Use editor.wordSeparators in order to classify characters.

[tekezo]

This change solves a problem that japanese-word-handler ignores editor.wordSeparators setting.

About an issue

The current japanese-word-handler does not check editor.wordSeparators setting in classifyChar.
Thus, the separators (e.g., underscore) are ignored.

For example, when editor.wordSeparators contains underscore (_),

editor.wordSeparators: "`~!@#$%^&*()-=+[{]}\\|;:'\",.<>/?_"

expected cursor move is:

‸classify_char
      ↓
classify‸_char

actual:

‸classify_char
      ↓
classify_char‸

About this change

This commit adds new character class Separator and classify editor.wordSeparators into Separator.

[sgryjp]

I've reviewed and tested, and it worked as expected. Thank you for contributing!

[tekezo]

Thank you for confirmation.
I cannot use VSCode without japanese-word-handler!