LGTM with a formatting suggestion.
Please have a close eye on performance; if there's other code somewhere that
needs to be updated to reflect these changes then that could cause tanking
in
some corner cases.
https://chromiumcodereview.appspot.com/23442016/diff/9001/src/ic.cc
File src/ic.cc (right):
https://chromiumcodereview.appspot.com/23442016/diff/9001/src/ic.cc#newcode1713
src/ic.cc:1713: if (!can_store &&
nit: I'm finding this heterogeneous if/else cascade a little hard to
read, and I think I'd prefer the following alternative formatting:
bool can_store = LookupForWrite(receiver, name, value, &lookup,
&state);
if (!can_store &&
strict_mode == kStrictMode &&
!(lookup.IsProperty() && lookup.IsReadOnly()) &&
IsUndeclaredGlobal(object)) {
// Strict mode doesn't allow setting non-existent global property.
return ReferenceError("not_defined", name);
}
if (FLAG_use_ic) {
if (state == UNINITIALIZED) {
Handle<Code> stub = (strict_mode == kStrictMode)
? pre_monomorphic_stub_strict()
: pre_monomorphic_stub();
set_target(*stub);
TRACE_IC("StoreIC", name, state, *stub);
} else if (can_store) {
UpdateCaches(&lookup, state, strict_mode, receiver, name, value);
} else if (!name->IsCacheable(isolate()) ||
lookup.IsNormal() ||
(lookup.IsField() && lookup.CanHoldValue(value))) {
Handle<Code> stub = (strict_mode == kStrictMode) ?
generic_stub_strict()
: generic_stub();
set_target(*stub);
}
}
I leave the decision up to you, though.
https://chromiumcodereview.appspot.com/23442016/
--
--
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/groups/opt_out.
