RFR: 8144030: [macosx] test java/awt/Frame/ShapeNotSetSometimes/ShapeNotSetSometimes.java fails (again)

Sergey Bylokhov serb at openjdk.java.net
Wed Apr 13 21:50:17 UTC 2022


On Wed, 13 Apr 2022 21:30:06 GMT, Alexander Zvegintsev <azvegint at openjdk.org> wrote:

>> This test has two undecorated frames: 
>> 
>> 1. background, 300x300
>> 2. shaped foreground, 200x200
>> 
>> After displaying this frames it tries to check color outside of shape but within frame 2 bounds.
>> 
>> Unfortunately, on Linux and macOS there is a shadow around the shaped frame:
>> ![image](https://user-images.githubusercontent.com/77687766/163270586-8dd97e93-582f-481f-999d-f08d4aaafba1.png)
>> (changed color to green to make it more noticeable).
>> 
>> One way to fix the test would be to add tolerance to color check.
>> But with old check points blue color may vary in 222-255 range, which is a bit too much.
>> 
>> Instead of this I made this shaped window  the size of background window. 
>> This allows us to move check points away from shape border and its shadow, but still be able to check whether shape is applied or not.
>> 
>> Testing is green on all platforms.
>
> test/jdk/java/awt/Frame/ShapeNotSetSometimes/ShapeNotSetSometimes.java line 144:
> 
>> 142:                         point.y,
>> 143:                         BACKGROUND_COLOR
>> 144:                 );
> 
> Dunno why the old version was not checking inner points, but I've added one just to check if the shaped frame is above background one.

Don't we need to check the places where the shadow is rendered that it is not the shape color(to prove that the shape is not simple rectangular)?

-------------

PR: https://git.openjdk.java.net/jdk/pull/8233



More information about the client-libs-dev mailing list