Hello Dmitry,
Did you run all cases of
java/awt/InputMethods/DiacriticsTest/DiacriticsTest.html after the fix?
Could you add the current case to DiacriticsTest.html as well?
--Semyon
On 11/7/2016 2:41 PM, dmitry markov wrote:
Hello,
Could you review a fix for jdk9, please?
bug: https://bugs.openjdk.java.net/browse/JDK-8158262
webrev: http://cr.openjdk.java.net/~dmarkov/8158262/webrev.00/
Problem description:
The new mechanism of dead keys detection and processing was introduced
by JDK-8139189 <https://bugs.openjdk.java.net/browse/JDK-8139189>. It
uses MapVirtualKey() function to detect dead key combination. Also an
output of MapVirtualKey() invocation is used as a value which is send
with Java's KEY_PRESSED/KEY_RELEASED events. However if the mechanism,
(i.e. WindowsKeyToJavaChar() function inside awtComponent.cpp) is
called from the Java layer via JNI right after a keyboard layout has
been changed, MapVirtualKey() returns a character value related to the
previous keyboard layout.
Fix:
The function WindowsKeyToJavaChar() has to take into account the
current keyboard layout, (i.e. MapVirtualKey() should be replaced with
MapVirtualKeyEx()).
Thanks,
Dmitry
