RFR: 8261891: sanity/client/SwingSet/src/EditorPaneDemoTest.java fails with Metal API Validation

[Jayathirth D V]

On Wed, 24 Feb 2021 09:45:45 GMT, Alexey Ushakov <avu at openjdk.org> wrote:

>> src/java.desktop/macosx/native/libawt_lwawt/java2d/metal/EncoderManager.m line 370:
>> 
>>> 368:   if (_encoder != nil) {
>>> 369:     if (_destination != dest || renderOptions->isAA != _encoderStates.aa ||
>>> 370:             renderOptions->isAAShader != _encoderStates.aaShader) {
>> 
>> Hi Alexey,
>> 
>> Is there any performance impact because of ending the encoder on switch?
>> 
>> Thanks,
>> Jay
>
> Hi Jay,
> 
> Thanks for the comment. I didn’t find any difference at least with RenderPerfTest. However, looks like we don’t need to end the encoder here - the appropriate state will be recovered later in the code (by updatePipelineState method).
> 
> So, I’m going to remove this condition from the if statement.
> 
> Best Regards,
> Alexey

Hi Alexey,

Yes i also think we dont need to explicit end encoder here and leave it to updatePipelineState. This was the next clarification i wanted to ask.

Thanks,
Jay

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

PR: https://git.openjdk.java.net/lanai/pull/201