quebert 3.0.2 → 3.0.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: fc2101289db4ff8fbc2ddc023b71b021ec6a2246
-  data.tar.gz: 766e6e57605b257f48fd0e97fdbe35b60ea75d2a
+  metadata.gz: 107451b4a0b3a15d3e42285f65c1c53be6700c52
+  data.tar.gz: ce2f9fd96496d6701eadad9abd94acc33a7fa6d4
 SHA512:
-  metadata.gz: 64744b734d54da02e310a292bebbf7f4eaf5c0a1f5f3d9f4d37299fad057edb4413480e723ea5ebec6f958f083d17825ee10a4b642d4cf28a962eb25d6f292cf
-  data.tar.gz: cab81b5345e3af905ebfedd1356a3a2ca186a2df4130a09e3e0eae202f5f4aeb44a98cfd8aae96e25354c88997e0cf5941a2d5f7e6016ba4264f510ce825e71b
+  metadata.gz: 06fe749f66920ee7ebb436b72d82c3d0f2f68049cf5dfbba10c9ca3e7ac76eebfb60644cab07634d23fb9d8863b9b5c16880c45b3191f91a6ee187604fc2adea
+  data.tar.gz: 3523257233d8d1f7732be791ae48a6bb7f359ba063231908c4c5957e9411333b48ea32ecd70daec03a0dd915847b087abf4247f21b33c78fadab3575e44e5b84

data/.ruby-version

@@ -1 +1 @@
-2.1.6
+2.2.5

data/.travis.yml

@@ -1,6 +1,6 @@
 language: ruby
 rvm:
-  - 2.0.0
+  - 2.2.5
 before_install:
   - curl -L https://github.com/kr/beanstalkd/archive/v1.9.tar.gz | tar xz -C /tmp
   - cd /tmp/beanstalkd-1.9/

data/README.md

@@ -1,35 +1,39 @@
 # Quebert
 
-[![Build Status](https://travis-ci.org/polleverywhere/quebert.png?branch=master)](https://travis-ci.org/polleverywhere/quebert) [![Code Climate](https://codeclimate.com/repos/555266fc6956805b9e0033b5/badges/008e51483e8e268f21db/gpa.svg)](https://codeclimate.com/repos/555266fc6956805b9e0033b5/feed)
+[![Build Status](https://travis-ci.org/polleverywhere/quebert.png?branch=master)](https://travis-ci.org/polleverywhere/quebert) [![Code Climate](https://codeclimate.com/github/polleverywhere/quebert/badges/gpa.svg)](https://codeclimate.com/github/polleverywhere/quebert)
 
-async_observer is great, but is dated and doesn't really support running jobs outside of the async_send idiom. Quebert is an attempt to mix how jobs are run in other popular worker queue frameworks, like resque and dj, with async_observer so that you can have it both ways.
+Quebert is a ruby background worker library that works with the very fast and simple [beanstalkd](http://kr.github.io/beanstalkd/) deamon.
 
-# Why Quebert (or how is it different from DJ and Resque)?
+## Why Quebert?
 
-Because it has really low latency. Other Ruby queuing frameworks, like DJ or Resque, have to poll their queue servers periodicly. You could think of it as a "pull" queue. Quebert is a "push" queue. It maintains a persistent connection with beanstalkd and when is enqueud, its instantly pushed to the workers and executed.
+Because it has really low latency. Other Ruby queuing frameworks, like [DJ](https://github.com/collectiveidea/delayed_job) or [Resque](https://github.com/resque/resque), have to poll their queue servers periodicly. You could think of it as a "pull" queue. Quebert is a "push" queue. It maintains a persistent connection with beanstalkd and when is enqueud, its instantly pushed to the workers and executed.
 
-# Who uses it?
+[Sidekiq](http://sidekiq.org) uses Redis's "push" primitives so it has low latency, but it doesn't support class reloading in a development environment. Sidekiq is also threaded, which means there are no guarantees of reliability when running non-threadsafe code.
 
-Quebert is a serious project. Its used in a production environment at Poll Everywhere to handle everything from SMS message processing to account downgrades.
+[Backburner](https://github.com/nesquena/backburner) is very similar to Quebert. It offers more options for concurrency (threading, forking, etc.) than queubert but lacks pluggable back-ends, which means you'll be stubbing and mocking async calls.
 
-# Features
+## Who uses it?
 
-* Multiple back-ends (InProcess, Sync, and Beanstalk)
-* Rails/ActiveRecord integration similar to async_observer
-* Pluggable exception handling (for Hoptoad integration)
-* Run workers with pid, log, and config files. These do not daemonize (do it yourself punk!)
-* Provide custom hooks to be called before, after & around jobs are run
+Quebert is a serious project. Its used in a production environment at [Poll Everywhere](https://www.polleverywhere.com/) to handle everything from SMS message processing to account downgrades.
 
-Some features that are currently missing that I will soon add include:
+## Features
+
+* **Multiple back-ends** (InProcess, Sync, and Beanstalk)
+* **Rails/ActiveRecord integration** similar to async_observer
+* **Pluggable exception handling** (for Hoptoad integration)
+* **Run workers with pid, log, and config files**. These do not daemonize (do it yourself punk!)
+* **Custom hooks** that may be called before, after & around jobs are run
+
+Some features that are currently *missing* that I will soon add include:
 
 * Rails plugin support (The AR integrations have to be done manually today)
 * Auto-detecting serializers. Enhanced ClassRegistry to more efficiently look up serializers for objects.
 
