On Mon, 12 Feb 2024 16:15:03 GMT, Andy Goryachev <ango...@openjdk.org> wrote:
> It looks like the original intent was to have one monospaced (Monaco) and one > proportional (Tahoma) font. Perhaps we could come up with a list of > candidates and pick a font to use based on availability? These would all have to be separate tests, as each font will have different widths. Since the specific font is not really relevant for alignment and wrapping code coverage, only that it is a real font, I think this isn't going to really add much (the code is already fully covered, a different font won't add any new branch coverage). Even a mono spaced font makes no difference for the tests (it would at most be a bit easier to test for). Other tests are free to use different fonts. The linux build environment could have Arial installed (it is freely available), Mac apparently has some slight differences with complex chars in Arial -- I don't know why that happens, but it doesn't really affect the tests. I can split the tests for those in two parts with different assumptions for platform. Anyway, IMHO, all that matters is that the test runs in the nightly, and that regressions can be spotted, no matter what platform. If there are any Mac/Linux/Windows specific issues reported (that are not related to the specific font), we can add more regression tests for those then. ------------- PR Comment: https://git.openjdk.org/jfx/pull/1236#issuecomment-1939200760
