Problem

We replace content protection tag with some private utf-8 character before sending to term tagger so that term tagger not include it into term and fail term in case if such a character found inside of tag.

Found out, that if such a character placed in the beginning of term - it will be included in term and then fails reapplying of tags in segment.

Solution

Add whitespace to character.

This will (probably) make term tagger think that this char is additional token and process it correctly.

Test

Create a task with segment that has protected content right before and right after term.

<cp n="1" /><term>integration</term><cp n="3" />

Run QA.

In master it will fail.

In fix branch will work.