[V8] Extend the set of types supported by SerializedScriptValue

[Vitaly Repeshko]

[V8] Extend the set of types supported by SerializedScriptValue

[Vitaly Repeshko]

Created attachment 52472 [details]
patch v1

[Yury Semikhatsky]

Comment on attachment 52472 [details]
patch v1

> -            push(new ObjectState(handleCast<v8::Object>(value)));
> -            // FIXME:
> -            // - check not a wrapper
> -            // - support File, ImageData, etc.
> -        }
> -        return true;
Please file a bug for the FIXME.

> +    static StateBase* newArrayState(v8::Handle<v8::Array> array, StateBase* next)
> +    {
> +        // FIXME: use plain Array state when we can quickly check that
> +        // an array is not sparse and has only indexed properties.
> +        return new SparseArrayState(array, next);
> +    }
> +
Ditto.

> +    static StateBase* newObjectState(v8::Handle<v8::Object> object, StateBase* next)
> +    {
> +        // FIXME:
> +        // - check not a wrapper
> +        // - support File, etc.
> +        return new ObjectState(object, next);
> +    }
Ditto.

[Yury Semikhatsky]

(In reply to comment #0)
> [V8] Extend the set of types supported by SerializedScriptValue

Can we write unit tests for serialize script values?

[Vitaly Repeshko]

(In reply to comment #3)
> (In reply to comment #0)
> > [V8] Extend the set of types supported by SerializedScriptValue
> 
> Can we write unit tests for serialize script values?

We can, but most interesting stuff is already covered by existing layout tests. (See fast/dom/Window/window-postmessage*.)

[Vitaly Repeshko]

Created attachment 52539 [details]
patch v1.1: rebased

[Vitaly Repeshko]

	M	WebCore/ChangeLog
	M	WebCore/bindings/v8/SerializedScriptValue.cpp
Committed r57079

[WebKit Review Bot]

http://trac.webkit.org/changeset/57079 might have broken Tiger Intel Release

[WebKit Review Bot]

http://trac.webkit.org/changeset/57079 might have broken Tiger Intel Release

[Eric Seidel (no email)]

bug 36798 seems to be the actual culprit.  Sheriffbot should not have warned twice.

[WebKit Review Bot]

http://trac.webkit.org/changeset/57079 might have broken Tiger Intel Release

[WebKit Review Bot]

http://trac.webkit.org/changeset/57079 might have broken Tiger Intel Release

[Adam Barth]

Sorry for the repeated messages.  I'm figuring out why he's spamming.

[Adam Barth]

The problem is [e-mail] is not in committers.py.  Fixing.

[Adam Barth]

Sorry, he's going to spam one more time when I bring him back online.

[WebKit Review Bot]

http://trac.webkit.org/changeset/57079 might have broken Tiger Intel Release

[Pavel Feldman]

Comment on attachment 52539 [details]
patch v1.1: rebased

Reverted for crashing layout test in:
fast/dom/Window/window-postmessage-clone.html = CRASH

Committing to http://svn.webkit.org/repository/webkit/trunk 
	M	WebCore/ChangeLog
	M	WebCore/bindings/v8/SerializedScriptValue.cpp
Committed r57133

[Pavel Feldman]

Reopening since has been rolled out.

[Vitaly Repeshko]

Created attachment 52625 [details]
patch v1.2: fixed delete bug in handleError

[Vitaly Repeshko]

	M	WebCore/ChangeLog
	M	WebCore/bindings/v8/SerializedScriptValue.cpp
Committed r57145