On Thu, 24 Feb 2022 09:31:57 GMT, Prasanta Sadhukhan <psadhuk...@openjdk.org> wrote:
>> DamonGuy has updated the pull request incrementally with one additional >> commit since the last revision: >> >> Fixed incomplete comment. > > test/jdk/javax/swing/JButton/HtmlButtonImageTest/HtmlButtonImageTest.java > line 93: > >> 91: Color rightClr = robot.getPixelColor(point.x + (SQUARE_WIDTH/2) >> - PIXEL_BUFFER, point.y); >> 92: Color topClr = robot.getPixelColor(point.x, point.y - >> (SQUARE_HEIGHT/2) + PIXEL_BUFFER); >> 93: Color botClr = robot.getPixelColor(point.x, point.y + >> (SQUARE_HEIGHT/2) - PIXEL_BUFFER); > > there was an issue of using getPixelColor on retina screen on mac laptop as > @honkar-jdk found out in > #[7219](https://git.openjdk.java.net/jdk/pull/7219)..She has to use > BufferedImage..Are you sure this is working on all mac CI system, mac > external monitor and mac laptop? Initially, I ran into this issue as well a few weeks ago when I made the test. My solution was to check for redness by checking for 250 or greater R values with 10 or less G and B values instead of checking for exactly 255,0,0. When tested in mach5, some Macs failed with values that had pixels that fit within this color range, but wasn't exactly 255,0,0. ------------- PR: https://git.openjdk.java.net/jdk/pull/7310
