more on hypermail-2a14-pl2

From: Craig A Summerhill <>
Date: Thu, 4 Mar 1999 16:18:28 -0500 (EST)
Message-Id: <>

This one is a bit odd. It seems to work fine, except I am not sure why hypermail is performing the way it is...

If you look at either of these messages:

You will see that hypermail has marked up the body of the message as an attachment:

   o text attachment: <a href="bin_something">stored</a>

The trouble is that in the case of both of these messages, the thing that is being stored as an attachment is simply ASCII text which could have just as easily been marked up as HTML.

I am guessing that the reason it is doing this is because the headers of the message contain a "Content-Type:" field. In fact, both of these messages came from me and were generated using ELM. ELM predates MIME by a decade, but it does insert "Content-Type: text" into the header of the message. See the following header, for instance:

   From Thu Jan 14 04:18:35 1999    Received: from by (5.65v3.2/

        id AA04257; Thu, 14 Jan 1999 04:18:35 -0500    Received: by id <>;

        Thu, 14 Jan 1999 04:18:34 -0500
   From: Craig A Summerhill <>    Message-Id: <>    Subject: Re: Recipes?
   Date: Thu, 14 Jan 1999 04:18:33 -0500 (EST)    X-Mailer: ELM [version 2.4 PL23]
   Content-Type: text

What hypermail is doing isn't the end of the world, but there two problems I have with it:

   o It does cause an extra click for everybody to get to the message,

      a minor annooyance

   o Because the text of the message is stored in a "bin*" attachment,

      a web crawler of local indexing app like SWISH-E is not going to 
      be able to index the text of the message.

      This problem qualifies as a major inconvenience for me.  On a 
      couple of lists we have, myself and other staff who use ELM 
      heavily are the primary posters to the list(s).  We use these 
      lists as a news distribution services from our organization to
      its members.  We want to index these lists as part of our general 
      web site index, so that people can find press releases, etc.
      The way hypermail is processing them, we're kinda screwed.  If 
      I setup SWISH-E to index files containing the filename "bin*" I 
      run the risk of trying to index real binary files which will 
      cause SWISH to crash -- or worse, corrupt indexes...

I believe a preferred action in this case, would be to check for the existence of the "Mime-Version:" header *first*. Then, if the "Mime-Version" declaration is positve, check for "Content-Type:" and MIME parts.


   Craig A. Summerhill, Systems Coordinator and Program Officer
   Coalition for Networked Information
   21 Dupont Circle, N.W., Washington, D.C.   20036
   Internet:   AT&Tnet (202) 296-5098
Received on Thu 04 Mar 1999 11:28:21 PM GMT

This archive was generated by hypermail 2.3.0 : Sat 13 Mar 2010 03:46:11 AM GMT GMT