avoid a certain type of bug

[edward.git] / edward

#! /usr/bin/env python3
# -*- coding: utf-8 -*-
"""*********************************************************************
* Edward is free software: you can redistribute it and/or modify       *
* it under the terms of the GNU Affero Public License as published by  *
* the Free Software Foundation, either version 3 of the License, or    *
* (at your option) any later version.                                  *
*                                                                      *
* Edward is distributed in the hope that it will be useful,            *
* but WITHOUT ANY WARRANTY; without even the implied warranty of       *
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the        *
* GNU Affero Public License for more details.                          *
*                                                                      *
* You should have received a copy of the GNU Affero Public License     *
* along with Edward.  If not, see <http://www.gnu.org/licenses/>.      *
*                                                                      *
* Copyright (C) 2014-2015 Andrew Engelbrecht                (AGPLv3+)  *
* Copyright (C) 2014      Josh Drake                        (AGPLv3+)  *
* Copyright (C) 2014      Lisa Marie Maginnis               (AGPLv3+)  *
* Copyright (C) 2009-2015 Tails developers <tails@boum.org> ( GPLv3+)  *
* Copyright (C) 2009      W. Trevor King <wking@drexel.edu> ( GPLv2+)  *
*                                                                      *
* Special thanks to Josh Drake for writing the original edward bot! :) *
*                                                                      *
************************************************************************

Code sourced from these projects:

  * http://agpl.fsf.org/emailselfdefense.fsf.org/edward/CURRENT/edward.tar.gz
  * https://git-tails.immerda.ch/whisperback/tree/whisperBack/encryption.py?h=feature/python3
  * http://www.physics.drexel.edu/~wking/code/python/send_pgp_mime
"""

import sys
import gpgme
import re
import io
import os

import email.parser
import email.message
import email.encoders

from email.mime.multipart       import MIMEMultipart
from email.mime.application     import MIMEApplication
from email.mime.nonmultipart    import MIMENonMultipart

import edward_config

match_types =  [('clearsign',
                '-----BEGIN PGP SIGNED MESSAGE-----.*?-----BEGIN PGP SIGNATURE-----.*?-----END PGP SIGNATURE-----'),
                ('message',
                '-----BEGIN PGP MESSAGE-----.*?-----END PGP MESSAGE-----'),
                ('pubkey',
                '-----BEGIN PGP PUBLIC KEY BLOCK-----.*?-----END PGP PUBLIC KEY BLOCK-----'),
                ('detachedsig',
                '-----BEGIN PGP SIGNATURE-----.*?-----END PGP SIGNATURE-----')]


class EddyMsg (object):
    def __init__(self):
        self.multipart          = False
        self.subparts           = []

        self.charset            = None
        self.payload_bytes      = None
        self.payload_pieces     = []

        self.filename           = None
        self.content_type       = None
        self.description_list   = None


class PayloadPiece (object):
    def __init__(self):
        self.piece_type         = None
        self.string             = None
        self.gpg_data           = None


class GPGData (object):
    def __init__(self):
        self.decrypted          = False

        self.plainobj           = None
        self.sigs               = []
        self.keys               = []


def main ():

    handle_args()

    gpgme_ctx = get_gpg_context(edward_config.gnupghome,
                              edward_config.sign_with_key)

    email_text = sys.stdin.read()
    result = parse_pgp_mime(email_text, gpgme_ctx)

    email_from, email_subject = email_from_subject(email_text)

#    plaintext, fingerprints = email_decode_flatten(email_text, gpgme_ctx, False)
#    encrypt_to_key = choose_reply_encryption_key(gpgme_ctx, fingerprints)
#
#    reply_message = generate_reply(plaintext, email_from, \
#                                   email_subject, encrypt_to_key,
#                                   gpgme_ctx)

    print(flatten_eddy(result))


def get_gpg_context (gnupghome, sign_with_key_fp):

    os.environ['GNUPGHOME'] = gnupghome

    gpgme_ctx = gpgme.Context()
    gpgme_ctx.armor = True

    try:
        sign_with_key = gpgme_ctx.get_key(sign_with_key_fp)
    except:
        error("unable to load signing key. is the gnupghome "
                + "and signing key properly set in the edward_config.py?")
        exit(1)

    gpgme_ctx.signers = [sign_with_key]

    return gpgme_ctx


def parse_pgp_mime (email_text, gpgme_ctx):

    email_struct = email.parser.Parser().parsestr(email_text)

    eddy_obj = parse_mime(email_struct)
    eddy_obj = split_payloads(eddy_obj)
    eddy_obj = gpg_on_payloads(eddy_obj, gpgme_ctx)

    return eddy_obj


def parse_mime(msg_struct):

    eddy_obj = EddyMsg()

    if msg_struct.is_multipart() == True:
        payloads = msg_struct.get_payload()

        eddy_obj.multipart = True
        eddy_obj.subparts = list(map(parse_mime, payloads))

    else:
        eddy_obj = get_subpart_data(msg_struct)

    return eddy_obj


def scan_and_split (payload_piece, match_type, pattern):

