2011/1/14 Michael M. Slusarz <slus...@curecanti.org> > You should only need the structure for each message, which you can get with > BODYSTRUCTURE command. >
Indeed I should be able to achieve what I'm looking for with this command. > > But beware that determination of what is an "attachment" is not so > straight-forward. The Content-Disposition header (RFC 2183), if it even > exists, is often not useful. > > Take these examples: > > * Message with a single application/pdf part. Is this an attachment? Many > users would say yes. But what if Content-Disposition does exist and is set > to 'inline'. > > * multipart/mixed part, with a single text/plain child part. This is most > likely not an attachment, at least by an end user definition, and shows that > you can't just search for 'multipart' in the Content-Type field. > > * multipart/related part with two embedded parts, text and HTML. Again, > most likely users would not expect this to be an attachment. > > * multipart/related part with two embedded parts, one text and one PDF. If > PDF can't be displayed in the MUA, should it be considered an attachment, > even though it lives in a multipart/related part? (e.g. the text part is > intended to be displayed as a quick view of the PDF, but loses all of the > formatting contained in the PDF.) > > And so on. > > michael > Ok, I understand. It seems I need to think my code as flexible as possible because I would certainly have to add unsupported combinations in the future... By the way, thanks for your answer. Antoine