Skip redundant UTF8 to UTF16 conversion follow-ups

[jordan-powers]

In PR #126492, we implemented an optimization to skip some redundant UTF8 to UTF16 conversions.

There were several follow-ups to that PR, which are tracked here.

• Support for escaped and/or non-ascii characters (Optimized text for full unicode and some escape sequences #129169)
• Support for match_only_text fields (Use optimized text in match_only_text fields #129371)
• Support for text fields
• Support for wildcard fields
• Support for other xcontent types (cbor Enable optimizedText for CBOR #132542, smile, yaml)
• Remove XContentParser#optimizedText() and instead have XContentParser#text() return XContentString instead of String

Maybe not even possible:

• Support for running normalizers on utf-8 encoded data instead of needing to convert to utf-16 strings

[elasticsearchmachine]

Pinging @elastic/es-storage-engine (Team:StorageEngine)