-
Task
-
Resolution: Delivered
-
P3
-
internal
-
x86_64
-
os_x
Tested build : EA-10 build available at - https://jdk.java.net/lanai/
Functionality testing :
----------------------------
Below automated tests were executed with Metal rendering pipeline.
- Automated JDK jtreg regression tests for client
- Automated JCK tests
Below manual tests were run with Metal rendering pipeline on macOS 10.14, macOs 10.15, macOS 11 with a mix of Metal API validation enabled/disabled. [Result : No functional regressions were observed when compared with OpenGL rendering pipeline.]
- Java 2D Demo - manual test
- Font 2D Demo - manual test
- SwingSet2 Demo - manual test - all look and feels
- JavaFX-Swing inter-op sanity tests - manual test
Test JFXPanel (JFXPanel is a component to embed JavaFX content into Swing applications)
Test SwingNode (This Node is used to embed a Swing content into a JavaFX application)
- Large-scale real world applications - manual test - Netbeans IDE usage
- JDK open jtreg regression manual tests - AWT, Swing and Java2D - (tests were distributed across mentioned OS versions)
- JCK manual tests - AWT, Swing and Java2D - (tests were distributed across mentioned OS versions)
Newly found issues -JDK-8262846 and JDK-8262882
Performance testing :
-----------------------------
Below Performance tests were run on Mac-mini (macOS 10.14) with uiScale=2.0 [ Test results attached as a screenshot ]
- RenderPerfTest - We haveJDK-8250658 and JDK-8259953 to address lower performance of 4 test cases - these are stress test cases which are uncommon in usual use.
- SwingMark for all Look-and-Feels
- J2DBench (identified test cases)
- A set of identified Netbeans IDE UI responsiveness tests
- XOR mode rendering - as expected, the performance is slower with metal pipeline
Functionality testing :
----------------------------
Below automated tests were executed with Metal rendering pipeline.
- Automated JDK jtreg regression tests for client
- Automated JCK tests
Below manual tests were run with Metal rendering pipeline on macOS 10.14, macOs 10.15, macOS 11 with a mix of Metal API validation enabled/disabled. [Result : No functional regressions were observed when compared with OpenGL rendering pipeline.]
- Java 2D Demo - manual test
- Font 2D Demo - manual test
- SwingSet2 Demo - manual test - all look and feels
- JavaFX-Swing inter-op sanity tests - manual test
Test JFXPanel (JFXPanel is a component to embed JavaFX content into Swing applications)
Test SwingNode (This Node is used to embed a Swing content into a JavaFX application)
- Large-scale real world applications - manual test - Netbeans IDE usage
- JDK open jtreg regression manual tests - AWT, Swing and Java2D - (tests were distributed across mentioned OS versions)
- JCK manual tests - AWT, Swing and Java2D - (tests were distributed across mentioned OS versions)
Newly found issues -
Performance testing :
-----------------------------
Below Performance tests were run on Mac-mini (macOS 10.14) with uiScale=2.0 [ Test results attached as a screenshot ]
- RenderPerfTest - We have
- SwingMark for all Look-and-Feels
- J2DBench (identified test cases)
- A set of identified Netbeans IDE UI responsiveness tests
- XOR mode rendering - as expected, the performance is slower with metal pipeline
- relates to
-
JDK-8251396 Test Plan for JEP 382: New macOS Rendering Pipeline
- Resolved