Use the Rails application logger

Setting the logger to be the Active Job logger changes the behavior
compared to using logger directly on the job.

The reason is that previously we ended up logging to Rails.logger,
because when ActiveJob initializes itself in the context of a Rails
application, it sets its logger to the Rails one:
https://github.com/rails/rails/blob/v7.0.7.2/activejob/lib/active_job/railtie.rb#L13-L15

Whereas saving the value of ActiveJob::Base.logger early will use the
default Active Job logger which logs to stdout:
https://github.com/rails/rails/blob/v7.0.7.2/activejob/lib/active_job/logging.rb#L11

This prevents early binding and keeps the pre-1.4.0 behavior of using
the current value of ActiveJob::Base.logger during jobs.

Gemfile.lock

@@ -7,7 +7,7 @@ GIT
 PATH
   remote: .
   specs:
-    job-iteration (1.4.0)
+    job-iteration (1.4.1)
       activejob (>= 5.2)
 
 GEM

lib/job-iteration.rb

@@ -13,9 +13,13 @@ module JobIteration
 
   extend self
 
-  attr_accessor :logger
+  attr_writer :logger
 
-  self.logger = ActiveJob::Base.logger
+  class << self
+    def logger
+      @logger || ActiveJob::Base.logger
+    end
+  end
 
   # Use this to _always_ interrupt the job after it's been running for more than N seconds.
   # @example

lib/job-iteration/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module JobIteration
-  VERSION = "1.4.0"
+  VERSION = "1.4.1"
 end