Oops, I stand corrected. Encoded-variable as currently defined does
indeed suffer from the limited-number-of-headers problem. However, it
could be otherwise -- in particular, we could probabaly allow it in any
quoted-phrase or comment, thus making it more open-ended.
Short comment:
Whatever is done here, please always qualify "any..." with "except
transport-specified trace fields". Otherwise, complex inter-layer
problems are going to, sooner or later, be introduced. A universally-
known character set for trace information will also save us a LOT of
trouble in debugging when something goes wrong.
john