The stringifier of URLSearchParams must use the
application/x-www-form-urlencoded byte serializer algorithm.

The "extract" algorithm of Body mixin defined in the Fetch Standard
should also use the algorithm for processing URLSearchParams.

This algorithm doesn't include CRLF normalization unlike <form>
submission algorithm.

BUG=557063
R=sigbjornf@opera.com,mkwst

Review-Url: https://codereview.chromium.org/1998563002
Cr-Commit-Position: refs/heads/master@{#396801}