On 27/07/16 01:19, Wietse Venema wrote: > Short answer: disable the after-220 tests: > > postscreen_bare_newline_enable = no > postscreen_non_smtp_command_enable = no > postscreen_pipelining_enable = no > > Long answer: fixing this requires changing the book-keeping of > passed tests, from per-session state into shared per-client state. > > With my current availability, that is likely to take longer than > most clients are willing to retry. Hence the "short answer" suggestion.
Would doing another cache read right before the update be an easy fix for this? I do realize this would add additional latency to postscreen, hopefully not too much. Peter
