Feature request: Use v-mode (unicodeSets) instead of the strictly-worse (less powerful) u-mode (unicode) for regex search.

For example, in a file containing the text "hello š𐓇", regex searching for \p{scx=Latn}[\p{L}--\p{scx=Latn}] should match š𐓇 (playground).

Current behavior is that the regex search fails with "Invalid regular expression: /\p{scx=Latn}[\p{L}--\p{scx=Latn}]/gu: Invalid character class".