gemini-code-assist[bot] commented on code in PR #38430:
URL: https://github.com/apache/beam/pull/38430#discussion_r3215954638
##########
runners/google-cloud-dataflow-java/worker/src/main/java/org/apache/beam/runners/dataflow/worker/WindowingWindmillReader.java:
##########
@@ -182,6 +182,11 @@ public boolean start() throws IOException {
@Override
public boolean advance() throws IOException {
Review Comment:
The anonymous iterator in `WindowingWindmillReader` should also check
`context.workIsFailed()` and throw `WorkItemCancelledException`, consistent
with the changes in `WindmillReaderIteratorBase`. This ensures that cancelled
work items are handled promptly.
```java
@Override
public boolean advance() throws IOException {
if (context.workIsFailed()) {
throw new
WorkItemCancelledException(context.getWorkItem().getShardingKey());
}
```
##########
runners/google-cloud-dataflow-java/worker/src/main/java/org/apache/beam/runners/dataflow/worker/StreamingModeExecutionContext.java:
##########
@@ -240,9 +243,11 @@ public void start(
Work work,
WindmillStateReader stateReader,
SideInputStateFetcher sideInputStateFetcher,
- Windmill.WorkItemCommitRequest.Builder outputBuilder) {
+ Windmill.WorkItemCommitRequest.Builder outputBuilder,
+ WorkExecutor workExecutor) {
this.key = key;
this.work = work;
+ this.workExecutor = workExecutor;
Review Comment:
Reset the `finishKeyCalled` flag when starting a new work item.
```suggestion
this.workExecutor = workExecutor;
this.finishKeyCalled = false;
```
##########
runners/google-cloud-dataflow-java/worker/src/main/java/org/apache/beam/runners/dataflow/worker/StreamingModeExecutionContext.java:
##########
@@ -157,6 +158,8 @@ public class StreamingModeExecutionContext extends
DataflowExecutionContext<Step
*/
private @Nullable UnboundedReader<?> activeReader;
+ private @Nullable WorkExecutor workExecutor;
Review Comment:
Add a boolean flag to track whether `finishKey()` has been called for the
current work item. This helps ensure idempotency, especially if multiple
readers are present or if `advance()` is called multiple times after exhaustion.
```java
private @Nullable WorkExecutor workExecutor;
private boolean finishKeyCalled;
```
##########
runners/google-cloud-dataflow-java/worker/src/main/java/org/apache/beam/runners/dataflow/worker/StreamingModeExecutionContext.java:
##########
@@ -270,6 +275,11 @@ public void start(
}
}
+ public void finishKey() throws Exception {
+ checkNotNull(workExecutor, "workExecutor must be set before calling
finishKey()");
+ workExecutor.finishKey();
+ }
Review Comment:
Use the `finishKeyCalled` flag to ensure that `workExecutor.finishKey()` is
only executed once per work item.
```java
public void finishKey() throws Exception {
if (finishKeyCalled) {
return;
}
checkNotNull(workExecutor, "workExecutor must be set before calling
finishKey()");
finishKeyCalled = true;
workExecutor.finishKey();
}
```
##########
runners/google-cloud-dataflow-java/worker/src/main/java/org/apache/beam/runners/dataflow/worker/WindmillReaderIteratorBase.java:
##########
@@ -54,9 +56,18 @@ public boolean start() throws IOException {
@Override
public boolean advance() throws IOException {
+ if (context.workIsFailed()) {
+ throw new
WorkItemCancelledException(context.getWorkItem().getShardingKey());
+ }
+
while (true) {
if (bundleIndex >= work.getMessageBundlesCount()) {
current = null;
+ try {
+ context.finishKey();
+ } catch (Exception e) {
+ throw new RuntimeException(e);
+ }
Review Comment:
Instead of wrapping all exceptions in `RuntimeException`, consider catching
`IOException` and `RuntimeException` separately to rethrow them directly, and
wrapping only other checked exceptions in `IOException`. This maintains
consistency with the `advance()` method's signature and allows callers to
handle specific exception types.
```java
} catch (IOException | RuntimeException e) {
throw e;
} catch (Exception e) {
throw new IOException(e);
}
```
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
