Error status wasn't correctly transmitted to UI after recent changes.

Oops, a zero retry_interval check had gone missing in scheduling reorganisation.

(0) -100 -30 -10 -2 +2 +10 +30 tip

mercurial