------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugs.exim.org/show_bug.cgi?id=1323 --- Comment #2 from Maciej Laskowski <[email protected]> 2012-12-16 12:36:40 --- Your proposed solution may solve the issue with 'mail' from within an Exim filter but then it breaks all 'regular' email sends from Exim clients. Regards, Maciej > -----Original Message----- > From: [email protected] [mailto:[email protected]] On Behalf Of > Phil Pennock > Sent: Sunday, December 16, 2012 4:37 AM > To: [email protected] > Subject: [Bug 1323] $sender_address still points to the original sender after > issuing 'mail' command in a filter > > ------- You are receiving this mail because: ------- You reported the bug. > > http://bugs.exim.org/show_bug.cgi?id=1323 > > > > > --- Comment #1 from Phil Pennock <[email protected]> 2012-12-16 10:36:41 --- > $sender_address is the address as _received_. There are two options for > messages generated by a "mail" filter command: > > 1. Use the original message's $sender_address 2. Use the same as the > $return_path, ie what is sent on the wire: this is <>, or empty > > DKIM-signing based on the envelope sender, for an empty envelope sender, > is broken, and conceptually I can only see $sender_address being one of > these two values. We should not add a case where $sender_address is > _not_ what was received _and_ is _not_ the same as the SMTP envelope > sender being sent, but is something faked from a From: header. > > My understanding is that the DKIM spec folks argue that the signing domain > should be the domain visible in From: or the like, not the envelope domain, > because the From: domain is what people normally see. > > This argues that you should be using: > > ${reduce{${addresses:$h_from:}}{}{$item}} > > as the lookup key. > > > -- > Configure bugmail: http://bugs.exim.org/userprefs.cgi?tab=email -- Configure bugmail: http://bugs.exim.org/userprefs.cgi?tab=email -- ## List details at https://lists.exim.org/mailman/listinfo/exim-dev Exim details at http://www.exim.org/ ##
