Singleton to help speed up lookups in an XML DOM tree.
The theory is that you often loop over occurrences of a certain tag in a Node.
If there are 20 occurrences, you go from index 0..19.
Every time we do the following
- found node 0
- found node 0, 1
- found node 0, 1, 2
- ...
So the time to search node index 19 is 20 times larger on average then index 0.
We can solve this by caching the position of index 18 and by starting back at that position.
This class is a singleton to keep everyting 100% compatible with the rest of the codebase.