Hurry.query

----------------

- Add new term: Difference. It does a difference between the first and
the following terms passed as arguments.

- Add new term: Objects. It creates a result out of the objects passed
in arguments. It let you mix real objects with existing catalog
queries (with And, Or or Difference for instance).

- Add an option start to searchResult to skip the first results in the
results set.

- Extend the result from searchResult. You have addition information
on the result, including the total number of results without
start/limit restriction. A method called first() return only the
first result if available or none.

- Add an option caching to searchResult to cache the result of each
terms within a Zope transaction, speeding similar queries. If
disabled, terms will still be cached within the same query.

----------------

* Add support for an All query.

------------------

* ExtentNone in set.py missed a parameter ``index_id``. Thanks to Danilo
Botelho for the bug report.

------------------

* Allow the searchResults method of a Query to take an additional keyword
argument `sort_field` that defines that defines (catalog_name, index_name) to
sort on. That index in that catalog should implement IIndexSort.

In addition to this keyword argument, `limit` and `reverse` keyword arguments
can be passed too, that will limit the sorted resultset and/or reverse its
order.

* Allow the searchResults method of a Query object to take an additional
optional context argument. This context will determine which catalog
the search is performed on.

------------------

* Refresh dependencies. Use zope.catalog and zope.intid instead of
zope.app.catalog and zope.app.intid respectively. Don't zope.app.zapi.

* Make package description more modern.

* Clean up the code style.

------------------

* BUG: NotEq query no longer fails when all values in the index
satisfy the NotEq condition.