DBMS_SCHEDULER implementation
The package uses tables under sys to store jobs, programs, schedules, and execution records. Job owners can inspect their job definitions through sys.user_scheduler_jobs and run records through sys.user_scheduler_job_run_details.
EVALUATE_CALENDAR_STRING computes the next occurrence from timestamps and a repeat expression. RUN_JOB(…, use_current_session ⇒ TRUE) executes the job in the calling backend. Calendar evaluation and synchronous execution can be used before enabling background scheduling.
Automatic execution requires the ivorysql_ora library in shared_preload_libraries and ivorysql.scheduler = on at server startup. The launcher dispatches work in databases configured in ivorysql.scheduler_databases. The package and catalog definitions are in dbms_scheduler—1.0.sql; the launcher and job worker are implemented in scheduler_launcher.c and scheduler_worker.c in the same source directory.
See DBMS_SCHEDULER usage for a synchronous example.