On Fri, 29 Oct 2021 06:43:38 GMT, Prasanta Sadhukhan <psadhuk...@openjdk.org> wrote:
>> As per JDK-8252813, some tests fails recurringly in CI macos system. This is >> an attempt to fix the swing tests. >> It was seen from the logs that we have color mismatch in these tests. >> >> For example, in PressedIcon test, we had following log >> >> ----------System.out:(1/75)---------- >> color java.awt.Color[r=55,g=55,b=55] COLOR_1Xjava.awt.Color[r=255,g=0,b=0] >> ----------System.err:(13/842)---------- >> java.lang.RuntimeException: Colors is different for scale=1! >> at PressedIconTest.main(PressedIconTest.java:97) >> at >> java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native >> Method) >> at >> java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:76) >> at >> java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:51) >> at java.base/java.lang.reflect.Method.invoke(Method.java:569) >> at >> com.sun.javatest.regtest.agent.MainWrapper$MainThread.run(MainWrapper.java:127) >> at java.base/java.lang.Thread.run(Thread.java:833) >> >> and JInternalFrame test, we had >> ----------System.err:(15/939)---------- >> FRAME_COLOR Red: 255; Green: 200; Blue: 0 >> Pixel color Red: 55; Green: 55; Blue: 55 >> java.lang.RuntimeException: Internal frame is not correctly dragged! >> at bug8069348.main(bug8069348.java:96) >> at >> java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native >> Method) >> at >> java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:76) >> at >> java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:51) >> at java.base/java.lang.reflect.Method.invoke(Method.java:569) >> at >> com.sun.javatest.regtest.agent.MainWrapper$MainThread.run(MainWrapper.java:127) >> at java.base/java.lang.Thread.run(Thread.java:833) >> >> where color is obtained as 55,55,55 instead of required color. The png image >> generated at failure also did not reveal anything abnormal apart from >> presence of mouse cursor around the same area where the getPixelColor >> location is pointing to. And since mouse cursor is also grayish black, it >> seems robot was wrongly picking up cursor pixels instead of background color. >> Modified tests to use location.x-10, location.y-10 to obtain the pixel >> color. It does not hamper the test as the whole area around the location is >> coloured with same color in the test so getting pixel color from a bit wide >> of the centre will not affect the test. >> Modified swing tests are passing in the affected system and also in other >> macos systems and platforms. Link in JBS. >> >> The awt tests that are failing seems to have different root cause and will >> be handled separately. > > Prasanta Sadhukhan has updated the pull request incrementally with two > additional commits since the last revision: > > - Review fix > - Fix I don't see a reason to discard the various stability improvements. The discussion is long so I may not be summarising correctly but - A product fix may resolve most of the issues and yet - Stability fixes are still valuable - The one test with a colour tolerance update may be better platform problem listed - The offset of -10,-10 for the cursor maybe better if it was parameterised and a bit bigger. ------------- PR: https://git.openjdk.java.net/jdk/pull/6140
