[Webkit-unassigned] [Bug 250382] Check for non-finite points before transforming them instead of after
bugzilla-daemon at webkit.org
bugzilla-daemon at webkit.org
Thu Feb 23 12:20:23 PST 2023
https://bugs.webkit.org/show_bug.cgi?id=250382
Said Abou-Hallawa <sabouhallawa at apple.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |sabouhallawa at apple.com
--- Comment #2 from Said Abou-Hallawa <sabouhallawa at apple.com> ---
The layout tests:
1) https://github.com/WebKit/WebKit/blob/28de25b653f64ca3fc8e87fbf1d1a963329a5342/LayoutTests/fast/canvas/canvas-path-isPointInPath.html
2) https://github.com/WebKit/WebKit/blob/28de25b653f64ca3fc8e87fbf1d1a963329a5342/LayoutTests/fast/canvas/canvas-path-isPointInStroke.html
test passing NaN values to isPointInPath() and isPointInStroke().
But we are fine because the complier ensures the math with NaN values are handled correctly all the way. However I think we should check
if (!std::isfinite(x) || !std::isfinite(y))
return false;
In CanvasRenderingContext2DBase::isPointInPathInternal() and CanvasRenderingContext2DBase::isPointInStrokeInternal() instead of checking the values of the transformedPoint.
We do not need to check whether the values transformedPoint.x() and transformedPoint.y() are finite because AffineTransform::mapPoint() does it math in double and then it clamp it to float.
--
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20230223/11358b9e/attachment.htm>
More information about the webkit-unassigned
mailing list