From: Matthias A. <mat...@gm...> - 2021-01-29 18:37:30
|
Am 29.01.21 um 08:59 schrieb Julian Bane: > Good morning Mattias and all, > > I have managed to suppress the X-Original-To header by > setting*enable_original_recipient = no*" in the postfix configuration > file. But still don't get duplicate emails suppressed, though the > X-Original-To header is. (Matthias - I can send you the mailbox file > if you want). > > I have cut out two emails and verified they still have differences in > the headers (some de-personalization done): > > 1,3c1,3 > < From XXX...@co... Fri Jan 29 07:38:31 2021 > < Return-Path: <XXX...@co...> > < Delivered-To: up...@my... > --- > > From jul...@co... Fri Jan 29 07:38:29 2021 > > Return-Path: <XXX...@co...> > > Delivered-To: do...@my... > 38c38 > < X-UID: 24 > --- > > X-UID: 23 > > It's hard to see how you could ever get two identical sets of headers > for de-duplication to work. Julian, As stated earlier, the deduplication code is from another era, when the envelope-recording headers such as X-Original-To, Envelope-To, Delivered-To were not in widespread use, and it was meant as a hack to avoid local spool explosions when fetchmail resorted to guessing recipients from the To:/Cc: headers (which breaks with Bcc: quite obviously) because then it did not have a 1:1 relationship between server-side message and message-to-ship-onwards. About X-UID, is it actually transmitted over the wire, or will your server software suppress it? I guess it goes hand in hand with the folder-internal status message that is also not sent over the wire. > > Not wishing to either make trouble or make work for others but could > the documentation be reworded to reflect how the de-duplication works. > Currently it says "/A piece of mail is considered duplicate if it has > the same message-ID as the message immediately preceding and more than > one addressee/". Definitely yes because what you quote is plainly does not refer to what the code does. I will see to fixing the manual page. I've filed an issue so I don't forget, because I can't get to it right away. This is 6.4 stuff. https://gitlab.com/fetchmail/fetchmail/-/issues/29 Best regards, MNatthias |