Add is_intrinsic and is_not_intrinsic bits to prevent repeated tests on the same function (inlining…

Add is_intrinsic and is_not_intrinsic bits to prevent repeated tests on the same function (inlining, recompilations). The bist are set lazily, initially both are false.
Committed: https://code.google.com/p/dart/source/detail?r=14444

[siva, 2012-11-01 23:04:06]

lgtm

https://codereview.chromium.org/11358047/diff/5001/runtime/vm/intrinsifier.cc
File runtime/vm/intrinsifier.cc (right):

https://codereview.chromium.org/11358047/diff/5001/runtime/vm/intrinsifier.cc...
runtime/vm/intrinsifier.cc:76: if (function.IsClosureFunction()) return false;
Should we do the closure function check first because
to always return false for these and the intrinsic bits will
never be set.

https://codereview.chromium.org/11358047/diff/5001/runtime/vm/intrinsifier.cc...
runtime/vm/intrinsifier.cc:90: }
Not related to your current change I was wondering if
we could move the line
const char*class_name = String::Handle(function_class.Name()).ToCString() to
after the if check it would save a handle creation for the negative case.

Also I don't think you need the core_lib, core_impl_lib handles, we could write
if (function_class.library() != Library::CoreLibrary() &&
    function_class.library() != Library::CoreImplLibrary()
    ....) {
  return false;
}

https://codereview.chromium.org/11358047/diff/5001/runtime/vm/object.h
File runtime/vm/object.h (right):

https://codereview.chromium.org/11358047/diff/5001/runtime/vm/object.h#newcod...
runtime/vm/object.h:1277: kUnknownIntrinsic,  // Initial value.
kUnknownIntrinsic = 0,

[srdjan, 2012-11-01 23:08:18]

Thanks for great suggestions.

https://codereview.chromium.org/11358047/diff/5001/runtime/vm/intrinsifier.cc
File runtime/vm/intrinsifier.cc (right):

https://codereview.chromium.org/11358047/diff/5001/runtime/vm/intrinsifier.cc...
runtime/vm/intrinsifier.cc:76: if (function.IsClosureFunction()) return false;
On 2012/11/01 23:04:06, siva wrote:
> Should we do the closure function check first because
> to always return false for these and the intrinsic bits will
> never be set.

Done.

https://codereview.chromium.org/11358047/diff/5001/runtime/vm/intrinsifier.cc...
runtime/vm/intrinsifier.cc:90: }
On 2012/11/01 23:04:06, siva wrote:
> Not related to your current change I was wondering if
> we could move the line
> const char*class_name = String::Handle(function_class.Name()).ToCString() to
> after the if check it would save a handle creation for the negative case.
> 
> Also I don't think you need the core_lib, core_impl_lib handles, we could
write
> if (function_class.library() != Library::CoreLibrary() &&
>     function_class.library() != Library::CoreImplLibrary()
>     ....) {
>   return false;
> }

Done.

https://codereview.chromium.org/11358047/diff/5001/runtime/vm/object.h
File runtime/vm/object.h (right):

https://codereview.chromium.org/11358047/diff/5001/runtime/vm/object.h#newcod...
runtime/vm/object.h:1277: kUnknownIntrinsic,  // Initial value.
On 2012/11/01 23:04:06, siva wrote:
> kUnknownIntrinsic = 0,

Done.