Github user spmallette commented on a diff in the pull request:
https://github.com/apache/tinkerpop/pull/899#discussion_r206485952
--- Diff:
gremlin-server/src/main/java/org/apache/tinkerpop/gremlin/server/op/AbstractOpProcessor.java
---
@@ -72,8 +73,19 @@ protected AbstractOpProcessor(final boolean
manageTransactions) {
* @param context The Gremlin Server {@link Context} object containing
settings, request message, etc.
* @param itty The result to iterator
* @throws TimeoutException if the time taken to serialize the entire
result set exceeds the allowable time.
+ * @see #handleIterator(ResponseHandlerContext, Iterator)
*/
protected void handleIterator(final Context context, final Iterator
itty) throws TimeoutException, InterruptedException {
+ handleIterator(new ResponseHandlerContext(context), itty);
+ }
+
+ /**
+ * A variant of {@link #handleIterator(Context, Iterator)} that is
suitable for use in situations when multiple
+ * threads may produce {@link ResponseStatusCode#isFinalResponse()
final} response messages concurrently.
+ * @see #handleIterator(Context, Iterator)
+ */
+ protected void handleIterator(final ResponseHandlerContext rhc, final
Iterator itty) throws TimeoutException, InterruptedException {
--- End diff --
same question here as on `evalOpInternal()` - is it reasonable for someone
to directly call this method? or is it more likely that they would continue to
call the original method which would let us control creation of the
`ResponseHandlerContext` ourselves?
---
