On Thu, 10 Nov 2022 06:01:39 GMT, Sergey Bylokhov <s...@openjdk.org> wrote:
>> Test intermittently fails in VM citing "Tree is not editing". Seems to be >> problem with mouse clicks not getting registered properly.. >> Similar test test/jdk/javax/swing/JTable/6263446/bug6263446.java is not >> affected, so made the test similar to it by using same safeguard using >> Robot.waitForIdle() and modifying clickpoint to tree cell midpoint. >> >> Several iterations of the test pass in the OCI VM and all other physical >> platforms (link in JBS) > > test/jdk/javax/swing/JTree/6263446/bug6263446.java line 131: > >> 129: robot.mousePress(InputEvent.BUTTON1_DOWN_MASK); >> 130: robot.mouseRelease(InputEvent.BUTTON1_DOWN_MASK); >> 131: robot.waitForIdle(); > > Why do we remove it here and then add it after each call of this "click" > method? why not move to the end of the method? I wanted to keep it same as JTable/626344 one and it worked in the OCI systems...I do not want to change it/optimise it with the access to test in these OCI systems being limited... > test/jdk/javax/swing/JTree/6263446/bug6263446.java line 186: > >> 184: frame.setAlwaysOnTop(true); >> 185: frame.setLocationRelativeTo(null); >> 186: frame.pack(); > > pack() after the setLocationRelativeTo() may move the frame out of the center > of the screen, is it intentional? it doesn't and I have checked ------------- PR: https://git.openjdk.org/jdk/pull/11057
