Reviewed by John, Maciej.
        
        Added test for whether a frame element reports its src attribute as a
        complete, rather than relative, URL.

        * fast/frames/frame-src-attribute-expected.txt: Added.
        * fast/frames/frame-src-attribute.html: Added.
        * fast/frames/resources/frame-src-attribute-subframe.html: Added.

WebCore:

        Reviewed by John, Maciej.
        
        Integrated some frame and iframe code. I'm trying to fix up frame ownership
        and loading. Reducing the number of different code paths involved
        seemed like a good first step.
        
        As a side effect, I fixed a bug where FRAME elements would report their
        src attributes as relative, rather than compelete, URLs. (IFRAME elements
        had the correct complete URL behavior.)

        * html/HTMLFrameElement.cpp:
        (WebCore::HTMLFrameElement::isURLAllowed): Fixed comment typo
        (WebCore::HTMLFrameElement::openURL):
            (1) Removed checks that requestFrame does for us
            (2) Added isURLAllowed check, to have one clear bottleneck for it
            (3) Added viewsource check, to have one clear bottleneck for it
        (WebCore::HTMLFrameElement::close): Changed to use the common contentFrame()
        method, instead of finding our content frame in our own unique way.
        (WebCore::HTMLFrameElement::setLocation): Removed isURLAllowed check,
        since openURL does this for us now.
        (WebCore::HTMLFrameElement::src): Return complete URL instead of relative.
        This is what FF does, and it made no sense to have different behaviors
        for FRAME and IFRAME elements.
        * html/HTMLIFrameElement.cpp:
        (WebCore::HTMLIFrameElement::HTMLIFrameElement): Removed duplicate init
        code.
        * html/HTMLIFrameElement.h: Removed src() and openURL() methods, since
        HTMLFrameElement now does everything we need.



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@16661 268f45cc-cd09-0410-ab3c-d52691b4dbfc