On Fri, 29 Oct 2021 06:00:54 GMT, Sergey Bylokhov <s...@openjdk.org> wrote:
> > We have same test failure in macos-aarch64 in JDK-8266245 and I guess > > tolerance level of 5 should not be an issue. I was told earlier tolerance > > level of upto 15 should be acceptable.. > > I do not like the tolerance approaches in general until we do not recognize > the root cause of the wrong color. There were many cases in the past when the > people said the issue was caused by the "broken driver"/"some non-default > configuration"/ etc...etc. But in reality, the problem was caused by our code > when we use the wrong color composite/ the wrong color profile/etc. ah colour tolerance. We should not need that here. Should be an exact match ------------- PR: https://git.openjdk.java.net/jdk/pull/6140
