The signature should be stripped from replies.
Specifically, anything following the string "CRLF--
CRLF" should be stripped.
The user can turn this behavior off by setting the
defcustom mh-reply-strip-signature-flag to nil
(default: t).
Ideally, the documentation (for
mh-reply-strip-signature-flag or mh-reply depending on
where it makes the most sense and the manual) should
cite an authorative source for doing this. The closest
we have so far include:
Usenet FAQ for signatures:
URL: http://www.faqs.org/faqs/usenet/signature-faq/
RFC 1855 does not actually specify the use of "\n-- \n"
just that the
signature should be no more than four lines.
URL: http://www.cis.ohio-state.edu/cgi-bin/rfc/rfc1855.html
The rocky-horror newsgroup charter specifies a
'standard signature
delimiter of "\n-- \n" ...
URL: http://www.rockynewsgroup.org/GetArticle/27636.html
A complaint about the standard format versus Microsoft
Outlook is here:
URL: http://www.okinfoweb.com/moe/bugs/bugs_047.htm
p.s. I cannot take credit for Peter's idea ;-). Thanks
to Mark for the references.
Logged In: YES
user_id=96862
Given the references I can't say that this is a ``standard''
more like
a BCP (Best Current Practice).
When striping the signature, we probably have to be MIME
aware. If the signature is in the first text part we should
only strip to the end of the part. Should we scan other
text parts or assume it is in the first? Maybe look in the
first and last?
Ticket moved from /p/mh-e/feature-requests/42/