Weird keyboard behavior with "Caps Lock" key

[Takuma Murakami]

> Check out the Test104/4.3.0-17 release.  I disabled processing of the 
> Win32 repeat count for VK_CAPITAL messages.  This may or may not help. 
> In fact, I suspect that it won't help.  However, it will at least give 
> me some incentive to fix it the right way when I find out that it isn't 
> correct.  Please try it and report back.

It does not help.  The main problem is not the repeat
count in WM_KEYDOWN messages but the number of WM_KEYDOWN
messages.  Following your idea, installing the piece of code
if (wParam == VK_CAPITAL && lParam & (1<<30))
  return 0
in the WM_KEYDOWN handler will help.  Of course the
repeat counter reset should make more robust solution.

Because the VK-based code does not work in Japanese
environments, I have tested with an equivalent
ScanCode-based code.

Takuma Murakami (murakami@ipl.t.u-tokyo.ac.jp)