Expose a way of retrieving lexbor tokenizer and tree errors

[nobodywasishere]

https://github.com/lexbor/lexbor/blob/91444fd2a4a5177a73a603944cc07a5c5a54a258/source/lexbor/html/tokenizer/error.h https://github.com/lexbor/lexbor/blob/91444fd2a4a5177a73a603944cc07a5c5a54a258/source/lexbor/html/tree/error.h

[kostya]

For what use case it needed? May be I not use too much, but never encountered these errors.

[nobodywasishere]

Am currently writing a linter for crinja and need a way to verify HTML is syntactically and semantically correct without going too crazy. My thought was using lexbor would be the easiest thing to build off of for this (doesn't mean you need to add this functionality, happy to PR if it'd be accepted).

[kostya]

lexbor autofix html while parsing, so meet error I think impossible. Only maybe rare use cases with broken utf8 or something

[nobodywasishere]

It may auto-fix them, but those errors are still reported / stored on the document. They're not rare ones like utf-8 encoding either (see the two error enums above). Trying to write bindings for the document and all of the types it uses though is proving to be extremely tedious.

Wonder if it'd be possible to add a little shim file to compile with lexbor-c that adds two methods that take the document and just returns the errors found.