cc#newcode543
src/scanner.cc:543: token = Token::ILLEGAL;
This can be restructured so that  you don't need this line

[caitp (gmail), 2015-02-25 20:52:59]

I think, despite being observable, the big I'd probably ok if this can't be
implemented in the compiler at this time. My personal opinion is, it is a really
small impact on the compiler, it makes more sense, it's smaller and it works
better. Realistically I think it will cause fewer headaches there, rather than
making the parser crazier.

If people aren't convinced of either of these, i guess removing the assignment
op is the thing to do

https://codereview.chromium.org/952243002/diff/1/src/parser.cc
File src/parser.cc (right):

https://codereview.chromium.org/952243002/diff/1/src/parser.cc#newcode524
src/parser.cc:524: AstValueFactory* ast_value_factory =
parser_->ast_value_factory();
On 2015/02/25 16:58:25, arv wrote:
> just inline this

Acknowledged.

[caitp (gmail), 2015-02-25 20:53:54]

On 2015/02/25 20:52:59, caitp wrote:
> I think, despite being observable, the big I'd probably ok if this can't be
> implemented in the compiler at this time. My personal opinion is, it is a
really
> small impact on the compiler, it makes more sense, it's smaller and it works
> better. Realistically I think it will cause fewer headaches there, rather than
> making the parser crazier.
> 
> If people aren't convinced of either of these, i guess removing the assignment
> op is the thing to do
> 
> https://codereview.chromium.org/952243002/diff/1/src/parser.cc
> File src/parser.cc (right):
> 
> https://codereview.chromium.org/952243002/diff/1/src/parser.cc#newcode524
> src/parser.cc:524: AstValueFactory* ast_value_factory =
> parser_->ast_value_factory();
> On 2015/02/25 16:58:25, arv wrote:
> > just inline this
> 
> Acknowledged.

The bug is* mobile =_=