Handlerton array is now created instead of using sys_table_types_st. All storage engines can now have inits and giant ifdef's are now gone for startup. No compeltely clean yet, handlertons will next be merged with sys_table_types. Federated and archive now have real cleanup if their inits fail. 


sql/examples/ha_archive.cc:
  Modifications for new ha_init code. The init method now checks for errors and will not start up if the errors occur.
sql/examples/ha_archive.h:
  Change for new init method.
sql/examples/ha_example.cc:
  New handlerton pieces.
sql/examples/ha_tina.cc:
  New handlerton pieces.
sql/ha_berkeley.cc:
  New handlerton pieces, plus changes for ha_init changes. I'm not happy with our current "skip" setup.
sql/ha_berkeley.h:
  Change in init return.
sql/ha_blackhole.cc:
  Changes for new handlerton pieces.
sql/ha_federated.cc:
  Changes for new handlerton and true cleanup code.
sql/ha_heap.cc:
  Changes for new handlerton returns.
sql/ha_innodb.cc:
  Changes for handlerton code.
sql/ha_innodb.h:
  Change in init.
sql/ha_myisam.cc:
  Changes for additional handlerton bits.
sql/ha_myisammrg.cc:
  Changes for  new handlerton bits.
sql/ha_ndbcluster.cc:
  Changes for new handlerton bits.
sql/ha_ndbcluster.h:
  Changes for handlerton bits.
sql/handler.cc:
  Changes for ditching show_table_type_st types, and collapsing it into a handlerton array. The ha_init now just loops through all handlers to init (much cleaner...). handlertons and sys_table_types should be merged next.
sql/handler.h:
  Additions for sys_table_types
sql/log.cc:
  Clean up of binlog for changes in handlerton
sql/mysql_priv.h:
  Removed unneeded define for binlog_init
sql/sql_show.cc:
  Changes for change in handlerton to sys_table_types