On Fri, 2 Sep 2022 07:18:13 GMT, Abhishek Kumar <d...@openjdk.org> wrote:
>> Another possible solution that I used in a previous test was a midpoint as >> suggested, but also one pixel near each edge. This eliminates the chance of >> the red still existing but not being detected at the midpoint because the >> red is at a different location. > > The test check has been changed to compare the pixels of entire width keeping > the height constant to half of tree height. Now it will check pixel color and > if it is not red , the test will fail. For me, test is not failing without the fix. Also, if you test entire width and fail if it is not red, it might encounter black pixel of text font too and might fail, so I think it's better to check width/2 to width -1 and not entire width. ------------- PR: https://git.openjdk.org/jdk/pull/10112
