CSC/ECE 517 Fall 2007/wiki1b 5 b4: Difference between revisions

From Expertiza_Wiki
Jump to navigation Jump to search
No edit summary
No edit summary
Line 5: Line 5:
One common use of AOP is profiling.  Profiling refers to the collection of data about the dynamic behavior of a system.  It is done to determine a breakdown of where time and memory are being consumed and make optimizations based on the result.  In this example, we create an aspect that will determine the number of (milli)seconds spent in a method.  It is kept very simple to clearly illustrate the functionality of AspectR, and would need extensive enhancement to actually be used.  A full-featured profiler would need to account for non-serialized and nested method invocations, and would need to aggregate the results, for example.
One common use of AOP is profiling.  Profiling refers to the collection of data about the dynamic behavior of a system.  It is done to determine a breakdown of where time and memory are being consumed and make optimizations based on the result.  In this example, we create an aspect that will determine the number of (milli)seconds spent in a method.  It is kept very simple to clearly illustrate the functionality of AspectR, and would need extensive enhancement to actually be used.  A full-featured profiler would need to account for non-serialized and nested method invocations, and would need to aggregate the results, for example.


<pre>
require 'aspectr'
require 'aspectr'
include AspectR
include AspectR

Revision as of 23:59, 30 September 2007

Hold for Robin Yehle Wiki Assignment 1b AspectR is a very useful Ruby module, but it is not easy to find documentation on it that is appropriate for students taking this class. Find, or construct, documentation that explains what it does without presuming previous knowledge of AspectJ, that describes many or all methods of the module and how they work. Also find or produce an easy-to-understand example that does not involve logging.

Example: Profiler Aspect

One common use of AOP is profiling. Profiling refers to the collection of data about the dynamic behavior of a system. It is done to determine a breakdown of where time and memory are being consumed and make optimizations based on the result. In this example, we create an aspect that will determine the number of (milli)seconds spent in a method. It is kept very simple to clearly illustrate the functionality of AspectR, and would need extensive enhancement to actually be used. A full-featured profiler would need to account for non-serialized and nested method invocations, and would need to aggregate the results, for example.

require 'aspectr'
include AspectR
class Profiler < Aspect
   
  def profiler_enter(method, object, exitstatus, *args) 
    @enterTime = Time.now
    puts "#{@enterTime.strftime('%Y-%m-%d %X')} #{self.class}##{method}: #{args.inspect}" 
   end
   
  def profiler_exit(method, object, exitstatus, *args) 
    @exitTime = Time.now
    print "#{@exitTime.strftime('%Y-%m-%d %X')} #{self.class}##{method}: exited " 
    if exitstatus.kind_of?(Array) 
      print "normally returning #{exitstatus[0].inspect} " 
    elsif exitstatus == true 
      print "with exception '#{$!}' " 
    else 
      print "normally " 
    end 
    puts "after #{@exitTime.to_f - @enterTime.to_f} seconds"
  end
end