On 18/12/2021 00.18, Tom Honermann via SG16 wrote:

 8. Specify how invalid code unit sequences are to be handled. This includes specifying, at least for self synchronizing encodings like UTF-8, UTF-16, and UTF-32, how such sequences are delimited. References to the Unicode standard (as indicated in the editor notes in the linked meeting summary) and/or WhatWG Encoding Standard are advised. This also includes specifying how wide strings are handled; presumably each wchar_t value in an ill-formed code unit sequence would be formatted as a single hex escape.

Since C++ is an ISO standard, normative references to other ISO standards
(e.g. ISO 10646) as opposed to third-party standards (e.g. Unicode) are
preferred, per ISO policy.

Thank you for the reminder, Jens. My comment about providing a reference was intended for the paper prose, not the wording.

If a wording reference is needed, the following from the terms and definitions section of ISO/IEC 10646:2020 (freely available from here) may be helpful. Unfortunately, the document does not appear to include any discussion of these terms.

• 3.32, ill-formed code unit sequence
  "UCS code unit sequence that purports to be in a UCS encoding form that does not conform to the specification of that encoding form"
  
• 3.33, ill-formed code unit subsequence
  "non-empty subsequence of a code unit sequence X that does not contain any code units that also belong to a minimal well-formed code unit subsequence of X"
  
• 3.39, minimal well-formed code unit sequence
  "well-formed code unit sequence that maps to a single UCS scalar value"
  
• 3.59, well-formed code unit sequence
  "UCS code unit sequence that purports to be in a UCS encoding form that conforms to the specification of that encoding form and contains no ill-formed code unit subsequence"
  

Tom.