Support all of HTML's character entities in WebVTT

This updates the VTTTokenizer to align with the updated WebVTT tokenizer
steps:

 https://w3c.github.io/webvtt/#html-character-reference-in-data-state

and

 https://w3c.github.io/webvtt/#html-character-reference-in-annotation-state

The old states for handling "escapes" are removed, adding new ones per
above, for invoking the HTML parser's entity matcher.
A new WEBVTT_SWITCH_TO operation is also added so that state transitions
can be performed without advancing the input position. (Matches
mechanics of the HTML parser.)

The WPT test webvtt/parsing/cue-text-parsing/tests/entities.html is
updated since it had an incorrect expected result for the "&notit;"
case.

Old, now incorrect and/or redundant, tests for entities are removed.

Bug: 761303
Change-Id: I8ffb4fb2da7169c2ba3c84eb8c3206daabefbb30
Reviewed-on: https://chromium-review.googlesource.com/647586
Reviewed-by: Philip Jägenstedt <foolip@chromium.org>
Commit-Queue: Fredrik Söderquist <fs@opera.com>
Cr-Commit-Position: refs/heads/master@{#499473}
6 files changed