[turbofan] Some javascript operators are globally shared singletons.

[turbofan] Some javascript operators are globally shared singletons.

Also cleanup the interface, and make the parameter class/accessors
explicit to work-around the type-unsafety of OpParameter<T>.

TEST=compiler-unittests,cctest,mjsunit
R=mstarzinger@chromium.org

Committed: https://code.google.com/p/v8/source/detail?r=24322

[Benedikt Meurer, 2014-09-29 12:11:13]

PTAL

[Michael Starzinger, 2014-09-29 15:24:06]

Looking good modulo one comment which I feel pretty strongly about, and that we
already talked offline about. I think we should avoid making construction of the
parameter-structures explicit.

https://codereview.chromium.org/613683002/diff/20001/src/compiler/ast-graph-b...
File src/compiler/ast-graph-builder.cc (right):

https://codereview.chromium.org/613683002/diff/20001/src/compiler/ast-graph-b...
src/compiler/ast-graph-builder.cc:79:
CallRuntimeParameters(Runtime::kTraceEnter, 0)));
I think we should continue to let the factory method materialize the
CallRuntimeParameters (and friends) for us instead of doing it explicitly.
Otherwise readability of these call-sites suffers.

[Benedikt Meurer, 2014-09-30 04:44:05]

Done

https://codereview.chromium.org/613683002/diff/20001/src/compiler/ast-graph-b...
File src/compiler/ast-graph-builder.cc (right):

https://codereview.chromium.org/613683002/diff/20001/src/compiler/ast-graph-b...
src/compiler/ast-graph-builder.cc:79:
CallRuntimeParameters(Runtime::kTraceEnter, 0)));
On 2014/09/29 15:24:06, Michael Starzinger wrote:
> I think we should continue to let the factory method materialize the
> CallRuntimeParameters (and friends) for us instead of doing it explicitly.
> Otherwise readability of these call-sites suffers.

Done.

[Michael Starzinger, 2014-09-30 10:39:38]

LGTM.

[Benedikt Meurer, 2014-09-30 10:43:04]

Committed patchset #6 (id:90001) manually as 24322 (presubmit successful).