by robert » Fri Jan 07, 2011 1:05 pm
On closer inspection, this would be a non-trivial modification to how RSTA tokenizes a document. For performance reasons, RSTA tokenizes a document on a per-line basis. This also simplifies the tokenizing code, and saves lots of memory over say, a JEditorPane approach. A side-effect of this is that it makes it very hard for RSTA to determine certain things while tokenizing, such as if a particular multi-line comment is ever closed. Unfortunately, for this reason, it's pretty unlikely you'll see this feature without a rewrite of the tokenizing code.