benchmark: add real-world examples for url parsing

We (@lemire and I) recently developed a small crawler to generate a large enough dataset (currently 100,000 URLs) for URL parser benchmarks.

Current benchmarks in Node.js uses Web Platform Tests, which creates a non-realistic output since it contains a lot of edge cases that no real-world application would ever encounter. This new benchmark creates a more realistic scenario with extensive data from top-visited websites.

The whole project for generating this dataset is available at https://github.com/ada-url/url-dataset.

cc @nodejs/url @nodejs/performance