The jump that tests if the value is too big isn't linked, so we end up setting the value to zero instead of 255.
<rdar://problem/10423888>
Created attachment 114404 [details] the patch Not quite sure how to test this since I've not used byte arrays… Oliver?
(In reply to comment #2) > Created an attachment (id=114404) [details] > the patch > > Not quite sure how to test this since I've not used byte arrays… Oliver? theoretically canvas = document.createElement("canvas") context = canvas.getContext("2d") imageData = context.createImageData(10,10); data = imageData.data; // data is a byte array
Created attachment 114605 [details] the patch Added test.
Landed in http://trac.webkit.org/changeset/99917
This test appears to be flaky on lion: http://build.webkit.org/old-results/Lion%20Intel%20Release%20(Tests)/r99919%20(2597)/fast/js/dfg-byte-array-put-pretty-diff.html