This makes two significant changes:

• When parsing srcset="foo.jpg", the result no longer includes the 1x density:

  - { url: "foo.jpg", density: 1 }
  + { url: "foo.jpg" }

• The validation checker now uses normalized values when checking for duplicates. It now recognizes that "foo.jpg 1x, bar.jpg" is invalid, as well as "foo.jpg 2x, bar.jpg 2.0x" because both contain two images with effectively duplicate descriptors.

@sindresorhus I had initially suggested putting this into a separate PR, but the more I think about it, the more I think it should just be included with sindresorhus#13. So, this PR targets the nfriedly:strict, meaning that if I merge it, then these changes will become part of that Pull Request. What do you think?

Either way, I think it makes sense to publish both sets of changes as a group.