This commit was manufactured by cvs2svn to create tag 'milter-0_8_4'.

Sprout from master 2005-10-20 23:36:11 UTC Stuart Gathman <stuart@gathman.org> 'Release 0.8.4'
Cherrypick from bmsi 2005-05-31 18:23:49 UTC Stuart Gathman <stuart@gathman.org> 'Development changes since 0.7.2':
    README
    cid2spf.py
    milter.rc
    milter.rc7
    rejects.py
    rhsbl.m4
    sample.py
    test.py
    test/amazon
    test/big5
    test/bounce
    test/bounce1
    test/bound
    test/honey
    test/missingboundary
    test/samp1
    test/spam44
    test/spam7
    test/spam8
    test/test1
    test/test8
    test/virus1
    test/virus13
    test/virus2
    test/virus3
    test/virus4
    test/virus5
    test/virus6
    test/virus7
    testsample.py

CREDITS

@@ -5,10 +5,16 @@ wrapper (Milter.py) that handles attachments, did lots of testing, packaged
 it with distutils, and generally transformed it from a quick hack to a
 real, usable Python extension.
 
-Other contributors:
+Other contributors (in random order):
 
+Dave MacQuigg 
+  for noticing that smfi_insheader wasn't supported, and creating
+  a template to help first time pymilter users create their own milter.
 Terence Way
   for providing a Python port of SPF
+Scott Kitterman
+  for doing lots of testing and debugging of SPF against draft standard,
+  and for putting up a web page that validates SPF records using spf.py
 Alexander Kourakos
   for plugging several memory leaks
 George Graf at Vienna University of Economics and Business Administration
@@ -22,6 +28,9 @@ John Draper
   then pointing out that it would be easier to just write the MTA in Python.
 Eric S. Johansson
   for helpful design discussions while working on camram
+Alex Savguira
+  for finding bugs with international headers and
+  suggesting the scan_zip option.
 Business Management Systems - http://www.bmsi.com
   for hosting the website, and providing paying clients who need milter service
   so I can work on it as part of my day job.

HOWTO

@@ -0,0 +1,136 @@
+	Step one.  Which DSPAM is right for you?
+
+The DSPAM project makes dspam part of the LDA (Local Delivery Agent).
+Pydspam puts dspam into the MTA (Mail Transfer Agent - sendmail with pymilter).
+
+The advantage of doing dspam in the LDA is that any aliasing has already been
+resolved.  You need only configure mailboxes.
+
+The advantage of doing dspam in the MTA is it can screen an entire 
+company as a gateway with multiple domains.  Unfortunately, this
+means you have to tell it about all the aliases that comprise each
+account.  (Also, pydspam is still uses dspam-2.6.5.2 - the Dspam API
+has changed for newer versions.)
+
+If the LDA is right for you, you'll want to use the official Dspam 
+package.  http://www.nuclearelephant.com/projects/dspam/
+
+If the MTA approach is what you want, then pydspam is what you want.
+
+In either case, you will still want pymilter to block forgeries, Windows 
+executables, etc.
+
+So, lets assume you want to install pymilter, and may or may not
+wish to install pydspam.
+
+	Step two.  Obtaining RPMS.
+
+For basic pymilter you'll need:
+
+python-2.4
+milter-0.8.2	(the RH9 rpm should work on Fedora Core - let me know)
+sendmail-8.13.x	(with milter support enabled)
+
+and for SPF you'll need:
+
+pydns-2.3.0-2.4
+
+and for SRS you'll need:
+
+pysrs-0.30.9-1.py24
+
+I'm pretty sure you will want to have SPF and SRS available.
+
+	Step three.  Activate basic milter.
+
+Activate the basic milter by editing /etc/mail/sendmail.mc and adding:
+
+INPUT_MAIL_FILTER(`pythonfilter', `S=local:/var/run/milter/pythonsock, F=T, T=C:5m;S:20s;R:5m;E:5m')
+
+You can then "make sendmail.cf" and restart sendmail.
+
+Tail /var/log/milter/milter.log while SMTP clients connect to your
+sendmail instance.  This should show you what the milter is doing.
+
+By default, milter-0.8.2 rejects on SPF fail, except for listed domains
+(that are known to be broken).  Some admins don't like that, and 0.8.3 will use
+the /etc/mail/access database to configure SPF responses.  For now,
+if you don't like SPF, you can disable spf by replacing "import spf"
+with "spf = None" around line 285 in /var/log/milter/bms.py.
+
+	Step four.  Tweaking the basic config.
+
+Most pymilter configuration is in /etc/mail/pymilter.cfg.
+
+By default, milter scans attachments for executable extensions.  You can
+turn this off by setting banned_exts to the empty list.  There are options
+to scan ZIP attachments and rfc822 attachments.  When it finds a banned
+file type, milter saves the original message in /var/log/milter/save,
+and replaces the attachment with a plain text warning message.
+
+Configure hello_blacklist with your own helo name and domains - which
+you know cannot legitimately be used by external MTAs.
+
+Configure trusted_relay with your secondary MX servers, if any.  These
+should also run pymilter with similar policies.  (But this isn't
+needed for initial testing.)
+
+Configure internal_connect with subnets of your internal SMTP clients.
+Internal connections skip SPF testing and other policies.
+
+Configure internal_domains with domains used by your internal SMTP clients.
+If they attempt to use any other domain, the attempt is blocked and the
+client is logged as a "zombie".  Conversely, any attempt by an external
+MTA to use one of your internal domains is treated as a forgery and
+blocked (a simplified form of local SPF).
+
+Adjust porn_words and spam_words - these block emails with a Subject
+containing the listed strings.  They can be empty to disable Subject
+string blocking.
+
+	Advanced SPF configuration.
+
+The sendmail access file, or another readonly database with that
+format, can be used for detail spf policy.  SPF access policy
+record are tagged with "SPF-{Result}:".  Results are
+Pass, Neutral, Softfail, Fail, PermError.  Currently supported
+policy keywords are OK, CBV, REJECT.  Currently, TempError always
+results in TEMPFAIL.
+
+The default policies are set in pymilter.cfg.  The defaults
+if none of the config options are set are as follows:
+
+SPF-Fail:	REJECT
+SPF-Softfail:	CBV
+SPF-Neutral:	OK
+SPF-PermError:	REJECT
+SPF-Pass:	OK
+
+The tag may be followed by a specific domain.  For instance, to
+require a Pass from aol.com:
+
+SPF-Neutral:aol.com	REJECT
+SPF-Softfail:aol.com	REJECT
+
+The CBV policy requires a valid HELO name.  If the EHLO name is 
+RFC2822 compliant, then a DSN is sent to the alleged sender.  The 
+template for the DSN is selected according to the SPF result:
+
+Fail:		softfail.txt
+SoftFail:	softfail.txt
+Neutral:	neutral.txt
+PermError:	permerror.txt
+None:		strike3.txt
+
+An SPF-Pass is always accepted by the milter.  Domains can be blacklisted 
+via sendmail in the access file or via a RHS DNS blacklist.
+
+	To be continued.
+
+Forthcoming topics:
+
+SRS config
+
+
+pydspam config
+wiretap config

MANIFEST.in

@@ -1,6 +1,7 @@
 include COPYING
 include TODO
 include NEWS
+include HOWTO
 include CREDITS
 include README
 include MANIFEST.in
@@ -11,6 +12,7 @@ include testdspam.py
 include rejects.py
 include bms.py
 include spf.py
+include cid2spf.py
 include spfquery.py
 include test.py
 include sample.py
@@ -22,5 +24,5 @@ include milter.rc
 include milter.rc7
 include milter.cfg
 include rhsbl.m4
-include softfail.txt
+include *.txt
-include strike3.txt
+include *.html

Milter.py

@@ -1,203 +0,0 @@
-
-# Author: Stuart D. Gathman <stuart@bmsi.com>
-# Copyright 2001 Business Management Systems, Inc.
-# This code is under GPL.  See COPYING for details.
-
-import os
-import milter
-import thread
-
-from milter import ACCEPT,CONTINUE,REJECT,DISCARD,TEMPFAIL,	\
-  set_flags, setdbg, setbacklog, settimeout, \
-  ADDHDRS, CHGBODY, ADDRCPT, DELRCPT, CHGHDRS,	\
-  V1_ACTS, V2_ACTS, CURR_ACTS
-
-try: from milter import QUARANTINE
-except: pass
-
-_seq_lock = thread.allocate_lock()
-_seq = 0
-
-def uniqueID():
-  """Return a sequence number unique to this process.
-  """
-  global _seq
-  _seq_lock.acquire()
-  seqno = _seq = _seq + 1
-  _seq_lock.release()
-  return seqno
-  
-class Milter:
-  """A simple class interface to the milter module.
-  """
-  def _setctx(self,ctx):
-    self.__ctx = ctx
-    if ctx:
-      ctx.setpriv(self)
-
-  # user replaceable callbacks
-  def log(self,*msg):
-    print 'Milter:',
-    for i in msg: print i,
-    print
-
-  def connect(self,hostname,unused,hostaddr):
-    "Called for each connection to sendmail."
-    self.log("connect from %s at %s" % (hostname,hostaddr))
-    return CONTINUE
-
-  def hello(self,hostname):
-    "Called after the HELO command."
-    self.log("hello from %s" % hostname)
-    return CONTINUE
-
-  def envfrom(self,f,*str):
-    """Called to begin each message.
-    f -> string		message sender
-    str -> tuple	additional ESMTP parameters
-    """
-    self.log("mail from",f,str)
-    return CONTINUE
-
-  def envrcpt(self,to,*str):
-    "Called for each message recipient."
-    self.log("rcpt to",to,str)
-    return CONTINUE
-
-  def header(self,field,value):
-    "Called for each message header."
-    self.log("%s: %s" % (field,value))
-    return CONTINUE
-
-  def eoh(self):
-    "Called after all headers are processed."
-    self.log("eoh")
-    return CONTINUE
-
-  def body(self,unused):
-    "Called to transfer the message body."
-    return CONTINUE
-
-  def eom(self):
-    "Called at the end of message."
-    self.log("eom")
-    return CONTINUE
-
-  def abort(self):
-    "Called if the connection is terminated abnormally."
-    self.log("abort")
-    return CONTINUE
-
-  def close(self):
-    "Called at the end of connection, even if aborted."
-    self.log("close")
-    return CONTINUE
-
-  # Milter methods which can be invoked from callbacks
-  def getsymval(self,sym):
-    return self.__ctx.getsymval(sym)
-
-  # If sendmail does not support setmlreply, then only the
-  # first msg line is used.
-  def setreply(self,rcode,xcode=None,msg=None,*ml):
-    return self.__ctx.setreply(rcode,xcode,msg,*ml)
-
-  # Milter methods which can only be called from eom callback.
-  def addheader(self,field,value):
-    return self.__ctx.addheader(field,value)
-
-  def chgheader(self,field,idx,value):
-    return self.__ctx.chgheader(field,idx,value)
-
-  def addrcpt(self,rcpt):
-    return self.__ctx.addrcpt(rcpt)
-
-  def delrcpt(self,rcpt):
-    return self.__ctx.delrcpt(rcpt)
-
-  def replacebody(self,body):
-    return self.__ctx.replacebody(body)
-
-  # When quarantined, a message goes into the mailq as if to be delivered,
-  # but delivery is deferred until the message is unquarantined.
-  def quarantine(self,reason):
-    return self.__ctx.quarantine(reason)
-
-  def progress(self):
-    return self.__ctx.progress()
-
-factory = Milter
-
-def connectcallback(ctx,hostname,family,hostaddr):
-  m = factory()
-  m._setctx(ctx)
-  return m.connect(hostname,family,hostaddr)
-
-def closecallback(ctx):
-  m = ctx.getpriv()
-  if not m: return CONTINUE
-  rc = m.close()
-  m._setctx(None)	# release milterContext
-  return rc
-
-def envcallback(c,args):
-  """Convert ESMTP parms to keyword parameters.
-  Can be used in the envfrom and/or envrcpt callbacks to process
-  ESMTP parameters as python keyword parameters."""
-  kw = {}
-  for s in args[1:]:
-    pos = s.find('=')
-    if pos > 0:
-      kw[s[:pos]] = s[pos+1:]
-  return apply(c,args,kw)
-
-def runmilter(name,socketname,timeout = 0):
-  # This bit is here on the assumption that you will be starting this filter
-  # before sendmail.  If sendmail is not running and the socket already exists,
-  # libmilter will throw a warning.  If sendmail is running, this is still
-  # safe if there are no messages currently being processed.  It's safer to
-  # shutdown sendmail, kill the filter process, restart the filter, and then
-  # restart sendmail.
-  pos = socketname.find(':')
-  if pos > 1:
-    s = socketname[:pos]
-    fname = socketname[pos+1:]
-  else:
-    s = "unix"
-    fname = socketname
-  if s == "unix" or s == "local":
-    print "Removing %s" % fname
-    try:
-      os.unlink(fname)
-    except:
-      pass
-
-  # The default flags set include everything
-  # milter.set_flags(milter.ADDHDRS)
-  milter.set_connect_callback(connectcallback)
-  milter.set_helo_callback(lambda ctx, host: ctx.getpriv().hello(host))
-  milter.set_envfrom_callback(lambda ctx,*str:
-  	ctx.getpriv().envfrom(*str))
-#  	envcallback(ctx.getpriv().envfrom,str))
-  milter.set_envrcpt_callback(lambda ctx,*str:
-  	ctx.getpriv().envrcpt(*str))
-#  	envcallback(ctx.getpriv().envrcpt,str))
-  milter.set_header_callback(lambda ctx,fld,val:
-  	ctx.getpriv().header(fld,val))
-  milter.set_eoh_callback(lambda ctx: ctx.getpriv().eoh())
-  milter.set_body_callback(lambda ctx,chunk: ctx.getpriv().body(chunk))
-  milter.set_eom_callback(lambda ctx: ctx.getpriv().eom())
-  milter.set_abort_callback(lambda ctx: ctx.getpriv().abort())
-  milter.set_close_callback(closecallback)
-
-  milter.setconn(socketname)
-  if timeout > 0: milter.settimeout(timeout)
-  # The name *must* match the X line in sendmail.cf (supposedly)
-  milter.register(name)
-  start_seq = _seq
-  try:
-    milter.main()
-  except milter.error:
-    if start_seq == _seq: raise	# couldn't start
-    # milter has been running for a while, but now it can't start new threads
-    raise milter.error("out of thread resources")

Milter/__init__.py

@@ -16,6 +16,8 @@ from milter import ACCEPT,CONTINUE,REJECT,DISCARD,TEMPFAIL,	\
 try: from milter import QUARANTINE
 except: pass
 
+__version__ = '0.8.4'
+
 _seq_lock = thread.allocate_lock()
 _seq = 0
 
@@ -42,7 +44,7 @@ class Milter:
     for i in msg: print i,
     print
 
-  def connect(self,hostname,unused,hostaddr):
+  def connect(self,hostname,family,hostaddr):
     "Called for each connection to sendmail."
     self.log("connect from %s at %s" % (hostname,hostaddr))
     return CONTINUE
@@ -104,8 +106,8 @@ class Milter:
     return self.__ctx.setreply(rcode,xcode,msg,*ml)
 
   # Milter methods which can only be called from eom callback.
-  def addheader(self,field,value):
+  def addheader(self,field,value,idx=-1):
-    return self.__ctx.addheader(field,value)
+    return self.__ctx.addheader(field,value,idx)
 
   def chgheader(self,field,idx,value):
     return self.__ctx.chgheader(field,idx,value)
@@ -141,16 +143,28 @@ def closecallback(ctx):
   m._setctx(None)	# release milterContext
   return rc
 
+def dictfromlist(args):
+  "Convert ESMTP parm list to keyword dictionary."
+  kw = {}
+  for s in args:
+    pos = s.find('=')
+    if pos > 0:
+      kw[s[:pos].upper()] = s[pos+1:]
+  return kw
+
 def envcallback(c,args):
-  """Convert ESMTP parms to keyword parameters.
+  """Call function c with ESMTP parms converted to keyword parameters.
   Can be used in the envfrom and/or envrcpt callbacks to process
   ESMTP parameters as python keyword parameters."""
   kw = {}
+  pargs = [args[0]]
   for s in args[1:]:
     pos = s.find('=')
     if pos > 0:
-      kw[s[:pos]] = s[pos+1:]
+      kw[s[:pos].upper()] = s[pos+1:]
-  return apply(c,args,kw)
+    else:
+      pargs.append(s)
+  return c(*pargs,**kw)
 
 def runmilter(name,socketname,timeout = 0):
   # This bit is here on the assumption that you will be starting this filter
@@ -177,14 +191,13 @@ def runmilter(name,socketname,timeout = 0):
   # milter.set_flags(milter.ADDHDRS)
   milter.set_connect_callback(connectcallback)
   milter.set_helo_callback(lambda ctx, host: ctx.getpriv().hello(host))
-  milter.set_envfrom_callback(lambda ctx,*str:
+  # For envfrom and envrcpt, we would like to convert ESMTP parms to keyword
-  	ctx.getpriv().envfrom(*str))
+  # parms, but then all existing users would have to include **kw to accept
-#  	envcallback(ctx.getpriv().envfrom,str))
+  # arbitrary keywords without crashing.  We do provide envcallback and
-  milter.set_envrcpt_callback(lambda ctx,*str:
+  # dictfromlist to make parsing the ESMTP args convenient.
-  	ctx.getpriv().envrcpt(*str))
+  milter.set_envfrom_callback(lambda ctx,*str: ctx.getpriv().envfrom(*str))
-#  	envcallback(ctx.getpriv().envrcpt,str))
+  milter.set_envrcpt_callback(lambda ctx,*str: ctx.getpriv().envrcpt(*str))
-  milter.set_header_callback(lambda ctx,fld,val:
+  milter.set_header_callback(lambda ctx,fld,val: ctx.getpriv().header(fld,val))
-  	ctx.getpriv().header(fld,val))
   milter.set_eoh_callback(lambda ctx: ctx.getpriv().eoh())
   milter.set_body_callback(lambda ctx,chunk: ctx.getpriv().body(chunk))
   milter.set_eom_callback(lambda ctx: ctx.getpriv().eom())
@@ -204,6 +217,6 @@ def runmilter(name,socketname,timeout = 0):
     raise milter.error("out of thread resources")
 
 __all__ = globals().copy()
-for priv in ('os','milter','thread','factory','_seq','_seq_lock'):
+for priv in ('os','milter','thread','factory','_seq','_seq_lock','__version__'):
   del __all__[priv]
 __all__ = __all__.keys()

Milter/dsn.py

@@ -9,6 +9,7 @@ import smtplib
 import spf
 import socket
 from email.Message import Message
+import Milter
 
 nospf_msg = """Subject: Critical mail server configuration error
 
@@ -95,12 +96,15 @@ Received-SPF: %(spf_result)s
 """
 
 def send_dsn(mailfrom,receiver,msg=None):
-  "Send DSN.  If msg is None, do callback verification."
+  """Send DSN.  If msg is None, do callback verification.
+     Mailfrom is original sender we are sending DSN or CBV to.
+     Receiver is the MTA sending the DSN.
+     Return None for success or (code,msg) for failure."""
   user,domain = mailfrom.split('@')
   q = spf.query(None,None,None)
   mxlist = q.dns(domain,'MX')
   if not mxlist:
-    mxlist = (0,domain),
+    mxlist = (0,domain),	# fallback to A record when no MX
   else:
     mxlist.sort()
   smtp = smtplib.SMTP()
@@ -109,10 +113,13 @@ def send_dsn(mailfrom,receiver,msg=None):
       smtp.connect(host)
       code,resp = smtp.helo(receiver)
       # some wiley spammers have MX records that resolve to 127.0.0.1
-      if resp.split()[0] == receiver:
+      a = resp.split()
-        return (553,'Fraudulent MX for %s' % domain)
+      if not a:
+        return (553,'MX for %s has no hostname in banner: %s' % (domain,host))
+      if a[0] == receiver:
+        return (553,'Fraudulent MX for %s: %s' % (domain,host))
       if not (200 <= code <= 299):
-	raise SMTPHeloError(code, resp)
+	raise smtplib.SMTPHeloError(code, resp)
       if msg:
         try:
 	  smtp.sendmail('<>',mailfrom,msg)
@@ -122,7 +129,7 @@ def send_dsn(mailfrom,receiver,msg=None):
       else:	# CBV
 	code,resp = smtp.docmd('MAIL FROM: <>')
 	if code != 250:
-	  raise SMTPSenderRefused(code, resp, '<>')
+	  raise smtplib.SMTPSenderRefused(code, resp, '<>')
 	code,resp = smtp.rcpt(mailfrom)
 	if code not in (250,251):
 	  return (code,resp)		# permanent error
@@ -131,9 +138,9 @@ def send_dsn(mailfrom,receiver,msg=None):
     except smtplib.SMTPRecipientsRefused,x:
       return x.recipients[mailfrom]	# permanent error
     except smtplib.SMTPSenderRefused,x:
-      return x		# does not accept DSN
+      return x.args[:2]			# does not accept DSN
     except smtplib.SMTPDataError,x:
-      return x				# permanent error
+      return x.args			# permanent error
     except smtplib.SMTPException:
       pass		# any other error, try next MX
     except socket.error:
@@ -148,13 +155,13 @@ def create_msg(q,rcptlist,origmsg=None,template=None):
   connectip = q.i
   receiver = q.r
   sender_domain = q.o
+  result = q.result
+  perm_error = q.perm_error
   rcpt = '\n\t'.join(rcptlist)
   try: subject = origmsg['Subject']
   except: subject = '(none)'
   try:
     spf_result = origmsg['Received-SPF']
-    if not spf_result.startswith('softfail'):
-      spf_result = None
   except: spf_result = None
 
   msg = Message()
@@ -162,11 +169,14 @@ def create_msg(q,rcptlist,origmsg=None,template=None):
   msg.add_header('To',sender)
   msg.add_header('From','postmaster@%s'%receiver)
   msg.add_header('Auto-Submitted','auto-generated (configuration error)')
+  msg.add_header('X-Mailer','PyMilter-'+Milter.__version__)
   msg.set_type('text/plain')
 
   if not template:
-    if spf_result: template = softfail_msg
+    if spf_result and spf_result.startswith('softfail'):
-    else: template = nospf_msg
+      template = softfail_msg
+    else:
+      template = nospf_msg
   hdrs,body = template.split('\n',1)
   for ln in hdrs.splitlines():
     name,val = ln.split(':',1)

NEWS

@@ -1,5 +1,27 @@
 Here is a history of user visible changes to Python milter.
 
