Also, how firm is the definition of text/plain?
Always an open question.
I mean ... if we have CT: text/plain and CTE: gzip64,
then are we assured that lines will be CR/LF delimited?
gzip64 should have neither more nor less problem in this regard than
base64. The object is to be converted to canonical form, and THEN encoded.
In this case, the encoding will also involve compression, but that doesn't
mean you are allowed to omit the canonicalization step.
--
Steve Dorner, Qualcomm Incorporated. "Oog make mission statement."