On 9/12/2016 10:36 PM, Vadim Pakhnushev wrote:
Looks good to me (have you submitted JPRT job just in case?)
Is JPRT really necessary? The native part is only changed for Windows on
which the build is deterministic.
--Semyon
Vadim
On 12/09/16 22:31, Semyon Sadetsky wrote:
I missed this file in the list of changes.
http://cr.openjdk.java.net/~ssadetsky/8155753/webrev.02/
--Semyon
On 9/12/2016 9:33 PM, Vadim Pakhnushev wrote:
Have you forgotten adding changes in AccelGraphicsConfig?
c:\Vadim\jdk9-client\jdk\src\java.desktop\windows\classes\sun\java2d\d3d\D3DGraphicsConfig.java:52:
error: D3DGraphicsConfig is not abstract and does not override
abstract method removeDeviceEventListener(AccelDeviceEventListener)
in AccelGraphicsConfig
public class D3DGraphicsConfig
^
c:\Vadim\jdk9-client\jdk\src\java.desktop\windows\classes\sun\java2d\opengl\WGLGraphicsConfig.java:59:
error: WGLGraphicsConfig is not abstract and does not override
abstract method removeDeviceEventListener(AccelDeviceEventListener)
in AccelGraphicsConfig
public class WGLGraphicsConfig
^
Also +#include "Devices.h" in the D3DContext.cpp is a leftover.
Vadim
On 12.09.2016 21:11, Semyon Sadetsky wrote:
http://cr.openjdk.java.net/~ssadetsky/8155753/webrev.01/
AccelDeviceEventNotifier is removed.
--Semyon
On 9/12/2016 6:56 PM, Semyon Sadetsky wrote:
Okay. I will remove AccelDeviceEventNotifier and all related code.
--Semyon
On 9/12/2016 6:43 PM, Vadim Pakhnushev wrote:
Hi Semyon,
Generally seems reasonable, it seems that you should use screen
instead of gdiScreen in the JNU_CallStaticMethodByName, otherwise
the code won't compile.
Not sure how the rest of the code handles monitor removal, seems
to me that there are no usages of this notifications anywhere, so
maybe we don't need this code at all?
Thanks,
Vadim
On 12.09.2016 17:24, Semyon Sadetsky wrote:
Hello,
Please review fix for JDK9:
bug: https://bugs.openjdk.java.net/browse/JDK-8155753
webrev: http://cr.openjdk.java.net/~ssadetsky/8155753/webrev.00/
The issue take place on Windows platform if Direct3d is on. The
notification routine about the monitor removal tries to get
screen number using the monitor handle which is obviously null
at this moment. As a fix the screen number is recorded in D3D
context for further possible notifications.
--Semyon
