Implement FetchEvent.navigationPreload

third_party/WebKit/Source/bindings/core/v8/V8Binding.h

 /*
 * Copyright (C) 2009 Google Inc. All rights reserved.
 * Copyright (C) 2012 Ericsson AB. All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions are
 * met:
 *
 *     * Redistributions of source code must retain the above copyright
 * notice, this list of conditions and the following disclaimer.
@@ skipping 353 matching lines @@
                         callbackInfo.Holder(), callbackInfo.GetIsolate()));
 }
 
 template <typename CallbackInfo, typename T, typename Wrappable>
 inline void v8SetReturnValueFast(const CallbackInfo& callbackInfo,
                                  PassRefPtr<T> impl,
                                  const Wrappable* wrappable) {
   v8SetReturnValueFast(callbackInfo, impl.get(), wrappable);
 }
 
+template <typename CallbackInfo, typename Wrappable>
+inline void v8SetReturnValueFast(const CallbackInfo& callbackInfo,

[yhirano, 2016/10/21 01:48:58]

Can you tell me why this is needed?

[horo, 2016/10/21 10:56:40]

navigationPreloadAttributeGetter() in V8FetchEvent.cpp which is generated from
FetchEvent.idl requires this method.

static void navigationPreloadAttributeGetter(const
v8::FunctionCallbackInfo<v8::Value>& info)
{
    v8::Local<v8::Object> holder = info.Holder();

    FetchEvent* impl = V8FetchEvent::toImpl(holder);

    ScriptState* scriptState = ScriptState::forReceiverObject(info);
    bool isNull = false;

    ScriptPromise cppValue(impl->navigationPreload(scriptState, isNull));

    if (isNull) {
        v8SetReturnValueNull(info);
        return;
    }

    v8SetReturnValueFast(info, cppValue, impl);
}
 

[yhirano, 2016/10/24 07:13:39]

+yukishiino@chromium.org

I think this is a bug in the generator; nullable Promise is treated as
ScriptWrappable. yukishiino@, what do you think?

[Yuki, 2016/10/24 08:43:21]

Is this spec-conformant?  I've never hard of nullable-Promise, and it looks a
little weird.

For DOM operations, they must be either of promise-returning function or
non-promise-returning function.  If it's a promise returning function, an
exception must be converted to a reject promise.

Having said that, there is no such rule for DOM attributes.  Also it's not
prohibited to make a promise type nullable seeing the following spec.
https://heycam.github.io/webidl/#idl-nullable-type
However, I'm afraid that it's just because the spec is not consistent nor
up-to-date.  It's possible that someone forgot to update the definition of
"nullable" when they introduced promise type.

Are you sure that it's okay to introduce a nullable promise type?  There is no
such definition at
https://www.w3.org/TR/service-workers/#fetch-event-section

+                                 blink::ScriptPromise promise,
+                                 const Wrappable* wrappable) {
+  v8SetReturnValue(callbackInfo, promise.v8Value());
+}
+
 // Convert v8::String to a WTF::String. If the V8 string is not already
 // an external string then it is transformed into an external string at this
 // point to avoid repeated conversions.
 inline String toCoreString(v8::Local<v8::String> value) {
   return v8StringToWebCoreString<String>(value, Externalize);
 }
 
 inline String toCoreStringWithNullCheck(v8::Local<v8::String> value) {
   if (value.IsEmpty() || value->IsNull())
     return String();
@@ skipping 765 matching lines @@
 
 // Freeze a V8 object. The type of the first parameter and the return value is
 // intentionally v8::Value so that this function can wrap toV8().
 // If the argument isn't an object, this will crash.
 CORE_EXPORT v8::Local<v8::Value> freezeV8Object(v8::Local<v8::Value>,
                                                 v8::Isolate*);
 
 }  // namespace blink
 
 #endif  // V8Binding_h