+0.8.4	Auto-whitelist recipients of outgoing email.
+0.8.3   Keep screened honeypot mail, but optionally discard honeypot only mail.
+	spf_accept_fail option for braindead SPF senders 
+	  (treats fail like softfail)
+	Option to set SPF policy via sendmail access map.
+	Option to supply Sender header from MAIL FROM when missing.
+	Consider SMTP AUTH connections internal.
+	Send DSN for SPF errors corrected by extended processing.
+	Send DSN before SCREENED mail is quarantined
+	Use logging package to keep log lines atomic.
+0.8.2	Strict processing limits per SPF RFC
+	Fixed several parsing bugs under RFC 
+	Support official IANA SPF record (type99)
+	Honeypot support (requires pydspam-1.1.9)
+	Extended SPF processing results beyond strict RFC limits
+	Support original SES for bounce protection (requires pysrs-0.30.10)
+	Callback exception processing option in milter module
+	Handle corrupt ZIP attachments
+0.8.1	Fix zip in zip loop in mime.py
+	Fix HeaderParseError in bms.py header callback
+	Check internal_domains for outgoing mail
+	Fix inconsistent results from send_dsn
 0.8.0	Move Milter module to subpackage.
 	DSN support for Three strikes rule and SPF SOFTFAIL
 	Move /*mime*/ and dynip to Milter subpackage

TODO

@@ -1,54 +1,43 @@
+Send DSN for permerror before processing extended result.  An additional
+DSN may be sent based on extended result.
+
+Rescind whitelist for banned extensions, in case sender is infected.
+
+Train honeypot on error only.
+
+Find rfc2822 policy for MFROM quoting.
+
+Support explicit errors for SPF policy in access file:
+SPF-Neutral:aol.com	ERROR:"550 AOL mail must get SPF PASS"
+
 Defer TEMPERROR in SPF evaluation - give precedence to security
 (only defer for PASS mechanisms).
 
 Option to add Received-SPF header, but never reject on SPF.
+  I think the above will handle this.
 
 Create null config that does nothing - except maybe add Received-SPF
 headers.  Many admins would like to turn features on one at a time.
 
-Checking in mime.py;
-/bms/cvs/milter/mime.py,v  <--  mime.py
-new revision: 1.56; previous revision: 1.55
-done
-Checking in spf.py;
-/bms/cvs/milter/spf.py,v  <--  spf.py
-new revision: 1.18; previous revision: 1.17
-done
-Checking in testmime.py;
-/bms/cvs/milter/testmime.py,v  <--  testmime.py
-new revision: 1.19; previous revision: 1.18
-
-Auto whitelist based on outgoing email - perhaps with magic subject
-or recipient prefix.
-
 Can't output messages with malformed rfc822 attachments.
 
-Example malformed SPF:
-onvunvuvvx.usafisnews.org text "v=spf1 mx ptr ip4:207.44.199.970 -all"
-
 Move milter,Milter,mime,spf modules to pymilter
 milter package will have bms.py application
 
-Support SMTP AUTH and disable SPF checks when connection is authorized.
 Web admin interface
-Check valid domains allowed by internal senders to detect PCs infected
-with spam trojans.
-Do CBV (callback verification) for mail with no published SPF record.
 message log for automated stats and blacklisting
-Skip dspam when SPF pass?
+Skip dspam when SPF pass? NO
 Report 551 with rcpt on SPF fail?
 check spam keywords with character classes, e.g.
 	{a}=[a@ãä], {i}=[i1í], {e}=[eë], {o}=[o0ö]
 
 Implement RRS - a backdoor for non-SRS forwarders.  User lists non-SRS 
 forwarder accounts, and a util provides a special local alias for the
-user to give to the forwarder.  Alias only works for mail from that
+user to give to the forwarder.  (Or user just adds arbitrary alias
+unique to that forwarder to a database.)  Alias only works for mail from that
 forwarder.  Milter gets forwarder domain from alias and uses it to
 SPF check forwarder.
 
-Another special dspam user, 'honeypot', can be listed in innoculations.
-All email to those addresses is treated as known spam.
-
 Framework for modular Python milter components within a single VM.
 Python milters can be already be composed through sendmail by running each in
 a separate process.  However, a significant amount of memory is wasted
@@ -57,8 +46,7 @@ is cumbersome (e.g., adding mail headers, writing external files).
 
 Backup copies for outgoing/incoming mail.
 
-Allow multiple wiretap groups, each with its own destination.  Perhaps 
+Copy incoming wiretap mail, even though sendmail alias works perfectly
-also copy incoming wiretap mail, even though sendmail alias works perfectly
 for the purpose, to avoid having to change two configs for a wiretap.
 
 Provide a way to reload milter.cfg without stopping/restarting milter.
@@ -72,10 +60,6 @@ Keep separate ismodified flag for headers and body.  This is important
 when rejecting outgoing mail with viruses removed (so as not to
 embarrass yourself), and also removing Received headers with hidepath.
 
-Wrap smfi_setbacklog(int) - but it is only available in sendmail >= 8.12.3,
-  so how can we detect whether to wrap it?
-
 Need a test module to feed sample messages to a milter though a live 
 sendmail and SMTP.  The mockup currently used is probably not very accurate,
 and doesn't test the threading code.
-

bms.py

@@ -1,6 +1,88 @@
 #!/usr/bin/env python
 # A simple milter that has grown quite a bit.
 # $Log$
+# Revision 1.34  2005/10/19 21:07:49  customdesigned
+# access.db stores keys in lower case
+#
+# Revision 1.33  2005/10/19 19:37:50  customdesigned
+# Train screener on whitelisted messages.
+#
+# Revision 1.32  2005/10/14 16:17:31  customdesigned
+# Auto whitelist refinements.
+#
+# Revision 1.31  2005/10/14 01:14:08  customdesigned
+# Auto whitelist feature.
+#
+# Revision 1.30  2005/10/12 16:36:30  customdesigned
+# Release 0.8.3
+#
+# Revision 1.29  2005/10/11 22:50:07  customdesigned
+# Always check HELO except for SPF pass, temperror.
+#
+# Revision 1.28  2005/10/10 23:50:20  customdesigned
+# Use logging module to make logging threadsafe (avoid splitting log lines)
+#
+# Revision 1.27  2005/10/10 20:15:33  customdesigned
+# Configure SPF policy via sendmail access file.
+#
+# Revision 1.26  2005/10/07 03:23:40  customdesigned
+# Banned users option.  Experimental feature to supply Sender when
+# missing and MFROM domain doesn't match From.  Log cipher bits for
+# SMTP AUTH.  Sketch access file feature.
+#
+# Revision 1.25  2005/09/08 03:55:08  customdesigned
+# Handle perverse MFROM quoting.
+#
+# Revision 1.24  2005/08/18 03:36:54  customdesigned
+# Don't innoculate with SCREENED mail.
+#
+# Revision 1.23  2005/08/17 19:35:27  customdesigned
+# Send DSN before adding message to quarantine.
+#
+# Revision 1.22  2005/08/11 22:17:58  customdesigned
+# Consider SMTP AUTH connections internal.
+#
+# Revision 1.21  2005/08/04 21:21:31  customdesigned
+# Treat fail like softfail for selected (braindead) domains.
+# Treat mail according to extended processing results, but
+# report any PermError that would officially result via DSN.
+#
+# Revision 1.20  2005/08/02 18:04:35  customdesigned
+# Keep screened honeypot mail, but optionally discard honeypot only mail.
+#
+# Revision 1.19  2005/07/20 03:30:04  customdesigned
+# Check pydspam version for honeypot, include latest pyspf changes.
+#
+# Revision 1.18  2005/07/17 01:25:44  customdesigned
+# Log as well as use extended result for best guess.
+#
+# Revision 1.17  2005/07/15 20:25:36  customdesigned
+# Use extended results processing for best_guess.
+#
+# Revision 1.16  2005/07/14 03:23:33  customdesigned
+# Make SES package optional.  Initial honeypot support.
+#
+# Revision 1.15  2005/07/06 04:05:40  customdesigned
+# Initial SES integration.
+#
+# Revision 1.14  2005/07/02 23:27:31  customdesigned
+# Don't match hostnames for internal connects.
+#
+# Revision 1.13  2005/07/01 16:30:24  customdesigned
+# Always log trusted Received and Received-SPF headers.
+#
+# Revision 1.12  2005/06/20 22:35:35  customdesigned
+# Setreply for rejectvirus.
+#
+# Revision 1.11  2005/06/17 02:07:20  customdesigned
+# Release 0.8.1
+#
+# Revision 1.10  2005/06/16 18:35:51  customdesigned
+# Ignore HeaderParseError decoding header
+#
+# Revision 1.9  2005/06/14 21:55:29  customdesigned
+# Check internal_domains for outgoing mail.
+#
 # Revision 1.8  2005/06/06 18:24:59  customdesigned
 # Properly log exceptions from pydspam
 #
@@ -216,6 +298,7 @@ import traceback
 import ConfigParser
 import time
 import re
+import anydbm
 import Milter.dsn as dsn
 from Milter.dynip import is_dynip as dynip
 
@@ -227,14 +310,16 @@ try:
   import SRS
   srsre = re.compile(r'^SRS[01][+-=]',re.IGNORECASE)
 except: SRS = None
+try:
+  import SES
+except: SES = None
 
 # Import spf if available
 try: import spf
 except: spf = None
 
 ip4re = re.compile(r'^[1-9][0-9]*\.[1-9][0-9]*\.[1-9][0-9]*\.[1-9][0-9]*$')
-#import syslog
+import logging
-#syslog.openlog('milter')
 
 # Thanks to Chris Liechti for config parsing suggestions
 
@@ -259,6 +344,7 @@ scan_rfc822 = True
 internal_connect = ()
 trusted_relay = ()
 internal_domains = ()
+banned_users = ()
 hello_blacklist = ()
 smart_alias = {}
 dspam_dict = None
@@ -266,33 +352,31 @@ dspam_users = {}
 dspam_userdir = None
 dspam_exempt = {}
 dspam_whitelist = {}
+whitelist_senders = {}
 dspam_screener = ()
 dspam_internal = True	# True if internal mail should be dspammed
 dspam_reject = ()
 dspam_sizelimit = 180000
 srs = None
+ses = None
 srs_reject_spoofed = False
-srs_fwdomain = None
+srs_domain = None
 spf_reject_neutral = ()
 spf_accept_softfail = ()
+spf_accept_fail = ()
 spf_best_guess = False
 spf_reject_noptr = False
-multiple_bounce_recipients = True
+supply_sender = False
-time_format = '%Y%b%d %H:%M:%S %Z'
+access_file = None
 timeout = 600
-cbv_cache = {}
+
-try:
+logging.basicConfig(
-  too_old = time.time() - 30*24*60*60	# 30 days
+	stream=sys.stdout,
-  for ln in open('send_dsn.log'):
+	level=logging.INFO,
-    try:
+	format='%(asctime)s %(message)s',
-      rcpt,ts = ln.strip().split(None,1)
+	datefmt='%Y%b%d %H:%M:%S'
-      l = time.strptime(ts,time_format)
+)
-      t = time.mktime(l)
+milter_log = logging.getLogger('milter')
-      if t > too_old:
-	cbv_cache[rcpt] = None
-    except:
-      cbv_cache[ln.strip()] = None
-except IOError: pass
 
 class MilterConfigParser(ConfigParser.ConfigParser):
 
@@ -354,6 +438,7 @@ def read_config(list):
     'hashlength': '8',
     'reject_spoofed': 'no',
     'reject_noptr': 'no',
+    'supply_sender': 'no',
     'best_guess': 'no',
     'dspam_internal': 'yes'
   })
@@ -407,7 +492,7 @@ def read_config(list):
   for sa in cp.getlist('wiretap','smart_alias'):
     sm = cp.getlist('wiretap',sa)
     if len(sm) < 2:
-      print 'malformed smart alias:',sa
+      milter_log.warning('malformed smart alias: %s',sa)
       continue
     if len(sm) == 2: sm.append(sa)
     key = (sm[0],sm[1])
@@ -416,6 +501,8 @@ def read_config(list):
   # dspam section
   global dspam_dict, dspam_users, dspam_userdir, dspam_exempt, dspam_internal
   global dspam_screener,dspam_whitelist,dspam_reject,dspam_sizelimit
+  global whitelist_senders
+  whitelist_senders = cp.getaddrset('dspam','whitelist_senders')
   dspam_dict = cp.getdefault('dspam','dspam_dict')
   dspam_exempt = cp.getaddrset('dspam','dspam_exempt')
   dspam_whitelist = cp.getaddrset('dspam','dspam_whitelist')
@@ -429,18 +516,21 @@ def read_config(list):
 
   # spf section
   global spf_reject_neutral,spf_best_guess,SRS,spf_reject_noptr
-  global spf_accept_softfail
+  global spf_accept_softfail,spf_accept_fail,supply_sender,access_file
   if spf:
     spf.DELEGATE = cp.getdefault('spf','delegate')
     spf_reject_neutral = cp.getlist('spf','reject_neutral')
     spf_accept_softfail = cp.getlist('spf','accept_softfail')
+    spf_accept_fail = cp.getlist('spf','accept_fail')
     spf_best_guess = cp.getboolean('spf','best_guess')
     spf_reject_noptr = cp.getboolean('spf','reject_noptr')
+    supply_sender = cp.getboolean('spf','supply_sender')
+    access_file = cp.getdefault('spf','access_file')
   srs_config = cp.getdefault('srs','config')
   if srs_config: cp.read([srs_config])
   srs_secret = cp.getdefault('srs','secret')
   if SRS and srs_secret:
-    global srs,srs_reject_spoofed,srs_fwdomain
+    global ses,srs,srs_reject_spoofed,srs_domain,banned_users
     database = cp.getdefault('srs','database')
     srs_reject_spoofed = cp.getboolean('srs','reject_spoofed')
     maxage = cp.getint('srs','maxage')
@@ -453,16 +543,40 @@ def read_config(list):
     else:
       srs = SRS.Guarded.Guarded(secret=srs_secret,
         maxage=maxage,hashlength=hashlength,separator=separator)
-    srs_fwdomain = cp.getdefault('srs','fwdomain')
+    if SES:
+      ses = SES.new(secret=srs_secret,expiration=maxage)
+      srs_domain = cp.getlist('srs','ses')
+    else:
+      srs_domain = []
+    srs_domain.append(cp.getdefault('srs','fwdomain'))
+    banned_users = cp.getlist('srs','banned_users')
+    #print srs_domain
 
 def parse_addr(t):
+  """Split email into user,domain.
+
+  >>> parse_addr('user@example.com')
+  ['user', 'example.com']
+  >>> parse_addr('"user@example.com"')
+  ['user@example.com']
+  >>> parse_addr('"user@bar"@example.com')
+  ['user@bar', 'example.com']
+  >>> parse_addr('foo')
+  ['foo']
+  """
   if t.startswith('<') and t.endswith('>'): t = t[1:-1]
+  if t.startswith('"'):
+    if t.endswith('"'): return [t[1:-1]]
+    pos = t.find('"@')
+    if pos > 0: return [t[1:pos],t[pos+2:]]
   return t.split('@')
 
 def parse_header(val):
+  """Decode headers gratuitously encoded to hide the content.
+  """
+  try:
     h = decode_header(val)
     if not len(h) or (not h[0][1] and len(h) == 1): return val
-  try:
     u = []
     for s,enc in h:
       if enc:
@@ -479,17 +593,128 @@ def parse_header(val):
       except UnicodeError: continue
   except UnicodeDecodeError: pass
   except LookupError: pass
+  except email.Errors.HeaderParseError: pass
   return val
 
+class SPFPolicy(object):
+  "Get SPF policy by result, defaulting to classic policy from pymilter.cfg"
+  def __init__(self,domain):
+    self.domain = domain.lower()
+    if access_file:
+      try: acf = anydbm.open(access_file,'r')
+      except: acf = None
+    else: acf = None
+    self.acf = acf
+
+  def getPolicy(self,pfx):
+    acf = self.acf
+    if not acf: return None
+    try:
+      return acf[pfx + self.domain]
+    except KeyError:
+      try:
+	return acf[pfx]
+      except KeyError:
+        return None
+
+  def getFailPolicy(self):
+    policy = self.getPolicy('spf-fail:')
+    if not policy:
+      if self.domain in spf_accept_fail:
+        policy = 'CBV'
+      else:
+	policy = 'REJECT'
+    return policy
+
+  def getNonePolicy(self):
+    policy = self.getPolicy('spf-none:')
+    if not policy:
+      if spf_reject_noptr:
+	policy = 'REJECT'
+      else:
+        policy = 'CBV'
+    return policy
+
+  def getSoftfailPolicy(self):
+    policy = self.getPolicy('spf-softfail:')
+    if not policy:
+      if self.domain in spf_accept_softfail:
+        policy = 'OK'
+      elif self.domain in spf_reject_neutral:
+        policy = 'REJECT'
+      else:
+        policy = 'CBV'
+    return policy
+
+  def getNeutralPolicy(self):
+    policy = self.getPolicy('spf-neutral:')
+    if not policy:
+      if self.domain in spf_reject_neutral:
+        policy = 'REJECT'
+      policy = 'OK'
+    return policy
+
+  def getPermErrorPolicy(self):
+    policy = self.getPolicy('spf-permerror:')
+    if not policy:
+      policy = 'REJECT'
+    return policy
+
+  def getPassPolicy(self):
+    policy = self.getPolicy('spf-pass:')
+    if not policy:
+      policy = 'OK'
+    return policy
+
+class AddrCache(object):
+  time_format = '%Y%b%d %H:%M:%S %Z'
+
+  def load(self,fname,age=7):
+    self.fname = fname
+    cache = {}
+    self.cache = cache
+    try:
+      too_old = time.time() - age*24*60*60	# max age in days
+      for ln in open(self.fname):
+	try:
+	  rcpt,ts = ln.strip().split(None,1)
+	  l = time.strptime(ts,AddrCache.time_format)
+	  t = time.mktime(l)
+	  if t > too_old:
+	    cache[rcpt.lower()] = None
+	except:
+	  cache[ln.strip().lower()] = None
+    except IOError: pass
+
+  def has_key(self,sender):
+    return self.cache.has_key(sender.lower())
+
+  def __getitem__(self,sender):
+    return self.cache[sender.lower()]
+
+  def __setitem__(self,sender,res):
+    lsender = sender.lower()
+    cached = lsender in self.cache
+    self.cache[lsender] = res
+    if not cached and not res:
+      s = time.strftime(AddrCache.time_format,time.localtime())
+      print >>open(self.fname,'a'),sender,s # log who we sent DSNs to
+
+  def __len__(self):
+    return len(self.cache)
+
+cbv_cache = AddrCache()
+cbv_cache.load('send_dsn.log',age=7)
+auto_whitelist = AddrCache()
+auto_whitelist.load('auto_whitelist.log',age=30)
+
 class bmsMilter(Milter.Milter):
   """Milter to replace attachments poisonous to Windows with a WARNING message,
      check SPF, and other anti-forgery features, and implement wiretapping
      and smart alias redirection."""
 
