On 10/30/2015 12:56 PM, Jeroen Scheerder wrote: > Quoth Jeroen Scheerder (30 Oct 2015, 12:46): > >> That would result in a >> >> 250-SIZE 10240000 >> >> helo message, *and* a true size limit of 10239918. > > I obviously omitted the evident edit. I had meant to write: > > "That would result in a > > 250-SIZE 10239918 > > helo message, *and* a true size limit of 10239918."
I probably misunderstand your requirements but is this not what you want? You want to accept the message and make sure the message is not refused later on because Postfix added some extra information (headers etc.). Then the true message size limit for submission should be less then the global message size limit. You need to make sure that the message size limit reported by the submission port is small enough to allow any overhead added by postfix (like the added headers). Kind regards, Martijn Brinkers > > Wietse's warning about extra metadata add to this issue. I'd like to > prevent clients from sending in messages that will certainly fail to > be accepted, because they will bump onto size limits a few steps > further (cleanup). It's not just my little header, the additional > 600 bytes of extra bookkeeping should also be subtracted from the > message_size_limit in the "250-SIZE" banner line. > > But that "250-SIZE" just reports the configured message_size_limit. > If I'm correct in interpreting Wietse's remark correctly, a message > within 600 bytes of the size reported in "250-SIZE" will exceed > message_size_limit, making cleanup fail and eventually resulting in a > rejected message and a failure notice to the client. My little > header just exacerbates this situation, making it happen 82 bytes > earlier. > > > Regards, Jeroen. > -- CipherMail email encryption Email encryption with support for S/MIME, OpenPGP, PDF encryption and secure webmail pull. https://www.ciphermail.com Twitter: http://twitter.com/CipherMail
