Looks fine.
On 11/06/2018 12:38, Dmitry Markov wrote:
Hello,
Could you review a fix for jdk11, please?
bug: https://bugs.openjdk.java.net/browse/JDK-8200353
webrev: http://cr.openjdk.java.net/~dmarkov/8200353/webrev.00/
The new mechanism of dead keys detection and processing was introduced
by JDK-8139189 [1]. According to that changes dead key input is
activated by WM_KEYDOWN and deactivated by WM_CHAR messages. However the
function WindowsKeyToJavaChar() (which actually sets the flag
deadKeyActive to true) is also called from _NativeHandleEvent() and in
this case target component doesn't receive WM_CHAR message, (i.e. dead
key input remains active).
At the same time _NativeHandleEvent() sends character back to the native
window using WM_AWT_FORWARD_CHAR message. So it is necessary to disable
dead key input, (i.e. set deadKeyActive to false) when
WM_AWT_FORWARD_CHAR is received.
I ran related AWT/Swing regression tests and did not observe any new
failures.
Thanks,
Dmitry
[1] - https://bugs.openjdk.java.net/browse/JDK-8139189
--
Best regards, Sergey.
