`prevent-abbreviations`: Smarter `whitelist` option possibly deprecating `ignore` option

There are now two options to exclude an identifier from linting: older whitelist and newer ignore. I think whitelist alone would suffice with the following modification.

We could make whitelist match any subsequence of the split identifier. For example (following up this https://github.com/sindresorhus/eslint-plugin-unicorn/issues/567#issuecomment-591007369): For fooE2eBar all of these would be checked against whitelist:

• foo
• e
• 2
• bar
• barE
• e2
• 2e
• eBar
• fooE2
• e2e
• 2eBar
• fooE2e
• e2eBar
• fooE2eBar

And if any of them is in the whitelist that subsequence is ignored.

But in my opinion it should not ignore the whole identifier if only subsequence matches whitelist.

For example, if the user had e2e in their whitelist, e2eErr should still be fixed to e2eError.

From https://github.com/sindresorhus/eslint-plugin-unicorn/pull/891#issuecomment-713729469