   def log(self,*msg):
-    print "%s [%d]" % (time.strftime('%Y%b%d %H:%M:%S'),self.id),
+    milter_log.info('[%d] %s',self.id,' '.join([str(m) for m in msg]))
-    for i in msg: print i,
-    print
 
   def __init__(self):
     self.tempname = None
@@ -536,10 +761,6 @@ class bmsMilter(Milter.Milter):
     else: ipaddr = ''
     self.connectip = ipaddr
     self.missing_ptr = dynip(hostname,self.connectip)
-    for pat in internal_connect:
-      if fnmatchcase(hostname,pat):
-	self.internal_connection = True
-	break
     if self.internal_connection:
       connecttype = 'INTERNAL'
     else:
@@ -603,6 +824,7 @@ class bmsMilter(Milter.Milter):
     self.hidepath = False
     self.discard = False
     self.dspam = True
+    self.whitelist = False
     self.reject_spam = True
     self.data_allowed = True
     self.trust_received = self.trusted_relay
@@ -612,8 +834,42 @@ class bmsMilter(Milter.Milter):
     self.new_headers = []
     self.recipients = []
     self.cbv_needed = None
-    t = parse_addr(f.lower())
+    self.whitelist_sender = False
+    t = parse_addr(f)
+    if len(t) == 2: t[1] = t[1].lower()
     self.canon_from = '@'.join(t)
+    # Some braindead MTAs can't be relied upon to properly flag DSNs.
+    # This heuristic tries to recognize such.
+    self.is_bounce = (f == '<>' or t[0].lower() in banned_users
+        #and t[1] == self.hello_name
+    )
+
+    # Check SMTP AUTH, also available:
+    #   auth_authen  authenticated user
+    #   auth_author  (ESMTP AUTH= param)
+    #   auth_ssf     (connection security, 0 = unencrypted)
+    #   auth_type    (authentication method, CRAM-MD5, DIGEST-MD5, PLAIN, etc)
+    # cipher_bits  SSL encryption strength
+    # cert_subject SSL cert subject
+    # verify       SSL cert verified
+
+    self.user = self.getsymval('{auth_authen}')
+    if self.user:
+      # Very simple SMTP AUTH policy by defaul:
+      #   any successful authentication is considered INTERNAL
+      # FIXME: configure allowed MAIL FROM by user
+      self.internal_connection = True
+      self.log(
+        "SMTP AUTH:",self.user, self.getsymval('{auth_type}'),
+        "sslbits =",self.getsymval('{cipher_bits}'),
+        "ssf =",self.getsymval('{auth_ssf}'), "INTERNAL"
+      )
+      if self.getsymval('{verify}'):
+	self.log("SSL AUTH:",
+	  self.getsymval('{cert_subject}'),
+	  "verify =",self.getsymval('{verify}')
+	)
+
     self.fp.write('From %s %s\n' % (self.canon_from,time.ctime()))
     if len(t) == 2:
       user,domain = t
@@ -623,17 +879,22 @@ class bmsMilter(Milter.Milter):
 	    self.log("REJECT: spam from self",pat)
 	    self.setreply('550','5.7.1','I hate talking to myself.')
 	    return Milter.REJECT
-      elif internal_domains:
+      else:
+        if internal_domains:
 	  for pat in internal_domains:
 	    if fnmatchcase(domain,pat): break
 	  else:
-	  self.log("REJECT: zombie PC at ",self.connectip," sending MAIL FROM ",
+	    self.log("REJECT: zombie PC at ",self.connectip,
-	  	self.canon_from)
+	    	" sending MAIL FROM ",self.canon_from)
-	  self.setreply('550','5.7.1','Get rid of your virus!',
+	    self.setreply('550','5.7.1',
 	    'Your PC is using an unauthorized MAIL FROM.',
 	    'It is either badly misconfigured or controlled by organized crime.'
 	    )
 	    return Milter.REJECT
+	wl_users = whitelist_senders.get(domain,())
+	if user in wl_users or '' in wl_users:
+	  self.whitelist_sender = True
+	  
       self.rejectvirus = domain in reject_virus_from
       if user in wiretap_users.get(domain,()):
         self.add_recipient(wiretap_dest)
@@ -652,20 +913,25 @@ class bmsMilter(Milter.Milter):
     if not (self.internal_connection or self.trusted_relay)	\
     	and self.connectip and spf:
       return self.check_spf()
-    if self.internal_connection:
+    self.spf = None
-      pass
     return Milter.CONTINUE
 
   def check_spf(self):
-    t = parse_addr(self.mailfrom)
-    if len(t) == 2: t[1] = t[1].lower()
     receiver = self.receiver
-    q = spf.query(self.connectip,'@'.join(t),self.hello_name,receiver=receiver)
+    q = spf.query(self.connectip,self.canon_from,self.hello_name,
-    q.set_default_explanation('SPF fail: see http://spf.pobox.com/why.html')
+    	receiver=receiver,strict=False)
+    q.set_default_explanation(
+      'SPF fail: see http://openspf.com/why.html?sender=%s&ip=%s' % (q.s,q.i))
     res,code,txt = q.check()
-    if res in ('none', 'softfail'):
+    q.result = res
+    if res in ('unknown','permerror') and q.perm_error and q.perm_error.ext:
+      self.cbv_needed = q	# report SPF syntax error to sender
+      res,code,txt = q.perm_error.ext	# extended (lax processing) result
+      txt = 'EXT: ' + txt
+    p = SPFPolicy(q.o)
+    if res not in ('pass','error','temperror'):
       if self.mailfrom != '<>':
-	# check hello name via spf
+	# check hello name via spf unless spf pass
 	h = spf.query(self.connectip,'',self.hello_name,receiver=receiver)
 	hres,hcode,htxt = h.check()
 	if hres in ('deny','fail','neutral','softfail'):
@@ -681,6 +947,7 @@ class bmsMilter(Milter.Milter):
 	  and not dynip(self.hello_name,self.connectip):
 	  hres,hcode,htxt = h.best_guess()
       else: hres = res
+      ores = res
       if spf_best_guess and res == 'none':
 	#self.log('SPF: no record published, guessing')
 	q.set_default_explanation(
@@ -692,29 +959,43 @@ class bmsMilter(Milter.Milter):
 	else:
 	  res,code,txt = q.best_guess()
 	receiver += ': guessing'
-      if self.missing_ptr and res in ('neutral', 'none') and hres != 'pass':
+        if q.perm_error:	# FIXME: should never happen?
-	if spf_reject_noptr:
+          res,code,txt = q.perm_error.ext	# extended result
+	  txt = 'EXT: ' + txt
+      if self.missing_ptr and ores == 'none' and res != 'pass' \
+      		and hres != 'pass':
+	policy = p.getNonePolicy()
+	if policy == 'CBV':
+	  if self.mailfrom != '<>':
+	    q.result = ores
+	    self.cbv_needed = q	# accept, but inform sender via DSN
+	elif policy != 'OK':
 	  self.log('REJECT: no PTR, HELO or SPF')
 	  self.setreply('550','5.7.1',
-    'You must have a reverse lookup or publish SPF: http://spf.pobox.com',
+    "You must have a reverse lookup or publish SPF: http://spf.pobox.com",
-    'Contact your mail administrator IMMEDIATELY!  Your mail server is',
+    "Contact your mail administrator IMMEDIATELY!  Your mail server is",
-    'severely misconfigured.  It has no PTR record (dynamic PTR records',
+    "severely misconfigured.  It has no PTR record (dynamic PTR records",
     "that contain your IP don't count), an invalid HELO, and no SPF record."
 	  )
 	  return Milter.REJECT
-        if self.mailfrom != '<>':
-	  q.result = res
-	  self.cbv_needed = q
     if res in ('deny', 'fail'):
+      policy = p.getFailPolicy()
+      if hres == 'pass' and policy == 'CBV':
+	if self.mailfrom != '<>':
+	  self.cbv_needed = q
+      elif policy != 'OK':
 	self.log('REJECT: SPF %s %i %s' % (res,code,txt))
 	self.setreply(str(code),'5.7.1',txt)
 	# A proper SPF fail error message would read:
 	# forger.biz [1.2.3.4] is not allowed to send mail with the domain
 	# "forged.org" in the sender address.  Contact <postmaster@forged.org>.
 	return Milter.REJECT
-    if res == 'softfail' and not q.o in spf_accept_softfail:
+    if res == 'softfail':
-      if self.missing_ptr and hres != 'pass':
+      policy = p.getSoftfailPolicy()
-        if spf_reject_noptr or q.o in spf_reject_neutral:
+      if policy == 'CBV' and hres == 'pass':
+	if self.mailfrom != '<>':
+	  self.cbv_needed = q
+      elif policy != 'OK':
 	self.log('REJECT: SPF %s %i %s' % (res,code,txt))
 	self.setreply('550','5.7.1',
 	  'SPF softfail: If you get this Delivery Status Notice, your email',
@@ -724,10 +1005,12 @@ class bmsMilter(Milter.Milter):
 	  'notify your administrator of the problem immediately.'
 	)
 	return Milter.REJECT
-      if self.mailfrom != '<>':
-        q.result = res
-	self.cbv_needed = q
     if res == 'neutral' and q.o in spf_reject_neutral:
+      policy = p.getNeutralPolicy()
+      if policy == 'CBV' and hres == 'pass':
+	if self.mailfrom != '<>':
+	  self.cbv_needed = q
+      elif policy != 'OK':
 	self.log('REJECT: SPF neutral for',q.s)
 	self.setreply('550','5.7.1',
 	  'mail from %s must pass SPF: http://spf.pobox.com/why.html' % q.o,
@@ -738,16 +1021,28 @@ class bmsMilter(Milter.Milter):
 	  'servers for %s should accomplish this.' % q.o
 	)
 	return Milter.REJECT
-    if res == 'error':
+    if res in ('unknown','permerror'):
-      if code >= 500:
+      policy = p.getPermErrorPolicy()
+      if policy == 'CBV' and hres == 'pass':
+	if self.mailfrom != '<>':
+	  self.cbv_needed = q
+      elif policy != 'OK':
 	self.log('REJECT: SPF %s %i %s' % (res,code,txt))
 	# latest SPF draft recommends 5.5.2 instead of 5.7.1
-	self.setreply(str(code),'5.5.2',txt)
+	self.setreply(str(code),'5.5.2',txt,
+	  'There is a fatal syntax error in the SPF record for %s' % q.o,
+	  'We cannot accept mail from %s until this is corrected.' % q.o
+	)
 	return Milter.REJECT
+    if res in ('error','temperror'):
       self.log('TEMPFAIL: SPF %s %i %s' % (res,code,txt))
       self.setreply(str(code),'4.3.0',txt)
       return Milter.TEMPFAIL
     self.add_header('Received-SPF',q.get_header(res,receiver))
+    self.spf = q
+    if res == 'pass' and auto_whitelist.has_key(self.canon_from):
+      self.whitelist = True
+      self.log("WHITELIST",self.canon_from)
     return Milter.CONTINUE
 
   # hide_path causes a copy of the message to be saved - until we
@@ -759,16 +1054,20 @@ class bmsMilter(Milter.Milter):
       self.log('DISCARD: RCPT TO:',to,str)
       return Milter.DISCARD
     self.log("rcpt to",to,str)
-    t = parse_addr(to.lower())
+    t = parse_addr(to)
     if len(t) == 2:
+      t[1] = t[1].lower()
       user,domain = t
-      if self.mailfrom == '<>' or self.canon_from.startswith('postmaster@') \
+      if self.is_bounce and srs and domain in srs_domain:
-      	or self.canon_from.startswith('mailer-daemon@'):
-        if self.recipients and not multiple_bounce_recipients:
-	  self.data_allowed = False
-        if srs and domain == srs_fwdomain:
 	oldaddr = '@'.join(parse_addr(to))
 	try:
+	  if ses:
+	    newaddr = ses.verify(oldaddr)
+	  else:
+	    newaddr = oldaddr,
+	  if len(newaddr) > 1:
+	    self.log("ses rcpt:",newaddr[0])
+	  else:
 	    newaddr = srs.reverse(oldaddr)
 	    # Currently, a sendmail map reverses SRS.  We just log it here.
 	    self.log("srs rcpt:",newaddr)
@@ -778,9 +1077,15 @@ class bmsMilter(Milter.Milter):
 	      self.log("REJECT: srs spoofed:",oldaddr)
 	      self.setreply('550','5.7.1','Invalid SRS signature')
 	      return Milter.REJECT
+	    if oldaddr.startswith('SES='):
+	      self.log("REJECT: ses spoofed:",oldaddr)
+	      self.setreply('550','5.7.1','Invalid SES signature')
+	      return Milter.REJECT
 	    self.data_allowed = not srs_reject_spoofed
+
       # non DSN mail to SRS address will bounce due to invalid local part
-      self.recipients.append('@'.join(t))
+      canon_to = '@'.join(t)
+      self.recipients.append(canon_to)
       users = check_user.get(domain)
       if self.discard:
         self.del_recipient(to)
@@ -796,6 +1101,14 @@ class bmsMilter(Milter.Milter):
         self.hidepath = True
       if not domain in dspam_reject:
         self.reject_spam = False
+      if self.internal_connection and self.whitelist_sender:
+	if internal_domains:
+	  for pat in internal_domains:
+	    if fnmatchcase(domain,pat): break
+	  else:
+	    auto_whitelist[canon_to] = None
+	else:
+	  auto_whitelist[canon_to] = None
     self.smart_alias(to)
     #rcpt = self.getsymval("{rcpt_addr}")
     #self.log("rcpt-addr",rcpt);
@@ -853,23 +1166,22 @@ class bmsMilter(Milter.Milter):
 	or mailer.find('optin') >= 0:
         self.log('REJECT: %s: %s' % (name,val))
         return Milter.REJECT
-    elif self.trust_received and lname == 'received':
-      self.trust_received = False
-      self.log('%s: %s' % (name,val.splitlines()[0]))
-    elif self.trust_spf and lname == 'received-spf':
-      self.trust_spf = False
-      self.log('%s: %s' % (name,val.splitlines()[0]))
     return Milter.CONTINUE
 
   def forged_bounce(self):
-    if len(self.recipients) > 1:
+    if self.mailfrom != '<>':
-      self.log('REJECT: Multiple bounce recipients')
+      self.log("REJECT: bogus DSN")
-      self.setreply('550','5.7.1','Multiple bounce recipients')
+      self.setreply('550','5.7.1',
+	"I do not accept mail from postmaster, mailer-daemon, or clamav.",
+	"All such mail has turned out to be Delivery Status Notifications",
+	"which failed to be marked as such.  Please send a real DSN if",
+	"you need to.  Use another MAIL FROM if you need to send me mail."
+      )
     else:
       self.log('REJECT: bounce with no SRS encoding')
       self.setreply('550','5.7.1',
 	"I did not send you that message. Please consider implementing SPF",
-      "(http://spf.pobox.com) to avoid bouncing mail to spoofed senders.",
+	"(http://openspf.com) to avoid bouncing mail to spoofed senders.",
 	"Thank you."
       )
     return Milter.REJECT
@@ -893,6 +1205,12 @@ class bmsMilter(Milter.Milter):
     # log selected headers
     if log_headers or lname in ('subject','x-mailer'):
       self.log('%s: %s' % (name,val))
+    elif self.trust_received and lname == 'received':
+      self.trust_received = False
+      self.log('%s: %s' % (name,val.splitlines()[0]))
+    elif self.trust_spf and lname == 'received-spf':
+      self.trust_spf = False
+      self.log('%s: %s' % (name,val.splitlines()[0]))
     if self.fp:
       try:
         val = val.encode('us-ascii')
@@ -908,8 +1226,28 @@ class bmsMilter(Milter.Milter):
     for name,val in self.new_headers:
       self.fp.write("%s: %s\n" % (name,val))	# add new headers to buffer
     self.fp.write("\n")				# terminate headers
+    # log when neither sender nor from domains matches mail from domain
+    if supply_sender and self.mailfrom != '<>' and not self.internal_connection:
+      mf_domain = self.canon_from.split('@')[-1]
       self.fp.seek(0)
+      msg = rfc822.Message(self.fp)
+      for rn,hf in msg.getaddrlist('from')+msg.getaddrlist('sender'):
+	t = parse_addr(hf)
+	if len(t) == 2 and t[1].lower() == mf_domain:
+	  break
+      else:
+	for f in msg.getallmatchingheaders('from'):
+	  self.log(f)
+	sender = msg.getallmatchingheaders('sender')
+	if sender:
+	  for f in sender:
+	    self.log(f)
+	else:
+	  self.log("NOTE: Supplying MFROM as Sender");
+	  self.add_header('Sender',self.mailfrom)
+      del msg
     # copy headers to a temp file for scanning the body
+    self.fp.seek(0)
     headers = self.fp.getvalue()
     self.fp.close()
     fd,fname = tempfile.mkstemp(".defang")
@@ -922,7 +1260,7 @@ class bmsMilter(Milter.Milter):
         dspam.DSF_CHAINED|dspam.DSF_CLASSIFY)
       try:
         ds.process(headers)
-        if ds.probability > 0.93 and self.dspam:
+        if ds.probability > 0.93 and self.dspam and not self.whitelist:
           self.log('REJECT: X-DSpam-HeaderScore: %f' % ds.probability)
 	  self.setreply('550','5.7.1','Your Message looks spammy')
 	  return Milter.REJECT
@@ -989,6 +1327,7 @@ class bmsMilter(Milter.Milter):
   #	   this will give a fast start to stats
 
   def check_spam(self):
+    "return True/False if self.fp, else return Milter.REJECT/TEMPFAIL/etc"
     if not dspam_userdir: return False
     ds = Dspam.DSpamDirectory(dspam_userdir)
     ds.log = self.log
@@ -1008,7 +1347,7 @@ class bmsMilter(Milter.Milter):
 		ds.add_spam(sender,txt)
 		txt = None
 		self.fp = None
-		return False
+		return Milter.DISCARD
 	    elif user == 'falsepositive' and self.internal_connection:
 	      sender = dspam_users.get(self.canon_from)
 	      if sender:
@@ -1023,6 +1362,34 @@ class bmsMilter(Milter.Milter):
 	      if len(txt) > dspam_sizelimit:
 		self.log("Large message:",len(txt))
 		return False
+	      if user == 'honeypot' and Dspam.VERSION >= '1.1.9':
+	        keep = False	# keep honeypot mail
+		self.fp = None
+	        if len(self.recipients) > 1:
+		  self.log("HONEYPOT:",rcpt,'SCREENED')
+		  if self.whitelist:
+		    # don't train when recipients includes honeypot
+		    return False
+		  if self.spf:
+		    # check that sender accepts quarantine DSN
+		    msg = mime.message_from_file(StringIO.StringIO(txt))
+		    rc = self.send_dsn(self.spf,msg,'quarantine.txt')
+		    del msg
+		    if rc != Milter.CONTINUE:
+		      return rc	
+		  ds.check_spam(user,txt,self.recipients,quarantine=True,
+		  	force_result=dspam.DSR_ISSPAM)
+		else:
+		  ds.check_spam(user,txt,self.recipients,quarantine=keep,
+		  	force_result=dspam.DSR_ISSPAM)
+		  self.log("HONEYPOT:",rcpt)
+		return Milter.DISCARD
+	      if self.whitelist:
+	        # Sender whitelisted: tag, but force as ham.  
+		# User can change if actually spam.
+	        txt = ds.check_spam(user,txt,self.recipients,
+			force_result=dspam.DSR_ISINNOCENT)
+	      else:
 		txt = ds.check_spam(user,txt,self.recipients)
 	      if not txt:
 	        # DISCARD if quarrantined for any recipient.  It
@@ -1030,12 +1397,12 @@ class bmsMilter(Milter.Milter):
 		# as a false positive.
 		self.log("DSPAM:",user,rcpt)
 		self.fp = None
-		return False
+		return Milter.DISCARD
 	      self.fp = StringIO.StringIO(txt)
 	      modified = True
 	  except Exception,x:
 	    self.log("check_spam:",x)
-	    traceback.print_exc()
+	    milter_log.error("check_spam: %s",x,exc_info=True)
     # screen if no recipients are dspam_users
     if not modified and dspam_screener and not self.internal_connection \
     	and self.dspam:
@@ -1046,14 +1413,34 @@ class bmsMilter(Milter.Milter):
 	return False
       screener = dspam_screener[self.id % len(dspam_screener)]
       if not ds.check_spam(screener,txt,self.recipients,
-      	classify=True,quarantine=not self.reject_spam):
+      	classify=True,quarantine=False):
-	self.fp = None
+	if self.whitelist:
+	  # messages is whitelisted but looked like spam, Train on Error
+	  self.log("TRAIN:",screener,'X-Dspam-Score: %f' % ds.probability)
+	  # user can't correct anyway if really spam, so discard tag
+	  ds.check_spam(screener,txt,self.recipients,
+		  force_result=dspam.DSR_ISINNOCENT)
+	  return False
 	if self.reject_spam:
 	  self.log("DSPAM:",screener,
 	  	'REJECT: X-DSpam-Score: %f' % ds.probability)
 	  self.setreply('550','5.7.1','Your Message looks spammy')
-	  return True
+	  self.fp = None
+	  return Milter.REJECT
 	self.log("DSPAM:",screener,"SCREENED")
+	if self.spf:
+	  # check that sender accepts quarantine DSN
+	  self.fp.seek(0)
+	  msg = mime.message_from_file(self.fp)
+	  rc = self.send_dsn(self.spf,msg,'quarantine.txt')
+	  if rc != Milter.CONTINUE:
+	    self.fp = None
+	    return rc
+	  del msg
+	if not ds.check_spam(screener,txt,self.recipients,classify=True):
+	  self.fp = None
+	  return Milter.DISCARD
+	# Message no longer looks spammy, deliver normally. We lied in the DSN.
     return modified
 
   def eom(self):
@@ -1064,8 +1451,7 @@ class bmsMilter(Milter.Milter):
       # analyze external mail for spam
       spam_checked = self.check_spam()	# tag or quarantine for spam
       if not self.fp:
-        if spam_checked: return Milter.REJECT
+        return spam_checked
-	return Milter.DISCARD	# message quarantined for all recipients
 
