Rules for detecting unnecessary header encoding in Quoted-Printable

[Philip Prindeville <ph...@redfish-solutions.com>]

I’m getting a fair amount of Spam where the To, From, and Subject lines have been unnecessarily encoded using QP even though they’re all ASCII.

This is a violation of RFC-2822 and Postel’s Law, and a characteristic of Spam more often than Ham (where legitimate MUA’s attempt to use the least complex encoding suitable in the interests of interoperability).

header __L_SUBJECT_QP_ENCODED	Subject:raw =~ /^ ?=\?(utf|UTF)-8\?Q\?/
header __L_SUBJECT_ASCII_ONLY	Subject =~ /^[[:print:]]*$/a

header __L_FROM_QP_ENCODED	From:raw =~ /^ ?=\?(utf|UTF)-8\?Q\?/
header __L_FROM_ASCII_ONLY	From =~ /^[[:print:]]*$/a

header __L_TO_QP_ENCODED	To:raw =~ /^ ?=\?(utf|UTF)-8\?Q\?/
header __L_TO_ASCII_ONLY	To =~ /^[[:print:]]*$/a

meta L_UNNECESSARY_QUOTING	(__L_SUBJECT_QP_ENCODED && __L_SUBJECT_ASCII_ONLY || __L_FROM_QP_ENCODED && __L_FROM_ASCII_ONLY || __L_TO_QP_ENCODED && __L_TO_ASCII_ONLY)
describe L_UNNECESSARY_QUOTING	Header lines have unnecessary Quoted-Printable encoding
score L_UNNECESSARY_QUOTING	20.0


Anyone want to sandbox this and see how it performs on a larger sample size?

We’ve been using it but it gets FP’s with GreenArrow Studio and Mailchimp, so we’ve had to whitelist those campaign mailers… they’ve been made aware of the issue.

Thanks