Thanks Tim for bringing that up :)
The previous URLValidator would consider some valid URLs as invalid.
That can cause a lot of problems, most people would expect an URL
Validator to return some false negatives sometimes (invalid URLs
considered valid) but not false positives (valid URLs considered
invalid). Furthermore, there are probably a few other projects that depend