On 01/08/2016 05:39, Sergey Bylokhov wrote:

Hello.
Please review the fix for jdk9.

In the fix the part of our internal api is opened via jawt library which is a jdk api, so the important part of the fix is in the jawt.h file. This api can be used by other toolkits to implement embedding of awt/swing components into some other native windows.

There are no tests in the fix since I plan to implement a number of tests via separate CR.(but I am not sure what is the best way to create a native dll, which will reuse jawt api -> manual/precompiled dll/ compilation on the fly?).

jprt job is passed, the ccc will be filed after the technical review.

 - JAWT_VERSION_9 is used intentionally instead of JAWT_VERSION_1_9.
- On OSX validateWithBounds() intentionally was changed to setBoundsPrivate() as on other platforms.

Bug: https://bugs.openjdk.java.net/browse/JDK-8148109
Webrev can be found at: http://cr.openjdk.java.net/~serb/8148109/webrev.00

At a high-level then this looks the right approach to me. I hope that Eclipse/SWT will be able to move to this easily and maybe there should be an updated comment added to their issue [1].

I see JAWT_VERSION_9 is proposed to have the value 0x00010009. Shouldn't this be0x00090000? In jni.h then JNI_VERSION_9 is defined as 0x00090000. JVM TI is similar (although it has an additional value encoded into its version).

-Alan

[1] https://bugs.eclipse.org/bugs/show_bug.cgi?id=482318

Reply via email to