python3 fixes

Author: jaredmauch

Mailman/Archiver/HyperArch.py

@@ -419,7 +419,7 @@ def decode_headers(self):
                     i18n.set_language(self._lang)
                     atmark = _(' at ')
                     subject = re.sub(r'([-+,.\w]+)@([-+.\w]+)',
-                              '\g<1>' + atmark + '\g<2>', subject)
+                              r'\g<1>' + atmark + r'\g<2>', subject)
                 finally:
                     i18n.set_translation(otrans)
             self.decoded['subject'] = subject
@@ -436,10 +436,10 @@ def strip_subject(self, subject):
             subject = re.sub(prefix_pat, '', subject)
         subject = subject.lstrip()
         # MAS Should we strip FW and FWD too?
-        strip_pat = re.compile('^((RE|AW|SV|VS)(\[\d+\])?:\s*)+', re.I)
+        strip_pat = re.compile(r'^((RE|AW|SV|VS)(\[\d+\])?:\s*)+', re.I)
         stripped = strip_pat.sub('', subject)
         # Also remove whitespace to avoid folding/unfolding differences
-        stripped = re.sub('\s', '', stripped)
+        stripped = re.sub(r'\s', '', stripped)
         return stripped
 
     def decode_charset(self, field):
@@ -589,7 +589,7 @@ def as_text(self):
                 if isinstance(atmark, bytes):
                     atmark = str(atmark, cset)
                 body = re.sub(r'([-+,.\w]+)@([-+.\w]+)',
-                              '\g<1>' + atmark + '\g<2>', body)
+                              r'\g<1>' + atmark + r'\g<2>', body)
             finally:
                 i18n.set_translation(otrans)
 

