[Webkit-unassigned] [Bug 37218] Implement FileStreamProxy that calls FileStream methods on FileThread for FileAPI

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=37218





--- Comment #16 from Kinuko Yasuda <kinuko at chromium.org>  2010-04-09 16:13:08 PST ---
(In reply to comment #13)
> (From update of attachment 52921 [details])
> I think the way to fix "new file mode..." problem in the patch is to create a
> new file
> from the scratch and then copy and paste the content over. You need to revert
> all new
> files just added and then create files and add them to the repository.

I tried something similar to this.  Hope it'll work this time.

> > diff --git a/WebCore/html/FileStreamProxy.cpp b/WebCore/html/FileStreamProxy.cpp
> > +FileStreamProxy::FileStreamProxy(ScriptExecutionContext* context, FileStreamClient* client)
> > +    ref();
> Could you please add the comment about why we need to keep its own ref here?

Done.

> > +void FileStreamProxy::stop()
> > +{
> > +    fileThread()->unscheduleTasks(m_stream.get());
> > +    fileThread()->postTask(createFileThreadTask(m_stream.get(), &FileStream::stop));
> > +}
> I still think we do not need both abort() and stop(). Since we are not going to
> resume
> after abort, can we simply call stop() from FileReader/FieWriter::abort()?

OK, I removed the abort method and added some more comments (chose stop() over
abort() just for consistency).

> > diff --git a/WebCore/html/FileStreamProxy.h b/WebCore/html/FileStreamProxy.h
> > +class FileStreamProxy : public RefCounted<FileStreamProxy>, public FileStreamClient {
> > +protected:
> If we're not going to have other child classes derived from FileStreamProxy,
> it would be better to make it "private".

Right, I've been missing to fix it from another merge... Done.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.