[Webkit-unassigned] [Bug 54618] New: Web Inspector: Need renderer hint on implementing hovered node highlight in WebCore.

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Wed Feb 16 22:35:15 PST 2011


https://bugs.webkit.org/show_bug.cgi?id=54618

           Summary: Web Inspector: Need renderer hint on implementing
                    hovered node highlight in WebCore.
           Product: WebKit
           Version: 528+ (Nightly build)
          Platform: All
        OS/Version: All
            Status: UNCONFIRMED
          Severity: Normal
          Priority: P2
         Component: Web Inspector
        AssignedTo: webkit-unassigned at lists.webkit.org
        ReportedBy: pfeldman at chromium.org
                CC: hyatt at apple.com, timothy at apple.com,
                    simon.fraser at apple.com, jamesr at chromium.org


Summarizing chats that I've had on the IRC. Background:

When you hover over an element in inspector, we paint node highlight. Different ports handle it differently, all ultimately calling inspector's drawNodeHighlight defined in WebCore:

- Mac uses floating NSView in WebNodeHighlightView
- Windows uses floating native window in WebNodeHighlight
- GTK and Chromium simply issue repaint and stick inspector's drawNodeHighlight into their paint routines.
- WebKit2 is pending implementation

I'd like to unify the approaches and keep all the code in the WebCore. There are two scenarios where naive approach of putting drawNodeHighlight into FrameView does not work:
- plugins that are on top of the screen
- accelerated compositing

We can give up the plugins one, but compositing still requires special handling.

I am a rendering idiot and reaching out for your hints. As I see, we could do two things:
1. Bump element's zindex on hover in its style and call highlight code from the RenderObject
2. Somehow have additional RenderLayer that would paint highlight

(1) has a side-effect of bringing node above other elements, but sounds simpler. It might be considered as either feature or a regression, I am not sure.
(2) would be simply perceived as no regressions.

On a side note, there is a pending request to make inspector highlight follow the moving / transforming elements, so (1) sounds more feasible to me, but I know nothing about rendering.

What do I miss?

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



More information about the webkit-unassigned mailing list