On Thu, 21 May 2026 22:25:15 GMT, Andy Goryachev <[email protected]> wrote:
>> This is a headless test, so the OS does not play any role here. >> I don't know 100% what the original purpose of this test was - as it never >> fulfilled any role. >> But this is close to the original test while we test a real table setup. >> >> In any OS it behaves correctly: Opening the `ContextMenu` from within the >> `TextField` will work and not stop the edit process (tested in the >> monkeytester). We try to reproduce that somewhat. >> >> EDIT: We could also e.g. send a `ContextMenu` event instead (like a real >> JavaFX app does). But that would change the test again compared to the OG >> test. But would be fine for me > > That's not what I meant. The test, for whatever reason, is testing a > perfectly valid scenario - mouse press followed by a release. The real > question is why did you have to change it? > > The test works fine with and without your changes on macOS. There might be > something in the fx code that handles platform differences between mac and > windows, so if the test failed on windows, it means this PR broke something. yep, the original test is broken on windows with this change. we shouldn't be changing test by replacing a legal sequence of events to an illegal one just to silence it. ------------- PR Review Comment: https://git.openjdk.org/jfx/pull/1935#discussion_r3284812889
