Rethink/rewrite rule 942130 with re2 support

Motivation

Allow porters to have more options for using CRS in additional engines.

Proposed solution

The rule 942130 (SQL Tautology) matches using these regexes:

[\s'\"`()]*?\b([\d\w]+)\b[\s'\"`()]*?=[\s'\"`()]*?(?:\b\1\b)
[\s'\"`()]*?\b([\d\w]+)\b[\s'\"`()]*?<=>[\s'\"`()]*?(?:\b\1\b)
[\s'\"`()]*?\b([\d\w]+)\b[\s'\"`()]*?like[\s'\"`()]*?(?:\b\1\b)
[\s'\"`()]*?\b([\d\w]+)\b[\s'\"`()]*?rlike[\s'\"`()]*?(?:\b\1\b)
[\s'\"`()]*?\b([\d\w]+)\b[\s'\"`()]*?sounds\s+like[\s'\"`()]*?(?:\b\1\b)
[\s'\"`()]*?\b([\d\w]+)\b[\s'\"`()]*?regexp[\s'\"`()]*?(?:\b\1\b)
[\s'\"`()]*?\b([\d\w]+)\b[\s'\"`()]*?!=[\s'\"`()]*?(?!\b\1\b)[\d\w]+
[\s'\"`()]*?\b([\d\w]+)\b[\s'\"`()]*?<=[\s'\"`()]*?(?!\b\1\b)[\d\w]+
[\s'\"`()]*?\b([\d\w]+)\b[\s'\"`()]*?>=[\s'\"`()]*?(?!\b\1\b)[\d\w]+
[\s'\"`()]*?\b([\d\w]+)\b[\s'\"`()]*?<>[\s'\"`()]*?(?!\b\1\b)[\d\w]+
[\s'\"`()]*?\b([\d\w]+)\b[\s'\"`()]*?<[\s'\"`()]*?(?!\b\1\b)[\d\w]+
[\s'\"`()]*?\b([\d\w]+)\b[\s'\"`()]*?>[\s'\"`()]*?(?!\b\1\b)[\d\w]+
[\s'\"`()]*?\b([\d\w]+)\b[\s'\"`()]*?\^[\s'\"`()]*?(?!\b\1\b)[\d\w]+
[\s'\"`()]*?\b([\d\w]+)\b[\s'\"`()]*?is\s+not[\s'\"`()]*?(?!\b\1\b)[\d\w]+
[\s'\"`()]*?\b([\d\w]+)\b[\s'\"`()]*?not\s+like[\s'\"`()]*?(?!\b\1\b)[\d\w]+
[\s'\"`()]*?\b([\d\w]+)\b[\s'\"`()]*?not\s+regexp[\s'\"`()]*?(?!\b\1\b)[\d\w]+

• The usage of \1 is not compatible with re2, but we can probably change it for a named group like this:

[\s'\"`()]*?\b([\d\w]+)\b[\s'\"`()]*?=[\s'\"`()]*?(?:\b\1\b)

to

[\s'\"`()]*?\b(?P<one>[\d\w]+)\b[\s'\"`()]*?=[\s'\"`()]*?(?:\b(?P=one)\b)

• still we need to think if there is an alternative for the “before text not matching” (?!\b that breaks the compatibility

Alternatives

Additional context

Tests are really simple for this rule in particular, I wonder if there is additional information on how this rule was crafted.

[fzipi]

I like the macro expansion. Gives us more flexibility. We need to check if this works 🤔

[fzipi]

Fixed by #2425