       # analyze all mail for dangerous attachments and scripts
       self.fp.seek(0)
@@ -1083,7 +1469,7 @@ class bmsMilter(Milter.Milter):
         if not exc_value.strerror:
 	  exc_value.strerror = exc_value.args[0]
 	if exc_value.strerror == 'Lock failed':
-	  self.log("LOCK: BUSY")	# log filename
+	  milter_log.warn("LOCK: BUSY")	# log filename
 	  self.setreply('450','4.2.0',
 		'Too busy discarding spam.  Please try again later.')
 	  return Milter.TEMPFAIL
@@ -1091,7 +1477,7 @@ class bmsMilter(Milter.Milter):
       os.rename(self.tempname,fname)
       self.tempname = None
       if exc_type == email.Errors.BoundaryError:
-	self.log("MALFORMED: %s" % fname)	# log filename
+	milter_log.warn("MALFORMED: %s",fname)	# log filename
         if self.internal_connection:
 	  # accept anyway for now
 	  return Milter.ACCEPT
@@ -1099,12 +1485,12 @@ class bmsMilter(Milter.Milter):
 		'Boundary error in your message, are you a spammer?')
         return Milter.REJECT
       if exc_type == email.Errors.HeaderParseError:
-	self.log("MALFORMED: %s" % fname)	# log filename
+	milter_log.warn("MALFORMED: %s",fname)	# log filename
 	self.setreply('554','5.7.7',
 		'Header parse error in your message, are you a spammer?')
         return Milter.REJECT
+      milter_log.error("FAIL: %s",fname)	# log filename
       # let default exception handler print traceback and return 451 code
-      self.log("FAIL: %s" % fname)	# log filename
       raise
     if rc == Milter.REJECT: return rc;
     if rc == Milter.DISCARD: return rc;
@@ -1132,39 +1518,17 @@ class bmsMilter(Milter.Milter):
 
     if self.cbv_needed:
       q = self.cbv_needed
-      sender = q.s
+      if q.result in ('softfail','fail','deny'):
-      cached = cbv_cache.has_key(sender)
+	template_name = 'softfail.txt'
-      if cached:
+      elif q.result in ('unknown','permerror'):
-	self.log('CBV:',sender,'(cached)')
+	template_name = 'permerror.txt'
-        res = cbv_cache[sender]
+      elif q.result == 'neutral':
+        template_name = 'neutral.txt'
       else:
-	self.log('CBV:',sender)
+	template_name = 'strike3.txt'
-	try:
+      rc = self.send_dsn(q,msg,template_name)
-	  if q.result == 'softfail':
-	    template = file('softfail.txt').read()
-	  else:
-	    template = file('strike3.txt').read()
-	except IOError: template = None
-	m = dsn.create_msg(q,self.recipients,msg,template)
-        m = m.as_string()
-        print >>open('last_dsn','w'),m
-	res = dsn.send_dsn(sender,self.receiver,m)
-      if res:
-        desc = "CBV: %d %s" % res[:2]
-        if 400 <= res[0] < 500:
-	  self.log('TEMPFAIL:',desc)
-          self.setreply('450','4.2.0',*desc.splitlines())
-	  return Milter.TEMPFAIL
-	if len(res) < 3: res += time.time(),
-	cbv_cache[sender] = res
-	self.log('REJECT:',desc)
-        self.setreply('550','5.7.1',*desc.splitlines())
-	return Milter.REJECT
-      cbv_cache[sender] = res
-      if not cached:
-        s = time.strftime(time_format,time.localtime())
-	print >>open('send_dsn.log','a'),sender,s # log who we sent DSNs to
       self.cbv_needed = None
+      if rc != Milter.CONTINUE: return rc
 
     if not defanged and not spam_checked:
       os.remove(self.tempname)
@@ -1176,6 +1540,8 @@ class bmsMilter(Milter.Milter):
     if defanged:
       if self.rejectvirus and not self.hidepath:
 	self.log("REJECT virus from",self.mailfrom)
+	self.setreply('550','5.7.1','Attachment type not allowed.',
+		'You attempted to send an attachment with a banned extension.')
 	self.tempname = None
 	return Milter.REJECT
       self.log("Temp file:",self.tempname)
@@ -1196,13 +1562,41 @@ class bmsMilter(Milter.Milter):
       out.close()
     return Milter.TEMPFAIL
 
+  def send_dsn(self,q,msg,template_name):
+    sender = q.s
+    cached = cbv_cache.has_key(sender)
+    if cached:
+      self.log('CBV:',sender,'(cached)')
+      res = cbv_cache[sender]
+    else:
+      self.log('CBV:',sender)
+      try:
+	template = file(template_name).read()
+      except IOError: template = None
+      m = dsn.create_msg(q,self.recipients,msg,template)
+      m = m.as_string()
+      print >>open('last_dsn','w'),m
+      res = dsn.send_dsn(sender,self.receiver,m)
+    if res:
+      desc = "CBV: %d %s" % res[:2]
+      if 400 <= res[0] < 500:
+	self.log('TEMPFAIL:',desc)
+	self.setreply('450','4.2.0',*desc.splitlines())
+	return Milter.TEMPFAIL
+      if len(res) < 3: res += time.time(),
+      cbv_cache[sender] = res
+      self.log('REJECT:',desc)
+      self.setreply('550','5.7.1',*desc.splitlines())
+      return Milter.REJECT
+    cbv_cache[sender] = res
+    return Milter.CONTINUE
+
   def close(self):
-    sys.stdout.flush()		# make log messages visible
     if self.tempname:
       os.remove(self.tempname)	# remove in case session aborted
     if self.fp:
       self.fp.close()
-    sys.stdout.flush()
+    
     return Milter.CONTINUE
 
   def abort(self):
@@ -1217,10 +1611,10 @@ def main():
   if srs or len(discard_users) > 0 or smart_alias or dspam_userdir:
     flags = flags + Milter.DELRCPT
   Milter.set_flags(flags)
-  print "%s bms milter startup" % time.strftime('%Y%b%d %H:%M:%S')
+  milter_log.info("bms milter startup")
   sys.stdout.flush()
   Milter.runmilter("pythonfilter",socketname,timeout)
-  print "%s bms milter shutdown" % time.strftime('%Y%b%d %H:%M:%S')
+  milter_log.info("bms milter shutdown")
 
 if __name__ == "__main__":
   read_config(["/etc/mail/pymilter.cfg","milter.cfg"])

faq.html

@@ -72,6 +72,9 @@ milter-0.4.5 or later to remove this dependency.
 <code>set_flags()</code> before calling <code>runmilter()</code>.  For
 instance, <code>Milter.set_flags(Milter.ADDRCPT)</code>.  You must add together
 all of <code>ADDHDRS, CHGBODY, ADDRCPT, DELRCPT, CHGHDRS</code> that apply.
+<p> NOTE - recent versions default flags to enabling all features.  You
+must now call <code>set_flags()</code> if you wish to disable features for
+efficiency.
 <p>
 
 <li> Q. Why does sendmail sometimes print something like:
@@ -94,14 +97,19 @@ for your specific needs.  We will of course continue to move generic
 code out of the sample as the project evolves.  Think of sample.py as
 an active config file.
 <p>
+If you are running bms.py, then the block_chinese option in 
+<code>/etc/mail/pymilter.cfg</code> controls this feature.
+<p>
 
 <li> Q. Why does sendmail coredump with milters on OpenBSD?
-<p>  A. Sendmail has a problem with unix sockets on OpenBSD.  Use
+<p>  A. Sendmail has a problem with unix sockets on old versions of OpenBSD.
-an internet domain socket instead.  For example, in <code>sendmail.cf</code> use
+Use an internet domain socket instead.  For example, in
+<code>sendmail.cf</code> use
 <pre>
 Xpythonfilter, S=inet:1234@localhost
 </pre>
 and change sample.py accordingly.
+<p> OpenBSD users report that this problem has been fixed.
 <p>
 
 <li> Q. How can I change the bounce message for an invalid recipient?
@@ -133,6 +141,36 @@ is a milter declaration for sendmail.cf with all timeouts specified:
 <pre>
 Xpythonfilter, S=local:/var/log/milter/pythonsock, F=T, T=C:5m;S:20s;R:60s;E:5m
 </pre>
+<li> Q. There is a Python traceback in the log file!  What happened to
+	my email?
+<p>  A. When the milter fails with an untrapped exception, a TEMPFAIL
+result (451) is returned to the sender.  The sender will then retry every
+hour or so for several days.  Hopefully, someone will notice the 
+traceback, and workaround or fix the problem.
+
+<li> Q. I read some notes such as "Check valid domains allowed by internal
+	senders to detect PCs infected with spam trojans." but could not
+	understand the idea. Could you clarify the content ?
+
+<p>  A. The <code>internal_domains</code> configuration specifies which
+MAIL FROM domains are used by internal connections.  If an internal
+PC tries to use some other domain, it is assumed to be a "Zombie".
+<p>
+Here is a sample log line:
+<pre>
+2005Jun22 12:01:04 [12430] REJECT: zombie PC at  192.168.100.171  sending MAIL FROM  debby@fedex.com
+</pre>
+No, fedex.com does not use pymilter, and there is no one named debby at my
+client.  But the idiot using the PC at 192.168.100.171 has downloaded and
+installed some stupid weatherbar/hotbar/aquariumscreensaver that is actually a
+spam bot.
+<p>
+The <code>internal_domains</code> option is simplistic, it assumes all
+valid senders of the domains are internal.  SPF provides a much more general
+check of IP and MAIL FROM for external email.  Pymilter should soon
+have a local policy feature for more general checking of internal mail.
+
+<h3> Using SPF </h3>
 
 <a name="spf">
 <li> Q. So how do I use the SPF support?  The sample.py milter doesn't seem
@@ -143,8 +181,8 @@ everything up for you.  For other systems:
 <li> Arrange to run bms.py in the background (as a service perhaps) and
      redirect output and errors to a logfile.  For instance, on AIX you'll want
      to use SRC (System Resource Controller).  
-<li> Copy milter.cfg to the directory you run bms.py in, and edit it.  The
+<li> Copy pymilter.cfg to the /etc/mail or the directory you run bms.py in,
-     comments should explain the options. 
+     and edit it.  The comments should explain the options. 
 <li> Start bms.py in the background as arranged.
 <li> Add Xpythonfilter to sendmail.cf or add an INPUT_MAIL_FILTER to
      sendmail.mc.  Regen sendmail.cf if you use sendmail.mc and restart 
@@ -156,4 +194,5 @@ everything up for you.  For other systems:
 </ol>
 
 </ol>
+</body>
 </html>

logmsgs.html

@@ -0,0 +1,76 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.1 Final//EN">
+<html>
+<head>
+<title>Python Milter Log Documentation</title>
+<style>
+DT              { font-weight: bolder; padding-top: 1em }
+</style>
+</head><body>
+
+<h1> Milter Log Documentation </h1>
+
+The milter log has a variety of "tags" in it that indicate what it did.
+
+<dl>
+<dt> DSPAM: honeypot SCREENED
+<dd> message was quarantined to the honeypot quarantine
+
+<dt> REJECT: hello SPF: fail 550 access denied
+<dt> REJECT: hello SPF: softfail 550 domain in transition
+<dt> REJECT: hello SPF: neutral 550 access neither permitted nor denied
+<dd> message was rejected because there was an SPF policy for the
+HELO name, and it did not pass.
+
+<dt> CBV: sender-17-44662668-643@bluepenmagic.com
+<dd> we performed a call back verification
+
+<dt> dspam
+<dd> dspam identifier was added to the message
+
+<dt> REJECT: spam from self: jsconnor.com
+<dd> message was reject because HELO was us (jsconnor.com)
+
+<dt> INNOC: richh
+<dd> message was used to update richh's dspam dictionary
+
+<dt> HONEYPOT: michaelb@jsconnor.com
+<dd> message was sent to a honeypot address (michaelb@jsconnor.com), the 
+message was added to the honeypot dspam dictionary as spam
+
+<dt> REJECT: numeric hello name: 63.217.19.146
+<dd> message was rejected because helo name was invalid (numeric)
+
+<dt> eom
+<dd> message was successfully received
+
+<dt> TEMPFAIL: CBV: 450 No MX servers available
+<dd> we tried to do a call back verification but could not look up 
+MX record, we told the sender to try again later
+
+<dt> CBV: info@emailpizzahut.com (cached)
+<dd> call back verification was needed, we had already done it recently
+
+<dt> abort after 0 body chars
+<dd> sender hung up on us
+
+<dt> REJECT: SPF fail 550 SPF fail: see 
+  http://openspf.com/why.html?sender=m.hendersonxk@163.net&ip=213.47.161.100
+<dd> message was reject because its sender's spf policy said to
+
+<dt> REJECT: Subject: Cialis - No prescription needed!
+<dd> message was rejected because its subject contained a bad expression
+
+<dt> DSPAM: tonyc tonyc@jsconnor.com
+<dd> message was sent to tonyc@jsconnor.com and it was identified as spam 
+and placed in the tonyc dspam quarantine
+
+<dt> REJECT: CBV: 550 calvinalstonis@ix.netcom.com...User unknown
+<dt> REJECT: CBV: 553 sorry, that domain isn't in my list
+<dt> REJECT: CBV: 554 delivery error: dd This user doesn't have an account 
+<dd> message was rejected because call back verification gave us a fatal
+error
+</dl>
+
+Please add more tags to this list if you know of any. Thanks.
+</body>
+</html>

milter.cfg

@@ -8,7 +8,7 @@ tempdir = /var/log/milter/save
 log_headers = 0
 # connection ips and hostnames are matched against this glob style list
 # to recognize internal senders.
-;internal_connect = 192.168.*.*
+;internal_connect = 192.168.*.*,127.*
 
 # mail that is not an internal_connect and claims to be from an
 # internal domain is rejected.  Furthermore, internal mail that
@@ -17,7 +17,7 @@ log_headers = 0
 # flexible.  However, SPF is not currently checked for outgoing
 # (internal_connect) mail because it doesn't yet handle authorizing 
 # internal IPs locally.
-;internal_domains = mycorp.com
+;internal_domains = mycorp.com,localhost.localdomain
 
 # connections from a trusted relay can trust the first Received header
 # SPF checks are bypassed for internal connections and trusted relays.
@@ -50,7 +50,7 @@ porn_words = penis, breast, pussy, horse cock, porn, xenical, diet pill, d1ck,
 	p-e-n-i-s, hydrocodone, vicodin, xanax, vicod1n, x@nax, diazepam,
 	v1@gra, xan@x, cialis, ci@lis, frëe, xãnax, valíum, vãlium, via-gra,
 	x@n3x, vicod3n, penís, c0d1n, phentermine, en1arge, dip1oma, v1codin,
-	valium, rolex, sexual, fuck
+	valium, rolex, sexual, fuck, adv1t
 # reject mail with these case sensitive strings in the subject
 spam_words = $$$, !!!, XXX, FREE, HGH
 # attachments with these extensions will be replaced with a warning
@@ -70,6 +70,10 @@ config=/etc/mail/pysrs.cfg
 ;fwdomain = mydomain.com
 # turn this on after a grace period to reject spoofed DSNs
 reject_spoofed = 0
+# Many braindead MTAs send DSNs with a non-DSN MFROM (e.g. to report that
+# some virus claiming to be sent by you).  This heuristic
+# refuses mail from user names commonly abused in that way.
+;banned_users = postmaster, mailer-daemon, clamav
 
 # See http://spf.pobox.com for more info on SPF.
 [spf]
@@ -85,6 +89,16 @@ reject_spoofed = 0
 ;reject_noptr = 0
 # always accept softfail from these domains, or send DSN otherwise
 ;accept_softfail = bounces.amazon.com
+# Treat fail from these domains like softfail: because their SPF record
+# or an important sender is screwed up.  Must have valid HELO, however.
+;accept_fail = custhelp.com
+# Use sendmail access map or similar format for detailed spf policy.
+# SPF entries in the access map will override any defaults set above.
+;access_file = /etc/mail/access.db
+# Add MAIL FROM as Sender when Sender is missing and From domain
+# doesn't match MAIL FROM.  Outlook and other email clients will then display
+# something like: "Sent by sender@domain.com on behalf of from@example.com"
+;supply_sender = 0
 
 # features intended to clean up outgoing mail
 [scrub]
@@ -101,6 +115,7 @@ blind = 1
 # (sendmail aliases let you monitor incoming mail)
 #
 ;users = disloyal@bigcorp.com, bigmouth@bigcorp.com
+# multiple destinations can use smart_alias
 ;dest = spy@bigcorp.com
 # discard outgoing mail without alerting sender
 # can be used in conjunction with wiretap to censor outgoing mail
@@ -108,7 +123,10 @@ blind = 1
 #
 # smart aliases trigger on both sender and recipient
 #
-;smart_alias = copycust,walter
+;smart_alias = copycust,walter,spy1,spy2
+# multiple wiretap monitors
+;spy1 = disloyal@bigcorp.com,spy@bigcorp.com
+;spy2 = bigmouth@bigcorp.com,spy@bigcorp.com
 # mail from client@clientcorp.com to sue@bigcorp.com is redirected to 
 # local alias copycust
 ;copycust = client@clientcorp.com,sue@bigcorp.com
@@ -128,7 +146,13 @@ blind = 1
 # only EXTERNAL messages are dspam filtered
 ;dspam_dict=/var/lib/dspam/moderator.dict
 
-# Opt-opt recipients from dspam screening and header triage
+# Recipients of mail sent from these senders are added to the auto_whitelist.
+# Auto_whitelisted senders with an SPF PASS are never rejected by dspam, and
+# messages from auto_whitelisted senders will be used to train screener
+# dictionaries as innocent mail.
+;whitelist_senders = @mycorp.com
+
+# Opt-out recipients entirely from dspam screening and header triage
 ;dspam_exempt=getitall@mycorp.com
 # Do not scan mail (ostensibly) from these senders
 ;dspam_whitelist=getitall@sender.com
@@ -152,6 +176,8 @@ blind = 1
 ;spam=spam@foocorp.com
 # address to forward false positives to.  milter will process and not deliver
 ;falsepositive=ham@foocorp.com
+# account which receives only spam: all received messages are marked as spam.
+;honeypot=spam-me@example.com
 # the dspam_screener is a list of dspam users who screen mail for all
 # recipients who are not dspam_users.  Spam goes to the screeners quarantine,
 # and the original recipients are saved so that false positives can be properly

milter.html

@@ -24,7 +24,7 @@ ALT="Viewable With Any Browser" BORDER="0"></A>
   Stuart D. Gathman</a><br>
 This web page is written by Stuart D. Gathman<br>and<br>sponsored by
 <a href="http://www.bmsi.com">Business Management Systems, Inc.</a> <br>
-Last updated Jun 09, 2005</h4>
+Last updated Oct 12, 2005</h4>
 
 See the <a href="faq.html">FAQ</a> | <a href="http://sourceforge.net/project/showfiles.php?group_id=139894">Download now</a> |
 <a href="/mailman/listinfo/pymilter">Subscribe to mailing list</a> |
@@ -47,9 +47,32 @@ efficient and secure.  I recommend upgrading.
 
