From 4c7c76fca48e03630be821f516c0faae5e60c58a Mon Sep 17 00:00:00 2001
From: "Stuart D. Gathman" <stuart@gathman.org>
Date: Wed, 17 Jun 2020 13:55:26 -0400
Subject: [PATCH] First cut at encoding error decorator

---
 Milter/__init__.py | 24 +++++++++++++++++++-----
 sample.py          |  4 +++-
 2 files changed, 22 insertions(+), 6 deletions(-)

diff --git a/Milter/__init__.py b/Milter/__init__.py
index 0ec4b38..f0d3347 100755
--- a/Milter/__init__.py
+++ b/Milter/__init__.py
@@ -181,15 +181,26 @@ def noreply(func):
   wrapper.milter_protocol = nr_mask
   return wrapper
 
+## Function decorator to set encoding error strategy.
+# Current RFCs define UTF-8 as the standard encoding for SMTP
+# envelope and header fields.  By default, Milter.Base decodes 
+# envelope and header values with errors='surrogateescape'.
+# This decorator can change the error strategy to, e.g., 'ignore' or 'replace'.
+def encodingerror(strategy):
+  def setstrategy(func):
+    func.error_strategy = strategy
+    return func
+  return setstrategy
+
 ## Function decorator to set macros used in a callback.
 # By default, the MTA sends all macros defined for a callback.
 # If some or all of these are unused, the bandwidth can be saved
 # by listing the ones that are used.
 # @since 1.0.2
 def symlist(*syms):
-  if len(syms) > 5:
-    raise ValueError('@symlist limited to 5 macros by MTA: '+func.__name__)
   def setsyms(func):
+    if len(syms) > 5:
+      raise ValueError('@symlist limited to 5 macros by MTA: '+func.__name__)
     if func.__name__ not in MACRO_CALLBACKS:
       raise ValueError('@symlist applied to non-symlist method: '+func.__name__)
     func._symlist = syms
@@ -328,7 +339,8 @@ class Base(object):
   # or trap utf-8 conversion exception, etc.
   def envfrom_bytes(self,*b):
     try:
-      s = (v.decode(encoding='utf-8',errors='surrogateescape') for v in b)
+      e = getattr(self.envfrom,'error_strategy','surrogateescape')
+      s = (v.decode(encoding='utf-8',errors=e) for v in b)
     except UnicodeDecodeError: s = b
     return self.envfrom(fld,*s)
   ## Called when the SMTP client says MAIL FROM. Called by the
@@ -349,7 +361,8 @@ class Base(object):
   # or trap utf-8 conversion exception, etc.
   def envrcpt_bytes(self,*b):
     try:
-      s = (v.decode(encoding='utf-8',errors='surrogateescape') for v in b)
+      e = getattr(self.envrcpt,'error_strategy','surrogateescape')
+      s = (v.decode(encoding='utf-8',errors=e) for v in b)
     except UnicodeDecodeError: s = b
     return self.envrcpt(fld,*s)
   ## Called when the SMTP client says RCPT TO. Called by the
@@ -377,7 +390,8 @@ class Base(object):
   # to pass bytes to @link #header the header callback @endlink instead.
   def header_bytes(self,fld,val):
     try:
-      s = val.decode(encoding='utf-8',errors='surrogateescape')
+      e = getattr(self.header,'error_strategy','surrogateescape')
+      s = val.decode(encoding='utf-8',errors=e)
     except UnicodeDecodeError: s = val
     return self.header(fld,s)
   ## Called for each header field in the message body.
diff --git a/sample.py b/sample.py
index 92d649e..6476ba9 100644
--- a/sample.py
+++ b/sample.py
@@ -67,13 +67,15 @@ class sampleMilter(Milter.Milter):
     self.log("rcpt to",to,str)
     return Milter.CONTINUE
 
+  @Milter.encodingerror('replace')
   def header(self,name,val):
     lname = name.lower()
     if lname == 'subject':
 
       # even if we wanted the Taiwanese spam, we can't read Chinese
       # (delete if you read chinese mail)
-      print('val=',val.encode(errors='surrogateescape'))
+      #print('val=',val.encode(errors='surrogateescape'))
+      print('val=',val)
       if val.startswith('=?big5') or val.startswith('=?ISO-2022-JP'):
         self.log('REJECT: %s: %s' % (name,val))
 	#self.setreply('550','','Go away spammer')