[Webkit-unassigned] [Bug 36520] [DRT/Chromium] Add LayoutTestController

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=36520





--- Comment #8 from TAMURA, Kent <tkent at chromium.org>  2010-03-26 00:09:38 PST ---
(In reply to comment #6)
> (From update of attachment 51600 [details])
> I am ok with checking this in as-is as long as we all understand this is far
> from final and is going in because we want to have something running to hack
> on. Please spread FIXMEs liberally and start filing bugs, so that we don't
> forget. Nate and I did this for V8 bindings refactoring, and it worked out
> really well
> (https://bugs.webkit.org/showdependencytree.cgi?id=32630&hide_resolved=0).
> 
> We will need to get rid of direct base dependencies, and convert all
> std::string uses to WebString, and remove STL use.

Removing base/ dependencies are understandable though I have no idea how to
remove Task interface for now.
Should we really remove std::string?  I thought WebString was just a bridge
between WebCore and Chromium code and we should have used String/AtomicString
in WebCore, string16/string in Chromium.  We need to add many functions to
WebString in order to replace string with WebString.

> On GURL::Resolve, we may have to poke this method through WebURL wrapper.

Same question as above.  Should we make WebKit API classes fully-functional
clases, not just wrappers?

> 
> > +    if (!m_queue.isEmpty()) {
> > +        // We delay processing queued work to avoid recursion problems.
> > +        m_timer.Start(base::TimeDelta(), this, &WorkQueue::processWork);
> 
> How did base/time squeak by here?

LayoutTestController.h still have base/timer.h.
I have added FIXME there.


I'll land this with some FIXME additions, and filed Bug#36641 for dependency
cleanup.

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