 <h2> Recent Changes </h2>
 
-Python milter is being moved to 
+Python milter has been moved to 
 <a href="http://sourceforge.net/projects/pymilter/">pymilter Sourceforge
-project</a> for development.
+project</a> for development and release downloads.
+<p>
+Release 0.8.3 uses the standard logging module, and supports configuring
+more detailed SPF policy via the sendmail access map.  SMTP AUTH connections
+are considered INTERNAL.  Preventing forgery between internal domains is
+just a matter of specifying the user-domain map - I'll define something
+for the next version.  We now send DSNs when mail is quarantined (rejecting
+if DSN fails) and for SPF syntax errors (PermError).  There is an
+experimental option to add a Sender header when it is missing and the From
+domain doesn't match the MAIL FROM domain.  Next release, we may start
+renaming and replacing an existing Sender header when neither it nor the
+From domain matches MAIL FROM.  Since bogus MAIL FROMs are rejected
+(to varying degrees depending on the configured SPF policy), and
+both Sender and From and displayed by default in many email clients,
+this provides some phishing protection without rejecting mail based
+on headers.
+<p>
+Release 0.8.2 has changes to <a href="http://openspf.net">SPF</a> to bring it
+in line with the newly official RFC.  It adds 
+<a href="http://ses.codeshare.ca/">SES</a>
+support (the original SES without body hash) for pysrs-0.30.10, and honeypot
+support for pydspam-1.1.9.  There is a new method in the base milter module.
+milter.set_exception_policy(i) lets you choose a policy of CONTINUE, REJECT, or
+TEMPFAIL (default) for untrapped exceptions encountered in a milter callback.
 <p>
 Release 0.8.0 is the first <a href="http://sourceforge.net/">Sourceforge</a>
 release.  It supports Python-2.4, and provides an option to accept mail
@@ -124,9 +147,9 @@ recommend ignoring it and continuing to implement and improve SPF until a
 working and unencumbered proposal for RFC2822 headers surfaces.
 
 <p>
-<a href="http://spf.pobox.com">
+<a href="http://openspf.com">
 <img src="SPF.gif" align=left alt="SPF logo"></a>
-Release 0.6.6 adds support for <a href="http://spf.pobox.com/">SPF</a>,
+Release 0.6.6 adds support for <a href="http://openspf.com/">SPF</a>,
 a protocol to prevent forging of the envelope from address.  
 SPF support requires <a href="http://pydns.sourceforge.net/">pydns</a>.
 The included spf.py module is an updated version of the original 1.6
@@ -222,7 +245,7 @@ methods that
 do nothing, and also provides wrappers for the libmilter methods to mutate
 the message.
 <p>
-The 'spf' module provides an implementation of <a href="http://spf.pobox.com">
+The 'spf' module provides an implementation of <a href="http://openspf.com">
 SPF</a> useful for detecting email forgery.
 <p>
 The 'mime' module provides a wrapper for the Python email package that
@@ -237,6 +260,8 @@ content filtering.  SPF checking
 requires <a href="http://pydns.sourceforge.net/">
 pydns</a>.  Configuration documentation is currently included as comments
 in the <a href="milter.cfg">sample config file</a> for the bms.py milter.
+See also the <a href="HOWTO">HOWTO</a> and <a href="logmsgs.html">
+Milter Log Message Tags</a>.
 <p>
 Python milter is under GPL.  The authors can probably be convinced to
 change this to LGPL if needed.

milter.spec

@@ -1,6 +1,6 @@
 %define name milter
-%define version 0.8.0
+%define version 0.8.4
-%define release 3.RH7
+%define release 1.RH7
 # what version of RH are we building for?
 %define redhat9 0
 %define redhat7 1
@@ -63,7 +63,7 @@ rm -rf $RPM_BUILD_ROOT
 mkdir -p $RPM_BUILD_ROOT/var/log/milter
 mkdir -p $RPM_BUILD_ROOT/etc/mail
 mkdir $RPM_BUILD_ROOT/var/log/milter/save
-cp bms.py strike3.txt softfail.txt $RPM_BUILD_ROOT/var/log/milter
+cp bms.py *.txt $RPM_BUILD_ROOT/var/log/milter
 cp milter.cfg $RPM_BUILD_ROOT/etc/mail/pymilter.cfg
 
 # logfile rotation
@@ -146,7 +146,7 @@ rm -rf $RPM_BUILD_ROOT
 
 %files -f INSTALLED_FILES
 %defattr(-,root,root)
-%doc README NEWS TODO CREDITS sample.py
+%doc README HOWTO NEWS TODO CREDITS sample.py
 /etc/logrotate.d/milter
 /etc/cron.daily/milter
 %ifos aix4.1
@@ -160,12 +160,48 @@ rm -rf $RPM_BUILD_ROOT
 %dir /var/log/milter/save
 %config /var/log/milter/start.sh
 %config /var/log/milter/bms.py
-%config /var/log/milter/strike3.txt
+%config(noreplace) /var/log/milter/strike3.txt
-%config /var/log/milter/softfail.txt
+%config(noreplace) /var/log/milter/softfail.txt
+%config(noreplace) /var/log/milter/neutral.txt
+%config(noreplace) /var/log/milter/quarantine.txt
+%config(noreplace) /var/log/milter/permerror.txt
 %config(noreplace) /etc/mail/pymilter.cfg
 /usr/share/sendmail-cf/hack/rhsbl.m4
 
 %changelog
+* Thu Oct 20 2005 Stuart Gathman <stuart@bmsi.com> 0.8.4-1
+- Fix SPF policy via sendmail access map (case insensitive keys).
+- Auto whitelist senders, train screener on whitelisted messages
+- Optional idx parameter to addheader to invoke smfi_insheader
+- Activate progress when SMFIR_PROGRESS defined
+* Wed Oct 12 2005 Stuart Gathman <stuart@bmsi.com> 0.8.3-1
+- Keep screened honeypot mail, but optionally discard honeypot only mail.
+- spf_accept_fail option for braindead SPF senders (treats fail like softfail)
+- Consider SMTP AUTH connections internal.
+- Send DSN for SPF errors corrected by extended processing.
+- Send DSN before SCREENED mail is quarantined
+- Option to set SPF policy via sendmail access map.
+- Option to supply Sender header from MAIL FROM when missing.
+- Use logging package to keep log lines atomic.
+* Fri Jul 15 2005 Stuart Gathman <stuart@bmsi.com> 0.8.2-4
+- Limit each CNAME chain independently like PTR and MX
+* Fri Jul 15 2005 Stuart Gathman <stuart@bmsi.com> 0.8.2-3
+- Limit CNAME lookups (regression)
+* Fri Jul 15 2005 Stuart Gathman <stuart@bmsi.com> 0.8.2-2
+- Handle corrupt ZIP attachments
+* Fri Jul 15 2005 Stuart Gathman <stuart@bmsi.com> 0.8.2-1
+- Strict processing limits per SPF RFC
+- Fixed several parsing bugs under RFC 
+- Support official IANA SPF record (type99)
+- Honeypot support (requires pydspam-1.1.9)
+- Extended SPF processing results beyond strict RFC limits
+- Support original SES for local bounce protection (requires pysrs-0.30.10)
+- Callback exception processing option in milter module
+* Thu Jun 16 2005 Stuart Gathman <stuart@bmsi.com> 0.8.1-1
+- Fix zip in zip loop in mime.py
+- Fix HeaderParseError in bms.py header callback
+- Check internal_domains for outgoing mail
+- Fix inconsistent results from send_dsn
 * Mon Jun 06 2005 Stuart Gathman <stuart@bmsi.com> 0.8.0-3
 - properly log pydspam exceptions
 * Sat Jun 04 2005 Stuart Gathman <stuart@bmsi.com> 0.8.0-2

miltermodule.c

@@ -34,6 +34,24 @@ $ python setup.py help
      libraries=["milter","smutil","resolv"]
 
  * $Log$
+ * Revision 1.7  2005/10/20 23:04:46  customdesigned
+ * Add optional idx for position of added header.
+ *
+ * Revision 1.6  2005/07/15 22:18:17  customdesigned
+ * Support callback exception policy
+ *
+ * Revision 1.5  2005/06/24 04:20:07  customdesigned
+ * Report context allocation error.
+ *
+ * Revision 1.4  2005/06/24 04:12:43  customdesigned
+ * Remove unused name argument to generic wrappers.
+ *
+ * Revision 1.3  2005/06/24 03:57:35  customdesigned
+ * Handle close called before connect.
+ *
+ * Revision 1.2  2005/06/02 04:18:55  customdesigned
+ * Update copyright notices after reading article on /.
+ *
  * Revision 1.1.1.2  2005/05/31 18:09:06  customdesigned
  * Release 0.7.1
  *
@@ -194,7 +212,7 @@ $ python setup.py help
 
 
 /* Yes, these are static.  If you need multiple different callbacks, */
-/* it's cleaner to use multiple filters. */
+/* it's cleaner to use multiple filters, or convert to OO method calls. */
 static PyObject *connect_callback = NULL;
 static PyObject *helo_callback    = NULL;
 static PyObject *envfrom_callback = NULL;
@@ -239,8 +257,11 @@ _get_context(SMFICTX *ctx) {
     PyEval_AcquireThread(t);	/* lock interp */
     self = PyObject_New(milter_ContextObject,&milter_ContextType);
     if (!self) {
-      /* Can't pass on exception since we are called from libmilter */
+      /* Report and clear exception since we are called from libmilter */
+      if (PyErr_Occurred()) {
+	PyErr_Print();
 	PyErr_Clear();
+      }
       PyThreadState_Clear(t);
       PyEval_ReleaseThread(t);
       PyThreadState_Delete(t);
@@ -331,7 +352,8 @@ CHGHDRS - filter may change/delete headers";
 
 static PyObject *
 milter_set_flags(PyObject *self, PyObject *args) {
-  if (!PyArg_ParseTuple(args, "i", &description.xxfi_flags)) return NULL;
+  if (!PyArg_ParseTuple(args, "i:set_flags", &description.xxfi_flags))
+    return NULL;
   Py_INCREF(Py_None);
   return Py_None;
 }
@@ -497,6 +519,28 @@ milter_set_close_callback(PyObject *self, PyObject *args) {
   return generic_set_callback(args, "O:set_close_callback", &close_callback);
 }
 
+static int exception_policy = SMFIS_TEMPFAIL;
+
+static char milter_set_exception_policy__doc__[] =
+"set_exception_policy(i) -> None\n\
+Sets the policy for untrapped Python exceptions during a callback.\n\
+Must be one of TEMPFAIL,REJECT,CONTINUE";
+
+static PyObject *
+milter_set_exception_policy(PyObject *self, PyObject *args) {
+  int i;
+  if (!PyArg_ParseTuple(args, "i:set_exception_policy", &i))
+    return NULL;
+  switch (i) {
+  case SMFIS_REJECT: case SMFIS_TEMPFAIL: case SMFIS_CONTINUE:
+    exception_policy = i;
+    Py_INCREF(Py_None);
+    return Py_None;
+  }
+  PyErr_SetString(MilterError,"invalid exception policy");
+  return NULL;
+}
+
 /** Report and clear any python exception before returning to libmilter. 
   The interpreter is locked when we are called, and we unlock it.  */
 static int _report_exception(milter_ContextObject *self) {
@@ -504,9 +548,16 @@ static int _report_exception(milter_ContextObject *self) {
     PyErr_Print();
     PyErr_Clear();	/* must clear since not returning to python */
     PyEval_ReleaseThread(self->t);
+    switch (exception_policy) {
+      case SMFIS_REJECT:
+	smfi_setreply(self->ctx, "554", "5.3.0", "Filter failure");
+	return SMFIS_REJECT;
+      case SMFIS_TEMPFAIL:
 	smfi_setreply(self->ctx, "451", "4.3.0", "Filter failure");
 	return SMFIS_TEMPFAIL;
     }
+    return SMFIS_CONTINUE;
+  }
   PyEval_ReleaseThread(self->t);
   return SMFIS_CONTINUE;
 }
@@ -616,7 +667,7 @@ milter_wrap_helo(SMFICTX *ctx, char *helohost) {
 }
 
 static int
-generic_env_wrapper(SMFICTX *ctx, PyObject*cb, char **argv, const char *name) {
+generic_env_wrapper(SMFICTX *ctx, PyObject*cb, char **argv) {
    PyObject *arglist;
    milter_ContextObject *self;
    int count = 0;
@@ -653,12 +704,12 @@ generic_env_wrapper(SMFICTX *ctx, PyObject*cb, char **argv, const char *name) {
 
 static int
 milter_wrap_envfrom(SMFICTX *ctx, char **argv) {
-  return generic_env_wrapper(ctx,envfrom_callback,argv,"milter_wrap_envfrom");
+  return generic_env_wrapper(ctx,envfrom_callback,argv);
 }
 
 static int
 milter_wrap_envrcpt(SMFICTX *ctx, char **argv) {
-  return generic_env_wrapper(ctx,envrcpt_callback,argv,"milter_wrap_envrcpt");
+  return generic_env_wrapper(ctx,envrcpt_callback,argv);
 }    
   
 static int
@@ -674,7 +725,7 @@ milter_wrap_header(SMFICTX *ctx, char *headerf, char *headerv) {
 }
 
 static int
-generic_noarg_wrapper(SMFICTX *ctx,PyObject *cb,const char *name) {
+generic_noarg_wrapper(SMFICTX *ctx,PyObject *cb) {
    PyObject *arglist;
    milter_ContextObject *c;
    if (cb == NULL) return SMFIS_CONTINUE;
@@ -686,7 +737,7 @@ generic_noarg_wrapper(SMFICTX *ctx,PyObject *cb,const char *name) {
 
 static int
 milter_wrap_eoh(SMFICTX *ctx) {
-  return generic_noarg_wrapper(ctx,eoh_callback,"milter_wrap_eoh");
+  return generic_noarg_wrapper(ctx,eoh_callback);
 }   
 
 static int
@@ -704,18 +755,31 @@ milter_wrap_body(SMFICTX *ctx, u_char *bodyp, size_t bodylen) {
 
 static int
 milter_wrap_eom(SMFICTX *ctx) {
-  return generic_noarg_wrapper(ctx,eom_callback,"milter_wrap_eom");
+  return generic_noarg_wrapper(ctx,eom_callback);
 }
 
 static int
 milter_wrap_abort(SMFICTX *ctx) {
   /* libmilter still calls close after abort */
-  return generic_noarg_wrapper(ctx,abort_callback,"milter_wrap_abort");
+  return generic_noarg_wrapper(ctx,abort_callback);
 }
 
 static int
 milter_wrap_close(SMFICTX *ctx) {
-  int r = generic_noarg_wrapper(ctx,close_callback,"milter_wrap_close");
+  /* xxfi_close can be called out of order - even before connect.  
+   * There may not yet be a private context pointer.  To avoid
+   * creating a ThreadContext and allocating a milter context only
+   * to destroy them, and to avoid invoking the python close_callback when
+   * connect has never been called, we don't use generic_noarg_wrapper here. */
+  PyObject *cb = close_callback;
+  milter_ContextObject *self = smfi_getpriv(ctx);
+  int r = SMFIS_CONTINUE;
+  if (self != NULL && cb != NULL && self->ctx == ctx) {
+    PyObject *arglist;
+    PyEval_AcquireThread(self->t);
+    arglist = Py_BuildValue("(O)", self);
+    r = _generic_wrapper(self, cb, arglist);
+  }
   /* FIXME: It is inefficient to have released the interp lock only to
      acquire it again in _clear_context.  We can tell _generic_return and
      friends not to release the lock by, for instance, setting self->t to NULL.
@@ -909,28 +973,31 @@ milter_setreply(PyObject *self, PyObject *args) {
 }
 
 static char milter_addheader__doc__[] =
-"addheader(field, value) -> None\n\
+"addheader(field, value, idx=-1) -> None\n\
 Add a header to the message. This header is not passed to other\n\
 filters. It is not checked for standards compliance;\n\
 the mail filter must ensure that no protocols are violated\n\
 as a result of adding this header.\n\
 field - header field name\n\
 value - header field value\n\
+idx - optional position in internal header list to insert new header\n\
 Both are strings.  This function can only be called from the EOM callback.";
 
 static PyObject *
 milter_addheader(PyObject *self, PyObject *args) {
   char *headerf;
   char *headerv;
+  int idx = -1;
   SMFICTX *ctx;
   PyThreadState *t;
 
-  if (!PyArg_ParseTuple(args, "ss:addheader", &headerf, &headerv)) return NULL;
+  if (!PyArg_ParseTuple(args, "ss|i:addheader", &headerf, &headerv, &idx))
+    return NULL;
   ctx = _find_context(self);
   if (ctx == NULL) return NULL;
   t = PyEval_SaveThread();
-  return _thread_return(t,smfi_addheader(ctx, headerf, headerv),
+  return _thread_return(t, (idx < 0) ? smfi_addheader(ctx, headerf, headerv) :
-			 "cannot add header");
+      smfi_insheader(ctx, idx, headerf, headerv), "cannot add header");
 }
 
 static char milter_chgheader__doc__[] =
@@ -1085,7 +1152,7 @@ milter_quarantine(PyObject *self, PyObject *args) {
 }
 #endif
 
-#if _FFR_SMFI_PROGRESS
+#ifdef SMFIR_PROGRESS
 static char milter_progress__doc__[] =
 "progress() -> None\n\
 Notify the MTA that we are working on a message so it will reset timeouts.";
@@ -1116,7 +1183,7 @@ static PyMethodDef context_methods[] = {
 #ifdef SMFIF_QUARANTINE
   { "quarantine",  milter_quarantine,  METH_VARARGS, milter_quarantine__doc__},
 #endif
-#if _FFR_SMFI_PROGRESS
+#ifdef SMFIR_PROGRESS
   { "progress",  milter_progress,  METH_VARARGS, milter_progress__doc__},
 #endif
   { NULL, NULL }
@@ -1155,6 +1222,8 @@ static PyMethodDef milter_methods[] = {
    { "set_eom_callback",     milter_set_eom_callback,     METH_VARARGS, milter_set_eom_callback__doc__},
    { "set_abort_callback",   milter_set_abort_callback,   METH_VARARGS, milter_set_abort_callback__doc__},
    { "set_close_callback",   milter_set_close_callback,   METH_VARARGS, milter_set_close_callback__doc__},
+   { "set_exception_policy",   milter_set_exception_policy,METH_VARARGS, milter_set_exception_policy__doc__},
+   { "register",             milter_register,             METH_VARARGS, milter_register__doc__},
    { "register",             milter_register,             METH_VARARGS, milter_register__doc__},
    { "main",                 milter_main,                 METH_VARARGS, milter_main__doc__},
    { "setdbg",               milter_setdbg,               METH_VARARGS, milter_setdbg__doc__},

mime.py

@@ -1,4 +1,10 @@
 # $Log$
+# Revision 1.4  2005/06/17 01:49:39  customdesigned
+# Handle zip within zip.
+#
+# Revision 1.3  2005/06/02 15:00:17  customdesigned
+# Configure banned extensions.  Scan zipfile option with test case.
+#
 # Revision 1.2  2005/06/02 04:18:55  customdesigned
 # Update copyright notices after reading article on /.
 #
@@ -87,6 +93,16 @@ from email import Errors
 
 from types import ListType,StringType
 
+def zipnames(txt):
+  fp =  StringIO.StringIO(txt)
+  zipf = zipfile.ZipFile(fp,'r')
+  names = []
+  for nm in zipf.namelist():
+    names.append(('zipname',nm))
+    if nm.lower().endswith('.zip'):
+      names += zipnames(zipf.read(nm))
+  return names
+
 class MimeGenerator(Generator):
     def _dispatch(self, msg):
         # Get the Content-Type: for the message, then try to dispatch to
@@ -177,13 +193,11 @@ class MimeMessage(Message):
       names.append((attr,val))
     names += [("filename",self.get_filename())]
     if scan_zip:
-      for key,name in names:
+      for key,name in tuple(names):	# copy by converting to tuple
 	if name and name.lower().endswith('.zip'):
 	  txt = self.get_payload(decode=True)
-	  fp =  StringIO.StringIO(txt)
+	  if txt.strip():
-	  zipf = zipfile.ZipFile(fp,'r')
+	    names += zipnames(txt)
-	  for nm in zipf.namelist():
-	    names.append(('zipname',nm))
     return names
 
   def ismodified(self):
@@ -294,19 +308,25 @@ See your administrator.
 
