zhuzhurk commented on a change in pull request #11213: [FLINK-16276][tests]
Introduce a builder and factory methods to create DefaultScheduler for testing
URL: https://github.com/apache/flink/pull/11213#discussion_r385534676
##########
File path:
flink-runtime/src/main/java/org/apache/flink/runtime/scheduler/SchedulerBase.java
##########
@@ -280,23 +281,8 @@ private ExecutionGraph createExecutionGraph(
failoverStrategy);
}
- /**
- * @deprecated Direct access to the execution graph by scheduler
implementations is discouraged
- * because currently the execution graph has various features and
responsibilities that a
- * scheduler should not be concerned about. The following specialized
abstractions to the
- * execution graph and accessors should be preferred over direct access:
- * <ul>
- * <li>{@link #getSchedulingTopology()}
- * <li>{@link #getFailoverTopology()}
- * <li>{@link #getInputsLocationsRetriever()}
- * <li>{@link #getExecutionVertex(ExecutionVertexID)}
- * <li>{@link #getExecutionVertexId(ExecutionAttemptID)}
- * <li>{@link #getExecutionVertexIdOrThrow(ExecutionAttemptID)}
- * </ul>
- * Currently, only {@link LegacyScheduler} requires direct access to
the execution graph.
- */
- @Deprecated
- protected ExecutionGraph getExecutionGraph() {
+ @VisibleForTesting
+ public ExecutionGraph getExecutionGraph() {
Review comment:
I will think again about whether we must expose `getExecutionGraph()` to the
tests.
Will drop this commit and may do it when reworking the tests. And as you
mentioned, it should be a non-hotfix commit then if we would have it.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
