removed rearrange() and throw() from AnalysisJobAdaptor by removing the unused CreateNewJob() method

Storable->new() now performs setting of all parameters without the use of rearrange (NB: change of call format!)

core-independent version of DBAdaptor. new() clean from URLFactory and self-reference, but still plays nice with optional Registry

fix: HiveGeneric_conf needs to declare that it in fact uses Registry (this could move into EnsemblGeneric_conf together with ApiVersion)

removed dependency on Core's DBSQL::BaseAdaptor and Storable by introducing own lightweight Hive::Storable

cosmetic: added deprecated message to CreateNewJob() and some documentation to store_jobs_and_adjust_counters()

store_jobs_and_adjust_counters() now also deals with semaphore_count corrections. It is now the preferred job-creation interface.

bugfix: no hope for other DBD drivers to support detection of auto-incrementing fields => use our inernal fixed rule

bugfix: stop printing overflow-related debug output to STDOUT; interface with overflow adaptor simplified

bugfix: substituting --password for --pass to pacify mysql 5.5.34+ that no longer recognizes options by their prefix

drop fetch_all_by_from_analysis_id_and_branch_code() method and rely on AUTOLOADed BaseAdaptor's methods instead

If greedy disoriented Workers overshoot analysis_capacity|hive_capacity, make sure they respecialize after the first batch

Individual Worker's specialization is now routed through Scheduler's main subroutine, to enforce more checks