Keyed stores to super where key is a name.

Keyed stores to super where key is a name.

R=arv@chromium.org, ishell@chromium.org
BUG=v:3330
LOG=N

Committed: https://code.google.com/p/v8/source/detail?r=24490

[Dmitry Lomov (no reviews), 2014-10-07 14:09:01]

PTAL, starting platform ports.

[arv (Not doing code reviews), 2014-10-07 15:05:38]

LGTM but I would want Igor to verify that the codegen is good too.

https://codereview.chromium.org/638623002/diff/20001/src/ia32/full-codegen-ia...
File src/ia32/full-codegen-ia32.cc (right):

https://codereview.chromium.org/638623002/diff/20001/src/ia32/full-codegen-ia...
src/ia32/full-codegen-ia32.cc:4419: if (assign_type == NAMED_PROPERTY) {
maybe switch here?

https://codereview.chromium.org/638623002/diff/20001/src/runtime/runtime-clas...
File src/runtime/runtime-classes.cc (right):

https://codereview.chromium.org/638623002/diff/20001/src/runtime/runtime-clas...
src/runtime/runtime-classes.cc:164: return ThrowUnsupportedSuper(isolate);
// TODO(dslomov): Implement

https://codereview.chromium.org/638623002/diff/20001/src/runtime/runtime.h
File src/runtime/runtime.h (right):

https://codereview.chromium.org/638623002/diff/20001/src/runtime/runtime.h#ne...
src/runtime/runtime.h:196: F(StoreKeyedToSuper_Strict, 4, 1)                    
   \
Same comment here as I commented on for StoreToSuper. Can we make this one
runtime function that takes a Smi for StrictMode? Maybe there is a reason to
prefer two runtime functions here and if so I would like to know what it is.

[Dmitry Lomov (no reviews), 2014-10-07 15:53:37]

Comments addressed + platform ports

https://codereview.chromium.org/638623002/diff/20001/src/ia32/full-codegen-ia...
File src/ia32/full-codegen-ia32.cc (right):

https://codereview.chromium.org/638623002/diff/20001/src/ia32/full-codegen-ia...
src/ia32/full-codegen-ia32.cc:4419: if (assign_type == NAMED_PROPERTY) {
On 2014/10/07 15:05:38, arv wrote:
> maybe switch here?

Done.

https://codereview.chromium.org/638623002/diff/20001/src/runtime/runtime-clas...
File src/runtime/runtime-classes.cc (right):

https://codereview.chromium.org/638623002/diff/20001/src/runtime/runtime-clas...
src/runtime/runtime-classes.cc:164: return ThrowUnsupportedSuper(isolate);
On 2014/10/07 15:05:38, arv wrote:
> // TODO(dslomov): Implement

Done.

https://codereview.chromium.org/638623002/diff/20001/src/runtime/runtime.h
File src/runtime/runtime.h (right):

https://codereview.chromium.org/638623002/diff/20001/src/runtime/runtime.h#ne...
src/runtime/runtime.h:196: F(StoreKeyedToSuper_Strict, 4, 1)                    
   \
On 2014/10/07 15:05:38, arv wrote:
> Same comment here as I commented on for StoreToSuper. Can we make this one
> runtime function that takes a Smi for StrictMode? Maybe there is a reason to
> prefer two runtime functions here and if so I would like to know what it is.

Less parameter passing on the stack => smaller generated code size + slightly
faster call.

[Igor Sheludko, 2014-10-09 11:07:20]

lgtm with minor suggestion:

https://codereview.chromium.org/638623002/diff/40001/src/full-codegen.h
File src/full-codegen.h (right):

https://codereview.chromium.org/638623002/diff/40001/src/full-codegen.h#newco...
src/full-codegen.h:523: enum LhsKind {
Since we make the enum "public" maybe "LvalueKind" would be a better name. WDYT?

[Dmitry Lomov (no reviews), 2014-10-09 11:17:57]

On 2014/10/09 11:07:20, Igor Sheludko wrote:
> lgtm with minor suggestion:
> 
> https://codereview.chromium.org/638623002/diff/40001/src/full-codegen.h
> File src/full-codegen.h (right):
> 
>
https://codereview.chromium.org/638623002/diff/40001/src/full-codegen.h#newco...
> src/full-codegen.h:523: enum LhsKind {
> Since we make the enum "public" maybe "LvalueKind" would be a better name.
WDYT?

As discussed offline, we don't really have a concept of Lvalue in Javascript.
Let's keep it as is until we find a better name

[Dmitry Lomov (no reviews), 2014-10-09 11:36:36]

Committed patchset #3 (id:40001) manually as 24490 (presubmit successful).