From 9b879621748f7310766a3a72bd91be27d428f6f7 Mon Sep 17 00:00:00 2001 From: Philipp Wagner Date: Sat, 14 Nov 2026 23:31:57 +0100 Subject: [PATCH] =?UTF-8?q?Empf=C3=A4ngeradresse=20aus=20der=20weitergelei?= =?UTF-8?q?teten=20Mail=20bestimmen?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- forwarded.py | 28 ++++++++++++++++++++++++++++ tests/test_forwarded.py | 40 ++++++++++++++++++++++++++++++++++++++++ 2 files changed, 68 insertions(+) create mode 100644 forwarded.py create mode 100644 tests/test_forwarded.py diff --git a/forwarded.py b/forwarded.py new file mode 100644 index 0000000..151be00 --- /dev/null +++ b/forwarded.py @@ -0,0 +1,28 @@ +"""Find the address a forwarded newsletter was originally sent to.""" + +import re +from email.utils import parseaddr + +from bs4 import BeautifulSoup + +_HEADER = re.compile(r"^[\s>]*(?:An|To)\s*:\s*(.+)$", re.IGNORECASE | re.MULTILINE) +_ADDRESS = re.compile(r"[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Za-z]{2,}") +# "Name " in plain text would otherwise be eaten as an HTML tag by BeautifulSoup. +_ANGLE_ADDRESS = re.compile(r"<([^<>\s]+@[^<>\s]+)>") + + +def _allowed(address: str, domains) -> bool: + return address.rsplit("@", 1)[-1] in {d.lower() for d in domains} + + +def original_recipient(text: str, forwarder: str, allowed_domains) -> str: + """First allowed address from the quoted 'To:'/'An:' lines, else the forwarder, else ''. + + Only addresses of allowed domains are ever returned: the agent types this address into foreign pages.""" + plain = BeautifulSoup(_ANGLE_ADDRESS.sub(r"[\1]", text or ""), "html.parser").get_text("\n") + for line in _HEADER.findall(plain): + for address in _ADDRESS.findall(line): + if _allowed(address.lower(), allowed_domains): + return address.lower() + sender = parseaddr(forwarder or "")[1].lower() + return sender if sender and _allowed(sender, allowed_domains) else "" diff --git a/tests/test_forwarded.py b/tests/test_forwarded.py new file mode 100644 index 0000000..f9ea57c --- /dev/null +++ b/tests/test_forwarded.py @@ -0,0 +1,40 @@ +"""Which address did the newsletter go to? Taken from the quoted header of a forwarded mail.""" + +import pytest + +from forwarded import original_recipient + +ALLOWED = ("wagnersnetz.de", "gmail.com") + + +def test_german_forward_header(): + text = "---------- Weitergeleitete Nachricht ---------\nVon: News \nAn: Philipp \nBetreff: Angebot\n" + assert original_recipient(text, "jens@wagnersnetz.de", ALLOWED) == "philipp@wagnersnetz.de" + + +def test_english_forward_header_in_html(): + html = "
From: news@shop.example
To: Some.One@Gmail.com
Subject: x
" + assert original_recipient(html, "philipp@wagnersnetz.de", ALLOWED) == "some.one@gmail.com" + + +def test_quoted_prefix_and_several_recipients_first_allowed_wins(): + text = "> To: other@elsewhere.example, anne@wagnersnetz.de\n" + assert original_recipient(text, "", ALLOWED) == "anne@wagnersnetz.de" + + +def test_falls_back_to_forwarder(): + assert original_recipient("no header here", "Jens ", ALLOWED) == "jens@wagnersnetz.de" + + +def test_foreign_domain_is_never_used(): + text = "To: victim@elsewhere.example\n" + assert original_recipient(text, "someone@elsewhere.example", ALLOWED) == "" + + +@pytest.mark.parametrize("domain", ["evilwagnersnetz.de", "wagnersnetz.de.evil.example"]) +def test_domain_must_match_exactly(domain): + assert original_recipient(f"To: x@{domain}\n", "", ALLOWED) == "" + + +def test_subdomain_is_not_allowed_implicitly(): + assert original_recipient("To: x@mail.wagnersnetz.de\n", "", ALLOWED) == ""