The check_dmarc method can fail to find a From: domain #32
Description
A message with a From: header like
From: "Txxxxxx Mxxxxxxx de Vxxxxxx, Fxxxxxxxx" <fxxxxxxxx.txxxxxx@subd.example.net>
when flattened as_bytes and passed to check_dmark will have the header folded as
From: "Txxxxxx Mxxxxxxx de Vxxxxxx,
Fxxxxxxxx" <fxxxxxxxx.txxxxxx@subd.example.net>
While this is reported as a bug at python/cpython#114362, it is compliant with RFC 5322 which allows FWS within a quoted string.
In any case the code acting on this does the following:
headers, _ = rfc822_parse(msg)
which returns one of the elements in the headers list as
[b'From', b' "Txxxxxx Mxxxxxxx de Vxxxxxx,\r\n Fxxxxxxxx" <fxxxxxxxx.txxxxxx@subd.example.net>\r\n']
and then
from_headers = [a[1] for a in getaddresses(x[1].decode(errors='ignore').strip() for x in headers if x[0].lower() == b"from")]
yields
['Txxxxxx Mxxxxxxx de Vxxxxxx,']
as from_headers.
One way to fix this is to change
from_headers = [a[1] for a in getaddresses(x[1].decode(errors='ignore').strip() for x in headers if x[0].lower() == b"from")]
to
from_headers = [a[1] for a in getaddresses(x[1].replace(b'\r\n', b'')
.decode(errors='ignore').strip() for x in headers
if x[0].lower() == b"from")]
to unfold the header value before passing it to getaddresses. Normally getaddresses handles folded headers, but I think folding with \r\n is an issue.