Migrating to eslint-plugin-html v3

eslint-plugin-html v3 introduces a new way of linting inline script tags. Previously, all HTML parts of the file content were replaced by a /* HTML */ comment. For example, if you had a file like this:

<!DOCTYPE html>
<html>
  <body>
    <script>
      console.log(1);
    </script>
    <script>
      console.log(2);
    </script>
  </body>
</html>

it was first transformed to something like that before being processed by ESLint:

/* HTML */
console.log(1);
/* HTML */
console.log(2);
/* HTML */

This caused many issues:

Error reported by ESLint outside of the JS code were ignored, so rules reporting issues at the beginning of the file were ignored (like max-len) ;
When run in the browser, each inline script gets its own context. Two main issues come from this:
- If you are using "use strict" in an inline script, the strict mode only applies in this script. Since the strict rule with the "global" option is executed only once, it only applies on the first inline script of the file ;
- Each module scripts gets its own variable scope. So if you declare a variable directly in the body of a module, it won't be share to other inline script, contrary to traditional scripts who are executed in the global scope. By linting scripts together, the no-undef rule will not warn if the variable is used in another script, and the no-redeclare will prevent declaring the same variable in multiple module scripts.
Some other unexpected behaviors occured because of those /* HTML */ comments, like breaking eslint-disable-line in some condition.

So, as of v3, eslint-plugin-html will lint inline scripts separately, meaning each script will be seen as a different file by ESLint. This implies backward incompatible changes for some rules, example:

max-len will report scripts with too much lines instead of the HTML file being too large ;
eol-last will report missing or unnecessary new lines at the end of sripts instead of the HTML file (you probably don't want that!) ;
no-undef will report undefined variables even if they are declared globally in another script (in this case you should use /*global ...*/ to declare a variable as global).

Please report issues you have with this new behavior. If some rules are causing too much trouble, I may add a setting to ignore messages coming from those in HTML files.

You'll find other (non-breaking) changes in the changelog.