Recurring Job Scheduler

Implement scheduled_job_runs(jobs, dependencies), which simulates a scheduler for recurring jobs and returns every run it performs.

Inputs

• jobs is a list of [job_id, start, interval, count]. job_id is a unique string. The job must run exactly count times.
• dependencies is a list of [before, after] pairs: job after may not run at all until job before has run at least once.

Scheduling rules

• A job's first run is due at time start. Every later run is due interval time units after the job's previous actual run (so a delayed run shifts all later runs).
• Runs take no time. A run is *ready* at its due time if each of its prerequisites has already run at least once. If not, it waits, and becomes ready at the exact moment its last missing prerequisite runs.
• Whenever several runs are ready at the same time, the scheduler executes the one with the smallest job_id (normal string comparison) first. A run that becomes ready during a timestamp (because a prerequisite just ran) joins the runs still waiting at that timestamp.

Return a list of [time, job_id] records, one per run, in the order the runs are executed.