Since the work in T387002: Try to reduce duplicate code in reference validation both parsers share the same validator for a lot of the <ref> as well as <references> tag validation. However, a few possible errors have been left untouched.
Possible TODOs:
- Reduce the list of currently excluded error cases at the beginning of References::extractRefFromNode() further.
- Identify and classify the error cases that are still outside of the validator. Try to move these into the validator as well, if it makes sense.
Benefits:
- Code duplication is reduced.
- Parsoid will behave more closely to the legacy parser.
- Additional test cases will help us making the two parsers behave the same.