Re: RIP: Crypto-Key header field

On 23 November 2016 at 18:51, Martin J. Dürst <> wrote:
>> I see lot of additional works and potential bugs by allowing non-UTF-8
>> octet sequences, and something like zero advantages.
> Yes indeed. Not knowing what character encoding is used for a string, or not
> knowing whether something is binary or text, it a lot of pain, and no
> "virtue" at all. We have a lot of experience with that; let's not repeat
> these same mistakes over and over again.

This presumes that the field is a string.  We already have a use for
it that treats it as an octet sequence.  If someone cares to designate
a usage that needs strings, then UTF-8 is available to them.

Received on Wednesday, 23 November 2016 23:27:47 UTC