在使用 python 的电子邮件和 smtplib 包发送的电子邮件中，EmailMessage 无法正确显示

Question

My emails can be correctly sent but don't show correctly in the receiver mails. It looks like this:

To: =?utf-8?b?..?= <....com> MIME-Version: 1.0 Content-Type: multipart/mixed; boundary="===============5404281335870522242=="

--===============5404281335870522242== Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: base64

5bCK5pWs55qE5a2U6LaF5YW...

--===============5404281335870522242== Content-Type: image/png Content-Transfer-Encoding: base64 Content-Disposition: attachment; filename="user.png" MIME-Version: 1.0

iVBORw0KGgo...

The MIME string is directly shown except the Subject and the From line(It is shown after the To ) as well as all bodies in plain text.

There's my code:

import smtplib
import ssl
import mimetypes
from pathlib import Path
from email.message import EmailMessage
from email.utils import formataddr
import time

class EmailSender:
    PORT = 465
    CONTEXT = ssl.create_default_context()

    def __init__(
        self,
        username,
        password,
        host,
    ):
        self.username = username
        self.password = password
        self.host = host
        self.mails = []

    def _add_name_header(self, name="", mail_addr=""):
        if name:
            return formataddr((name, mail_addr))
        else:
            return mail_addr

    def add_mail(
        self,
        from_email="",
        from_name="",
        to_email="",
        to_name="",
        subject="",
        message_txt="",
        files=None,
    ):
        msg = EmailMessage()
        msg["Subject"] = subject
        msg["From"] = self._add_name_header(from_name, from_email)
        msg["To"] = self._add_name_header(to_name, to_email)
        msg.set_content(message_txt)

        if not files is None:
            for file_obj in files:
                if file_obj.exists():
                    file = str(file_obj)
                    ctype, encoding = mimetypes.guess_type(file)
                    if ctype is None or encoding is not None:
                        # No guess could be made, or the file is encoded (compressed), so use a generic bag-of-bits type.
                        ctype = "application/octet-stream"
                    maintype, subtype = ctype.split("/", 1)
                    with file_obj.open("rb") as fp:
                        msg.add_attachment(
                            fp.read(),
                            maintype=maintype,
                            subtype=subtype,
                            filename=file_obj.name,
                        )

        self.mails.append(msg)

    def send(self, time_interval=1):
        with smtplib.SMTP_SSL(
            host=self.host, port=self.PORT, context=self.CONTEXT
        ) as server:
            try:
                server.login(user=self.username, password=self.password)
            except Exception as e:
                # Need process errors
                raise e
            for msg in self.mails:
                server.send_message(msg)
                time.sleep(time_interval)

And I just do:

sender = EmailSender(
        username, password, host="smtp.163.com"
)

files = list(Path("D:/").glob("*.pdf"))

sender.add_mail(
        from_email, from_name, to_email, to_name, subject, message_txt, files=None
)
sender.send(time_interval=10)

Answer 1

I'm the OP of the question. I just solved this problem by myself and I'd share the solution.

TLNR: Non-Ascii chars are used in my mails so use msg = EmailMessage(EmailPolicy(utf8=True)) instead of msg = EmailMessage() .

I misunderstood these sentences in the doc of SMTP.send_message :

If any of the addresses in from_addr and to_addrs contain non-ASCII characters and the server does not advertise SMTPUTF8 support, an SMTPNotSupported error is raised. Otherwise the Message is serialized with a clone of its policy with the utf8 attribute set to True, and SMTPUTF8 and BODY=8BITMIME are added to mail_options.

Since I add a non-ASCII header to my address, I believe that smtplib will automatically use the utf8 policy for me. But in the file smtplib.py I saw this:

if from_addr is None:
    # Some code
    from_addr = email.utils.getaddresses([from_addr])[0][1]
if to_addrs is None:
    # Some code
    to_addrs = [a[1] for a in email.utils.getaddresses(addr_fields)]
# Some code
international = False
try:
    "".join([from_addr, *to_addrs]).encode("ascii")
except UnicodeEncodeError:
    # Some code
    international = True

That is, the function only checks if the address parts have non-ASCII chars but not along with the header names.

After that, the message is dealt with as pure ASCII content, that's maybe no problem, I have no idea why, but somohow, many extra /r chars are inserted before and after the To:xxx line , which makes the smtp server think this as a separator maybe? And finally caused the problem.