Improve regex to not match empty html entities

[denisu]

Thank you for this module! I have used it to decode millions of old forum posts, it works well, but I have discovered an error with malformed html entities, specifically with &#;.

Example:

htmlEntities.decode('hello world &#;')

Leads to:

htmlentities.lua:2317: bad argument #1 to 'abs' (number expected, got string)

To solve this bug, I have changed the matching regex from * to + to match only entities with at least one occurrence.