[Webkit-unassigned] [Bug 39582] [Gtk] very slow page scrolling on big -webkit-box-shadow areas

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=39582





--- Comment #23 from Alejandro G. Castro <alex at igalia.com>  2010-08-18 04:03:33 PST ---
(In reply to comment #20)
> (From update of attachment 64004 [details])
>
> [...]
>
> A great patch, but maybe a bit more complicated than it should be.
> This should be set at once:
>    tileRect = FloatRect(0, smallBufferSize.height() - 2 * radius - bottomRightRadius.height(), ...)
> Of course you can use multiple line for that.
>

I just used the sets because with FloatRect we are creating a new object each time, but not sure which option is better wrt performance or if it make sense, I agree with you creating the objects is clearer, I¡ll change them.

> You should use the logic of FloatRect/Size/Point much more in your code. and also it's funtcions. E.g. inflateX and inflateY extends the rect. Please use them to enlarge your rect to respect the blurring.

I think in this case (if it is the case I'm checking :) is not an inflate but a size change, we do not want to move the location.

> Don't mix up the enlargement of your rects with the offset. You can apply the blur radius to the with and position of the rects and move the rect by the shadowOffset.
> 

Yeah, I'll review all the calculations to make them clearer

> 
> WebCore/platform/graphics/cairo/GraphicsContextCairo.cpp:1296
>  +      if (offsetWidth > 0)
> use if (offsetWidth)

In this case we have to use > 0 because it is not just a boolean check, in the else we want to check the case where offsetWidth is less than 0.

Thanks for the comments, I'll try to have a patch ready, I will also use the new drawPattern added by Hyatt to avoid the copies of the image.

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