ExtractText2 (#929)

* New proposal for spacing/lining with deeper analysis of font data and text positionning

Includes :
* XObject Processing, 
* choice between encoding and tounicode fields
* partial compliance with Identify-H/V encoding (missing processing on 2-bytes)
*legacy conversion reintroduced as old for comparison

increase test and refactory depreciation warning ignore in test

* take into account Tm scale and improve space calculation