Updating find element by link text atom to handle non-breaking spaces

The WebDriver W3C Specification states that when finding by link text, the
text of a link element (<a>) should be trimmed of leading and trailing
whitespace after calling getVisibleText on the element before comparison.
Links with text that contains leading or trailing non-breaking spaces
(&nbsp) were having the non-breaking spaces converted to regular spaces,
but not trimmed after the fact. This commit corrects that oversight.