https://bugs.winehq.org/show_bug.cgi?id=53739
--- Comment #4 from 399989567@qq.com --- (In reply to Nikolay Sivov from comment #1)
Ideally, it will be great to see bisect result - specific commit that causes problems for you. The process is described here https://wiki.winehq.org/Regression_Testing.
One of the recent changes to dwrite was an addition of fallback font list. That list does use various Noto Sans fonts.
Are you sure QQ is using dwrite in a first place? WINEDEBUG=+dwrite will show it.
Dear Nikolay Sivov： Through my attempts I found that the cause of the problem may be the following: (everything below is based on the commit "26e717c5653d2c6fcb29846e7d22aa97e37048e1")
'''bash $ git reset 26e717c5653d2c6fcb29846e7d22aa97e37048e1 '''
In "dwrite/analyzer.c", there is a problem with the "fallback_map_characters" function, where the problem occurs as follows：
'''C 2246: data = fallback->data.count == ~0u ? &system_fallback : &fallback->data; '''
When I force "data = fallback->data;" the problem is solved.
So I guess there might be a problem with the assignment of "system_fallback" or with data.count, and I'm still looking into the cause of the problem.
For now, it seems that the brute force solution is to force "data = &fallback->data;", but a more elegant solution is yet to be thought of.
I look forward to hearing from you.
Yours: SvenL