Also includes whitespace at the beginning of a text node. Without this, we throw away information needed to create an exact round-trip copy of the input, as well as information needed to create a canonicalized version of the XML Document. Should only affect the .nodes() interface, as well as .contents(). .elements() and XML::Text.string should be unaffected. Note that XML::Text.Str may be affected.
XML::Element. These methods support set-wise operations on attribute values. These operations are useful when, for example, the attribute values are type NMTOKENS rather than CDATA or for applications that like to treat CDATA attribute values as white space separated list of values. The values are still stored as strings. I modified README.md to document the new methods.