Add bcp47_strict_language_tag validator

[bfabio]

Add strict BCP47 language tag validator enforcing RFC5646 and rejecting Unicode extensions unlike language.Parse().

Fixes Or Enhances

Fix #1221.

Make sure that you've checked the boxes below before you submit PR:

• Tests exist or have been written that cover this particular change.

@go-playground/validator-maintainers

[coveralls]

coverage: 73.815% (+0.09%) from 73.73%
when pulling ba8f1be on bfabio:bcp47_strict_language_tag
into 6a38036 on go-playground:master.

[bfabio]

@nodivbyzero rebased

[nodivbyzero]

This regex compiled every call. That means compiling a huge regex every invocation, which is very expensive.
It's better to move it to a package-level var.

[bfabio]

You're right, fixed

[nodivbyzero]

@bfabio Thanks for contributing!

Feels like isBCP47StrictLanguageTag still needs a bit more refactoring:

1. Move regex to global
2. Reduce ToLower/ToUpper
3. Early rejection

Something like that:

var languageTagRe = regexp.MustCompile(strings.Join([]string{ ... }, ""))

func isBCP47StrictLanguageTag(fl FieldLevel) bool {
	field := fl.Field()

	if field.Kind() != reflect.String {
		panic(fmt.Sprintf("Bad field type %s", field.Type()))
	}

	raw := field.String()
	upper := strings.ToUpper(raw)
	lower := strings.ToLower(raw)
	lowerTagDash := lower + "-"

	m := languageTagRe.FindStringSubmatch(upper)
	if m == nil {
		return false
	}

	// reuse lower/upper everywhere instead of recomputing
	...
}

Also, is it possible to completely replace regex?

tag, err := language.Parse(languageTag)

and then validate against IANA constraints manually.

[bfabio]

Thanks for the review

Feels like isBCP47StrictLanguageTag still needs a bit more refactoring:

1. Move regex to global

Done

2. Reduce ToLower/ToUpper

I refactored it a little bit more so they're not necessary. I don't even know why I did upcase it, but it was a while ago 🤷

3. Early rejection

Done.

Also, is it possible to completely replace regex?

tag, err := language.Parse(languageTag)

and then validate against IANA constraints manually.

I'm not sure what you mean here, a custom parser?