Introducing: ActiveJob, and support for other queue adapters (#18)

* Add new ActiveJobs to replace DJs
* Fix test error (wasn't resolving constant for me)
* Enqueue ActiveJob instead of DJ (+spec coverage)
* Remove 'enqueue' abstraction since it doesn't really map to AJs
* Enable appraisal when running top-level rake
* Rework README to be more adapter-agnostic (as long as it's DB-backed, i.e. supports cotransactionality!)
* Raise an exception in production unless the queue adapter is supported
* Decouple entirely from delayed_job dependency
* No need for binding.pry when we have binding.irb
* Remove unused file
* Switch to sqlite3 in-memory DB for testing
* Trim down deps - railties, activejob, activerecord
* Increment version to 3.0.0
* Regenerate coach config (no more postgres DB)
* Remove unused sprockets test dep
* Mention priority support being only Rails 6+
* Apply README suggestion
Co-authored-by: Sam Moore <sam.moore@hey.com>
* Push queue adapter safety check out until after initialization
Co-authored-by: Sam Moore <sam.moore@hey.com>

Appraisals

@@ -1,11 +1,15 @@
 appraise 'rails-5-1' do
-  gem 'rails', '~> 5.1.0'
+  gem 'railties', '~> 5.1.0'
 end
 
 appraise 'rails-5-2' do
-  gem 'rails', '~> 5.2.0'
+  gem 'railties', '~> 5.2.0'
 end
 
 appraise 'rails-6-0' do
-  gem 'rails', '~> 6.0.0'
+  gem 'railties', '~> 6.0.0'
+end
+
+appraise 'rails-6-1' do
+  gem 'railties', '~> 6.1.0'
 end

README.md

@@ -1,16 +1,16 @@
 # Journaled
 
-A Rails engine to durably deliver schematized events to Amazon Kinesis via DelayedJob.
+A Rails engine to durably deliver schematized events to Amazon Kinesis via ActiveJob.
 
 More specifically, `journaled` is composed of three opinionated pieces:
 schema definition/validation via JSON Schema, transactional enqueueing
-via Delayed::Job (specifically `delayed_job_active_record`), and event
+via ActiveJob (specifically, via a DB-backed queue adapter), and event
 transmission via Amazon Kinesis. Our current use-cases include
 transmitting audit events for durable storage in S3 and/or analytical
 querying in Amazon Redshift.
 
 Journaled provides an at-least-once event delivery guarantee assuming
-Delayed::Job is configured not to delete jobs on failure.
+ActiveJob's queue adapter is not configured to delete jobs on failure.
 
 Note: Do not use the journaled gem to build an event sourcing solution
 as it does not guarantee total ordering of events. It's possible we'll
@@ -20,9 +20,20 @@ durable, eventually consistent record that discrete events happened.
 
 ## Installation
 
-1. [Install `delayed_job_active_record`](https://github.com/collectiveidea/delayed_job_active_record#installation)
-if you haven't already.
+1. If you haven't already,
+[configure ActiveJob](https://guides.rubyonrails.org/active_job_basics.html)
+to use one of the following queue adapters:
 
+- `:delayed_job` (via `delayed_job_active_record`)
+- `:que`
+- `:good_job`
+- `:delayed`
+
+Ensure that your queue adapter is not configured to delete jobs on failure.
+
+**If you launch your application in production mode and the gem detects that
+`ActiveJob::Base.queue_adapter` is not in the above list, it will raise an exception
+and prevent your application from performing unsafe journaling.**
 
 2. To integrate Journaled into your application, simply include the gem in your
 app's Gemfile.
@@ -85,9 +96,11 @@ Journaling provides a number of different configuation options that can be set i
 
 #### `Journaled.job_priority` (default: 20)
 
-  This can be used to configure what `priority` the Delayed Jobs are enqueued with. This will be applied to all the Journaled::Devivery jobs that are created by this application.
+  This can be used to configure what `priority` the ActiveJobs are enqueued with. This will be applied to all the `Journaled::DeliveryJob`s that are created by this application.
   Ex: `Journaled.job_priority = 14`
 
+  _Note that job priority is only supported on Rails 6.0+. Prior Rails versions will ignore this parameter and enqueue jobs with the underlying ActiveJob adapter's default priority._
+
 #### `Journaled.http_idle_timeout` (default: 1 second)
 
   The number of seconds a persistent connection is allowed to sit idle before it should no longer be used.
@@ -100,9 +113,9 @@ Journaling provides a number of different configuation options that can be set i
 
   The number of seconds before the :http_handler should timeout while waiting for a HTTP response.
 
-#### DJ `enqueue` options
+#### ActiveJob `set` options
 
-Both model-level directives accept additional options to be passed into DelayedJob's `enqueue` method:
+Both model-level directives accept additional options to be passed into ActiveJob's `set` method:
 
 ```ruby
 # For change journaling:

Rakefile

@@ -28,7 +28,13 @@ if %w(development test).include? Rails.env
   RuboCop::RakeTask.new
 
   task(:default).clear
-  task default: %i(rubocop spec)
+  if ENV['APPRAISAL_INITIALIZED'] || ENV['CI']
+    task default: %i(rubocop spec)
+  else
+    require 'appraisal'
+    Appraisal::Task.new
+    task default: :appraisal
+  end
 
   task 'db:test:prepare' => 'db:setup'
 end

app/jobs/journaled/application_job.rb

@@ -0,0 +1,4 @@
+module Journaled
+  class ApplicationJob < Journaled.job_base_class_name.constantize
+  end
+end