shorten_text sometimes omits close tags

The core function shorten_text is supposed to close any HTML tags that it cuts off. However this only occurs sometimes, depending on the parameters. The sequence of events that occurs is:

• Function divides input into blocks starting with each tag (open/close/empty)
• Loops through these input 'lines' until there is more text than the shorten_text allows

At this point, any tags that are not closed (ie that would be closed in future 'lines') are safe - they will be closed.

• Goes back looking for a space or suitable breakpoint to chop the text so it is under the limit

During this process close tags may be chopped off. These are not re-added.