RFR: 8339974: Graphics2D.drawString doesn't always work with Font derived from AffineTransform [v3]

[Phil Race]

On Mon, 7 Oct 2024 18:08:58 GMT, Daniel Gredler <duke at openjdk.org> wrote:

>> There is a bug in `TextLayout.getOutline(AffineTransform)`, where the transformation is applied too early (before the layout path applies its mapping). The most obvious error caused by this bug is that `Graphics2D.drawString(...)` fails to draw the string when using a rotated + scaled `AffineTransform` (see minimal test case in the bug report).
>> 
>> There are two uses of `TextLayout.getOutline(AffineTransform)`: the first one, triggered via a simple `Graphics2D.drawString(...)` was the one I ran into; the second one requires the use of a transformed font with a `PSPrinterJob`. Both are broken. This PR includes two test classes verifying that both scenarios were broken before the fix and are working after the fix.
>> 
>> Two points which might require some discussion:
>> 1. I've changed the signature of method `TextLine.getOutline(AffineTransform)`, to just `TextLine.getOutline()`; I'm assuming this is OK since `TextLine` is a package-private class.
>> 2. I've added a fast path in `TextLayout.getOutline(AffineTransform)` to transform the shape in place if it is a `GeneralPath` (it should usually be, and avoids an extra `Shape` copy vs. the current code, which makes no copy). A bit more complex, but efficient.
>
> Daniel Gredler has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Add info to test failure messages, increase allowed variance

With your updated tolerance I now have 10 passes in a row. No failures seen.

-------------

PR Comment: https://git.openjdk.org/jdk/pull/20993#issuecomment-2397912900