Github user kl0u commented on a diff in the pull request:
https://github.com/apache/flink/pull/3390#discussion_r103405219
--- Diff:
flink-libraries/flink-cep/src/main/java/org/apache/flink/cep/nfa/NFA.java ---
@@ -254,7 +255,18 @@ public int hashCode() {
states.push(stateTransition.getTargetState());
break;
case IGNORE:
-
resultingComputationStates.add(computationState);
+ final
DeweyNumber version;
+ if (branched) {
--- End diff --
Here we assume that the `IGNORE` is going to be processed after the `TAKE`.
This is true because of the way the `NFACompiler` creates the pattern graph,
but different implementations may change in the future so we should somehow
impose this ordering. It could be by sorting the `stateTransitions` list in
the `State` class, or by sorting them just before processing them in the
`NFA::computeNextStates()`.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
