khk-globus
commented
3 weeks ago I was thinking less the "wait in increments" and more like "the tests assume this is exists, so we wait until this is in place." I had made the try* methods (following Qt's implementation of similar in their test framework) a couple of years ago, so they were easily at-hand to solve this. If there's a better way ... do let me know!
Ensure that infrastructure is in place before we let the tests poke at it. The flakiness is a very rare occurrence — I've only encountered it once, recently — but flaky == no good.
Type of change