PCRE impl.

[leodido]

The hypothesis is that processing (validation + group extraction) the URN with PCRE regex the performance will shine (respect to the ANTLR4 version).

Not to mention the maintainability (less code, less is more).

TODOs.

• [x] PCRE with capturing groups
• [x] Depreacte lexer and parser
• [x] Slight adapt parsing test
• [x] Introduce tests about <hex>es
• [x] Normalization task
• [x] Lexical equivalence logic
• [x] Benchmark

Obs.

Without the ANTLR4 all(*) generated parser is not comfortable to create a CST (concrete syntax tree) the user can navigate with a listener or visitor pattern.

[leodido]

For future reference: is not possibile to capture all the <hex>es (eg., a%1f%2Cbcdse%21) within the specific string part putting them in a group; this because groups do not accumulate.

The last group match overrides the content of the previous.

[leodido]

At the moment the normalization task is done. Anyway unit test for it are missing. Same for the lexical equivalence (which depends on normalization task).

Very close ...