Add like_regex predicate to JSON path

[piotrrzysko]

Description

This PR introduces the like_regex predicate to JSON path. It aims to comply with the ISO/IEC 9075-2:2016(E) standard. The only feature defined in the standard that is not implemented here is the case-insensitive mode (the i flag). The reason for this is described in this issue: #24615.

Relevant links to the specifications referenced by the standard:

• https://www.w3.org/TR/xquery-operators/ (section 7.6.1 Regular Expression Syntax)
• https://www.w3.org/TR/xmlschema-2/ (section F Regular Expressions)
• https://www.unicode.org/reports/tr18/#Line_Boundaries

I want to highlight a specific part of the specification for reviewers to verify if my interpretation is correct.

The ISO/IEC 9075-2:2016(E) standard, in Section 9.22, specifies the following:

1) Let S be the STRING, let PAT be the PATTERN, let SP be the POSITION, let U be the UNITS, and let FL 
   be the FLAG in an application of the General Rules of this Subclause. The result of the application of 
   this Subclause is LOMV, which is returned as LIST.
2) Case:
    a) If the character repertoire of PAT is UCS, then let P be PAT.
    b) Otherwise, let P be the result of an implementation-defined transliteration of PAT to UCS.
3) Case:
    a) If the character repertoire of FL is UCS, then let F be FL.
    b) Otherwise, let F be the result of an implementation-defined transliteration of FL to UCS.
...

9) An XQuery regular expression match is defined as follows:
   Case:
   a) If the character repertoire of S is UCS, then an XQuery regular expression match is a match as defined
      by [XQueryFO] for the function fn:matches(), with F as the value of the parameter $flags, with
      the following modifications:
      ...
   b) Otherwise, the definition of XQuery regular expression match is implementation-defined.

Points 2 and 3 specify that if the provided arguments contain characters outside the UCS repertoire, they should be transliterated to UCS. My understanding is that if the arguments are valid UTF-8 sequences, they contain only characters within the UCS repertoire. Even if the sequence includes code points that do not have characters currently assigned to them, they are still part of the repertoire, as specified in Unicode Technical Report #17:

For the Unicode Standard, on the other hand, the repertoire is inherently open. Because Unicode is intended to be the universal encoding, any abstract character that ever could be encoded is potentially a member of the set to be encoded, whether that character is currently known or not.

Point 9 states that the definition of a regular expression match is implementation-defined when the character repertoire of the provided subject string is not UCS. The approach in this PR is similar to what we do for string functions: if the string is not valid UTF-8, the result is undefined.

Additional context and related issues

Release notes

( ) This is not user-visible or is docs only, and no release notes are required.
( ) Release notes are required. Please propose a release note for me.
(x) Release notes are required, with the following suggested text:

## Section
* Add like_regex predicate to the JSON path language. ({issue}`issuenumber`)