* [15.Ağu.07 14:16 +0200] Thomas Roessler: > On 2007-08-15 14:52:45 +0300, Sertaç Ö. Yıldız wrote: > > Not the text part, the multipart/alternative part itself is labeled > > as base64. And AFAIK, that's not permitted for multipart types. > > That base64 has nothing at all to do on the multipart/alternative > body part: (a) the content of that body part isn't actually base64 > encoded, and (b) if it was, that would be breaking a MUST NOT in the > relevant spec.
But mutt is trying to decode base64 in contrary to (a) and (b) and failing (IMHO) unnecessarily with an unhelpful message. > Indeed, it's serious garbage; anything mutt could do to deal > "correctly" with this particular message would cause breakage > elsewhere. Is it "incorrect" to skip decoding if body part is multipart? I guess it just tolerates the error here, and if the multipart was actually encoding transformed (in spite of being expressly forbidden by the spec), it would fail later anyway. -- ~sertaç