Mailman/Bouncers/SimpleMatch.py

@@ -49,11 +49,11 @@ def _c(pattern):
     # robanal.demon.co.uk
     (_c('this message was created automatically by mail delivery software'),
      _c('original message follows'),
-     _c('rcpt to:\s*<(?P<addr>[^>]*)>')),
+     _c(r'rcpt to:\s*<(?P<addr>[^>]*)>')),
     # s1.com (InterScan E-Mail VirusWall NT ???)
     (_c('message from interscan e-mail viruswall nt'),
      _c('end of message'),
-     _c('rcpt to:\s*<(?P<addr>[^>]*)>')),
+     _c(r'rcpt to:\s*<(?P<addr>[^>]*)>')),
     # Smail
     (_c('failed addresses follow:'),
      _c('message text follows:'),
@@ -65,35 +65,35 @@ def _c(pattern):
     # turbosport.com runs something called `MDaemon 3.5.2' ???
     (_c('The following addresses did NOT receive a copy of your message:'),
      _c('--- Session Transcript ---'),
-     _c('[>]\s*(?P<addr>.*)$')),
+     _c(r'[>]\s*(?P<addr>.*)$')),
     # usa.net
-    (_c('Intended recipient:\s*(?P<addr>.*)$'),
+    (_c(r'Intended recipient:\s*(?P<addr>.*)$'),
      _c('--------RETURNED MAIL FOLLOWS--------'),
-     _c('Intended recipient:\s*(?P<addr>.*)$')),
+     _c(r'Intended recipient:\s*(?P<addr>.*)$')),
     # hotpop.com
-    (_c('Undeliverable Address:\s*(?P<addr>.*)$'),
+    (_c(r'Undeliverable Address:\s*(?P<addr>.*)$'),
      _c('Original message attached'),
-     _c('Undeliverable Address:\s*(?P<addr>.*)$')),
+     _c(r'Undeliverable Address:\s*(?P<addr>.*)$')),
     # Another demon.co.uk format
     (_c('This message was created automatically by mail delivery'),
      _c('^---- START OF RETURNED MESSAGE ----'),
      _c("addressed to '(?P<addr>[^']*)'")),
     # Prodigy.net full mailbox
     (_c("User's mailbox is full:"),
      _c('Unable to deliver mail.'),
-     _c("User's mailbox is full:\s*<(?P<addr>[^>]*)>")),
+     _c(r"User's mailbox is full:\s*<(?P<addr>[^>]*)>")),
     # Microsoft SMTPSVC
     (_c('The email below could not be delivered to the following user:'),
      _c('Old message:'),
      _c('<(?P<addr>[^>]*)>')),
     # Yahoo on behalf of other domains like sbcglobal.net
-    (_c('Unable to deliver message to the following address\(es\)\.'),
-     _c('--- Original message follows\.'),
+    (_c(r'Unable to deliver message to the following address\(es\)\.'),
+     _c(r'--- Original message follows\.'),
      _c('<(?P<addr>[^>]*)>:')),
     # googlemail.com
     (_c('Delivery to the following recipient(s)? failed'),
      _c('----- Original message -----'),
-     _c('^\s*(?P<addr>[^\s@]+@[^\s@]+)\s*$')),
+     _c(r'^\s*(?P<addr>[^\s@]+@[^\s@]+)\s*$')),
     # kundenserver.de, mxlogic.net
     (_c('A message that you( have)? sent could not be delivered'),
      _c('^---'),
@@ -108,39 +108,39 @@ def _c(pattern):
      # unique to stop on, so stop on the first line of at least 3 characters
      # that doesn't start with 'D' (to not stop immediately) and has no '@'.
      _c('^[^D][^@]{2,}$'),
-     _c('^\s*(. )?(?P<addr>[^\s@]+@[^\s@]+)\s*$')),
+     _c(r'^\s*(. )?(?P<addr>[^\s@]+@[^\s@]+)\s*$')),
     # and another thehartfod.com/hartfordlife.com
-    (_c('^Your message\s*$'),
+    (_c(r'^Your message\s*$'),
      _c('^because:'),
-     _c('^\s*(?P<addr>[^\s@]+@[^\s@]+)\s*$')),
+     _c(r'^\s*(?P<addr>[^\s@]+@[^\s@]+)\s*$')),
     # kviv.be (InterScan NT)
     (_c('^Unable to deliver message to'),
      _c(r'\*+\s+End of message\s+\*+'),
      _c('<(?P<addr>[^>]*)>')),
     # earthlink.net supported domains
     (_c('^Sorry, unable to deliver your message to'),
      _c('^A copy of the original message'),
-     _c('\s*(?P<addr>[^\s@]+@[^\s@]+)\s+')),
