--- Comment #2 from Tore Anderson <> ---
(In reply to Jeremy Harris from comment #1)
> Created attachment 1163 [details]
> proposed patch
> I had to assume the command-string format.
> Not being an rspamd user, I can't test this.  If somebody could, we should
> gain enough confidence to commit it.

Hi Jeremy. I haven't tested the patch yet, but I've skimmed through it and have
one comment.

As far as I can tell, with the patch, the Deliver-To header will always be sent
to Rspamd, even for the typical "spam = nobody:true" case (Deliver-To =
nobody). Correct?

If so, that'll be a possibly undesired change in behaviour for folks that
prefers the current behaviour where Rspamd takes care of deducing the username
from the e-mail contents itself (which happens when no explicit username is
communicated using the Deliver-To header).

Not sure how to best handle that. Maybe handle "nobody" specially and omit the
Deliver-To header, but at the same time it feels like a cop-out as "nobody"
could very well be a valid username.


You are receiving this mail because:
You are on the CC list for the bug.
## List details at Exim 
details at ##

Reply via email to