CHANGELOG.md

Rails 7.1.0 (October 05, 2023)

• No changes.

Rails 7.1.0.rc2 (October 01, 2023)

• Make sure scheduled_at is a Time object when asserting enqueued jobs.

  Rafael Mendonça França

Rails 7.1.0.rc1 (September 27, 2023)

• Set scheduled_at attribute as a Time object instead of epoch seconds, and serialize and deserialize the value when enqueued. Assigning a numeric/epoch value to scheduled_at= is deprecated; use a Time object instead.

  Deserializes enqueued_at as a Time instead of ISO8601 String.

  Ben Sheldon

• Clarify the backoff strategy for the recommended :wait option when retrying jobs

  wait: :exponentially_longer is waiting polynomially longer, so it is now recommended to use wait: :polynomially_longer to keep the same behavior.

  Victor Mours

Rails 7.1.0.beta1 (September 13, 2023)

• Fix Active Job log message to correctly report a job failed to enqueue when the adapter raises an ActiveJob::EnqueueError.

  Ben Sheldon

• Add after_discard method.

  This method lets job authors define a block which will be run when a job is about to be discarded. For example:

  class AfterDiscardJob < ActiveJob::Base
    after_discard do |job, exception|
      Rails.logger.info("#{job.class} raised an exception: #{exception}")
    end
  
    def perform
      raise StandardError
    end
  end

  The above job will run the block passed to after_discard after the job is discarded. The exception will still be raised after the block has been run.

  Rob Cardy

• Fix deserialization of ActiveSupport::Duration

  Previously, a deserialized Duration would return an array from Duration#parts. It will now return a hash just like a regular Duration.

  This also fixes an error when trying to add or subtract from a deserialized Duration (eg duration + 1.year).

  Jonathan del Strother

• perform_enqueued_jobs is now compatible with all Active Job adapters

  This means that methods that depend on it, like Action Mailer's assert_emails, will work correctly even if the test adapter is not used.

  Alex Ghiculescu

• Allow queue adapters to provide a custom name by implementing queue_adapter_name

  Sander Verdonschot

• Log background job enqueue callers

  Add verbose_enqueue_logs configuration option to display the caller of background job enqueue in the log to help with debugging.

  Example log line:

  Enqueued AvatarThumbnailsJob (Job ID: ab528951-41fb-4c48-9129-3171791c27d6) to Sidekiq(default) with arguments: 1092412064
  ↳ app/models/user.rb:421:in `generate_avatar_thumbnails'
  

  Enabled in development only for new and upgraded applications. Not recommended for use in the production environment since it relies on Ruby's Kernel#caller which is fairly slow.

  fatkodima

• Set provider_job_id for Backburner jobs

  Cameron Matheson

• Add perform_all_later to enqueue multiple jobs at once

  This adds the ability to bulk enqueue jobs, without running callbacks, by passing multiple jobs or an array of jobs. For example:

  ActiveJob.perform_all_later(MyJob.new("hello", 42), MyJob.new("world", 0))
  
  user_jobs = User.pluck(:id).map { |id| UserJob.new(user_id: id) }
  ActiveJob.perform_all_later(user_jobs)

  This can greatly reduce the number of round-trips to the queue datastore. For queue adapters that do not implement the new enqueue_all method, we fall back to enqueuing jobs individually. The Sidekiq adapter implements enqueue_all with push_bulk.

  This method does not use the existing enqueue.active_job event, but adds a new event enqueue_all.active_job.

  Sander Verdonschot

• Don't double log the job when using ActiveRecord::QueryLog

  Previously if you set config.active_record.query_log_tags to an array that included :job, the job name would get logged twice. This bug has been fixed.

  Alex Ghiculescu

• Add support for Sidekiq's transaction-aware client

  Jonathan del Strother

• Remove QueAdapter from Active Job.

  After maintaining Active Job QueAdapter by Rails and Que side to support Ruby 3 keyword arguments and options provided as top level keywords, it is quite difficult to maintain it this way.

  Active Job Que adapter can be included in the future version of que gem itself.

  Yasuo Honda

• Fix BigDecimal (de)serialization for adapters using JSON.

  Previously, BigDecimal was listed as not needing a serializer. However, when used with an adapter storing the job arguments as JSON, it would get serialized as a simple String, resulting in deserialization also producing a String (instead of a BigDecimal).

  By using a serializer, we ensure the round trip is safe.

  To ensure applications using BigDecimal job arguments are not subject to race conditions during deployment (where a replica running a version of Rails without BigDecimalSerializer fails to deserialize an argument serialized with it), ActiveJob.use_big_decimal_serializer is disabled by default, and can be set to true in a following deployment..

  Sam Bostock

• Preserve full-precision enqueued_at timestamps for serialized jobs, allowing more accurate reporting of how long a job spent waiting in the queue before it was performed.

  Retains IS08601 format compatibility.

  Jeremy Daer

• Add --parent option to job generator to specify parent class of job.

  Example:

  bin/rails g job process_payment --parent=payment_job generates:

  class ProcessPaymentJob < PaymentJob
    # ...
  end

  Gannon McGibbon

• Add more detailed description to job generator.

  Gannon McGibbon

• perform.active_job notification payloads now include :db_runtime, which is the total time (in ms) taken by database queries while performing a job. This value can be used to better understand how a job's time is spent.

  Jonathan Hefner

• Update ActiveJob::QueueAdapters::QueAdapter to remove deprecation warning.

  Remove a deprecation warning introduced in que 1.2 to prepare for changes in que 2.0 necessary for Ruby 3 compatibility.

  Damir Zekic and Adis Hasovic

• Add missing bigdecimal require in ActiveJob::Arguments

  Could cause uninitialized constant ActiveJob::Arguments::BigDecimal (NameError) when loading Active Job in isolation.

  Jean Boussier

• Allow testing discard_on/retry_on ActiveJob::DeserializationError

  Previously in perform_enqueued_jobs, deserialize_arguments_if_needed was called before calling perform_now. When a record no longer exists and is serialized using GlobalID this led to raising an ActiveJob::DeserializationError before reaching perform_now call. This behavior makes difficult testing the job discard_on/retry_on logic.

  Now deserialize_arguments_if_needed call is postponed to when perform_now is called.

  Example:

  class UpdateUserJob < ActiveJob::Base
    discard_on ActiveJob::DeserializationError
  
    def perform(user)
      # ...
    end
  end
  
  # In the test
  User.destroy_all
  assert_nothing_raised do
    perform_enqueued_jobs only: UpdateUserJob
  end

  Jacopo Beschi

Please check 7-0-stable for previous changes.