This is the side effect of https://bugs.webkit.org/show_bug.cgi?id=71225. When page scale factor is greater than one, the invalidation area is not scale correctly. This caused some area are not updated correct when it needs.
Created attachment 117074 [details] Patch
Comment on attachment 117074 [details] Patch Need a test!
Comment on attachment 117074 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=117074&action=review > Source/WebCore/platform/graphics/chromium/GraphicsLayerChromium.cpp:684 > - return pageScaleFactor() * deviceScaleFactor(); > + return pageScaleFactor(); This is going to break ChromeOS, where pageScaleFactor is 1 but deviceScaleFactor is not (see NonCompositedContentHost.h)
Comment on attachment 117074 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=117074&action=review >> Source/WebCore/platform/graphics/chromium/GraphicsLayerChromium.cpp:684 >> + return pageScaleFactor(); > > This is going to break ChromeOS, where pageScaleFactor is 1 but deviceScaleFactor is not (see NonCompositedContentHost.h) Also, note that deviceScaleFactor() would be 1 on non-ChromeOS already (see WebViewImpl.cpp:3685)
I wonder if this is related: https://bugs.webkit.org/show_bug.cgi?id=88762
This patch appears to no longer be needed.