Re: Patch: Don't add attachment icon quite as often

[Toralf Lund <toralf kscanners com>]

On 2001.08.22 16:15 Brian Stafford wrote:
> On Wed, 22 August 15:12 Toralf Lund wrote:
> > On 2001.08.22 15:31 Brian Stafford wrote:
> > > On Wed, 22 August 13:38 Toralf Lund wrote:
> > > 
> > > > The patch introduces the distinction between a multipart message
> and a
> > > > "message with attachments", which I think is important to keep. It
> is
> > > > obvious that "multipart" does not mean the same as "having
> > > attachments",
> > > > since there is something called "multipart/alternative"
> > > 
> > > Whether a message is considered to have attachments is determined by
> the
> > > Content-Disposition: MIME header described in RFC 2183 and should not
> be
> > > inferred from the MIME type.
> > > 
> > > Brian Stafford
> > You're right, of course. I didn't even look into this, though, as the
> point
> > I was trying to make was that Balsa's original assumption that
> everything
> > multipart has attachments, and also that there are multiple parts only
> if
> > attachments exist, is wrong.
> 
> Fair comment.  The real difficulty with attachments is that whether a
> part is
> an attachment depends to some extent on the intentions of the human
> message
> sender.
Exactly.
> 
> > I will see if I can make another quick fix of
> > libbalsa_message_has_attachment(), though...
Looks like implementing the test properly would require a bit too many
changes (for now, anyway) - unless I've overlooked something,
"Content-disposition" is only saved as a part of "mutt_body", and the body
list may be empty when building the message index (which is when the
"attachment" icon is added.)

- Toralf