25 Jun
2024
25 Jun
'24
10:03 a.m.
On Tue Jun 25 09:51:20 2024 +0000, Marc-Aurel Zent wrote:
I guess it is also an option to only use the unicode codepath if `buffer[0] >= 'a' && buffer[0] <= 'z'` is false, if it is a performance bottleneck... Ideally this would just be a `CharUpperBuffW(buffer, len)` operation, if possible.
Well, I can easily believe that applications rely on single letter ASCII key names being uppercase, but I'm more doubtful that they do for any unicode ones, or that Windows really enforces anything (it's probably just a default table that happens to be uppercase for ASCII keys).
A quick check for instance with a German keyboard, shows that the Ü key name is lowercase ü.