    # don't try to re-split pieces containing gpg data
    if payload_piece.piece_type != "text":
        return [payload_piece]

    flags = re.DOTALL | re.MULTILINE
    matches = re.search("(?P<beginning>.*?)(?P<match>" + pattern +
                        ")(?P<rest>.*)", payload_piece.string, flags=flags)

    if matches == None:
        pieces = [payload_piece]

    else:

        beginning               = PayloadPiece()
        beginning.string        = matches.group('beginning')
        beginning.piece_type    = payload_piece.piece_type

        match                   = PayloadPiece()
        match.string            = matches.group('match')
        match.piece_type        = match_type

        rest                    = PayloadPiece()
        rest.string             = matches.group('rest')
        rest.piece_type         = payload_piece.piece_type

        more_pieces = scan_and_split(rest, match_type, pattern)
        pieces = [beginning, match ] + more_pieces

    return pieces


def get_subpart_data (part):

    obj = EddyMsg()

    obj.charset             = part.get_content_charset()
    obj.payload_bytes       = part.get_payload(decode=True)

    obj.filename            = part.get_filename()
    obj.content_type        = part.get_content_type()
    obj.description_list    = part['content-description']

    # your guess is as good as a-myy-ee-ine...
    if obj.charset == None:
        obj.charset = 'utf-8'

    if obj.payload_bytes != None:
        try:
            payload = PayloadPiece()
            payload.string = obj.payload_bytes.decode(obj.charset)
            payload.piece_type = 'text'

            obj.payload_pieces = [payload]
        except UnicodeDecodeError:
            pass

    return obj


def do_to_eddys_pieces (function_to_do, eddy_obj, data):

    if eddy_obj == None:
        return []

    if eddy_obj.multipart == True:
        result_list = []
        for sub in eddy_obj.subparts:
            result_list += do_to_eddys_pieces(function_to_do, sub, data)
    else:
        result_list = [function_to_do(eddy_obj, data)]

    return result_list


def split_payloads (eddy_obj):

    for match_type in match_types:
        do_to_eddys_pieces(split_payload_pieces, eddy_obj, match_type)

    return eddy_obj


def split_payload_pieces (eddy_obj, match_type):

    (match_name, pattern) = match_type

    new_pieces_list = []
    for piece in eddy_obj.payload_pieces:
        new_pieces_list += scan_and_split(piece, match_name, pattern)

    eddy_obj.payload_pieces = new_pieces_list


def gpg_on_payloads (eddy_obj, gpgme_ctx):

    do_to_eddys_pieces(gpg_on_payload_pieces, eddy_obj, gpgme_ctx)

    return eddy_obj


def gpg_on_payload_pieces (eddy_obj, gpgme_ctx):

    for piece in eddy_obj.payload_pieces:

        if piece.piece_type == "text":
            # don't transform the plaintext.
            pass

        elif piece.piece_type == "message":
            (plaintext, sigs) = decrypt_block (piece.string, gpgme_ctx)

            if plaintext:
                piece.gpg_data = GPGData()
                piece.gpg_data.sigs = sigs
                # recurse!
                piece.gpg_data.plainobj = parse_pgp_mime(plaintext, gpgme_ctx)

        elif piece.piece_type == "pubkey":
            fingerprints = add_gpg_key(piece.string, gpgme_ctx)

            if fingerprints != []:
                piece.gpg_data = GPGData()
                piece.gpg_data.keys = fingerprints

        elif piece.piece_type == "clearsign":
            (plaintext, fingerprints) = verify_clear_signature(piece.string, gpgme_ctx)

            if fingerprints != []:
                piece.gpg_data = GPGData()
                piece.gpg_data.sigs = fingerprints
                piece.gpg_data.plainobj = parse_pgp_mime(plaintext, gpgme_ctx)

        else:
            pass


def flatten_eddy (eddy_obj):

    string = "\n".join(do_to_eddys_pieces(flatten_payload_pieces, eddy_obj, None))

    return string


def flatten_payload_pieces (eddy_obj, _ignore):

    string = ""
    for piece in eddy_obj.payload_pieces:
        if piece.piece_type == "text":
            string += piece.string
        elif piece.piece_type == "message":
            # recursive!
            string += flatten_eddy(piece.gpg_data.plainobj)
        elif piece.piece_type == "pubkey":
            string += "thanks for your public key:"
            for key in piece.gpg_data.keys:
                string += "\n" + key
        elif piece.piece_type == "clearsign":
            string += "*** Begin signed part ***\n"
            string += flatten_eddy(piece.gpg_data.plainobj)
            string += "\n*** End signed part ***"

    return string


def email_from_subject (email_text):

    email_struct = email.parser.Parser().parsestr(email_text)

    email_from      = email_struct['From']
    email_subject   = email_struct['Subject']

    return email_from, email_subject


def add_gpg_key (key_block, gpgme_ctx):

    fp = io.BytesIO(key_block.encode('ascii'))

    result = gpgme_ctx.import_(fp)
    imports = result.imports

    fingerprints = []

    if imports != []:
        for import_ in imports:
            fingerprint = import_[0]
            fingerprints += [fingerprint]

            debug("added gpg key: " + fingerprint)

    return fingerprints


def verify_clear_signature (sig_block, gpgme_ctx):