-# How to use
+## How to use
 
 There are two ways to enqueue jobs with Quebert: through the Job itself, provided you set a default back-end for the job, or put it on the backend.
 
-## Jobs
+### Jobs
 
 Quebert includes a Job class so you can implement how you want certain types of Jobs performed.
 
@@ -52,7 +56,10 @@ Quebert.backend.put WackyMathWizard.new(1, 2, 3)
 Or drop it in right from the job:
 
 ```ruby
+# Run job right away!
 WackyMathWizard.new(4, 5, 6).enqueue
+# Run a lower priority job in 10 seconds for a max of 120 seconds
+WackyMathWizard.new(10, 10, 10).enqueue(ttr: 120, priority: 100, delay: 10)
 ```
 
 Then perform the jobs!
@@ -60,9 +67,10 @@ Then perform the jobs!
 ```ruby
 Quebert.backend.reserve.perform # => 6
 Quebert.backend.reserve.perform # => 15
+Quebert.backend.reserve.perform # => 30
 ```
 
-## Rails integration
+### Rails integration
 
 config/quebert.yml:
 
@@ -83,7 +91,7 @@ Quebert.config.from_hash(Rails.application.config.quebert)
 Quebert.config.logger = Rails.logger
 ```
 
-## Before/After/Around Hooks
+### Job & Worker hooks
 
 Quebert has support for providing custom hooks to be called before, after & around your jobs are being run.
 A common example is making sure that any active ActiveRecord database connections are put back on the connection pool after a job is done:
@@ -104,7 +112,7 @@ Quebert.config.around_job do |job|
 end
 ```
 
-## Async Sender
+### Async sender
 
 Take any ol' class and include the Quebert::AsyncSender.
 
@@ -162,13 +170,13 @@ Quebert.backend.reserve.perform # => "waazup Coraline!"
 
 * Only basic data types are included for serialization. Serializers may be customized to include support for different types.
 
-## Backends
+### Backends
 
 * Beanstalk: Enqueue jobs in a beanstalkd service. The workers run in a separate process. Typically used in production environments.
 * Sync: Perform jobs immediately upon enqueuing. Typically used in testing environments.
 * InProcess: Enqueue jobs in an in-memory array. A worker will need to reserve a job to perform.
 
-## Using multiple queues
+### Multiple queues
 
 To start a worker pointed at a non-default queue (e.g., a Quebert "tube"), start the process with `-q`:
 
@@ -190,16 +198,29 @@ class FooJob < Quebert::Job
 end
 ```
 
-## Beanstalk: Changing a job's TTR
+### Setting job defaults
+
+A `Quebert::Job` is a Plain Ol' Ruby Object. The defaults of a job, including its `ttr`, `queue_name`, and `delay` may be overridden in a super class as follows:
 
 ```ruby
+# Assuming you're in Rails or using ActiveSupport
 class FooJob < Quebert::Job
   def ttr
     5.minutes
   end
 
+  def delay
+    30.seconds
+  end
+
+  def queue_name
+    "long-running-delayed-jobs"
+  end
+
   def perform(args)
     # ...
   end
 end
 ```
+
+Take a look at the [`Quebert::Job` class](https://github.com/polleverywhere/quebert/blob/master/lib/quebert/job.rb) code for more details on methods you may ovveride.

data/lib/quebert/controller/beanstalk.rb

@@ -22,7 +22,7 @@ module Quebert
       end
 
       def perform
-        logger.error(job) { "Performing with args #{job.args.inspect}" }
+        logger.error(job) { "Performing #{job.class.name} with args #{job.args.inspect}" }
         logger.error(job) { "Beanstalk Job Stats: #{beanstalk_job.stats.inspect}" }
 
         result = false
@@ -31,7 +31,7 @@ module Quebert
           beanstalk_job.delete
         end
 
-        logger.error(job) { "Completed in #{(time*1000*1000).to_i/1000.to_f} ms\n" }
+        logger.info(job) { "Completed in #{(time*1000*1000).to_i/1000.to_f} ms\n" }
         result
       rescue Job::Delete
         logger.info(job) { "Deleting job" }

data/lib/quebert/version.rb

@@ -1,3 +1,3 @@
 module Quebert
-  VERSION = "3.0.2"
+  VERSION = "3.0.3"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: quebert
 version: !ruby/object:Gem::Version
-  version: 3.0.2
+  version: 3.0.3
 platform: ruby
 authors:
 - Brad Gessler
@@ -10,7 +10,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-06-12 00:00:00.000000000 Z
+date: 2016-11-29 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: json
@@ -131,7 +131,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: quebert
-rubygems_version: 2.2.3
+rubygems_version: 2.4.5.1
 signing_key: 
 specification_version: 4
 summary: A worker queue framework built around beanstalkd
@@ -149,4 +149,3 @@ test_files:
 - spec/support/jobs.rb
 - spec/support_spec.rb
 - spec/worker_spec.rb
-has_rdoc: