On Tue Sep 16 18:49:46 2025 +0000, Dmitry Timoshkov wrote:
I guess you mean to make tests less depend on Windows vs Wine font metrics difference? Sure, we could. That won't help much I'm afraid, in my testing I dropped exactly same Windows fonts to Wine prefix and even that produced pretty large differences in 135 degrees, so that I removed 135 degrees from test data. That made the tests in current version of the patch have very little difference between Wine and Windows, and that shouldn't be a problem anymore.
If there's a difference, doesn't that mean there's something wrong with this implementation?