MailAttachment.filename parse non-ascii filename

README.rst

@@ -59,7 +59,7 @@ MailBox, MailBoxTls, MailBoxUnencrypted - for create mailbox client. `TLS exampl
 
 BaseMailBox.<auth> - login, login_utf8, xoauth2, logout - authentication functions, they support context manager.
 
-BaseMailBox.fetch - first searches email nums by criteria in current folder, then fetch and yields `MailMessage <#email-attributes>`_:
+BaseMailBox.fetch - first searches email uids by criteria in current folder, then fetch and yields `MailMessage <#email-attributes>`_, args:
 
 * *criteria* = 'ALL', message search criteria, `query builder <#search-criteria>`_
 * *charset* = 'US-ASCII', indicates charset of the strings that appear in the search criteria. See rfc2978
@@ -69,7 +69,7 @@ BaseMailBox.fetch - first searches email nums by criteria in current folder, the
 * *headers_only* = False, get only email headers (without text, html, attachments)
 * *bulk* = False, False - fetch each message separately per N commands - low memory consumption, slow; True - fetch all messages per 1 command - high memory consumption, fast
 
-BaseMailBox.uids - search mailbox for matching message uids in current folder, returns [str]
+BaseMailBox.uids - search mailbox for matching message uids in current folder, returns [str | None], None when MailMessage.from_bytes used, args:
 
 * *criteria* = 'ALL', message search criteria, `query builder <#search-criteria>`_
 * *charset* = 'US-ASCII', indicates charset of the strings that appear in the search criteria. See rfc2978
@@ -92,7 +92,7 @@ MailMessage and MailAttachment public attributes are cached by functools.lru_cac
 .. code-block:: python
 
     for msg in mailbox.fetch():  # generator: imap_tools.MailMessage
-        msg.uid          # str: '123'
+        msg.uid          # str | None: '123'
         msg.subject      # str: 'some subject 你 привет'
         msg.from_        # str: 'Bartö[email protected]'
         msg.to           # tuple: ('[email protected]', '[email protected]', )

docs/release_notes.rst

@@ -1,3 +1,7 @@
+1.5.0
+=====
+* Fixed: MailAttachment.filename parse non-ascii filename
+
 1.4.0
 =====
 * [Breaking] MailMessage.html replacing charset to utf-8 in html meta for consistency

imap_tools/__init__.py

@@ -11,4 +11,4 @@
 from .utils import EmailAddress
 from .errors import *
 
-__version__ = '1.4.0'
+__version__ = '1.5.0'

imap_tools/consts.py

@@ -4,6 +4,7 @@
 
 UID_PATTERN = re.compile(r'(^|\s+|\W)UID\s+(?P<uid>\d+)')
 
+CODECS_OFFICIAL_REPLACEMENT_CHAR = '�'
 
 class MailMessageFlags:
     """

imap_tools/message.py

@@ -6,10 +6,11 @@
 from itertools import chain
 from functools import lru_cache
 from email.header import decode_header
+from email.message import _parseparam, _unquotevalue  # noqa
 from typing import Tuple, Dict, Optional, List
 
 from .utils import decode_value, parse_email_addresses, parse_email_date, EmailAddress, replace_html_ct_charset
-from .consts import UID_PATTERN
+from .consts import UID_PATTERN, CODECS_OFFICIAL_REPLACEMENT_CHAR
 
 
 class MailMessage:
@@ -238,8 +239,37 @@ def filename(self) -> str:
             forwarded message (Content-Type = message/rfc822)
         :return: filename
         """
+        # attempt 1
         raw = self.part.get_filename() or ''
-        return ''.join(decode_value(*head_part) for head_part in decode_header(raw))
+        attempt_1_filename = ''.join(decode_value(*head_part) for head_part in decode_header(raw))
+        if CODECS_OFFICIAL_REPLACEMENT_CHAR not in attempt_1_filename:
+            return attempt_1_filename
+
+        # attempt 2 - non-ascii filename
+        for header_name, target_param_name in (('content-disposition', 'filename'), ('content-type', 'name')):
+            header_obj = self.part.get(header_name, None)
+            if not header_obj:
+                continue
+            for header_item in decode_header(header_obj):
+                if header_item[1] == 'unknown-8bit':
+                    try:
+                        # suppose encoded utf8
+                        parsed_params = _parseparam(header_item[0].decode(errors='replace'))  # ['',]
+                    except Exception:  # noqa
+                        continue
+                    for parsed_param_item in parsed_params:
+                        try:
+                            name, val = parsed_param_item.split('=', 1)
+                            name = name.strip()
+                            val = val.strip()
+                        except ValueError:
+                            # Must have been a bare attribute
+                            name = parsed_param_item.strip()
+                            val = ''
+                        if name == target_param_name and val and CODECS_OFFICIAL_REPLACEMENT_CHAR not in val:
+                            return val.strip('"')
+
+        return attempt_1_filename
 
     @property
     @lru_cache()

tests/messages_data/attachment_emails/attachment_nonascii_filename.py

@@ -15,7 +15,7 @@
     headers={'mime-version': ('1.0 (Apple Message framework v730)',), 'content-type': ('multipart/mixed; boundary=Apple-Mail-13-196941151',), 'message-id': ('<[email protected]>',), 'from': ('[email protected]',), 'subject': ('testing',), 'date': ('Mon, 6 Jun 2005 22:21:22 +0200',), 'to': ('[email protected]',)},
     attachments=[
         dict(
-            filename='ci��le.txt',
+            filename='ciële.txt',
             content_id='',
             content_disposition='attachment',
             content_type='text/plain',