Saxon XSLT and XQuery Processor

I am somewhat puzzled by the following statement in the list of XSLT
changes:

"The use of a separate name pool for use at compile time has been
discontinued."

I can't see where this has changed in the source code. In particular,
in LiteralResultElement.java, there is still the following code:

          // Now translate the list of namespace codes to use the target namepool,
            // getting rid of the -1 entries while we go.

            int count = namespaceCodes.length - numberExcluded;
            NamePool oldPool = getNamePool();
            NamePool newPool = getTargetNamePool();
            int[] newNamespaceCodes = new int[count];
            count = 0;
            for (int i=0; i<namespaceCodes.length; i++) {
                if (namespaceCodes[i] != -1) {
                    String prefix = oldPool.getPrefixFromNamespaceCode(namespaceCodes[i]);
                    String uri = oldPool.getURIFromNamespaceCode(namespaceCodes[i]);
                    newNamespaceCodes[count++] = newPool.allocateNamespaceCode(prefix, uri);
                }
            }

which seems to belie the statement.
-- 
Colin Paul Adams
Preston Lancashire