Ian Eiloart wrote: > On 23 May 2011, at 23:10, Franck Martin wrote: > >> There is an interesting post today on >> http://chilli.nosignal.org/mailman/listinfo/mailop about exim and 8bit >> >> It seems they will stop to downgrade. > > Exim doesn't downgrade. It doesn't advertise 8bitmime either, by default. > If you switch on 8bitmime advertising, it still doesn't downgrade. I think > it just tries to deliver the mail as 8bit, regardless of what the receiving > MTA does. I think postfix and sendmail do the same, but I'm not sure.
Look, the general rule of thumb is PASSTHRU mail is untouched (except for adding network control/trace header lines). Only upon final delivery begins the idea of any transformations and gateways so you expect under normal circumstances, the payload will be delivered as it was created. That is why, overall, the system has worked over the last two scores of years and allowed us to get to this point. Honestly, if the Internet mail network was that chaotic, I highly doubt I will be doing this work since 1982, nor do I think RFC822/RFC821 would of taking over the existing mail networks at the time, commercial or otherwise. There were at least 4-6 competing mail frameworks existing and in development in the 80s and RFC822/RFC821 won out for many obvious reasons - its flexibility was one key reason, and mainly, it (predecessor, X.400) was government funded and already used in government and academia. This is no different than the early telecommunications days when dealing with half/full duplex issues, client/host LF/CR translations issues or 7bit vs 8 bits file transfer protocols and which worked better over the different available wires, including X.25 networks (i.e. PAD configurations, etc). How do you think ZMODEM got invented? As well as KERMIT? Before then, you has ASCII (7bit), then XMODEM, YMODEM, then ZMODEM and KERMIT, and it *not* ironic, the Postel principle: Be conservative in what you send; be liberal in what you accept. always applied in data communications: Receive HIGH (buffers), Send Low (Buffers) I always used the Bucket Brigade idea to illustrate this; A fireman is passing water in a bigger bucket than the next fireman in the brigade has - you will have "flow control" issues or spillage. Someone has to slow down. The system works because we have an expectation for an optimal behavior across the board. When one or more node begins to do things differently with an unrealistic unknown expectation for downlinks, its no surprise problems develop for some or many. The only reason we are seeing it now, its because of this DKIM integrity invention highlight the issues. -- Hector Santos, CTO http://www.santronics.com http://santronics.blogspot.com _______________________________________________ NOTE WELL: This list operates according to http://mipassoc.org/dkim/ietf-list-rules.html