allthingssecurity commented on PR #29374: URL: https://github.com/apache/camel/pull/29374#issuecomment-6704325918
Thanks Claus. On merge timing: the PR is a draft now, and stays one until the JMH result is in (the run is going now; I'll post the numbers here). The R1 prototype comes in the next PR, and if it shows the SPI needs a change, that change comes before the 4.24 release, as the description says. The other three are in 1720ca3: - `Synchronization`: agreed. `SedaProducer` gives the completions back from a copy that has no unit of work, so the sentence was stricter than what holds. The Javadoc now says that a commit *in a unit of work* moves only by a handover through that unit of work, made by the exchange the unit of work was created for. A commit held by an exchange without a unit of work moves with that exchange: into the unit of work it gets, or back to the exchange that handed it over when the copy is not routed (the seda give-back). One note: the rule sentence of R2 in the design uses the same strict words. Its list of reports (item 2, "handed over through a unit of work") is what narrows it. I left that sentence for a separate small doc change, so this PR only touches the one design line below. - `ErrorHandlerAwareStepListenerProcessor`: `StepListenerTest` has a new multicast route. One branch fails on its first call, and the error handler that the multicast wraps its branches in redelivers it. The test asserts that the listener sees the multicast step once and the failing branch twice (once per attempt), that the call of the multicast channel is the `ErrorHandlerAware` wrapper, and that `getErrorHandler()` and `setErrorHandler()` pass through to the `MulticastProcessor`. A wrapper that is not `ErrorHandlerAware`, or one that does not pass the error handler through, fails the test. The three step listener test classes pass 20 times each (360/360). - Design nit: the line now says "an equal listener is ignored". The code is unchanged and still uses `equals`, as the `ExtendedCamelContext` Javadoc says. _Claude Code on behalf of allthingssecurity_ -- 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]
