On Mon, 25 Apr 2022 09:37:10 GMT, Prasanta Sadhukhan <psadhuk...@openjdk.org> wrote:
> Test used to fail in specific CI macos M1 system owing to miniscule color > difference > > > x 0 y 0 red1 171 red2 171 green1 174 green2 175 blue1 184 blue2 184 > x 0 y 1 red1 172 red2 173 green1 177 green2 177 blue1 185 blue2 185 > x 0 y 2 red1 173 red2 174 green1 177 green2 178 blue1 187 blue2 187 > x 0 y 6 red1 0 red2 0 green1 1 green2 0 blue1 0 blue2 0 > x 0 y 15 red1 1 red2 0 green1 0 green2 0 blue1 0 blue2 0 > > > SInce we already have color-tolerance check present, there is no need of > exact color value check. > Also, made the frame undecorated and remove unneeded library being built. > Several iterations of the test passed in the same system (where it used to > fail 4/10) along with other platforms (link in JBS) BTW did you have a chance to check why to images are not the same? Is it a problem in the robot or in the rendering? Can it be related to upscaling/downscaling? ------------- PR: https://git.openjdk.java.net/jdk/pull/8380