Hello,

I’m working on a port of webkit currently and haven’t been able to 
appropriately address hi-dpi device scale factors while using accelerated 
compositing with TextureMapper/TextureMapperGL (note i’m not using coordinated 
graphics and its running on a WebKit1 type interface). 

The issue is I don’t seem to see how to adjust the scale factor without 
creating all of the root graphics layers at twice the size, handle invalidation 
and setDisplayInRect in the device pixel size and subsequently start polluting 
or introducing the always fun issue of “is this pixel in device or logical 
pixels?” problem.

Is there some sort of way of implementing device scale factors in 
TextureMapper/TextureMapperGL? I’ve tried Page::setDeviceScaleFactor, 
Frame::tiledBackingStore()->setContentsScale.

I’ve also tried a patch I found recently for TextureMapper, but it required 
VERY carefully creating graphics layers, painting them and invalidating rects 
in different pixel coordinates (logical/device) depending on recursive 
conditions which makes things uber complicated.  In addition I never was able 
to get it working correctly with CSS3 animations (they’d typically render with 
half sized clips, or the x/y would be half size when the device scale factor 
was 2). 

http://fishsoup.net/misc/webkit-hidpi-patches/0004-GTK-WK2-Implement-hi-dpi-support-for-accelerated-com.patch

At anyrate, i’m hoping someone here can help give me some background or 
guidance on how to start implementing this.  If the only option at the moment 
is facing the uphill “implement in texture mapper” problem i’m fine with it, i 
just want to make sure its the only approach to the issue since i’m not using 
coordinated compositor.

You can find my webcoresupport here:
https://github.com/trevorlinton/webkit.js/tree/master/src/WebCoreSupport

And the platform API here:
https://github.com/trevorlinton/webkit.js/blob/master/src/WebView.cpp

- Trevor


_______________________________________________
webkit-help mailing list
[email protected]
https://lists.webkit.org/mailman/listinfo/webkit-help

Reply via email to