Mime Parsing

[Judy Northrop]

I wasn't sure if the malformed header was a result of the pdf file being
attached or a result of the filter.  I just disabled
MailScanner/SpamAssassin and that seems to have corrected things.  My
filter rule-set is updated with a cron, so I will have to debug that. 
Thanks again.

Judy



Judy Northrop wrote:
> Peter,
>
> Thanks for helping.
>
> Yes I did truncate the message.  I just looked at the delivered message,
> it appears to be properly terminated.  Here is the complete header and i
> will include the ending terminator. Also, word documents (filename.doc)
> seem to work ok, to date I am only having problems with pdf files.
>   
Well, at a glance my guess would be that the attachment is perfectly 
fine, but no mail app can see it because the main message header is 
malformed. For example, the main message header specifies a MIME version 
but has no Content-Type header. Then after the header separator there is 
a MIME boundary string that hasn't even been declared. From my 
recollection, this is a violation of the MIME RFCs. There is one header 
in there that is rather informative:

X-jathome-MailScanner-SpamCheck: spam, SpamAssassin (not cached,
score=4.168,
        required 4, MISSING_HB_SEP 2.50, MISSING_HEADERS 0.19,
        MISSING_SUBJECT 1.34, TO_CC_NONE 0.13)

So it would seem that the spam check service found problems with the 
message as well. The MISSING_HB_SEP is particularly disturbing. If I had 
to guess what was going on I would say there is a poorly written client 
or script sending these, or some filtering layer is altering the message 
content due to bugs or bad choice of rules.

Peter