Adapt hydrogen-based Array constructor to also support InternalArray and function call

Adapt hydrogen-based Array constructor to also support InternalArray and function call
versions.

BUG=
R=mstarzinger@chromium.org

Committed: https://code.google.com/p/v8/source/detail?r=14948

[mvstanton, 2013-05-16 09:58:28]

Hi Danno, could you have a look?
thanks,
--Michael

[mvstanton, 2013-05-16 10:46:25]

And/or if Michael could have a look, all the best,
--Michael

[Michael Starzinger, 2013-05-23 09:16:49]

LGTM from my end with a few nits. The one big comment about runtime function
moving from builtins.cc to runtime.cc we already discussed offline. Also
architecture ports are still missing.

https://codereview.chromium.org/14576005/diff/6001/src/builtins-decls.h
File src/builtins-decls.h (right):

https://codereview.chromium.org/14576005/diff/6001/src/builtins-decls.h#newco...
src/builtins-decls.h:36: DECLARE_RUNTIME_FUNCTION(MaybeObject*,
ArrayConstructor_StubFailure);
As discussed offline: I would really like to see these two runtime functions be
moved into runtime.cc, but that is a job for a follow-up CL. But we might want
to leave a TODO here in that regard.

https://codereview.chromium.org/14576005/diff/6001/src/builtins.cc
File src/builtins.cc (right):

https://codereview.chromium.org/14576005/diff/6001/src/builtins.cc#newcode197
src/builtins.cc:197: static MaybeObject* CommonArrayConstructorStubFailure(
nit: Can we call this "ArrayConstructorStubFailureCommon", that would be more in
line with the rest of the code base.

https://codereview.chromium.org/14576005/diff/6001/src/code-stubs-hydrogen.cc
File src/code-stubs-hydrogen.cc (right):

https://codereview.chromium.org/14576005/diff/6001/src/code-stubs-hydrogen.cc...
src/code-stubs-hydrogen.cc:563: if (argument_class == NONE) {
Can we model that as a switch without a default case? This way the compiler will
puke when we extend the enum above.

https://codereview.chromium.org/14576005/diff/6001/src/code-stubs-hydrogen.cc...
src/code-stubs-hydrogen.cc:580: if (argument_class == NONE) {
Likewise.

https://codereview.chromium.org/14576005/diff/6001/src/hydrogen.cc
File src/hydrogen.cc (right):

https://codereview.chromium.org/14576005/diff/6001/src/hydrogen.cc#newcode1895
src/hydrogen.cc:1895: builder_(builder),
nit: Indentation is off.

https://codereview.chromium.org/14576005/diff/6001/src/hydrogen.cc#newcode1910
src/hydrogen.cc:1910: builder_(builder),
nit: Indentation is off.

https://codereview.chromium.org/14576005/diff/6001/src/hydrogen.h
File src/hydrogen.h (right):

https://codereview.chromium.org/14576005/diff/6001/src/hydrogen.h#newcode1271
src/hydrogen.h:1271: HValue* InternalEmitMapCode();
nit: Can we call this "EmitInternalMapCode()", that would be more in line with
the other naming schemes.

[mvstanton, 2013-05-23 14:46:05]

Hi Michael, thx for the review. I have added the ports now and addressed your
comments, all the best!
--Michael

https://codereview.chromium.org/14576005/diff/6001/src/builtins-decls.h
File src/builtins-decls.h (right):

https://codereview.chromium.org/14576005/diff/6001/src/builtins-decls.h#newco...
src/builtins-decls.h:36: DECLARE_RUNTIME_FUNCTION(MaybeObject*,
ArrayConstructor_StubFailure);
On 2013/05/23 09:16:49, Michael Starzinger wrote:
> As discussed offline: I would really like to see these two runtime functions
be
> moved into runtime.cc, but that is a job for a follow-up CL. But we might want
> to leave a TODO here in that regard.

TODO placed, and will follow up in a 2nd CL.

https://codereview.chromium.org/14576005/diff/6001/src/builtins.cc
File src/builtins.cc (right):

https://codereview.chromium.org/14576005/diff/6001/src/builtins.cc#newcode197
src/builtins.cc:197: static MaybeObject* CommonArrayConstructorStubFailure(
On 2013/05/23 09:16:49, Michael Starzinger wrote:
> nit: Can we call this "ArrayConstructorStubFailureCommon", that would be more
in
> line with the rest of the code base.

Done.

https://codereview.chromium.org/14576005/diff/6001/src/code-stubs-hydrogen.cc
File src/code-stubs-hydrogen.cc (right):

https://codereview.chromium.org/14576005/diff/6001/src/code-stubs-hydrogen.cc...
src/code-stubs-hydrogen.cc:563: if (argument_class == NONE) {
On 2013/05/23 09:16:49, Michael Starzinger wrote:
> Can we model that as a switch without a default case? This way the compiler
will
> puke when we extend the enum above.

Cool, didn't think about that good trick. Done.

https://codereview.chromium.org/14576005/diff/6001/src/code-stubs-hydrogen.cc...
src/code-stubs-hydrogen.cc:580: if (argument_class == NONE) {
On 2013/05/23 09:16:49, Michael Starzinger wrote:
> Likewise.

Done.

https://codereview.chromium.org/14576005/diff/6001/src/hydrogen.cc
File src/hydrogen.cc (right):

https://codereview.chromium.org/14576005/diff/6001/src/hydrogen.cc#newcode1895
src/hydrogen.cc:1895: builder_(builder),
On 2013/05/23 09:16:49, Michael Starzinger wrote:
> nit: Indentation is off.

Done.

https://codereview.chromium.org/14576005/diff/6001/src/hydrogen.cc#newcode1910
src/hydrogen.cc:1910: builder_(builder),
On 2013/05/23 09:16:49, Michael Starzinger wrote:
> nit: Indentation is off.

Done.

https://codereview.chromium.org/14576005/diff/6001/src/hydrogen.h
File src/hydrogen.h (right):

https://codereview.chromium.org/14576005/diff/6001/src/hydrogen.h#newcode1271
src/hydrogen.h:1271: HValue* InternalEmitMapCode();
On 2013/05/23 09:16:49, Michael Starzinger wrote:
> nit: Can we call this "EmitInternalMapCode()", that would be more in line with
> the other naming schemes.

Done.

[mvstanton, 2013-06-04 16:04:41]

Hi Michael, thanks for the lgtm a week ago. I want to call your attention to the
recent update, could you look at the delta between patches 6 and 7? This is to
deal with the chrome issue experienced last week. Basically I need to correctly
handle Smi type feedback in Call sites as well as CallNew sites.

If that looks okay I'll check this in now.
All the best,
--Michael

[Michael Starzinger, 2013-06-05 09:59:13]

Still LGTM (just looked at patch set 7)

[mvstanton, 2013-06-05 10:43:32]

Committed patchset #7 manually as r14948 (presubmit successful).