v0.1.12

• add worker.RunOnce. This is useful for external schedulers like k8s. The k8s scheduler will start a pod, and this pod only runs a job.
• add custom retry strategy
• update dependencies

v0.1.11

• breaking change: sidekiq queue needs to explicitly pull to move jobs from sidekiq ({ns1}:low) to work-compatible queue ({ns1}:low:SomeWorker).

	q.Pull(&PullOptions{
		Namespace:        "{ns1}",
		SidekiqNamespace: "{ns1}",
		SidekiqQueue:     "low",
	})

SidekiqNamespace can be empty if you don't use redis-namespace. In that case, the sidekiq queue source is low instead of {ns1}:low.

v0.1.10

• if context handler is used, worker.Stop will be able to quickly stop current job execution.
• add external enqueuer interface to enqueue job into other queue systems.

// ExternalEnqueuer enqueues a job with other queue protocol.
// Queue adaptor that implements this can publish jobs directly to other types of queue systems.
type ExternalEnqueuer interface {
	ExternalEnqueue(*Job, *EnqueueOptions) error
}

• sidekiq.Queue implements work.ExternalEnqueuer

v0.1.9

• use error func only for redis errors

v0.1.8

• add redis cluster tests
• use errors.is instead of direct comparison

v0.1.7

• add discard max retry middleware

v0.1.6

• revert to previous max backoff interval

v0.1.5

• add more sidekiq fields related to retries
• fix backoff after upgrading to cenkalti/backoff/v4

v0.1.4

• redis queue supports bulk job finder interface (#40)
• support context.Context in job handler (#41)
• test against go 1.15 and redis 6 (#42)
• add "Do not retry" error (#39) (handler returns an error, but it will not trigger default exponential backoff. you can use it to implement custom backoff algorithm by reenqueuing)
• add redislock package for future custom ratelimit package (#45)

v0.1.3

• dependency update
• remove pkg/errors from dependencies