Fixes LP #628448

Avoid conflict error hotspot in PluginIndexes' Unindex class by using IITreeSets instead of simple ints from the start. Idea taken from ``enfold.fixes``. This optimizes the common case of FieldIndexes being used for a low number of unique values. The edge-case of having a FieldIndex for an unique value per document gets penalized, as it uses many more persistent objects. This really warrants a separate type of index.

copy the list of packages to initialize. install_package calls package_initialized which removes from the list

Windows, because xrange uses 32bit integers and __getslice__ gets 64bit integers

Changed policy on handling exceptions during ZCML parsing in ``Products``. We no longer catch any exceptions in non-debug mode. These days most products won't actually do anything useful if their ZCML is not loaded - it's no longer helpful to pretend to get a working instance. We'd rather fail early and visible.

Added a new BooleanIndex to the standard PluginIndexes, sponsored by Jarn AS and based on work in the Products.BooleanIndex distribution.

Fixed ``testZODBCompat`` tests in ZopeTestCase to match modern ZODB semantics. There is no explicit contract for the time _v attributes actually get removed. The tests in question where bogus, as they didn't test database load but a class variable not affected by the transaction machinery. There's other tests in the same module testing more abort semantics that have a proper demo storage setup.