Previously, these tests simply compared against an equivalent regular expression pattern by generating a large string containing either all code points up to U+FFFF or U+10FFFF, depending on the `u` flag, and then performing:
str.replace(re, '') === str.replace(otherRe, '');
These two `String#replace` calls can be optimized into a single `RegExp#test` call by following the pattern used in the property escapes tests at https://github.com/mathiasbynens/unicode-property-escapes-tests.