Section 7.6.1 "Regular expression syntax" says that two metacharacters,
'^' and '$' are added, and it alters XML Schema part 2 rule [10] to
list these metacharacters.  However, this serves only to exclude 
'^' and '$' as normal characters (and thereby create escapes for them)
but it does not make them part of the regular expression syntax.
Using the modified rules of XML Schema part 2, there is no way to 
derive that, for example, '^abc' is a regular expression.  I think 
the anchors are supposed to enter the regular expression rules at
the precedence level of rule [9] atom.  Probably a modification of 
this rule is needed.