[sf-lug] Lists: plain text only please :-) [ugh: ("An HTML attachment was scrubbed...")]

[Rick Moen]

Quoting Jesse Zbikowski (embeddedlinuxguy at gmail.com):

> OK just throwing out ideas here, but why not have the list server deal
> with HTML attachments?  Some piece of code figured out that there is a
> scrubbable HTML attachment (Mailman or Pipermail) so I would think the
> Right Thing might be one of the following:
> 
> 1. List server should silently remove HTML attachments with no trace
> 
> 2. List server should reject mail with HTML attachments (idiot filter)
> 
> 3. List server should auto-send "nag-mail" to the offender saying
> essentially what Michael said above

4.  List server (actually, the pipermail archiver, in this case) will
archive only the ASCII part of any multipart-alternative posting, and 
run any HTML-only post through "/usr/bin/lynx -dump -force_html '%s'"
or equivalent, before archiving.

People spill a lot of virtual ink over objecting to HTML mail, and I
_do_ concur with Michael on his basic stance, but at the MUA (user mail
program) level, long ago I decided the smart thing to do (as a _main_
remedy, at least) was to autoconvert incoming HTML and other oddly
formatted e-mails to something more reasonable.  Therefore, my
/home/rick/.mailcap file includes a bunch of entries like these:


application/msword; wvWare --nographics -x /usr/share/wv/wvHtml.xml '%s' | lynx -dump -force_html -nolist /dev/stdin; copiousoutput; description=Microsoft Word Document; print=wvWare --nographics -x /usr/share/wv/wvHtml.xml '%s' | lynx -dump -force_html -nolist /dev/stdin | print text/plain:-

text/html; /usr/bin/lynx -dump -force_html '%s'; copiousoutput; description=HTML Text; nametemplate=%s.html

application/ms-tnef; /usr/bin/tnef -w -f '%s'; needsterminal; description=Microsoft's Exchange Encapsulation; nametemplate=%s.data



So, basically, instead of putting my primary faith in educating people
sending me e-mail to be clueful, I make my mail software automatically
repair those mails upon receipt -- retrofitting clue, as it were.  Most
of the time, I'm only barely aware that "Oh, that must originally have
been HTML-only, and lynx fixed it on the fly."  In some cases, I
literally never notice.

Of course, I'll note that hacking pipermail is pretty non-trivial,
unlike writing a few new .mailcap entries.  So, trying to teach users to
be clueful might have some mileage in it, yet.  ;->