On Fri, 11 Sep 2026 11:12:46 GMT, Prasanta Sadhukhan <[email protected]> wrote:
>> Test used to fail on linux in CI in older releases but with upgrade in >> harfbuzz and linux, with current jdk, it is not failing now in several >> iterations in linux and other platforms (CI test job in JBS) so >> deproblemlisting >> >> >> >> >> --------- >> - [x] I confirm that I make this contribution in accordance with the >> [OpenJDK Interim AI Policy](https://openjdk.org/legal/ai). > > Prasanta Sadhukhan has updated the pull request incrementally with one > additional commit since the last revision: > > Use Dialog font The problem I see is that it is the checkColors test that always fails. And I can see why. The test uses bg.fill(font.getStringBounds(STR, bg.getFontRenderContext())); to color over where the text was drawn and expects NO pixels from the text to be left. However getStringBounds() does NOT promise to tell you about all the pixels that are touched. It tells you about the "LOGICAL" bounds - for where to place the text. At the very least this needs to get VISUAL bounds to be remotely correct. So I see black text pixels which the checkColors isn't expecting and the test fails. It fails with AA and non-AA so using AA isn't relevant, but I don't know why the text uses KEY_ANTIALIASING instead of KEY_TEXT_ANTIALIASING ------------- PR Comment: https://git.openjdk.org/jdk/pull/32742#issuecomment-5768558278
