Added new fundamental filter 'xslt'. This filter reads XML records

and uses LibXSLT for both indexing (extract) and retrieval (present).
During indexing the filter generates a Zebra indexing record via XSLT
which describes how Zebra is to index the record. Because the driver is
XSLT driven it can use any X-Path plus logic behind the scenes and is thus
more powerful than xelm/elm in .abs. The XSLT can accept parameters from
Zebra. For example, if a date is received the filter could make a date
index. The filter also uses allows splitting of XML records during
indexing, so that MARC collections can be indexed directly (but it
is quite limited and takes place before XSLT is invoked: XSLT requires
a DOM structure in memory). Refer ot example test case in in test/xslt.