On Tue Sep 16 19:39:42 2025 +0000, Elizabeth Figura wrote:
Okay, I guess I was thrown off by your earlier saying it was "almost identical", but in that later comment "pretty large differences".
probably because it doesn't use floating point math for calculations.
Maybe we shouldn't either, then?
We do use floating point to scale font metrics everywhere in gdi32/win32u, why should we stop doing that and make scaling results worse just because rotated font metrics scaling is bad in Windows? The test application attached to https://bugs.winehq.org/show_bug.cgi?id=33190 demonstrates that rotated fonts match pretty closely in Windows and Wine, probably that could be a better measure of the patch quality? Besides, I have a real world application that uses rotated fonts and rotated text in it looks pretty good, and matches what I observe under Windows.