Re: draft-klensin-unicode-escapes-02.txt

[Frank Ellermann <[e-mail]>]

John C Klensin wrote:

> Anyone who prefers the latter should please send the ABNF they would
> like to see.

Here's what I'd do with the existing ABNF:

   EmbeddedUnicodeChar =  BMP-form / Full-form
   BMP-form  =  %x5C.75 4HEXDIG   ; starting with lower case "\u"
   Full-form =  %x5C.55 8HEXDIG   ; starting with upper case "\U"

IOW no more <Hex-quad> because you didn't use it elsewhere, and
replacing <HexDigit> by <HEXDIG>, because the latter is already
defined in RFC 4234.

For the XML version I propose to adopt and fix the RFC 4646 ABNF:

   UNICHAR    = %x26.23.78 2*6HEXDIG ";"    ; starts with "&#x"

With a remark in the prose, that a literal "&" can be expressed by
&#x26; when using this style.  Otherwise folks could be tempted to
use &amp; - but we don't want them to try that.

You don't need ABNF in 5.4, it's in essence the same as in 5.1.

For 5.3 (perl) I don't know the correct syntax, if it's like 5.2:

  UNICODEPOINT = %x5C.78 "{" 2*6HEXDIG "}"  ; starts with "\x"

If the x is case insensitive it's simply:

  UNICODEPOINT = "\x{" 2*6HEXDIG "}"

The name "UNICODEPOINT" is horrible, please find something better, I
tried to use a new name.

Please add the [CharMod] reference (informative) with a note about
its C042..C048 conformance criteria in ch. 4.6 (character escaping).

Frank