+     _c(r'\s*(?P<addr>[^\s@]+@[^\s@]+)\s+')),
     # ademe.fr
     (_c('^A message could not be delivered to:'),
      _c('^Subject:'),
-     _c('^\s*(?P<addr>[^\s@]+@[^\s@]+)\s*$')),
+     _c(r'^\s*(?P<addr>[^\s@]+@[^\s@]+)\s*$')),
     # andrew.ac.jp
     (_c('^Invalid final delivery userid:'),
      _c('^Original message follows.'),
-     _c('\s*(?P<addr>[^\s@]+@[^\s@]+)\s*$')),
+     _c(r'\s*(?P<addr>[^\s@]+@[^\s@]+)\s*$')),
     # E500_SMTP_Mail_Service@lerctr.org and similar
     (_c('---- Failed Recipients ----'),
      _c(' Mail ----'),
      _c('<(?P<addr>[^>]*)>')),
     # cynergycom.net
     (_c('A message that you sent could not be delivered'),
      _c('^---'),
-     _c('(?P<addr>[^\s@]+@[^\s@)]+)')),
+     _c(r'(?P<addr>[^\s@]+@[^\s@)]+)')),
     # LSMTP for Windows
-    (_c('^--> Error description:\s*$'),
+    (_c(r'^--> Error description:\s*$'),
      _c('^Error-End:'),
-     _c('^Error-for:\s+(?P<addr>[^\s@]+@[^\s@]+)')),
+     _c(r'^Error-for:\s+(?P<addr>[^\s@]+@[^\s@]+)')),
     # Qmail with a tri-language intro beginning in spanish
     (_c('Your message could not be delivered'),
      _c('^-'),
@@ -152,49 +152,49 @@ def _c(pattern):
     # dadoservice.it
     (_c('Your message has encountered delivery problems'),
      _c('Your message reads'),
-     _c('addressed to\s*(?P<addr>[^\s@]+@[^\s@)]+)')),
+     _c(r'addressed to\s*(?P<addr>[^\s@]+@[^\s@)]+)')),
     # gomaps.com
     (_c('Did not reach the following recipient'),
      _c('^\s*$'),
-     _c('\s(?P<addr>[^\s@]+@[^\s@]+)')),
+     _c(r'\s(?P<addr>[^\s@]+@[^\s@]+)')),
     # EYOU MTA SYSTEM
     (_c('This is the deliver program at'),
      _c('^-'),
      _c('^(?P<addr>[^\s@]+@[^\s@<>]+)')),
     # A non-standard qmail at ieo.it
     (_c('this is the email server at'),
      _c('^-'),
-     _c('\s(?P<addr>[^\s@]+@[^\s@]+)[\s,]')),
+     _c(r'\s(?P<addr>[^\s@]+@[^\s@]+)[\s,]')),
     # pla.net.py (MDaemon.PRO ?)
     (_c('- no such user here'),
      _c('There is no user'),
-     _c('^(?P<addr>[^\s@]+@[^\s@]+)\s')),
+     _c(r'^(?P<addr>[^\s@]+@[^\s@]+)\s')),
     # fastdnsservers.com
     (_c('The following recipient.*could not be reached'),
      _c('bogus stop pattern'),
-     _c('^(?P<addr>[^\s@]+@[^\s@]+)\s*$')),
+     _c(r'^(?P<addr>[^\s@]+@[^\s@]+)\s*$')),
     # lttf.com
     (_c('Could not deliver message to'),
      _c('^\s*--'),
-     _c('^Failed Recipient:\s*(?P<addr>[^\s@]+@[^\s@]+)\s*$')),
+     _c(r'^Failed Recipient:\s*(?P<addr>[^\s@]+@[^\s@]+)\s*$')),
     # uci.edu
     (_c('--------Message not delivered'),
      _c('--------Error Detail'),
-     _c('^\s*(?P<addr>[^\s@]+@[^\s@]+)\s*$')),
+     _c(r'^\s*(?P<addr>[^\s@]+@[^\s@]+)\s*$')),
     # Dovecot LDA Over quota MDN (bogus - should be DSN).
     (_c('^Your message'),
      _c('^Reporting'),
      _c(
-        'Your message to (?P<addr>[^\s@]+@[^\s@]+) was automatically rejected'
+        r'Your message to (?P<addr>[^\s@]+@[^\s@]+) was automatically rejected'
        )),
     # mail.ru
     (_c('A message that you sent was rejected'),
      _c('This is a copy of your message'),
-     _c('\s(?P<addr>[^\s@]+@[^\s@]+)')),
+     _c(r'\s(?P<addr>[^\s@]+@[^\s@]+)')),
     # MailEnable
     (_c('Message could not be delivered to some recipients.'),
      _c('Message headers follow'),
-     _c('Recipient: \[SMTP:(?P<addr>[^\s@]+@[^\s@]+)\]')),
+     _c(r'Recipient: \[SMTP:(?P<addr>[^\s@]+@[^\s@]+)\]')),
     # This one is from Yahoo but dosen't fit the yahoo recognizer format
     (_c(r'wasn\'t able to deliver the following message'),
      _c(r'---Below this line is a copy of the message.'),

