https://codereview.chromium.org/1149133005/diff/100001/src/compiler/ast-graph-builder.cc
File src/compiler/ast-graph-builder.cc (right):

https://codereview.chromium.org/1149133005/diff/100001/src/compiler/ast-graph-builder.cc#newcode2313
src/compiler/ast-graph-builder.cc:2313: if (!property->IsSuperAccess())
{
On 2015/06/05 13:50:26, arv wrote:
On 2015/06/05 13:28:19, Michael Starzinger wrote:
> Could this be modeled as a switch over the LhsKind as well? If so,
can we
leave
> a TODO here in that regard?

I tried it but reverted back to the ifs. I found the switch cleaner
but it does
mean more code duplication.

Acknowledged. OK, then let's leave it as it is.

https://codereview.chromium.org/1149133005/diff/100001/src/compiler/ast-graph-builder.cc#newcode3461
src/compiler/ast-graph-builder.cc:3461: return Record(js_type_feedback_,
value, feedback.slot());
On 2015/06/05 13:50:26, arv wrote:
On 2015/06/05 13:28:19, Michael Starzinger wrote:
> Just for clarification: My understanding is that we don't gather
type-feedback
> for any of these super loads and stores yet. Correct? I am still
fine with
> recording the nodes in the JSTypeFeedbackTable, because the overhead
should be
> negligible. Just wanted to check.

I think that is right. At some point we probably want this, at least
for loads,
I think we can skip it for super-store since it is not nearly as
common as
super-load.

Acknowledged.

https://codereview.chromium.org/1149133005/

--
--
v8-dev mailing list
[email protected]
http://groups.google.com/group/v8-dev
--- You received this message because you are subscribed to the Google Groups "v8-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to