I saw the ICU4C intltest idna/TestIDNAMonkeyTest fail once in a test but it does not fail any more and it does not seem obvious how to recreate the precise conditions for the one failure. We know that in general the IDNA code works fine for "R" unlike what the test failure suggests.
This test, like any random-input monkey test, needs to output enough to make it possible to trace any reported situation, and, if possible, to recreate it.
The one-time failure should also be investigated. Its log is: