Added fragment percent encode set for URL fragments.

[ulf5]

Hello!

I noticed that the spec here defines a fragment percent encode set:
https://url.spec.whatwg.org/#percent-encoded-bytes

The fragment percent-encode set is the C0 control percent-encode set and U+0020 SPACE, U+0022 ("), U+003C (<), U+003E (>), and U+0060 (`).

It seemed to me that rust-url was using the SIMPLE_ENCODE_SET while parsing the fragment and that it should be using the fragment encode set instead:
https://url.spec.whatwg.org/#url-parsing

UTF-8 percent encode c using the fragment percent-encode set and append the result to url’s fragment.

I made this PR which should fix it, or have I got it wrong?

---

This change is 

[bors-servo]

☔ The latest upstream changes (presumably #517) made this pull request unmergeable. Please resolve the merge conflicts.

[nox]

Your assessment seems to be correct. I'll check upstream if the tests changed, given the tests are supposed to follow the spec and your interpretation of the spec seems correct to me.

[nox]

thread '"http://f:21/ b ? d # e " @ base "http://example.org/foo/bar"' panicked at '
"http://f:21/%20b%20?%20d%20# e"
!= expected.href
"http://f:21/%20b%20?%20d%20#%20e"
for URL "http://f:21/%20b%20?%20d%20# e"
'

Yep, upstream tests were changed and we are just not aligned anymore.