https://bugs.winehq.org/show_bug.cgi?id=44410
--- Comment #7 from Nikolay Sivov bunglehead@gmail.com --- (In reply to Thomas Faber from comment #5)
Yeah I'm looking at that. The fix itself should be simple, but it'll require a bunch of char vs glyph tests that I'll need to spend some time on.
Any news on that? I think we should either skip cache on fNoGlyphIndex, or do GetGlyphIndicesW() on every call with fNoGlyphIndex.
How did you plan to test this? Let me know if you don't have time for this, it's something we need to get fixed for 4.0 I think.