Fix: Handling error situation in multi-instance call activities when in async completion job #3264
Conversation
| @@ -67,8 +67,13 @@ protected static void populateEventWithCurrentContext(FlowableEngineEventImpl ev | |||
| if (persistedObject instanceof Job) { | |||
| Job jobObject = (Job) persistedObject; | |||
| if (jobObject.getScopeType() == null) { | |||
| event.setExecutionId(jobObject.getExecutionId()); | |||
| event.setProcessInstanceId(jobObject.getProcessInstanceId()); | |||
| if ("async-complete-call-actiivty".equals(jobObject.getJobHandlerType())) { | |||
There was a problem hiding this comment.
Wouldn't it be better to use the public static variable AsyncCompleteCallActivityJobHandler.TYPE instead of a literal string? At least it would hide the incorrect spelling of activity 😄
There was a problem hiding this comment.
Wouldn't it be better to use the public static variable
AsyncCompleteCallActivityJobHandler.TYPEinstead of a literal string? At least it would hide the incorrect spelling ofactivity😄
Hi @dbmalkovsky, thanks for your feedback. The typo in existing flowable code is quite unfortunate... excellent intention to cover it...
However, the change here is in FlowableJobEventBuilder which is located in maven module flowable-job-service, from here the AsyncCompleteCallActivityJobHandler class (and its constant) is not reachable as it is located in flowable-engine without dependency in that direction.
…n-asynch-execution_TEST Tests: failing subflow completion w/ completeAsync true/false
The flowable-engine dispatches a JOB_EXECUTION_FAILURE event when an exception is thrown during the EndExecutionOperation of a process instance. The failing job is converted to a DeadLetterJob.
We use a CallActivity to trigger (one or multiple) SubProcess instance(s). We have an event listener which handles failing subprocess instances.
We discovered that the context of the JOB_EXECUTION_FAILURE event is set differently depending on the asyncComplete flag of the CallActivity:
From our perspective the JOB_EXECUTION_FAILURE event should always be dispatched in the context of the execution which caused the exception.
Currently this is the case for CallActivity triggered with asyncComplete=false, but not for asyncComplete=true.
Following PR corrects this behaviour.
Please let us know your opinion/feedback. Based on this we will finalise this PR.