Since the document's contents are read in 16K blocks and Document.insertString() is done for each block the lexer gets
invoked after each block's insert. For large documents with complex embeddings with possibly long top-level tokens such
as JSP it may be costly. It would be better to update hierarchy at the end of the whole reading.
The clients of token hierarchy should be prepared that the hierarchy becomes inactive at any time but it's necessary to
verify whether they really are in reality.
Integrated into 'main-golden', will be available in build *200811080201* on http://bits.netbeans.org/dev/nightly/ (upload may still be in progress)
User: Miloslav Metelka <email@example.com>
Log: #150204 - Switch off Lexer's Token Hierarchy during document loading.