Mailman/Handlers/Approve.py

@@ -122,7 +122,7 @@ def process(mlist, msg, msgdata):
             # If we don't find the pattern in the decoded part, but we do
             # find it after stripping HTML tags, we don't know how to remove
             # it, so we just reject the post.
-            pattern = name + ':(\xA0|\s| )*' + re.escape(passwd)
+            pattern = name + r':(\xA0|\s| )*' + re.escape(passwd)
             for part in typed_subpart_iterator(msg, 'text'):
                 if part is not None and part.get_payload() is not None:
                     lines = part.get_payload(decode=True)

Mailman/Handlers/CookHeaders.py

@@ -38,7 +38,7 @@
 COMMASPACE = ', '
 MAXLINELEN = 78
 
-nonascii = re.compile('[^\s!-~]')
+nonascii = re.compile(r'[^\s!-~]')
 
 def uheader(mlist, s, header_name=None, continuation_ws=' ', maxlinelen=None):
     # Get the charset to encode the string in. Then search if there is any
@@ -183,7 +183,7 @@ def process(mlist, msg, msgdata):
             uvia = str(via, lcs, errors='replace')
 
         # Replace the dummy replacements.
-        uvia = re.sub(u'%\(lrn\)s', ulrn, re.sub(u'%\(realname\)s', urn, uvia))
+        uvia = re.sub(u'%\\(lrn\\)s', ulrn, re.sub(u'%\\(realname\\)s', urn, uvia))
         # And get an RFC 2047 encoded header string.
         dn = str(Header(uvia, lcs))
         change_header('From',
@@ -424,7 +424,7 @@ def prefix_subject(mlist, msg, msgdata):
     # leading space after stripping the prefix. It is not known what MUA would
     # create such a Subject:, but the issue was reported.
     rematch = re.match(
-                       '(\s*(RE|AW|SV|VS)\s*(\[\d+\])?\s*:\s*)+',
+                       r'(\s*(RE|AW|SV|VS)\s*(\[\d+\])?\s*:\s*)+',
                         subject, re.I)
     if rematch:
         subject = subject[rematch.end():]

Mailman/Handlers/Decorate.py

@@ -210,7 +210,7 @@ def decorate(mlist, template, what, extradict=None):
     # `what' is just a descriptive phrase used in the log message
 
     # If template is only whitespace, ignore it.
-    if len(re.sub('\s', '', template)) == 0:
+    if len(re.sub(r'\s', '', template)) == 0:
         return ''
 
     # BAW: We've found too many situations where Python can be fooled into

Mailman/Handlers/SMTPDirect.py

@@ -31,6 +31,7 @@
 import time
 import socket
 import smtplib
+from smtplib import SMTPException
 from base64 import b64encode
 
 from Mailman import mm_cfg
@@ -66,7 +67,11 @@ def __connect(self):
                     self.quit()
                     raise
                 try:
-                    self.__conn.ehlo(mm_cfg.SMTP_HELO_HOST)
+                    # Use a valid hostname for EHLO, fallback to localhost if SMTP_HELO_HOST is empty or invalid
+                    helo_host = mm_cfg.SMTP_HELO_HOST
+                    if not helo_host or helo_host.startswith('.') or helo_host == '@URLHOST@':
+                        helo_host = 'localhost'
+                    self.__conn.ehlo(helo_host)
                 except SMTPException as e:
                     syslog('smtp-failure', 'SMTP EHLO error: %s', e)
                     self.quit()

Mailman/Handlers/SpamDetect.py

@@ -74,7 +74,7 @@ def getDecodedHeaders(msg, cset='utf-8'):
         uvalue = u''
         try:
             if isinstance(v, str):
-                v = decode_header(re.sub('\n\s', ' ', v))
+                v = decode_header(re.sub(r'\n\s', ' ', v))
             else:
                 continue
         except HeaderParseError:

Mailman/Handlers/ToDigest.py

@@ -208,7 +208,7 @@ def send_i18n_digests(mlist, mboxfp):
     print(mastheadtxt, file=plainmsg)
     print(file=plainmsg)
     # Now add the optional digest header but only if more than whitespace.
-    if re.sub('\s', '', mlist.digest_header):
+    if re.sub(r'\s', '', mlist.digest_header):
         lc_digest_header_msg = _('digest header')
         if isinstance(lc_digest_header_msg, bytes):
             lc_digest_header_msg = str(lc_digest_header_msg)
@@ -371,7 +371,7 @@ def send_i18n_digests(mlist, mboxfp):
             print(file=plainmsg)
 
     # Now add the footer but only if more than whitespace.
-    if re.sub('\s', '', mlist.digest_footer):
+    if re.sub(r'\s', '', mlist.digest_footer):
         lc_digest_footer_msg = _('digest footer')
         if isinstance(lc_digest_footer_msg, bytes):
             lc_digest_footer_msg = str(lc_digest_footer_msg)

Mailman/SecurityManager.py

@@ -367,7 +367,7 @@ def __checkone(self, c, authcontext, user):
 
 
 
-splitter = re.compile(';\s*')
+splitter = re.compile(r';\s*')
 
 def parsecookie(s):
     c = {}

Mailman/Utils.py

@@ -1263,7 +1263,7 @@ def strip_verbose_pattern(pattern):
         elif c == ']' and inclass:
             inclass = False
             newpattern += c
-        elif re.search('\s', c):
+        elif re.search(r'\s', c):
             if inclass:
                 if c == NL:
                     newpattern += '\\n'
@@ -1746,7 +1746,7 @@ def check_eq_domains(email, domains_list):
     except ValueError:
         return []
     domain = domain.lower()
-    domains_list = re.sub('\s', '', domains_list).lower()
+    domains_list = re.sub(r'\s', '', domains_list).lower()
     domains = domains_list.split(';')
     domains_list = []
     for d in domains: