On Fri, 30 Sep 2022 20:44:03 GMT, Phil Race <p...@openjdk.org> wrote:
> And test - which appears to be a straight copy of the sample code in the bug > report is sufficient to show there's a bug here and that you are fixing it > for that one case of ONE unicode code point but I think the test needs to be > more comprehensive - inserting and then undoing at the beginning and end of > some existing text. I suspect we have no existing test which would uncover a > regression being introduced here, so this test needs to be more comprehensive. And with some text being inserted in the middle of existing text in the document. I also wonder if changing component orientation breaks undo/redo with left-to-right text only since RTL component orientation enables bidi support. ------------- PR: https://git.openjdk.org/jdk/pull/10446