    # FIXME: this might require the un-decoded bytes
    # or the correct re-encoding with the carset of the mime part.
    msg_fp = io.BytesIO(sig_block.encode('utf-8'))
    ptxt_fp = io.BytesIO()

    result = gpgme_ctx.verify(msg_fp, None, ptxt_fp)

    # FIXME: this might require using the charset of the mime part.
    plaintext = ptxt_fp.getvalue().decode('utf-8')

    fingerprints = []
    for res_ in result:
        fingerprints += [res_.fpr]

    return plaintext, fingerprints


def decrypt_block (msg_block, gpgme_ctx):

    block_b = io.BytesIO(msg_block.encode('ascii'))
    plain_b = io.BytesIO()

    try:
        sigs = gpgme_ctx.decrypt_verify(block_b, plain_b)
    except:
        return ("",[])

    plaintext = plain_b.getvalue().decode('utf-8')
    return (plaintext, sigs)


def choose_reply_encryption_key (gpgme_ctx, fingerprints):

    reply_key = None
    for fp in fingerprints:
        try:
            key = gpgme_ctx.get_key(fp)

            if (key.can_encrypt == True):
                reply_key = key
                break
        except:
            continue


    return reply_key


def generate_reply (plaintext, email_from, email_subject, encrypt_to_key,
                    gpgme_ctx):


    reply  = "To: " + email_from + "\n"
    reply += "Subject: " + email_subject + "\n"

    if (encrypt_to_key != None):
        plaintext_reply  = "thanks for the message!\n\n\n"
        plaintext_reply += email_quote_text(plaintext)

        # quoted printable encoding lets most ascii characters look normal
        # before the decrypted mime message is decoded.
        char_set = email.charset.Charset("utf-8")
        char_set.body_encoding = email.charset.QP

        # MIMEText doesn't allow setting the text encoding
        # so we use MIMENonMultipart.
        plaintext_mime = MIMENonMultipart('text', 'plain')
        plaintext_mime.set_payload(plaintext_reply, charset=char_set)

        encrypted_text = encrypt_sign_message(plaintext_mime.as_string(),
                                              encrypt_to_key,
                                              gpgme_ctx)

        control_mime = MIMEApplication("Version: 1",
                                       _subtype='pgp-encrypted',
                                       _encoder=email.encoders.encode_7or8bit)
        control_mime['Content-Description'] = 'PGP/MIME version identification'
        control_mime.set_charset('us-ascii')

        encoded_mime = MIMEApplication(encrypted_text,
                                       _subtype='octet-stream; name="encrypted.asc"',
                                       _encoder=email.encoders.encode_7or8bit)
        encoded_mime['Content-Description'] = 'OpenPGP encrypted message'
        encoded_mime['Content-Disposition'] = 'inline; filename="encrypted.asc"'
        encoded_mime.set_charset('us-ascii')

        message_mime = MIMEMultipart(_subtype="encrypted", protocol="application/pgp-encrypted")
        message_mime.attach(control_mime)
        message_mime.attach(encoded_mime)
        message_mime['Content-Disposition'] = 'inline'

        reply += message_mime.as_string()

    else:
        reply += "\n"
        reply += "Sorry, i couldn't find your key.\n"
        reply += "I'll need that to encrypt a message to you."

    return reply


def email_quote_text (text):

    quoted_message = re.sub(r'^', r'> ', text, flags=re.MULTILINE)

    return quoted_message


def encrypt_sign_message (plaintext, encrypt_to_key, gpgme_ctx):

    plaintext_bytes = io.BytesIO(plaintext.encode('ascii'))
    encrypted_bytes = io.BytesIO()

    gpgme_ctx.encrypt_sign([encrypt_to_key], gpgme.ENCRYPT_ALWAYS_TRUST,
            plaintext_bytes, encrypted_bytes)

    encrypted_txt = encrypted_bytes.getvalue().decode('ascii')
    return encrypted_txt


def error (error_msg):

    sys.stderr.write(progname + ": " + str(error_msg) + "\n")


def debug (debug_msg):

    if edward_config.debug == True:
        error(debug_msg)


def handle_args ():
    if __name__ == "__main__":

        global progname
        progname = sys.argv[0]

        if len(sys.argv) > 1:
            print(progname + ": error, this program doesn't " \
                    "need any arguments.", file=sys.stderr)
            exit(1)


main()