Tony Finch wrote: >
> Has anyone here tried using require verify = header_syntax on a system
> with lots of users? If you found that the test has too many false
> positives, do you have any opinions on how it could be modified to
> create a slightly more lenient version that is usable in practise?
>
> I've been experimenting with it and many problematic rejections are
> caused by (in terms of RFC 2822 syntax) display-names with dots in, or
> which are in the form of an addr-spec.
Lots of problems with that on a large system, lots of MUA's are broken
in this respect. The biggest problem is bad formations of 'Undisclosed
Recipients' when all of the "to's" are in the BCC field. This of
course could be fixed if users would not try to use MUA address-books
as mailing lists :) OR, if those that wrote the MUA software would just
follow the specs.
About the only usefull thing I have been able to use this check on is
NDR email "MAIL FROM:<>".
Also, just a side note, nearly all SPAM passed this check, go figure :P
