what is the most elegant way to manipulate an arbitrary unicode character in perl regex?

Question

Consider a unicode character, such as zero-width space, which is not on any conventional keyboard and is not part of any human writing system. Suppose one wants to use perl to r