tests: adds test function for IdnaTestV2.json

[miguelteixeiraa]

ref: #222

The good news is that almost all tests cases are passing.

[lemire]

@miguelteixeiraa Would you add it to https://github.com/ada-url/idna ?

We eventually want to rely on ada-url/idna for ada.

[miguelteixeiraa]

Makes sense, good idea!

[miguelteixeiraa]

@miguelteixeiraa Would you add it to https://github.com/ada-url/idna ?

We eventually want to rely on ada-url/idna for ada.

Done!

[lemire]

@miguelteixeiraa Why did you close it here? In my view, it belongs at both places. Your work here was good and relevant. Of course, here you were testing our ICU-based implementation, but no matter. My understanding is that your tests ought to pass everywhere (or, else, we need to figure something out).

[anonrig]

There is 1 test case we are failing at the moment.

FAIL: 
  Returned: xn--8-qfa.xn---?-261a
  Test case: {
    "comment": "P1; V6; V3 (ignored)",
    "input": "8\u00df.\udb40\udd10-?\u2d0f",
    "output": null
  }

[anonrig]

We moved is_forbidden_domain_code_point check to outside of to_ascii method. I assume this is why it's breaking this pull request. We just need to add the following lines to wpt_tests with a TODO to fix this when we get rid of ICU.

if (output.has_value()) {
  if (std::any_of(output.value().begin(), output.value().end(), ada::unicode::is_forbidden_domain_code_point)) {
    output = std::nullopt;
  }
}

[miguelteixeiraa]

Is it okay to remove the ada_really_inline constexpr from is_forbidden_domain_code_point? I'm getting linkage problems.

[lemire]

@miguelteixeiraa This will need to be updated because we did major surgery on the code.

[lemire]

Let me see if I can integrate the code in the other PR.

[lemire]

I am going to push the content of this PR into #216

It will be less trouble for everyone.

You may close this PR.