[Webkit-unassigned] [Bug 43779] RenderImage::imageChanged invalidates wrong area

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=43779





--- Comment #17 from Patrick R. Gansterer <paroga at paroga.com>  2010-08-12 00:27:07 PST ---
(In reply to comment #15)
> (From update of attachment 64162 [details])
> WebCore/rendering/RenderImage.cpp:61
>  +      m_cachedImage = RenderImageResource::create(this);
> No need to move this into the constructor body.
Is it safe to use this in initialization list? Some compiler throw warnings?

> WebCore/rendering/RenderImage.h:46
>  +      void setCachedImage(CachedImage* image) { m_cachedImage->setCachedImage(image); }
> Who calls these methods? It seems to me, all callers of these functions could now operate on RenderImageResource. Please check twice.
They are called on all other places where images are used (e.g. Pasteboard, DragDrop, ...)
If you prefere it that way, I'll change it.

(In reply to comment #16)
> (From update of attachment 64163 [details])
> Also some comments:
> WebCore/rendering/RenderSVGImage.cpp:51
>  +      m_cachedImage = RenderImageResource::create(this);
> Can be moved in the initialization list.
Same as in RenderImage.

> WebCore/rendering/RenderSVGImage.cpp:104
>  +                  if (svgStyle->shapeRendering() == SR_CRISPEDGES)
> Is that necessary for images? I don't think so.
I copied fomr RenderPath. I'll remove it.

> WebCore/rendering/RenderSVGImage.cpp:169
>  +      // This empty image may be cached by SVG filter effects which must be invalidated.
> That change is not correct, you basically reverted to an older version of the imageChanged version.
> The SVGResourcesCache::... call is correct.
I think that happened when I rebased to head. I'll check it.

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