Expose Object::DefineOwnProperty on the API

include/v8.h

Index: include/v8.h
diff --git a/include/v8.h b/include/v8.h
index cdb23b4655d6b2caadde4a9bd16f45d99902e4fd..6bda3f4a0300c42a8e53e039b5353d83d0e48872 100644
--- a/include/v8.h
+++ b/include/v8.h
@@ -2686,6 +2686,10 @@ class V8_EXPORT Object : public Value {
   Maybe<bool> ForceSet(Local<Context> context, Local<Value> key,
                        Local<Value> value, PropertyAttribute attribs = None);
 
+  V8_WARN_UNUSED_RESULT MaybeLocal<Value> DefineObjectProperty(

[adamk, 2015/04/15 16:31:53]

Why not "DefineOwnProperty"? Also, it seems like it might be more useful to make
this non-throwing (in normal operation) and return a boolean, with true meaning
success and false meaning failure. This would better match up with
[[DefineOwnProperty]], which is what I was thinking of when I filed this bug.
The callers of this in the embedder are more like "spec code" than user code,
and the boolean is going to be easier to handle than sprinkling TryCatches and
ASSERTs around.

[dcarney, 2015/04/15 19:02:56]

i think a  maybe<bool> is better than a meaningless maybelocal, but in either
case, the result is meaningless. only the presence of a result signals non
exceptional execution, whereas the absence signal something thrown.  what we
would really need here would be maybe<void>

+      Local<Context> context, Local<Value> key, Local<Value> value,

[adamk, 2015/04/15 16:31:53]

Why not add a version that takes a uint32_t key? That'd match up with Set/Get,
and make this more convenient to use with Arrays (which is one of the places
where ForceSet currently doesn't work well).

+      PropertyAttribute attribute = None);
+
   V8_DEPRECATE_SOON("Use maybe version", Local<Value> Get(Handle<Value> key));
   V8_WARN_UNUSED_RESULT MaybeLocal<Value> Get(Local<Context> context,
                                               Local<Value> key);