 def check_name(msg,savname=None,ckname=check_ext,scan_zip=False):
   "Replace attachment with a warning if its name is suspicious."
+  try:
     for key,name in msg.getnames(scan_zip):
       badname = ckname(name)
       if badname:
-      hostname = socket.gethostname()
         if key == 'zipname':
           badname = msg.get_filename()
+	break
+    else:
+      return Milter.CONTINUE
+  except zipfile.BadZipfile:
+    # a ZIP that is not a zip is very suspicious
+    badname = msg.get_filename()
+  hostname = socket.gethostname()
   msg.set_payload(virus_msg % (badname,hostname,savname))
   del msg["content-type"]
   del msg["content-disposition"]
   del msg["content-transfer-encoding"]
   name = "WARNING.TXT"
   msg["Content-Type"] = "text/plain; name="+name
-      break
   return Milter.CONTINUE
 
 import email.Iterators

neutral.txt

@@ -0,0 +1,34 @@
+Subject: SPF %(result)s (POSSIBLE FORGERY)
+
+This is an automatically generated Delivery Status Notification.
+
+THIS IS A WARNING MESSAGE ONLY.
+
+YOU DO *NOT* NEED TO RESEND YOUR MESSAGE.
+
+Delivery to the following recipients has been delayed.
+
+       %(rcpt)s
+
+Subject: %(subject)s
+Received-SPF: %(spf_result)s
+
+Your sender policy (or lack thereof) indicated that the above email was not
+sent via an authorized SMTP server, but may still be legitimate.  Since there
+is no positive confirmation that the message is really from you, we have
+to give it extra scrutiny - including verifying that the sender really
+exists by sending you this DSN.  We will remember this sender and not
+bother you again for a while.  You can avoid this message entirely for
+legitimate mail by using an authorized SMTP server.  Contact your mail
+administrator and ask how to configure your email client to use an
+authorized server.  
+
+If you never sent the above message, then your domain has been forged.
+Your mail admin needs to publish a strict SPF record so that I can reject
+those forgeries instead of bugging you about them.
+
+If you need further assistance, please do not hesitate to contact me.
+
+Kind regards,
+
+postmaster@%(receiver)s

permerror.txt

@@ -0,0 +1,31 @@
+Subject: Critical SPF configuration error
+
+This is an automatically generated Delivery Status Notification.
+
+THIS IS A WARNING MESSAGE ONLY.
+
+YOU DO *NOT* NEED TO RESEND YOUR MESSAGE.
+
+Delivery to the following recipients has been delayed.
+
+	%(rcpt)s
+
+Subject: %(subject)s 
+
+Your spf record has a permanent error.  The error was:
+
+	%(perm_error)s
+
+We will reinterpret your record using "lax" processing heuristics
+which may result in your mail being accepted anyway.  But you or your
+mail administrator need to fix your SPF record as soon as possible.
+
+We are sending you this message to alert you to the fact that
+you have problems with your email configuration.
+
+If you need further assistance, please do not hesitate to
+contact me again.
+
+Kind regards,
+
+postmaster@%(receiver)s

policy.html

@@ -0,0 +1,237 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">
+<html>
+<head>
+<title>Python Milter Mail Policy </title>
+</head><body>
+
+<h1> Python Milter Mail Policy </h1>
+
+<h3> Classify connection </h3>
+
+When the SMTP client connects, the connection IP address is
+saved for later verification, and the connection
+is classified as INTERNAL or EXTERNAL by matching the ip
+address against the <code>internal_connect</code> configuration.
+IP addresses with no PTR, and PTR names that look like
+the kind assigned to dynamic IPs (as determined by a heuristic
+algorithm) are flagged as DYNAMIC.  IPs that match the
+<code>trusted_relay</code> configuration are flagged as TRUSTED.
+<p>
+Examples from the log file (<i>not</i> the SMTP error message returned):
+<pre>
+2005Jul29 13:56:53 [71207] connect from p50863492.dip0.t-ipconnect.de at ('80.134.52.146', 1858) EXTERNAL DYN
+2005Jul29 18:10:15 [74511] connect from foopub at ('1.2.3.4', 46513) EXTERNAL TRUSTED
+2005Jul29 14:41:00 [71805] connect from foobar at ('192.168.0.1', 41205) INTERNAL
+2005Jul29 14:41:15 [71806] connect from cncln.online.ln.cn at ('218.25.240.137', 35992) EXTERNAL
+</pre>
+<p>
+Certain obviously evil PTR names are blocked at this point:
+"localhost" (when IP is not 127.*) and ".".
+<pre>
+2005Jul29 14:49:50 [71918] connect from localhost at ('221.132.0.6', 50507) EXTERNAL
+2005Jul29 14:49:50 [71918] REJECT: PTR is localhost
+</pre>
+
+<h3> HELO Check </h3>
+
+The HELO name provided by the client is saved for later verification
+(for example by SPF).  We could validate the HELO at this point
+by verifying that an A record for the HELO name matches the connect ip.
+However, currently we only block certain obvious problems.  
+HELO names that look like an IP4 address 
+and ones that match the <code>hello_blacklist</code> configuration
+are immediately rejected.  The hello_blacklist typically contains
+the current MTAs own HELO name or email domains.
+Clients that attempt to skip HELO are immediately rejected.
+<pre>
+2005Jul29 18:10:15 [74512] hello from example.com
+2005Jul29 18:10:15 [74512] REJECT: spam from self: example.com
+2005Jul29 18:17:09 [74581] hello from 80.191.244.69
+2005Jul29 18:17:09 [74581] REJECT: numeric hello name: 80.191.244.69
+</pre>
+
+<h3> MAIL FROM Check </h3>
+
+Before calling our milter, sendmail checks a DNS blacklist to 
+block banned sender domains.  We never see a blocked domain.
+<p>
+The MAIL FROM address is saved for possible use by the smart-alias
+feature.  First, the <code>internal_domains</code> is used for
+a simple screening if defined.  If the MAIL FROM for an INTERNAL connection 
+is NOT in <code>internal_domains</code>, then it is rejected (the
+PC is most likely infected and attempting to send out spam).
+If the MAIL FROM for an EXTERNAL connection IS in
+<code>internal_domains</code>, then the message is immediately rejected.
+This is quick and effective for most small company MTAs.  For more
+complex mail networks, it is too simplistic, and should not be defined.
+SPF will handle the complex cases.
+
+<h4> wiretap </h4>
+
+The wiretap feature can screen and/or monitor mail to/from certain
+users.  If the MAIL FROM is being wiretapped, the recipients are
+altered accordingly.
+
+<h4> SPF check </h4>
+
+Finally, the MAIL FROM, connect IP, and HELO name are checked against
+any SPF records published via DNS for the alleged sender (MAIL FROM).
+If there is no SPF record, we check for a local substitute under the
+domain defined in the <code>[spf]delegate</code> configuration.
+Further checks depend on the result.
+
+<table border=1>
+<tr><th>NONE</th><td>
+If there is no SPF record (official or delegated), then we 
+initiate a "three strikes and your out" regime, which looks for
+<b>some</b> form of validated identification.
+<ol>
+<li>We try a "best guess" SPF record of "v=spf1 a/24 mx/24 ptr".  If this
+    passes, good.  
+<li> We try to validate the HELO name. First check for an SPF record.
+   Otherwise, check whether the connect IP matches any A record for
+   the HELO name, or any A record for any MX name for the HELO name,
+   or is at least in the same /24 subnet as any of the above.
+   (In other words, a HELO SPF "best guess" of "v=spf1 a/24 mx/24".)
+   If so, good.  We consider the HELO validated.  If the HELO SPF
+   check fails, we reject the email.
+</ol>
+<pre>
+2005Jul30 19:45:16 [93991] connect from [221.200.41.54] at ('221.200.41.54', 3581) EXTERNAL DYN
+2005Jul30 19:45:18 [93991] hello from adelphia.net
+2005Jul30 19:45:19 [93991] mail from <wendy.stubbsua@link-it.com> ()
+2005Jul30 19:45:19 [93991] REJECT: hello SPF: fail 550 access denied
+</pre>
+<ol>
+<li> If there is a validated PTR name, and it doesn't look
+   like a dynamic name, good.  We consider the connection validated.
+</ol>
+If any of the above can be validated, we continue on.
+If none of the above can be validated, and the <code>[SPF]reject_noptr</code>
+option is true, we reject the message immediately with the explanation
+that we need some form of valid identification before we accept an email.
+If <code>[SPF]reject_noptr</code> is false, we flag the message as
+needing Call Back Validation.  
+The Call Back Valildation sends a DSN to the purported sender informing
+them of the lack of identification.  If the message is legitimate, the
+sender needs to know that their email setup is broken and should be corrected.
+If the message is forged, the sender is informed of the forgery, 
+and their need to publish an SPF record or at least use a valid HELO name.
+If the purported sender does not accept the DSN,
+then the message is rejected.  The CBV status is cached to avoid
+annoying the purported sender with too many DSNs.  Currently, the DSN
+is repeated to the same sender once per month.
+<p>
+In this example, although 3com.com has no SPF record, we assume that 
+any legitimate mail from them will at least have a valid HELO or PTR.
+<pre>
+2005Jul30 23:52:03 [96777] connect from [222.252.233.200] at ('222.252.233.200', 29934) EXTERNAL DYN
+2005Jul30 23:52:03 [96777] hello from 3mail.3com.com
+2005Jul30 23:52:04 [96777] mail from <etec_nic_family@3mail.3com.com> ()
+2005Jul30 23:52:04 [96777] REJECT: no PTR, HELO or SPF
+</pre>
+</td></tr>
+
+<tr><th>PASS</th><td>
+A pass result normally lets the email continue on, but the domain is 
+tracked for reputation (and may be blocked), and may skip content scanning if
+it matches a whitelist.
+<pre>
+2005Jul24 17:44:26 [2104] mail from <gnucash-devel-bounces@gnucash.org> ('SIZE=4410',)
+2005Jul24 17:44:26 [2104] Received-SPF: pass (mail.bmsi.com: domain of gnucash.org
+	designates 204.107.200.65 as permitted sender)
+	client-ip=204.107.200.65; envelope-from=gnucash-devel-bounces@gnucash.org; helo=cvs.gnucash.org;
+</pre>
+</td></tr>
+
+<tr><th>NEUTRAL</th><td>
+A neutral result normally lets the email continue on, but the domain is not
+tracked for reputation or matched against any whitelists. 
+Highly forged domains listed in <code>[SPF]reject_neutral</code> are
+rejected.
+<pre>
+2005Jul24 17:41:37 [2070] connect from cp500627-a.dbsch1.nb.home.nl at ('84.27.225.3', 3465) EXTERNAL
+2005Jul24 17:41:37 [2070] hello from cp500627-a.dbsch1.nb.home.nl
+2005Jul24 17:41:38 [2070] mail from <nwarjejkw@yahoo.com> ()
+2005Jul24 17:41:38 [2070] REJECT: SPF neutral for nwarjejkw@yahoo.com
+</pre>
+</td></tr>
+
+<tr><th>SOFTFAIL</th><td>
+A softfail result normally lets the email continue on, but the domain is not
+tracked for reputation or matched against any whitelists.  Furthermore,
+the message is flagged as needing Call Back Validation,
+and the highly forged domains listed in <code>[SPF]reject_neutral</code> are
+rejected as well.  
+<p>
+At present, we also require a valid HELO or PTR to avoid rejecting 
+a softfail.  But this should probably change to only require a
+successful CBV.
+<p>
+The Call Back Valildation sends a DSN to the purported sender informing
+them of the softfail.  If the message is legitimate, the sender needs
+to know about the softfail so that their email setup can be corrected.
+If the message is forged, the sender is informed of the forgery, confirming
+that SPF is protecting their reputation and encouraging a rapid transition
+to a strict policy.  If the purported sender does not accept the DSN,
+then the message is rejected.  The CBV status is cached to avoid
+annoying the purported sender with too many DSNs.  Currently, the DSN
+is repeated to the same sender once per month.
+<pre>
+2005Jul24 15:41:33 [801] mail from <Aitp@horafeliz.com> ()
+2005Jul24 15:41:33 [801] Received-SPF: softfail (mail.bmsi.com: transitioning domain of horafeliz.com
+	does not designate 221.184.83.185 as permitted sender)
+	client-ip=221.184.83.185; envelope-from=Aitp@horafeliz.com;
+	helo=p8185-ipad30funabasi.chiba.ocn.ne.jp;
+2005Jul24 15:41:33 [801] rcpt to <david@example.com> ()
+2005Jul24 15:41:35 [801] Subject: Microsoft, Adobe, Macromedia, Corel software. Up to 80% discount.
+2005Jul24 15:41:35 [801] X-Mailer: Microsoft Outlook, Build 10.0.2605
+2005Jul24 15:41:35 [801] CBV: Aitp@horafeliz.com
+2005Jul24 15:41:38 [801] REJECT: CBV: 550 <Aitp@horafeliz.com>: User unknown
+</pre>
+</td></tr>
+
+<tr><th>FAIL</th><td>
+The message is rejected with a reference the SPF why page.
+<pre>
+2005Jul30 19:53:27 [94070] connect from [212.70.52.16] at ('212.70.52.16', 3192) EXTERNAL DYN
+2005Jul30 19:53:27 [94070] hello from winzip.com
+2005Jul30 19:53:27 [94070] mail from <dan@winzip.com> ()
+2005Jul30 19:53:27 [94070] REJECT: SPF fail 550 SPF fail:
+	see http://openspf.com/why.html?sender=dan@winzip.com&ip=212.70.52.16
+</pre>
+</td></tr>
+
+<tr><th>PERMERROR</th><td>
+Permanent errors were called "unknown", and are still show that way
+in the log.  The message is rejected.  Previously, we enabled "lax" parsing
+of the SPF record, but rejecting is better because it informs the
+sender about their problem.  The next milter version will 
+look for a local substitute SPF record (as for a missing SPF record)
+before rejecting.  This will inform the sender of their problem, but
+also let the receiver install a temporary workaround.
+<pre>
+2005Jul24 18:05:37 [2312] mail from <b-mihdbcgaacaa-becibijh-000-@msg.euxiphipops.com> () 
+2005Jul24 18:05:37 [2312] REJECT: SPF unknown 550 SPF Permanent Error:
+	include mechanism missing domain: include
+</pre>
+The SPF record for msg.euxiphipops.com looked like this at the time of the
+above error:
+<pre>
+msg.euxiphipops.com	TXT "v=spf1 mx ptr a include"
+</pre>
+</td></tr>
+
+<tr><th>TEMPERROR</th><td>
+Temporary errors result in a 451 "Try again later" response.  The sender
+should retry the message at a later time.
+<pre>
+2005Jul24 07:33:13 [29846] mail from <quickenloans@rate.quicken.com> ('SIZE=73775', 'BODY=8BITMIME')
+2005Jul24 07:33:43 [29846] TEMPFAIL: SPF error 450 SPF Temporary Error: DNS Timeout
+</pre>
+</td></tr>
+
+</table>
+
+</body>
+</html>

quarantine.txt

@@ -0,0 +1,26 @@
+Subject: DELIVERY STATUS (POSSIBLE SPAM)
+
+This is an automatically generated Delivery Status Notification.
+
+THIS IS A WARNING MESSAGE ONLY.
+
+YOU DO *NOT* NEED TO RESEND YOUR MESSAGE.
+
+Delivery to the following recipients has been delayed.
+
+       %(rcpt)s
+
+Subject: %(subject)s
+Received-SPF: %(spf_result)s
+
+A statistical analysis of your message has classified it as junk mail,
+and it has been quarantined.  Eventually, the recipients will review
+their quarantined mail and may notice your message.  If your message is
+important, please contact them via other means.  You may also try sending
+them a simple plain text message.
+
+If you need further assistance, please do not hesitate to contact me.
+
+Kind regards,
+
+postmaster@%(receiver)s

setup.cfg

@@ -1,5 +1,5 @@
 [bdist_rpm]
-python=python2
+python=python2.4
 doc_files=README NEWS TODO
 packager=Stuart D. Gathman <stuart@bmsi.com>
-release=2.4
+release=1

setup.py

@@ -1,6 +1,7 @@
 import os
 import sys
 from distutils.core import setup, Extension
+import Milter
 
 # FIXME: on some versions of sendmail, smutil is renamed to sm
 libs = ["milter", "smutil"]
@@ -12,7 +13,7 @@ if sys.version < '2.2.3':
   DistributionMetadata.classifiers = None
   DistributionMetadata.download_url = None
 
-setup(name = "milter", version = "0.8.0",
+setup(name = "milter", version = Milter.__version__,
 	description="Python interface to sendmail milter API",
 	long_description="""\
 This is a python extension module to enable python scripts to

softfail.txt

@@ -1,4 +1,4 @@
-Subject: SPF softfail (POSSIBLE FORGERY)
+Subject: SPF %(result)s (POSSIBLE FORGERY)
 
 This is an automatically generated Delivery Status Notification.
 
@@ -14,7 +14,9 @@ Subject: %(subject)s
 Received-SPF: %(spf_result)s
 
 Your sender policy indicated that the above email was likely forged and that
-feedback was desired.
+feedback was desired.  If you are sending from a foreign ISP,
+then you may need to follow your home ISPs instructions for configuring
+your outgoing mail server.
 
 If you need further assistance, please do not hesitate to contact me.
 

spf.py

@@ -1,5 +1,5 @@
 #!/usr/bin/env python
-"""SPF (Sender-Permitted From) implementation.
+"""SPF (Sender Policy Framework) implementation.
 
 Copyright (c) 2003, Terence Way
 Portions Copyright (c) 2004,2005 Stuart Gathman <stuart@bmsi.com>
@@ -19,10 +19,11 @@ AND THERE IS NO OBLIGATION WHATSOEVER TO PROVIDE MAINTENANCE,
 SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
 
 For more information about SPF, a tool against email forgery, see
-	http://spf.pobox.com
+	http://spf.pobox.com/
 
 For news, bugfixes, etc. visit the home page for this implementation at
 	http://www.wayforward.net/spf/
+	http://sourceforge.net/projects/pymilter/
 """
 
 # Changes:
@@ -46,6 +47,130 @@ For news, bugfixes, etc. visit the home page for this implementation at
 # Terrence is not responding to email.
 #
 # $Log$
+# Revision 1.13  2005/07/22 16:00:23  customdesigned
+# Limit CNAME chains independently of DNS lookup limit
+#
+# Revision 1.31  2005/07/22 02:11:50  customdesigned
+# Use dictionary to check for CNAME loops.  Check limit independently for
+# each top level name, just like for PTR.
+#
+# Revision 1.30  2005/07/21 20:07:31  customdesigned
+# Translate DNS error in DNSLookup.  This completely isolates DNS
+# dependencies to the DNSLookup method.
+#
+# Revision 1.29  2005/07/21 17:49:39  customdesigned
+# My best guess at what RFC intended for limiting CNAME loops.
+#
+# Revision 1.28  2005/07/21 17:37:08  customdesigned
+# Break out external DNSLookup method so that test suite can
+# duplicate CNAME loop bug.  Test zone data dictionary now
+# mirrors structure of real DNS.
+#
+# Revision 1.27  2005/07/21 15:26:06  customdesigned
+# First cut at updating docs.  Test suite is obsolete.
+#
+# Revision 1.26  2005/07/20 03:12:40  customdesigned
+# When not in strict mode, don't give PermErr for bad mechanism until
+# encountered during evaluation.
+#
+# Revision 1.25  2005/07/19 23:24:42  customdesigned
+# Validate all mechanisms before evaluating.
+#
+# Revision 1.24  2005/07/19 18:11:52  kitterma
+# Fix to change that compares type TXT and type SPF records.  Bug in the change
+# prevented records from being returned if it was published as TXT, but not SPF.
+#
+# Revision 1.23  2005/07/19 15:22:50  customdesigned
+# MX and PTR limits are MUST NOT check limits, and do not result in PermErr.
+# Also, check belongs in mx and ptr specific methods, not in dns() method.
+#
+# Revision 1.22  2005/07/19 05:02:29  customdesigned
+# FQDN test was broken.  Added test case.  Move FQDN test to after
+# macro expansion.
+#
+# Revision 1.21  2005/07/18 20:46:27  kitterma
+# Fixed reference problem in 1.20
+#
+# Revision 1.20  2005/07/18 20:21:47  kitterma
+# Change to dns_spf to go ahead and check for a type 99 (SPF) record even if a
+# TXT record is found and make sure if type SPF is present that they are
+# identical when using strict processing.
+#
+# Revision 1.19  2005/07/18 19:36:00  kitterma
+# Change to require at least one dot in a domain name.  Added PermError
+# description to indicate FQDN should be used.  This is a common error.
+#
+# Revision 1.18  2005/07/18 17:13:37  kitterma
+# Change macro processing to raise PermError on an unknown macro.
+# schlitt-spf-classic-02 para 8.1.  Change exp modifier processing to ignore
+# exp strings with syntax errors.  schlitt-spf-classic-02 para 6.2.
+#
+# Revision 1.17  2005/07/18 14:35:34  customdesigned
+# Remove debugging printf
+#
+# Revision 1.16  2005/07/18 14:34:14  customdesigned
+# Forgot to remove debugging print
+#
+# Revision 1.15  2005/07/15 21:17:36  customdesigned
+# Recursion limit raises AssertionError in strict mode, PermError otherwise.
+#
+# Revision 1.14  2005/07/15 20:34:11  customdesigned
+# Check whether DNS package already supports SPF before patching
+#
+# Revision 1.13  2005/07/15 20:01:22  customdesigned
+# Allow extended results for MX limit
+#
+# Revision 1.12  2005/07/15 19:12:09  customdesigned
+# Official IANA SPF record (type 99) support.
+#
+# Revision 1.11  2005/07/15 18:03:02  customdesigned
+# Fix unknown Received-SPF header broken by result changes
+#
+# Revision 1.10  2005/07/15 16:17:05  customdesigned
+# Start type99 support.
+# Make Scott's "/" support in parse_mechanism more elegant as requested.
+# Add test case for "/" support.
+#
+# Revision 1.9  2005/07/15 03:33:14  kitterma
+# Fix for bug 1238403 - Crash if non-CIDR / present.  Also added
+# validation check for valid IPv4 CIDR range.
+#
+# Revision 1.8  2005/07/14 04:18:01  customdesigned
+# Bring explanations and Received-SPF header into line with
+# the unknown=PermErr and error=TempErr convention.
+# Hope my case-sensitive mech fix doesn't clash with Scotts.
+#
+# Revision 1.7  2005/07/12 21:43:56  kitterma
+# Added processing to clarify some cases of unknown
+# qualifier errors (to distinguish between unknown qualifier and
+# unknown mechanism).
+# Also cleaned up comments from previous updates.
+#
+# Revision 1.6  2005/06/29 14:46:26  customdesigned
+# Distinguish trivial recursion from missing arg for diagnostic purposes.
+#
+# Revision 1.5  2005/06/28 17:48:56  customdesigned
+# Support extended processing results when a PermError should strictly occur.
+#
+# Revision 1.4  2005/06/22 15:54:54  customdesigned
+# Correct spelling.
+#
+# Revision 1.3  2005/06/22 00:08:24  kitterma
+# Changes from draft-mengwong overall DNS lookup and recursion
+# depth limits to draft-schlitt-spf-classic-02 DNS lookup, MX lookup, and
+# PTR lookup limits.  Recursion code is still present and functioning, but
+# it should be impossible to trip it.
+#
+# Revision 1.2  2005/06/21 16:46:09  kitterma
+# Updated definition of SPF, added reference to the sourceforge project site,
+# and deleted obsolete Microsoft Caller ID for Email XML translation routine.
+#
+# Revision 1.1.1.1  2005/06/20 19:57:32  customdesigned
+# Move Python SPF to its own module.
+#
+# Revision 1.5  2005/06/14 20:31:26  customdesigned
+# fix pychecker nits
+#
 # Revision 1.4  2005/06/02 04:18:55  customdesigned
 # Update copyright notices after reading article on /.
 #
@@ -144,135 +269,21 @@ import struct  # for pack() and unpack()
 import time    # for time()
 
 import DNS	# http://pydns.sourceforge.net
-import xml.sax
+if not hasattr(DNS.Type,'SPF'):
+  # patch in type99 support
+  DNS.Type.SPF = 99
+  DNS.Type.typemap[99] = 'SPF'
+  DNS.Lib.RRunpacker.getSPFdata = DNS.Lib.RRunpacker.getTXTdata
 
