On Mon, 22 Jun 2026 14:51:51 GMT, Marius Hanl <[email protected]> wrote:
>> Right, but the test failures are a separate problem that you've addressed in >> JDK-8387020. This test is correct in its setting the specific stylesheet, >> while the rest are probably ok because they are supposed to work with the >> default stylesheet (it happens to be modena.css today, but who knows what >> happens tomorrow). >> Right? > > Yes correct. > > But what I mean is that the preexisting test > `TextInputControlModenaTest#testHighlightTextInput` failed as well without > JDK-8387020. > And this test does not set the modena stylesheet with > `Application.setUserAgentStylesheet(Application.STYLESHEET_MODENA);` but > should do as well because it will fail if another stylesheet is used tomorrow. > > Thats why I propose to either remove > `Application.setUserAgentStylesheet(Application.STYLESHEET_MODENA);` or set > it in an `@BeforeEach` method - here in `TextInputControlModenaTest` and > reset in an `@AfterEach`, because not just this new test depends on modena, > but the existing one does as well. But a false cleanup from another test (in this case the test mentioned in JDK-8387020) that changes the Application user agent stylesheet value doesn't really mean that our test should fail no ? as it is a wrong implementation i guess in the that test. I tried it too and yes `TextInputControlModenaTest#testHighlightTextInput` fails without JDK-8387020 as it uses the previous value. Maybe with that new fix it is fine to remove `Application.setUserAgentStylesheet(Application.STYLESHEET_MODENA);` ------------- PR Review Comment: https://git.openjdk.org/jfx/pull/2184#discussion_r3454370664
