The intent tests which test channel results contain a race condition that often fails in a local run and always fails on a lower-performance CI machine. There are also a number of minor issues the app definitions that should be fixed (but don't cause failures, but might in some implementations if they require the appD record to be accurate) and excessive use of magic strings rather than string constants.
This PR:
Corrects the Intent Channel result tests,
Ensures Errors are thrown on timeouts rather than strings,
Makes sure the intent tests use constants for control message types, rather than magic strings.
RemovesbTestingIntent from the Intent B app's appD definition as it is not used and no listener is added for it by the v2.0 test app.
Adds ConformanceListener to Conformance1's declared intents as it adds a listener to it
Removing aTestingIntent listener from metadata app as its not used and was likely a copy/paste error
intent-e and intent-f are the main files to review.
Tests re-run against Finsemble which passes all (previously failing the 2 intent channel tests, which had the race conditions):
The intent tests which test channel results contain a race condition that often fails in a local run and always fails on a lower-performance CI machine. There are also a number of minor issues the app definitions that should be fixed (but don't cause failures, but might in some implementations if they require the appD record to be accurate) and excessive use of magic strings rather than string constants.
This PR:
bTestingIntentfrom the Intent B app's appD definition as it is not used and no listener is added for it by the v2.0 test app.ConformanceListenerto Conformance1's declared intents as it adds a listener to itaTestingIntentlistener from metadata app as its not used and was likely a copy/paste errorintent-e and intent-f are the main files to review.
Tests re-run against Finsemble which passes all (previously failing the 2 intent channel tests, which had the race conditions):