[Webkit-unassigned] [Bug 103513] Provide the backend for exposing the layer tree to the Web Inspector

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=103513





--- Comment #38 from Antoine Quint <graouts at apple.com>  2012-12-01 03:10:53 PST ---
(In reply to comment #37)
> (In reply to comment #36)
> > (In reply to comment #34)
> > > (From update of attachment 176977 [details] [details] [details])
> > > View in context: https://bugs.webkit.org/attachment.cgi?id=176977&action=review
> > > 
> > > >> Source/WebCore/rendering/RenderLayerCompositor.cpp:514
> > > >> +    InspectorInstrumentation::layerTreeDidChange(this->page());
> > > > 
> > > > just page() ?
> > > 
> > > So the layer tree managed by InspectorLayerTreeAgent has regular render layers, yet you only update it when in the accelerated compositing mode -- this looks a bit confusing. Isn't there use in the layer tree if the accelerated compositing is not used? Should we make it explicit that the tree is only available in accelerated compositing mode?
> > 
> > The entire agent is guarded with #if USE(ACCELERATED_COMPOSITING).
> 
> That was another thing that I was curious about -- does it have to be? My original question, though, was rather about the effect of run-time compositing mode, not the compile-time ability to do accelerated compositing.

If we go back to the bug's description, the real intent of this agent is provide information about layers composited in hardware. What we're trying to address in the Safari Web Inspector are issues related to the increased memory footprint caused by layers and the ability to easily see what gets composited in a hardware-backed RenderLayer. As such, we only want to show this feature when ACCELERATED_COMPOSITING is on. We can revisit this later as needed, but I think in the context of this bug we should keep the flags as they are.

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