A workaround for treating long address splitted into lines

[naoki-fukuta]

---Summary:---
 Following is a workaround for a problem when
processing very long email address field that is
splitted into two or more lines.

I tested it on Mailman 2.1.8.

 Please evaluate this patch and feedback to the latest
build.

I appreciate if this is a duplicate of already reported
one.

---Problem Description:---
 The mail address parser program
(pythonlib/email/_parseaddr.py) has a problem when the
field is splitted into two or more lines by using
continuations(eg. "\r\n " or "\r\n\t").

For example, following "Cc:" field can be splitted into
 lines when the mail is sent to the server.

Cc:
=?ISO-2022-JP?B?IktCU0UbJEI4JjVmMnEbKEIgGyRCNDQ7dkNEGyhC?=
 =?ISO-2022-JP?B?Ig==?= <email address hidden>,
 =?ISO-2022-JP?B?IhskQkVFO1I+cEpzREw/LjNYMnEbKEIgGyRCOCY1ZhsoQg==?=

 =?ISO-2022-JP?B?GyRCMnFIL0k9Pz05fiU3JTklRiVgGyhCIg==?=
<email address hidden>

The main problem is there are no actual email address
at line 1 and 3.

Current code in _parseaddr.py expects there is at least
one valid email address in a line. So what will happen?
See below. When posting the mail to members, two
addresses are newly made, with a default domain name
(@mailman.example3.org).

Cc: =?ISO-2022-JP?B?Ig==?= <email address hidden>,
 =?ISO-2022-JP?B?IktCU0UbJEI4JjVmMnEbKEIgGyRCNDQ7dkNEGyhC?=@mailman.example3.org,
 =?ISO-2022-JP?B?IhskQkVFO1I+cEpzREw/LjNYMnEbKEIgGyRCOCY1ZhsoQg==?=@mailman.example3.org,
 =?ISO-2022-JP?B?GyRCMnFIL0k9Pz05fiU3JTklRiVgGyhCIg==?=
<email address hidden>

---Patch---
I made a workaroud patch by adding a darty code into
getphraselist() method.

----------------------------------------------------
    def getphraselist(self):
        """Parse a sequence of RFC 2822 phrases.

        A phrase is a sequence of words, which are in
turn either RFC 2822
        atoms or quoted-strings. Phrases are
canonicalized by squeezing all
        runs of continuous whitespace into one space.
        """
        plist = []

        while self.pos < len(self.field):
            if self.field[self.pos] in self.LWS:
                self.pos += 1
            elif self.field[self.pos] == '"':
                plist.append(self.getquote())
            elif self.field[self.pos] == '(':
                self.commentlist.append(self.getcomment())
+ elif (self.pos + 4) < len(self.field) and
self.field[self.pos] in '\r' and self.field[self.pos+1]
in '\n' and self.field[self.pos+2] in '\t':
+ self.pos += 3
+ elif (self.pos + 4) < len(self.field) and
self.field[self.pos] in '\r' and self.field[self.pos+1]
in '\n' and self.field[self.pos+2] in ' ':
+ self.pos += 3
+ elif (self.pos + 4) < len(self.field) and
self.field[self.pos] in '\n' and self.field[self.pos+1]
in '\t':
+ self.pos += 2
+ elif (self.pos + 4) < len(self.field) and
self.field[self.pos] in '\n' and self.field[self.pos+1]
in ' ':
+ self.pos += 2
+ elif (self.pos + 4) < len(self.field) and
self.field[self.pos] in '\r' and self.field[self.pos+1]
in '\t':
+ self.pos += 2
            elif self.field[self.pos] in self.phraseends:
                break
            else:
                plist.append(self.getatom(self.phraseends))

        return plist
----------------------------------------