Re: IRIs, IDNAbis, and HTTP [i74]

Julian Reschke wrote:

> if we *would* introduce a way for new headers to use UTF-8
> (similarly to the "TEXT" rule in RFC2616), would that break
> anything?

Using "raw" UTF-8 would be dubious if the other side hates C1
controls.  Or it could break in interesting ways if the other
side likes some C1 controls (NEL, SS2, SS3).  But I guess you
mean something more sophisticated like BCP 113.

 Frank