email.message.Message - as_string fails

Helmut Jarausch jarausch at
Sat Dec 29 11:26:49 CET 2012

On Fri, 28 Dec 2012 20:57:46 -0500, Terry Reedy wrote:

> On 12/28/2012 7:22 AM, Helmut Jarausch wrote:
>> Hi,
>> I'm trying to filter an mbox file by removing some messages.
>> For that I use Parser= FeedParser(policy=policy.SMTP)
>> and 'feed' any lines to it.
>> If the mbox file contains a white line followed by '^From ',
>> I do
>> Msg= Parser.close()
>> (lateron I delete the Parser and create a new one by Parser=
>> FeedParser(policy=policy.SMTP)
>> )
>> I can access parts of the message by  Msg['Message-ID'], e.g.
>> but even for the very first message, trying to print it or convert it
>> to a string by  MsgStr=Msg.as_string(unixfrom=True)
>> lets Python (3.3.1_pre20121209) die with
>> Traceback (most recent call last):
>>    File "", line 35, in <module>
>>      MsgStr=Msg.as_string(unixfrom=True)
>>    File "/usr/lib64/python3.3/email/", line 151, in as_string
>>      g.flatten(self, unixfrom=unixfrom)
>>    File "/usr/lib64/python3.3/email/", line 112, in flatten
>>      self._write(msg)
>>    File "/usr/lib64/python3.3/email/", line 171, in _write
>>      self._write_headers(msg)
>>    File "/usr/lib64/python3.3/email/", line 198, in
>>    _write_headers
>>      self.write(self.policy.fold(h, v))
>>    File "/usr/lib64/python3.3/email/", line 153, in fold
>>      return self._fold(name, value, refold_binary=True)
>>    File "/usr/lib64/python3.3/email/", line 176, in _fold
>>      (len(lines[0])+len(name)+2 > maxlen or
>> IndexError: list index out of range
> The only list index visible is 0 in lines[0]. If this raises, lines is
> empty. You could trace back to see where lines is defined. I suspect it
> is all or part of the Msg you started with.
> I believe that some of email was rewritten for 3.3, so it is possible
> that you found a bug based on an untrue assumption. It is also possible
> that you missed a limitation in the doc, or tripped over an intended but
> not written limitation. So I hope you do the tracing, so if doc or code
> need a fix, a tracker issue can be opened.

Thanks Terry,
I've debugged it and it smells like a bug.
I have created


More information about the Python-list mailing list