-# -------------------------------------------------------------------------
+def DNSLookup(name,qtype):
-# Convert a MS Caller-ID entry (XML) to a SPF entry
+  try:
-#
+    req = DNS.DnsRequest(name, qtype=qtype)
-# (c) 2004 by Ernesto Baschny
+    resp = req.req()
-# (c) 2004 Python version by Stuart Gathman
+    #resp.show()
-#
+    # key k: ('wayforward.net', 'A'), value v
-# Date: 2004-02-25
+    return [((a['name'], a['typename']), a['data']) for a in resp.answers]
-#
+  except DNS.DNSError,x:
-# A complete reverse translation (SPF -> CID) might be impossible, since
+    raise TempError,'DNS ' + str(x)
-# there are no ways to handle:
-# - PTR and EXISTS mechanism 
-# - MX mechanism with an different domain as argument
-# - macros
-# 
-# References:
-# http://www.microsoft.com/mscorp/twc/privacy/spam_callerid.mspx
-# http://spf.pobox.com/
-#
-# Known bugs:
-# - Currently it won't handle the exclusions provided in the A and R
-#   tags (prefix '!'). They will show up "as-is" in the SPF record
-# - I really haven't read the MS-CID specs in-depth, so there are probably
-#   other bugs too :)
-#
-# Ernesto Baschny <ernst@baschny.de>
-#
-
-class CIDParser(xml.sax.ContentHandler):
-  "Convert a MS Caller-ID entry (XML) to a SPF entry."
-
-  def __init__(self,q=None):
-    self.spf = []
-    self.action = '-all'
-    self.has_servers = None
-    self.spf_entry = None
-    if q:
-      self.spf_query = q
-    else:
-      self.spf_query = query(i='127.0.0.1', s='localhost', h='unknown')
-
-  def startElement(self,tag,attr):
-      if tag == 'm':
-	if self.has_servers != None and not self.has_servers:
-	  raise ValueError(
-    "Declared <noMailServers\> and later <m>, this CID entry is not valid."
-	  )
-	self.has_servers = True
-      elif tag == 'noMailServers':
-	if self.has_servers:
-	  raise ValueError(
-    "Declared <m> and later <noMailServers\>, this CID entry is not valid."
-	  )
-	self.has_servers = False
-      elif tag == 'ep':
-	if attr.has_key('testing') and attr.getValue('testing') == 'true':
-	  # A CID with 'testing' found:
-	  # From the MS-specs:
-	  #  "Documents in which such attribute is present with a true
-	  #  value SHOULD be entirely ignored (one should act as if the
-	  #  document were absent)"
-	  # From the SPF-specs:
-	  #  "Neutral (?): The SPF client MUST proceed as if a domain did
-	  #  not publish SPF data."
-	  # So we set SPF action to "neutral":
-	  self.action = '?all'
-      elif tag == 'mx':
-	  # The empty MX-tag, same as SPF's MX-mechanism
-	  self.spf.append('mx')
-      self.tag = tag
-
-  def characters(self,text):
-	tag = self.tag
-	# Remove starting and trailing spaces from text:
-	text = text.strip()
-
-	if tag == 'a' or tag == 'r':
-	    # The A and R tags from MS-CID are both handled by the 
-	    # ipv4/6-mechanisms from SPF:
-	    if text.find(':') < 0:
-	      mechanism = 'ip4'
-	    else:
-	      mechanism = 'ip6'
-	    self.spf.append(mechanism + ':' + text)
-	elif tag == 'indirect':
-	    # MS-CID's indirect is "sort of" the include from SPF:
-	    # Not really true, because the <indirect> tag from MS-CID also 
-	    # provides a fallback in case the included domain doesn't provide
-	    # _ep-records: The inbound MX-servers of the included domains
-	    # are added to the list of allowed outgoing mailservers for the
-	    # domain that declared the _ep-record with the <indirect> tag.
-	    # In SPF you would use the 'mx:domain' to handle this, but this
-	    # wouldn't depend on referred domain having or not SPF-records.
-	    cid_xml = self.cid_txt(text)
-	    if cid_xml:
-	      p = CIDParser()
-	      xml.sax.parseString(cid_xml,p)
-	      if p.has_servers != False:
-		self.spf += p.spf
-	    else:
-	      self.spf.append('mx:' + text)
-
-  def cid_txt(self,domain):
-    q = self.spf_query
-    domain='_ep.' + domain
-    a = q.dns_txt(domain)
-    if not a: return None
-    if a[0].lower().startswith('<ep ') and a[-1].lower().endswith('</ep>'):
-      return ''.join(a)
-    return None
-
-  def endElement(self,tag):
-      if tag == 'ep':
-	# This is the end... assemble what we've got
-	spf_entry = ['v=spf1']
-	if self.has_servers != False:
-	  spf_entry += self.spf
-	spf_entry.append(self.action)
-	self.spf_entry = ' '.join(spf_entry)
-
-  def spf_txt(self,cid_xml):
-    if not cid_xml.startswith('<'):
-      cid_xml = self.cid_txt(cid_xml)
-      if not cid_xml: return None
-    # Parse the beast. Any XML-problem will be reported by xlm.sax
-    self.spf_entry = None
-    xml.sax.parseString(cid_xml,self)
-    return self.spf_entry
 
 # 32-bit IPv4 address mask
 MASK = 0xFFFFFFFFL
@@ -286,6 +297,8 @@ RE_CHAR = re.compile(r'%(%|_|-|(\{[a-zA-Z][0-9]*r?[^\}]*\}))')
 # Regular expression to break up a macro expansion
 RE_ARGS = re.compile(r'([0-9]*)(r?)([^0-9a-zA-Z]*)')
 
+RE_CIDR = re.compile(r'/([1-9]|1[0-9]*|2[0-9]*|3[0-2]*)$')
+
 # Local parts and senders have their delimiters replaced with '.' during
 # macro expansion
 #
@@ -293,11 +306,12 @@ JOINERS = {'l': '.', 's': '.'}
 
 RESULTS = {'+': 'pass', '-': 'fail', '?': 'neutral', '~': 'softfail',
            'pass': 'pass', 'fail': 'fail', 'unknown': 'unknown',
-	   'neutral': 'neutral', 'softfail': 'softfail',
+	   'error': 'error', 'neutral': 'neutral', 'softfail': 'softfail',
 	   'none': 'none', 'deny': 'fail' }
 
 EXPLANATIONS = {'pass': 'sender SPF verified', 'fail': 'access denied',
-                'unknown': 'SPF unknown',
+                'unknown': 'permanent error in processing',
+                'error': 'temporary error in processing',
 		'softfail': 'domain in transition',
 		'neutral': 'access neither permitted nor denied',
 		'none': ''
@@ -316,22 +330,28 @@ except NameError:
 	def bool(x): return not not x
 # ...pre 2.2.1
 
-# standard default SPF record
+# standard default SPF record for best_guess
 DEFAULT_SPF = 'v=spf1 a/24 mx/24 ptr'
 
 # maximum DNS lookups allowed
-MAX_LOOKUP = 100
+MAX_LOOKUP = 10 #draft-schlitt-spf-classic-02 Para 10.1
+MAX_MX = 10 #draft-schlitt-spf-classic-02 Para 10.1
+MAX_PTR = 10 #draft-schlitt-spf-classic-02 Para 10.1
+MAX_CNAME = 10 # analogous interpretation to MAX_PTR
 MAX_RECURSION = 20
+ALL_MECHANISMS = ('a', 'mx', 'ptr', 'exists', 'include', 'ip4', 'ip6', 'all')
+COMMON_MISTAKES = { 'prt': 'ptr', 'ip': 'ip4', 'ipv4': 'ip4', 'ipv6': 'ip6' }
 
 class TempError(Exception):
 	"Temporary SPF error"
 
 class PermError(Exception):
 	"Permanent SPF error"
-	def __init__(self,msg,mech=None):
+	def __init__(self,msg,mech=None,ext=None):
 	  Exception.__init__(self,msg,mech)
 	  self.msg = msg
 	  self.mech = mech
+	  self.ext = ext
 	def __str__(self):
 	  if self.mech:
 	    return '%s: %s'%(self.msg,self.mech)
@@ -372,7 +392,7 @@ class query(object):
 
 	Also keeps cache: DNS cache.
 	"""
-	def __init__(self, i, s, h,local=None,receiver=None):
+	def __init__(self, i, s, h,local=None,receiver=None,strict=True):
 		self.i, self.s, self.h = i, s, h
 		if not s and h:
 		  self.s = 'postmaster@' + h
@@ -387,6 +407,8 @@ class query(object):
 		self.exps = dict(EXPLANATIONS)
 		self.local = local	# local policy
     		self.lookups = 0
+		# strict can be False, True, or 2 for harsh
+		self.strict = strict
 
 	def set_default_explanation(self,exp):
 		exps = self.exps
@@ -408,10 +430,47 @@ class query(object):
 
 	def check(self, spf=None):
 		"""
-	Returns (result, mta-status-code, explanation) where
+	Returns (result, mta-status-code, explanation) where result
-	result in ['fail', 'softfail', 'neutral' 'unknown', 'pass', 'error']
+	in ['fail', 'softfail', 'neutral' 'unknown', 'pass', 'error', 'none']
+
+	Examples:
+	>>> q = query(s='strong-bad@email.example.com',
+	...           h='mx.example.org', i='192.0.2.3')
+	>>> q.check(spf='v=spf1 ?all')
+	('neutral', 250, 'access neither permitted nor denied')
+
+	>>> q.check(spf='v=spf1 ip4:192.0.0.0/8 ?all moo')
+	('unknown', 550, 'SPF Permanent Error: Unknown mechanism found: moo')
+
+	>>> q.check(spf='v=spf1 =a ?all moo')
+	('unknown', 550, 'SPF Permanent Error: Unknown qualifier, IETF draft para 4.6.1, found in: =a')
+
+	>>> q.check(spf='v=spf1 ip4:192.0.0.0/8 ~all')
+	('pass', 250, 'sender SPF verified')
+
+	>>> q.strict = False
+	>>> q.check(spf='v=spf1 ip4:192.0.0.0/8 -all moo')
+	('pass', 250, 'sender SPF verified')
+
+	>>> q.check(spf='v=spf1 ip4:192.1.0.0/16 moo -all')
+	('unknown', 550, 'SPF Permanent Error: Unknown mechanism found: moo')
+
+	>>> q.check(spf='v=spf1 ip4:192.1.0.0/16 ~all')
+	('softfail', 250, 'domain in transition')
+
+	>>> q.check(spf='v=spf1 -ip4:192.1.0.0/6 ~all')
+	('fail', 550, 'access denied')
+
+	# Assumes DNS available
+	>>> q.check()
+	('none', 250, '')
 		"""
 		self.mech = []		# unknown mechanisms
+		# If not strict, certain PermErrors (mispelled
+		# mechanisms, strict processing limits exceeded)
+		# will continue processing.  However, the exception
+		# that strict processing would raise is saved here
+		self.perm_error = None
 		if self.i.startswith('127.'):
 			return ('pass', 250, 'local connections always pass')
 
@@ -421,31 +480,104 @@ class query(object):
 			    spf = self.dns_spf(self.d)
 			if self.local and spf:
 			    spf += ' ' + self.local
-			return self.check1(spf, self.d, 0)
+			rc = self.check1(spf, self.d, 0)
-		except DNS.DNSError,x:
+			if self.perm_error:
-			return ('error', 450, 'SPF DNS Error: ' + str(x))
+			  # extended processing succeeded, but strict failed
+			  self.perm_error.ext = rc
+			  raise self.perm_error
+			return rc
 		except TempError,x:
 			return ('error', 450, 'SPF Temporary Error: ' + str(x))
 		except PermError,x:
 		    self.prob = x.msg
+		    if x.mech:
 		      self.mech.append(x.mech)
 		    # Pre-Lentczner draft treats this as an unknown result
 		    # and equivalent to no SPF record.
-		    # return ('unknown', 550, 'SPF Permanent Error: ' + str(x))
+		    return ('unknown', 550, 'SPF Permanent Error: ' + str(x))
-		    return ('error', 550, 'SPF Permanent Error: ' + str(x))
 
 	def check1(self, spf, domain, recursion):
 		# spf rfc: 3.7 Processing Limits
 		#
 		if recursion > MAX_RECURSION:
-			self.prob =  'Too many levels of recursion'
+			# This should never happen in strict mode
-			return ('unknown', 250, 'SPF recursion limit exceeded')
+			# because of the other limits we check,
+			# so if it does, there is something wrong with
+			# our code.  It is not a PermError because there is not
+			# necessarily anything wrong with the SPF record.
+			if self.strict:
+			  raise AssertionError('Too many levels of recursion')
+			# As an extended result, however, it should be
+			# a PermError.
+			raise PermError('Too many levels of recursion')
 		try:
 			tmp, self.d = self.d, domain
 			return self.check0(spf,recursion)
 		finally:
 			self.d = tmp
 
+	def validate_mechanism(self,mech):
+		"""Parse and validate a mechanism.
+	Returns mech,m,arg,cidrlength,result
+
+	Examples:
+	>>> q = query(s='strong-bad@email.example.com',
+	...           h='mx.example.org', i='192.0.2.3')
+	>>> q.validate_mechanism('A')
+	('A', 'a', 'email.example.com', 32, 'pass')
+
+	>>> q.validate_mechanism('?mx:%{d}/27')
+	('?mx:%{d}/27', 'mx', 'email.example.com', 27, 'neutral')
+
+	>>> q.validate_mechanism('-mx::%%%_/.Clara.de/27')
+	('-mx::%%%_/.Clara.de/27', 'mx', ':% /.Clara.de', 27, 'fail')
+
+	>>> q.validate_mechanism('~exists:%{i}.%{s1}.100/86400.rate.%{d}')
+	('~exists:%{i}.%{s1}.100/86400.rate.%{d}', 'exists', '192.0.2.3.com.100/86400.rate.email.example.com', 32, 'softfail')
+		"""
+		# a mechanism
+		m, arg, cidrlength = parse_mechanism(mech, self.d)
+		# map '?' '+' or '-' to 'unknown' 'pass' or 'fail'
+		if m:
+		  result = RESULTS.get(m[0])
+		  if result:
+			# eat '?' '+' or '-'
+			m = m[1:]
+		  else:
+			# default pass
+			result = 'pass'
+		if m in COMMON_MISTAKES:
+		  try:
+		    raise PermError('Unknown mechanism found',mech)
+		  except PermError, x:
+		    if self.strict: raise
+		    m = COMMON_MISTAKES[m]
+		    if not self.perm_error:
+		      self.perm_error = x
+		  
+		if m in ('a', 'mx', 'ptr', 'exists', 'include'):
+		  arg = self.expand(arg)
+		  if not (0 < arg.find('.') < len(arg) - 1):
+		    raise PermError('Invalid domain found (use FQDN)',
+			  arg)
+		  if m == 'include':
+		    if arg == self.d:
+		      if mech != 'include':
+			raise PermError('include has trivial recursion',mech)
+		      raise PermError('include mechanism missing domain',mech)
+		  return mech,m,arg,cidrlength,result
+		if m in ALL_MECHANISMS:
+		  return mech,m,arg,cidrlength,result
+		try:
+		  if m[1:] in ALL_MECHANISMS:
+		    raise PermError(
+		      'Unknown qualifier, IETF draft para 4.6.1, found in',
+		      mech)
+		  raise PermError('Unknown mechanism found',mech)
+		except PermError, x:
+		  if self.strict: raise
+		  return mech,m,arg,cidrlength,x
+
 	def check0(self, spf,recursion):
 		"""Test this query information against SPF text.
 
