Position offsets for Phrase queries

[Mattew Johnstein]

Hi all,

I am trying to figure out how Lucene gets the offset positionsinformation for 
Phrase/MultiPhrase queries when it is doing the highlighting with 
"hl.useFastVectorHighlighter=true" . 
I know that using FVH, Lucene is able to get the term positions of the termSet 
from the TermVectorComponent. But for phrase queries how does Lucene get the 
offset position of only the "exact phrase" match?

Thanks,
Mattew