[Mailman-Developers] Two bugs in email 0.96 that you may want to fix sooner rather than later

[Dan Mick]

I've reported these to Barry, and so real fixes are on the way
eventually; however, they're both annoying enough that I thought
I might drop them here too.  AFAIK, these bugs are still present
in email 1.0 as well.

1) MIME messages with spaces around = sign in parameters of Content-Type
aren't handled

I think this is a non-compliant message, but I get a bunch of bounces
(mostly from Hotmail..big shock that *they'd* be noncompliant, huh?)
that have Content-Type headers of the form:

Content-Type: Multipart/mixed;
	boundary = "CPIMSSMTPC06p5f3tG"
	
The spaces around the = are the problem.  This fixes them, I 
think without doing any other harm.

(Message.py in the email package):

--- Message.py  Mon Dec  3 18:13:50 2001
***************
*** 304,309 ****
--- 304,311 ----
          for p in paramre.split(value):
              try:
                  name, val = p.split('=', 1)
+                 name = name.strip()
+                 val = val.strip()
              except ValueError:
                  # Must have been a bare attribute
                  name = p
                  
Without this, qfiles/shunt fills up full of these messages.

2) MIME messages with long header lines have those headers corrupted
on regeneration of messages.

I'm pretty sure Bob Puff noticed this, but I can't find the thread anymore.
Anyway, you end up with Content-Type (typically) that's been weirdly
formatted into things like

Content-Type: multipart/report;;
		Content-Type: multipart/report;;boundary=foo
		
instead of the proper

Content-Type: multipart/report;
		boundary=foo
		
because of some full-part/subpart variable name confusion in
email's Generator.py:

Printing long header lines that require split_header() in Generator.py
can go wrong because of confusion between the original line and
the sub-lines.  Here's a unified diff.  For the most part, it's
replacing 'text' with 'line' in the right places.


--- /export/home/dmick/pymod/Generator.py       Mon Dec  3 17:59:58 2001
+++ Generator.py        Mon Dec  3 18:57:02 2001
@@ -175,22 +175,22 @@
                 rtn.append(line)
                 SEMINLTAB.join(rtn)
             else:
-                oldlen = len(text)
+                oldlen = len(line)
                 # Try to break the line on semicolons, but if that doesn't
                 # work, try to split on folding whitespace.
-                while len(text) > maxheaderlen:
-                    i = text.rfind(';', 0, maxheaderlen)
+                while len(line) > maxheaderlen:
+                    i = line.rfind(';', 0, maxheaderlen)
                     if i < 0:
                         break
-                    rtn.append(text[:i])
-                    text = text[i+1:].lstrip()
-                if len(text) <> oldlen:
+                    rtn.append(line[:i])
+                    line = line[i+1:].lstrip()
+                if len(line) <> oldlen:
                     # Splitting on semis worked
-                    rtn.append(text)
+                    rtn.append(line)
                     return SEMINLTAB.join(rtn)
                 # Splitting on semis didn't help, so try to split on
                 # whitespace.
-                parts = re.split(r'(\s+)', text)
+                parts = re.split(r'(\s+)', line)
                 # Watch out though for "Header: longnonsplittableline"
                 if parts[0].endswith(':') and len(parts) == 3:
                     return text