@@ -468,17 +600,23 @@ class query(object):
 		# overridden with 'default=' modifier
 		#
 		default = 'neutral'
+		mechs = []
 
 		# Look for modifiers
 		#
-		for m in spf:
+		for mech in spf:
-			m = RE_MODIFIER.split(m)[1:]
+		    m = RE_MODIFIER.split(mech)[1:]
-			if len(m) != 2: continue
+		    if len(m) != 2:
+		      mechs.append(self.validate_mechanism(mech))
+		      continue
 
 		    if m[0] == 'exp':
-				exps['fail'] = exps['unknown'] = \
+                        try:
-					self.get_explanation(m[1])
+                            self.set_default_explanation(self.get_explanation(m[1]))
+                        except PermError:
+                            pass
 		    elif m[0] == 'redirect':
+		        self.check_lookups()
 			redirect = self.expand(m[1])
 		    elif m[0] == 'default':
 			# default=- is the same as default=fail
@@ -486,77 +624,65 @@ class query(object):
 
 		    # spf rfc: 3.6 Unrecognized Mechanisms and Modifiers
 
-		# Look for mechanisms
+		# Evaluate mechanisms
 		#
-		for mech in spf:
+		for mech,m,arg,cidrlength,result in mechs:
-		    if RE_MODIFIER.match(mech): continue
-		    m, arg, cidrlength = parse_mechanism(mech, self.d)
-
-		    # map '?' '+' or '-' to 'unknown' 'pass' or 'fail'
-		    if m:
-		      result = RESULTS.get(m[0])
-		      if result:
-			    # eat '?' '+' or '-'
-			    m = m[1:]
-		      else:
-			    # default pass
-			    result = 'pass'
-
-		    if m in ['a', 'mx', 'ptr', 'prt', 'exists', 'include']:
-			    arg = self.expand(arg)
 
 		    if m == 'include':
-		      if arg != self.d:
+		      self.check_lookups()
 		      res,code,txt = self.check1(self.dns_spf(arg),
 					arg, recursion + 1)
 		      if res == 'pass':
 			break
 		      if res == 'none':
+			try:
+			  if self.strict or not self.perm_error:
 			    raise PermError(
 			      'No valid SPF record for included domain: %s'%arg,
 			      mech)
+			except PermError,x:
+			  if self.strict:
+			    raise x
+			  self.perm_error = x
 		      continue
-		      else:
-			raise PermError('include mechanism missing domain',mech)
 		    elif m == 'all':
 			    break
 
 		    elif m == 'exists':
+		        self.check_lookups()
 			if len(self.dns_a(arg)) > 0:
 				break
 
 		    elif m == 'a':
-			    if cidrmatch(self.i, self.dns_a(arg),
+		        self.check_lookups()
-					 cidrlength):
+			if cidrmatch(self.i, self.dns_a(arg), cidrlength):
 			      break
 
 		    elif m == 'mx':
-			    if cidrmatch(self.i, self.dns_mx(arg),
+		        self.check_lookups()
-					 cidrlength):
+			if cidrmatch(self.i, self.dns_mx(arg), cidrlength):
 			      break
 
-		    elif m in ('ip4', 'ipv4', 'ip') and arg != self.d:
+		    elif m == 'ip4' and arg != self.d:
 			try:
 			    if cidrmatch(self.i, [arg], cidrlength):
 				break
 			except socket.error:
 			    raise PermError('syntax error',mech)
 			    
-		    elif m in ('ip6', 'ipv6'):
+		    elif m == 'ip6':
 			# Until we support IPV6, we should never
 			# get an IPv6 connection.  So this mech
 			# will never match.
 			pass
 
-		    elif m in ('ptr', 'prt'):
+		    elif m == 'ptr':
-			    if domainmatch(self.validated_ptrs(self.i),
+		        self.check_lookups()
-					   arg):
+			if domainmatch(self.validated_ptrs(self.i), arg):
 				break
 
 		    else:
-		      # unknown mechanisms cause immediate unknown
+		      raise result
-		      # abort results
-		      raise PermError('Unknown mechanism found',mech)
 		else:
 		    # no matches
 		    if redirect:
@@ -570,6 +696,17 @@ class query(object):
 		else:
 		    return (result, 250, exps[result])
 
+	def check_lookups(self):
+	    self.lookups = self.lookups + 1
+	    if self.lookups > MAX_LOOKUP:
+	      try:
+		if self.strict or not self.perm_error:
+		  raise PermError('Too many DNS lookups')
+	      except PermError,x:
+		if self.strict or self.lookups > MAX_LOOKUP*4:
+		  raise x
+		self.perm_error = x
+
 	def get_explanation(self, spec):
 		"""Expand an explanation."""
 		if spec:
@@ -661,8 +798,10 @@ class query(object):
 				letter = macro[2].lower()
 				if letter == 'p':
 					self.getp()
-				expansion = getattr(self, letter, '')
+				expansion = getattr(self, letter, 'Macro Error')
 				if expansion:
+                                        if expansion == 'Macro Error':
+                                            raise PermError('Unknown Macro Encountered') 
 					result += expand_one(expansion,
 						macro[3:-1],
 					        JOINERS.get(letter))
@@ -675,23 +814,27 @@ class query(object):
 		name.  Returns None if not found, or if more than one record
 		is found.
 		"""
+		# for performance, check for most common case of TXT first
 		a = [t for t in self.dns_txt(domain) if t.startswith('v=spf1')]
-		if not a:
+		if len(a) == 1 and self.strict < 2:
-		  if DELEGATE:
+		    return a[0]   			
+		# check official SPF type first when it becomes more popular
+		b = [t for t in self.dns_99(domain) if t.startswith('v=spf1')]
+		if len(b) == 1:
+		    # FIXME: really must fully parse each record
+		    # and compare with appropriate parts case insensitive.
+		    if self.strict >= 2 and len(a) == 1 and a[0] != b[0]:
+		        raise PermError(
+'v=spf1 records of both type TXT and SPF (type 99) present, but not identical')
+		    return b[0]
+		if len(a) == 1:
+		    return a[0]	# return TXT if SPF wasn't found
+		if DELEGATE:	# use local record if neither found
 		    a = [t
 		      for t in self.dns_txt(domain+'._spf.'+DELEGATE)
 			if t.startswith('v=spf1')
 		    ]
-		  if not a:
+		    if len(a) == 1: return a[0]
-		    # No SPF record: convert and return CID if present
-		    p = CIDParser(q=self)
-		    try:
-		      return p.spf_txt(domain)
-		    except xml.sax._exceptions.SAXParseException:
-		      raise PermError("Caller-ID parse error",domain)
-
-		if len(a) == 1:
-			return a[0]
 		return None
 
 	def dns_txt(self, domainname):
@@ -699,12 +842,23 @@ class query(object):
 		if domainname:
 		  return [''.join(a) for a in self.dns(domainname, 'TXT')]
 		return []
+	def dns_99(self, domainname):
+		"Get a list of type SPF=99 records for a domain name."
+		if domainname:
+		  return [''.join(a) for a in self.dns(domainname, 'SPF')]
+		return []
 
 	def dns_mx(self, domainname):
 		"""Get a list of IP addresses for all MX exchanges for a
 		domain name.
 		"""
-		return [a for mx in self.dns(domainname, 'MX') \
+# draft-schlitt-spf-classic-02 section 5.4 "mx"
+# To prevent DoS attacks, more than 10 MX names MUST NOT be looked up
+		if self.strict:
+		  max = MAX_MX
+		else:
+		  max = MAX_MX * 4
+		return [a for mx in self.dns(domainname, 'MX')[:max] \
 		          for a in self.dns_a(mx[1])]
 
 	def dns_a(self, domainname):
@@ -719,13 +873,18 @@ class query(object):
 		"""Figure out the validated PTR domain names for a given IP
 		address.
 		"""
-		return [p for p in self.dns_ptr(i) if i in self.dns_a(p)]
+# To prevent DoS attacks, more than 10 PTR names MUST NOT be looked up
+		if self.strict:
+		  max = MAX_PTR
+		else:
+		  max = MAX_PTR * 4
+		return [p for p in self.dns_ptr(i)[:max] if i in self.dns_a(p)]
 
 	def dns_ptr(self, i):
 		"""Get a list of domain names for an IP address."""
 		return self.dns(reverse_dots(i) + ".in-addr.arpa", 'PTR')
 
-	def dns(self, name, qtype):
+	def dns(self, name, qtype, cnames=None):
 		"""DNS query.
 
 		If the result is in cache, return that.  Otherwise pull the
@@ -739,26 +898,29 @@ class query(object):
 		pre: qtype in ['A', 'AAAA', 'MX', 'PTR', 'TXT', 'SPF']
 		post: isinstance(__return__, types.ListType)
 		"""
-		self.lookups += 1
-		if self.lookups > MAX_LOOKUP:
-			raise PermError('Too many DNS lookups')
 		result = self.cache.get( (name, qtype) )
 		cname = None
 		if not result:
-			req = DNS.DnsRequest(name, qtype=qtype)
+			for k,v in DNSLookup(name,qtype):
-			resp = req.req()
-			for a in resp.answers:
-				# key k: ('wayforward.net', 'A'), value v
-				k, v = (a['name'], a['typename']), a['data']
 			    if k == (name, 'CNAME'):
 				cname = v
 			    self.cache.setdefault(k, []).append(v)
 			result = self.cache.get( (name, qtype), [])
 		if not result and cname:
-			result = self.dns(cname, qtype)
+			if not cnames:
+			  cnames = {}
+			elif len(cnames) >= MAX_CNAME:
+			  raise PermError(
+			    'Length of CNAME chain exceeds %d' % MAX_CNAME)
+			cnames[name] = cname
+			if cname in cnames:
+			  raise PermError,'CNAME loop'
+			result = self.dns(cname, qtype, cnames=cnames)
 		return result
 
-	def get_header(self,res,receiver):
+	def get_header(self,res,receiver=None):
+	  if not receiver:
+	    receiver = self.r
 	  if res in ('pass','fail','softfail'):
 	    return '%s (%s: %s) client-ip=%s; envelope-from=%s; helo=%s;' % (
 	  	res,receiver,self.get_header_comment(res),self.i,
@@ -789,10 +951,10 @@ class query(object):
 		    	% (self.i,sender)
 		    #"%s does not designate permitted sender hosts" % sender
 		elif res == 'unknown': return \
-		    "error in processing during lookup of domain of %s: %s" \
+		    "permanent error in processing domain of %s: %s" \
 		    	% (sender, self.prob)
 		elif res == 'error': return \
-		    "error in processing during lookup of %s" % sender
+		    "temporary error in processing during lookup of %s" % sender
 		elif res == 'fail': return \
 		    "domain of %s does not designate %s as permitted sender" \
 			% (sender,self.i)
@@ -836,20 +998,32 @@ def parse_mechanism(str, d):
 	>>> parse_mechanism('a/24', 'foo.com')
 	('a', 'foo.com', 24)
 
-	>>> parse_mechanism('a:bar.com/16', 'foo.com')
+	>>> parse_mechanism('A:foo:bar.com/16', 'foo.com')
-	('a', 'bar.com', 16)
+	('a', 'foo:bar.com', 16)
+
+	>>> parse_mechanism('-exists:%{i}.%{s1}.100/86400.rate.%{d}','foo.com')
+	('-exists', '%{i}.%{s1}.100/86400.rate.%{d}', 32)
+
+	>>> parse_mechanism('mx::%%%_/.Claranet.de/27','foo.com')
+	('mx', ':%%%_/.Claranet.de', 27)
+
+	>>> parse_mechanism('mx:%{d}/27','foo.com')
+	('mx', '%{d}', 27)
+
+	>>> parse_mechanism('iP4:192.0.0.0/8','foo.com')
+	('ip4', '192.0.0.0', 8)
 	"""
-	a = str.split('/')
+	a = RE_CIDR.split(str)
-	if len(a) == 2:
+	if len(a) == 3:
 		a, port = a[0], int(a[1])
 	else:
 		a, port = str, 32
 
-	b = a.split(':')
+	b = a.split(':',1)
 	if len(b) == 2:
-		return b[0], b[1], port
+		return b[0].lower(), b[1], port
 	else:
-		return a, d, port
+		return a.lower(), d, port
 
 def reverse_dots(name):
 	"""Reverse dotted IP addresses or domain names.
@@ -1033,7 +1207,9 @@ if __name__ == '__main__':
 			receiver=socket.gethostname())
 	elif len(sys.argv) == 5:
 		i, s, h = sys.argv[2:]
-		q = query(i=i, s=s, h=h, receiver=socket.gethostname())
+		q = query(i=i, s=s, h=h, receiver=socket.gethostname(),
+			strict=False)
 		print q.check(sys.argv[1])
+		if q.perm_error: print q.perm_error.ext
 	else:
 		print USAGE

strike3.txt

@@ -23,7 +23,7 @@ SMTP (email) servers to prevent criminals from forging your
 domain.  The simplest step is usually to publish an SPF record
 with your Sender Policy.  
 
-For more information, see: http://spfhelp.net
+For more information, see: http://openspf.com
 
 I hate to annoy you with a DSN (Delivery Status
 Notification) from a possibly forged email, but since you

test/zip2

@@ -0,0 +1,49 @@
+From paulp@go2net.com  Wed Jun  1 22:35:12 2005
+Return-Path: <paulp@go2net.com>
+Received: from mail.bmsi.com (spidey.bmsi.com [192.168.9.81])
+	by bmsred.bmsi.com (8.13.1/8.12.10) with ESMTP id j522ZCQg014058
+	for <stuart@bmsred.bmsi.com>; Wed, 1 Jun 2005 22:35:12 -0400
+Received: from 127.0.0.1 ([220.117.92.241])
+	by mail.bmsi.com (8.13.1/8.13.1) with ESMTP id j522Ynjm028604
+	for stuart@bmsi.com; Wed, 1 Jun 2005 22:34:51 -0400
+Message-Id: <200506020234.j522Ynjm028604@mail.bmsi.com>
+SUBJECT: urgent
+FROM: paulp@go2net.com
+TO: stuart@bmsi.com
+DATE: [[ ¸ñ, 02 6 2005 ¿ÀÀü 11:34:47 ]]
+MIME-Version: 1.0
+Content-Type: multipart/mixed; boundary="--------bound--"
+X-DSpam-Score: 0.081200
+Received-SPF: neutral (mail.bmsi.com: guessing: 220.117.92.241 is neither permitted nor denied by domain of go2net.com)
+Status: RO
+X-Status: 
+X-Keywords: NonJunk         
+
+----------bound--
+Content-Type: text/plain; charset=us-ascii
+Content-Transfer-Encoding: 7bit
+
+Hi 
+
+Sorry, I forgot to send an important
+document to you in that last email. I had an important phone call.
+Please checkout attached doc file when you have a moment.
+
+Best Regards
+
+<!DSPAM:1043AE6B6492860536935410>
+
+
+----------bound--
+Content-Type: application/octet-stream;
+	name="Readme.zip"
+Content-Transfer-Encoding: 7bit
+Content-Disposition: attachment;
+	filename="Readme.zip"
+
+
+----------bound--
+
+
+----------bound----
+

test/zip3

@@ -0,0 +1,51 @@
+From paulp@go2net.com  Wed Jun  1 22:35:12 2005
+Return-Path: <paulp@go2net.com>
+Received: from mail.bmsi.com (spidey.bmsi.com [192.168.9.81])
+	by bmsred.bmsi.com (8.13.1/8.12.10) with ESMTP id j522ZCQg014058
+	for <stuart@bmsred.bmsi.com>; Wed, 1 Jun 2005 22:35:12 -0400
+Received: from 127.0.0.1 ([220.117.92.241])
+	by mail.bmsi.com (8.13.1/8.13.1) with ESMTP id j522Ynjm028604
+	for stuart@bmsi.com; Wed, 1 Jun 2005 22:34:51 -0400
+Message-Id: <200506020234.j522Ynjm028604@mail.bmsi.com>
+SUBJECT: urgent
+FROM: paulp@go2net.com
+TO: stuart@bmsi.com
+DATE: [[ ¸ñ, 02 6 2005 ¿ÀÀü 11:34:47 ]]
+MIME-Version: 1.0
+Content-Type: multipart/mixed; boundary="--------bound--"
+X-DSpam-Score: 0.081200
+Received-SPF: neutral (mail.bmsi.com: guessing: 220.117.92.241 is neither permitted nor denied by domain of go2net.com)
+Status: RO
+X-Status: 
+X-Keywords: NonJunk         
+
+----------bound--
+Content-Type: text/plain; charset=us-ascii
+Content-Transfer-Encoding: 7bit
+
+Hi 
+
+Sorry, I forgot to send an important
+document to you in that last email. I had an important phone call.
+Please checkout attached doc file when you have a moment.
+
+Best Regards
+
+<!DSPAM:1043AE6B6492860536935410>
+
+
+----------bound--
+Content-Type: application/x-msdownload; name="zip.zip"
+Content-Transfer-Encoding: base64
+Content-Disposition: attachment; filename="zip.zip"
+
+USsDBAoBAAAAADVVwjLaV2nEGgAAABoAAAAzABUAemlwLmRvYyAgICAgICAgICAgICAgICAg
+ICAgICAgICAgICAgICAgICAgICAgICAuZXhlVVQJAAOmGp9CphqfQlV4BACGA2UAVGhpcyBw
+cm9ncmFtIHdhcyBhIHZpcnVzLgpQSwECFwMKAAAAAAA1VcIy2ldpxBoAAAAaAAAAMwANAAAA
+AAABAAAAtIEAAAAAemlwLmRvYyAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
+ICAgICAuZXhlVVQFAAOmGp9CVXgAAFBLBQYAAAAAAQABAG4AAACAAAAAAAA=
+----------bound--
+
+
+----------bound----
+

test/ziploop

@@ -0,0 +1,47 @@
+From ttaie1@thfalcon.com Thu Jun 16 10:23:13 2005
+Received: from thfalcon.com (unknown [202.90.113.150])
+	by thfalcon.com (Postfix) with ESMTP id 32F0DD819C
+	for <stuart@bmsi.com>; Thu, 16 Jun 2005 15:42:08 +0700 (ICT)
+From: ttaie1@thfalcon.com
+To: stuart@bmsi.com
+Subject: Returned mail: see transcript for details
+Date: Thu, 16 Jun 2005 15:50:10 +0700
+MIME-Version: 1.0
+Content-Type: multipart/mixed;
+	boundary="----=_NextPart_000_0014_E4E04420.5619685C"
+X-Priority: 3
+X-MSMail-Priority: Normal
+X-Mailer: Microsoft Outlook Express 6.00.2600.0000
+X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
+Message-Id: <20050616084208.32F0DD819C@thfalcon.com>
+Received-SPF: pass (mail.bmsi.com: guessing: domain of thfalcon.com designates 203.147.3.44 as permitted sender) client-ip=203.147.3.44; envelope-from=ttaie1@thfalcon.com; helo=thfalcon.com;
+
+This is a multi-part message in MIME format.
+
+------=_NextPart_000_0014_E4E04420.5619685C
+Content-Type: text/plain;
+	charset=us-ascii
+Content-Transfer-Encoding: 7bit
+
+Message could not be delivered
+
+
+------=_NextPart_000_0014_E4E04420.5619685C
+Content-Type: application/octet-stream;
+	name="stuart@bmsi.com.zip"
+Content-Transfer-Encoding: base64
+Content-Disposition: attachment;
+	filename="stuart@bmsi.com.zip"
+
+UEsDBAoAAAAAAM6r0DL7SfbCBAEAAAQBAAAFABUAdC56aXBVVAkAA7MnskK4J7JCVXgEAIYD
+ZQBQSwMECgAAAAAANVXCMtpXacQaAAAAGgAAADMAFQB6aXAuZG9jICAgICAgICAgICAgICAg
+ICAgICAgICAgICAgICAgICAgICAgICAgIC5leGVVVAkAA6Yan0KmGp9CVXgEAIYDZQBUaGlz
+IHByb2dyYW0gd2FzIGEgdmlydXMuClBLAQIXAwoAAAAAADVVwjLaV2nEGgAAABoAAAAzAA0A
+AAAAAAEAAAC0gQAAAAB6aXAuZG9jICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
+ICAgICAgIC5leGVVVAUAA6Yan0JVeAAAUEsFBgAAAAABAAEAbgAAAIAAAAAAAFBLAQIXAwoA
+AAAAAM6r0DL7SfbCBAEAAAQBAAAFAA0AAAAAAAAAAAC0gQAAAAB0LnppcFVUBQADsyeyQlV4
+AABQSwUGAAAAAAEAAQBAAAAAPAEAAAAA
+
+------=_NextPart_000_0014_E4E04420.5619685C--
+
+

testbms.py

@@ -1,4 +1,5 @@
 import unittest
+import doctest
 import Milter
 import bms
 import mime
@@ -22,7 +23,7 @@ class TestMilter(bms.bmsMilter):
 
   def getsymval(self,name):
     if name == 'j': return 'test.milter.org'
-    return bms.bmsMilter.getsymval(self,name)
+    return ''
 
   def replacebody(self,chunk):
     if self._body:
@@ -284,7 +285,10 @@ class BMSMilterTestCase(unittest.TestCase):
 #    self.failUnless(rc == Milter.REJECT)
 #    milter.close();
 
-def suite(): return unittest.makeSuite(BMSMilterTestCase,'test')
+def suite(): 
+  s = unittest.makeSuite(BMSMilterTestCase,'test')
+  s.addTest(doctest.DocTestSuite(bms))
+  return s
 
 if __name__ == '__main__':
   if len(sys.argv) > 1:
@@ -296,4 +300,5 @@ if __name__ == '__main__':
       fp = milter._body
       sys.stdout.write(fp.getvalue())
   else:
-    unittest.main()
+    #unittest.main()
+    unittest.TextTestRunner().run(suite())

testmime.py

@@ -1,4 +1,10 @@
 # $Log$
+# Revision 1.3  2005/06/17 01:49:39  customdesigned
+# Handle zip within zip.
+#
+# Revision 1.2  2005/06/02 15:00:17  customdesigned
+# Configure banned extensions.  Scan zipfile option with test case.
+#
 # Revision 1.1.1.2  2005/05/31 18:23:49  customdesigned
 # Development changes since 0.7.2
 #
@@ -126,10 +132,19 @@ class MimeTestCase(unittest.TestCase):
     self.failUnless(name == "Jim&Girlz.jpg","name=%s"%name)
 
   def testZip(self,vname="zip1",fname='zip.zip'):
-    self.testDefang('zip1',1,'zip.zip')
+    self.testDefang(vname,1,'zip.zip')
+    # test scan_zip flag
     msg = mime.message_from_file(open('test/'+vname,"r"))
     mime.defang(msg,scan_zip=False)
     self.failIf(msg.ismodified())
+    # test ignoring empty zip (often found in DSNs)
+    msg = mime.message_from_file(open('test/zip2','r'))
+    mime.defang(msg,scan_zip=True)
+    self.failIf(msg.ismodified())
+    # test corrupt zip (often an EXE named as a ZIP)
+    self.testDefang('zip3',1,'zip.zip')
+    # test zip within zip
+    self.testDefang('ziploop',1,'stuart@bmsi.com.zip')
 
   def testHTML(self,fname=""):
     result = StringIO.StringIO()
@@ -153,3 +168,5 @@ if __name__ == '__main__':
     for fname in sys.argv[1:]:
       fp = open(fname,'r')
       msg = mime.message_from_file(fp)
+      mime.defang(msg,scan_zip=True)
+      print msg.as_string()