fix scheduler restart

[27149chen]

What does this do?

There is an issue introduced by #819

Steps to reproduce:

1. use a job function starts with ctx and do not pass an ctx
2. stop the scheduler by running .StopJobs()
3. start the scheduler by running .Start()
4. the job function does not work properly since ctx is already cancelled

This pr fix it by updating the ctx after job is stopped

Which issue(s) does this PR fix/relate to?

related to #819

List any changes that modify/break current functionality

Have you included tests for your changes?

yes

Did you document any new/modified functionality?

• Updated example_test.go
• Updated README.md

Notes

[JohnRoesler]

hey @27149chen thanks for the PR. Right now, this causes a race condition because we haven't waited for the jobs to fully stopped. If you duplicate and move this for loop block down right above s.stopErrCh <- err then we'll have waited for the <-s.exec.done chan to report back and update the job params race free.

Here is the full stopScheduler func with that change:

func (s *scheduler) stopScheduler() {
	s.logger.Debug("gocron: stopping scheduler")
	if s.started {
		s.exec.stopCh <- struct{}{}
	}

	for _, j := range s.jobs {
		j.stop()
	}
	for _, j := range s.jobs {
		<-j.ctx.Done()
	}
	var err error
	if s.started {
		t := time.NewTimer(s.exec.stopTimeout + 1*time.Second)
		select {
		case err = <-s.exec.done:
			t.Stop()
		case <-t.C:
			err = ErrStopExecutorTimedOut
		}
	}
	for id, j := range s.jobs {
		oldCtx := j.ctx
		if j.parentCtx == nil {
			j.parentCtx = s.shutdownCtx
		}
		j.ctx, j.cancel = context.WithCancel(j.parentCtx)

		// also replace the old context with the new one in the parameters
		if len(j.parameters) > 0 && j.parameters[0] == oldCtx {
			j.parameters[0] = j.ctx
		}

		s.jobs[id] = j
	}

	s.stopErrCh <- err
	s.started = false
	s.logger.Debug("gocron: scheduler stopped")
}

[27149chen]

@JohnRoesler thanks, updated