Re: [Cbor] [Technical Errata Reported] RFC8610 (6278)

[Carsten Bormann <[e-mail]>]

The below is exactly what should never happen: having to encode detailed (and somewhat arcane) knowledge about a standard in the bowels of another, mostly unrelated standard. I think we hit the right balance  with VCHAR and SCHAR, we just forget to fix BCHAR.  But we can discuss that when we do the bis...

Sent from mobile, sorry for terse

> On 4. Sep 2020, at 05:03, Doug Ewell <[e-mail]> wrote:
> 
> 
>> 
>> Corrected Text
>> --------------
>> BCHAR = %x20-26 / %x28-5B / %x5D-7E / %x80-10FFFD / SESC / CRLF
> 
> Actually, if the range is meant to end at U+10FFFD (thus excluding the noncharacters U+10FFFE and U+10FFFF), then it should also exclude the following:
> 
> 1. the U+xxFFFE and U+xxFFFF noncharacters in every other plane
> 2. noncharacters U+FDD0 through U+FDEF
> 3. surrogate code points 0xD800 through 0xDFFF, which aren't even Unicode scalar values (hence not marked with "U+" notation)
> 
> Of course this goes for the definitions of PCHAR, SCHAR, and SESC as well.
> 
> A new symbol, NONASCII, could resolve all this:
> 
> NONASCII = %x80-D7FF / %xE000-FDCF / %xFDF0-FFFD
>         / %x10000-1FFFD
>         / %x20000-2FFFD
>         / %x30000-3FFFD
>         / %x40000-4FFFD
>         / %x50000-5FFFD
>         / %x60000-6FFFD
>         / %x70000-7FFFD
>         / %x80000-8FFFD
>         / %x90000-9FFFD
>         / %xA0000-AFFFD
>         / %xB0000-BFFFD
>         / %xC0000-CFFFD
>         / %xD0000-DFFFD
>         / %xE0000-EFFFD
>         / %xF0000-FFFFD
>         / %x100000-10FFFD
> [...]
> SCHAR = %x20-21 / %x23-5B / %x5D-7E / NONASCII / SESC
> SESC = "\" (%x20-7E / NONASCII)
> [...]
> BCHAR = %x20-26 / %x28-5B / %x5D-7E / NONASCII / SESC / CRLF
> [...]
> PCHAR = %x20-7E / NONASCII
> 
> --
> Doug Ewell | Thornton, CO, US | ewellic.org
> 
>