Another fix from Hung-Te Lin; apparently in some IMEs (such as
authorjacob <jacob@cda61777-01e9-0310-a592-d414129be87e>
Wed, 23 Mar 2005 01:08:18 +0000 (01:08 +0000)
committerjacob <jacob@cda61777-01e9-0310-a592-d414129be87e>
Wed, 23 Mar 2005 01:08:18 +0000 (01:08 +0000)
"MS NewPhonetics"), move events (arrow keys) were being doubled up,
apparently because we turned both KEYDOWN and KEYUP events into new
KEYDOWN events.

I don't claim to understand the precise effect of this patch :( but
I'm reasonably confident that it only affects IME users, and experimentally
it doesn't seem to break anything obvious, so if piaip says it makes
things better that's good enough for me :)

git-svn-id: svn://svn.tartarus.org/sgt/putty@5545 cda61777-01e9-0310-a592-d414129be87e

windows/window.c

index ac54ada..8448ca3 100644 (file)
@@ -2734,13 +2734,15 @@ static LRESULT CALLBACK WndProc(HWND hwnd, UINT message,
            unsigned char buf[20];
            int len;
 
-           if (wParam == VK_PROCESSKEY) {
-               MSG m;
-               m.hwnd = hwnd;
-               m.message = WM_KEYDOWN;
-               m.wParam = wParam;
-               m.lParam = lParam & 0xdfff;
-               TranslateMessage(&m);
+           if (wParam == VK_PROCESSKEY) { /* IME PROCESS key */
+               if (message == WM_KEYDOWN) {
+                   MSG m;
+                   m.hwnd = hwnd;
+                   m.message = WM_KEYDOWN;
+                   m.wParam = wParam;
+                   m.lParam = lParam & 0xdfff;
+                   TranslateMessage(&m);
+               } else break; /* pass to Windows for default processing */
            } else {
                len = TranslateKey(message, wParam, lParam, buf);
                if (len == -1)