Request for review: 7124552 [macosx] NullPointerException in getBufferStrategy()

[Andrew Brygin]

The change looks fine to me.

Alex, please note that there we just have to flush current buffer content
from the image buffer to a surface related to canvas.

The buffer strategy machinery is responsible to verify that the buffer
state is valid before flip.

Thanks,
Andrew

On 2/15/12 4:38 PM, Sergey Bylokhov wrote:
> 15.02.2012 16:28, Alexander Potochkin пишет:
>> Hello Sergey
>>> 15.02.2012 15:39, Alexander Potochkin wrote:
>>>> Hello Sergey
>>>>
>>>> LWCanvasPeer.flip()
>>>>
>>>> I don't think you can work withVolatileImage as with an ordinary 
>>>> Image,
>>>> you take a buffer and then draw it to the graphics
>>>> when the buffer may be already expired.
>>>>
>>>> look at RepaintManager.PaintManager.paint()
>>>>
>>>> it tries to repaint the image if its content is lost,
>>>> after the fixed amount of attempts it falls back to the BufferedImage
>>> Buffer strategy has its own contentLost()(see 
>>> FlipBufferStrategy.contentLost()) method which should be used by the 
>>> user.
>>
>> Could you give me a test case which uses this strategy?
> test.sun.java2d.DirectX.RenderingToCachedGraphicsTest.RenderingToCachedGraphicsTest.java 
>
>>
>> Thanks
>> alexp
>>
>>>>
>>>> Here is a good blog about using VolatileImage
>>>> http://weblogs.java.net/blog/chet/archive/2003/09/volatileimage_n.html
>>>>
>>>> LWWindowPeer - it would be better to move the comment to the 
>>>> method's javadoc
>>>>
>>>> Thanks
>>>> alexp
>>>>
>>>>> Hi Everyone,
>>>>> This is the first implementation for canvas backbuffer. npe was 
>>>>> eliminated.
>>>>>
>>>>> Bug: http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=7124552
>>>>> Webrev can be found at: 
>>>>> http://cr.openjdk.java.net/~serb/7124552/webrev.00/
>>>>>
>>>>
>>>>
>>>>
>>>
>>>
>>
>
>