https://bugs.exim.org/show_bug.cgi?id=2314
--- Comment #12 from Matthias Kurz <m.kurz@???> ---
> Could you supply a comment (for the first patch) with the rationale for
> raising the event in the new place and not the old?
If I remember correctly not all (failed) messages reach the failure_log(...)
function (Which itself is called by post_process_one(...)). Some failed
messages "stop" sooner and don't even execute post_process_one(...).
That's why I had to raise the event (for ignored errors) earlier. And to avoid
duplicate events I had to add the new "if (!addr->prop.ignore_error)" condition
in the original place within the failure_log(...) function (because obviously
the event was fired already).
Does that answer your question?
--
You are receiving this mail because:
You are on the CC list for the bug.