Perforce integration

.gitignore

@@ -6,3 +6,4 @@ pkg/*
 .bundle
 .rvmrc
 turbulence/*
+tmp/*

README.md

@@ -14,7 +14,15 @@ In your project directory, run:
 
     $ bule
 
-For now it just dumps out a hash of churn + flog metrics
+and it will generate (and open) turbulence/turbulence.html 
+
+Supported SCM systems
+---------------------
+Currently, bule defaults to using git. If you are using Perforce, call it like so:
+
+    $ bule --scm p4
+
+You need to have an environment variable P4CLIENT set to the name of your client workspace.
 
 WARNING
 -------

lib/turbulence/calculators/churn.rb

@@ -5,7 +5,7 @@ class Churn
 
       class << self
         attr_accessor :scm, :compute_mean, :commit_range
-        
+
         def for_these_files(files)
           changes_by_ruby_file.each do |filename, count|
             yield filename, count if files.include?(filename)
@@ -14,14 +14,21 @@ def for_these_files(files)
 
         def changes_by_ruby_file
           ruby_files_changed_in_scm.group_by(&:first).map do |filename, stats|
-            churn = stats[0..-2].map(&:last).inject(0){|n, i| n + i}
-            if compute_mean && stats.size > 1
-              churn /= (stats.size-1)
-            end
-            [filename, churn]
+            churn_for_file(filename,stats)
           end
         end
 
+        def churn_for_file(filename,stats)
+          churn = stats[0..-2].map(&:last).inject(0){|running_total, changes| running_total + changes}
+          churn = calculate_mean_of_churn(churn, stats.size - 1) if compute_mean
+          [filename, churn]
+        end
+
+        def calculate_mean_of_churn(churn, sample_size) 
+          return churn if sample_size < 1
+          churn /= sample_size
+        end
+
         def ruby_files_changed_in_scm
           counted_line_changes_by_file_by_commit.select do |filename, _|
             filename.end_with?(RUBY_FILE_EXTENSION) && File.exist?(filename)

lib/turbulence/scatter_plot_generator.rb

@@ -35,23 +35,31 @@ def mangle
     end
 
     def to_js
+      clean_metrics_from_missing_data
+      directory_series = {}
+      grouped_by_directory.each_pair do |directory, metrics_hash|
+        directory_series[directory] = file_metrics_for_directory(metrics_hash)      end
+
+      "var directorySeries = #{directory_series.to_json};"
+    end
+
+    def clean_metrics_from_missing_data
       metrics_hash.reject! do |filename, metrics|
         metrics[x_metric].nil? || metrics[y_metric].nil?
       end
-
-      grouped_by_directory = metrics_hash.group_by do |filename, _|
+    end
+
+    def grouped_by_directory
+      metrics_hash.group_by do |filename, _|
         directories = File.dirname(filename).split("/")
         directories[0..1].join("/")
       end
+    end
 
-      directory_series = {}
-      grouped_by_directory.each_pair do |directory, metrics_hash|
-        directory_series[directory] = metrics_hash.map do |filename, metrics|
-          {:filename => filename, :x => metrics[x_metric], :y => metrics[y_metric]}
-        end
+    def file_metrics_for_directory(metrics_hash)
+      metrics_hash.map do |filename, metrics|
+        {:filename => filename, :x => metrics[x_metric], :y => metrics[y_metric]}
       end
-
-      "var directorySeries = #{directory_series.to_json};"
     end
   end
 end

lib/turbulence/scm/git.rb

@@ -1,17 +1,17 @@
-class Turbulence
-  module Scm
-    class Git
-      class << self
-        def log_command(commit_range)
-          `git log --all -M -C --numstat --format="%n" #{commit_range}`
-        end
-
-        def is_repo?(directory)
-          FileUtils.cd(directory) {
-            return !(`git status 2>&1` =~ /Not a git repository/)
-          }
-        end
-      end
-    end
-  end
-end
+class Turbulence
+  module Scm
+    class Git
+      class << self
+        def log_command(commit_range = "")
+          `git log --all -M -C --numstat --format="%n" #{commit_range}`
+        end
+
+        def is_repo?(directory)
+          FileUtils.cd(directory) {
+            return !(`git status 2>&1` =~ /Not a git repository/)
+          }
+        end
+      end
+    end
+  end
+end

lib/turbulence/scm/perforce.rb

@@ -1,15 +1,90 @@
-class Turbulence
-  module Scm
-    class Perforce 
-      class << self
-        def log_command(commit_range)
-        end
-
-        def is_repo?(directory)
-          p4client = ENV['P4CLIENT']
-          return !(p4client.nil? or p4client.empty?)
-        end
-      end
-    end
-  end
-end
+require 'fileutils'
+require 'pathname'
+
+class Turbulence
+  module Scm
+    class Perforce 
+      class << self
+        def log_command(commit_range = "")
+          full_log = ""
+          changes.each do |cn|
+            files_per_change(cn).each do |file|
+              full_log << transform_for_output(file)
+            end
+          end
+          return full_log
+        end
+
+        def is_repo?(directory)
+          p4client = ENV['P4CLIENT']
+          return !((p4client.nil? or p4client.empty?) and not self.has_p4?)
+        end
+
+        def has_p4?
+          ENV['PATH'].split(File::PATH_SEPARATOR).any? do |directory|
+            File.executable?(File.join(directory, 'p4'))
+          end
+        end
+
+        def changes(commit_range = "")
+          p4_list_changes.each_line.map do |change|
+            change.match(/Change (\d+)/)[1]
+          end
+        end
+
+        def depot_to_local(depot_file)
+          abs_path = extract_clientfile_from_fstat_of(depot_file) 
+          Pathname.new(abs_path).relative_path_from(Pathname.new(FileUtils.pwd)).to_s
+        end
+
+        def extract_clientfile_from_fstat_of(depot_file)
+          p4_fstat(depot_file).each_line.select {
+            |line| line =~ /clientFile/
+          }[0].split(" ")[2].tr("\\","/")
+        end
+
+        def files_per_change(change)
+          describe_output = p4_describe_change(change).split("\n")
+          map = []
+          describe_output.each_index do |index|
+            if describe_output[index].start_with?("====")
+              fn = filename_from_describe(describe_output, index)
+              churn = sum_of_changes(describe_output[index .. index + 4].join("\n"))
+              map << [churn,fn]
+            end
+          end
+          return map
+        end
+
+        def filename_from_describe(output,index)
+          depot_to_local(output[index].match(/==== (\/\/.*)#\d+/)[1])
+        end
+
+        def transform_for_output(arr)
+          "#{arr[0]}\t0\t#{arr[1]}\n"
+        end
+
+        def sum_of_changes(p4_describe_output) 
+          churn = 0
+          p4_describe_output.each_line do |line|
+            next unless line =~ /(add|deleted|changed) .* (\d+) lines/
+            churn += line.match(/(\d+) lines/)[1].to_i
+          end
+          return churn
+        end
+
+        def p4_list_changes(commit_range = "")
+          `p4 changes -s submitted ...#{commit_range}`
+        end
+
+        def p4_fstat(depot_file)
+          `p4 fstat #{depot_file}`
+        end
+
+        def p4_describe_change(change)
+          `p4 describe -ds #{change}`
+        end
+      end
+    end
+  end
+end

spec/turbulence/calculators/churn_spec.rb

@@ -86,6 +86,16 @@
       calculator.compute_mean = false
       calculator.changes_by_ruby_file.should =~ [ ['lib/eddies.rb', 25], ['lib/turbulence.rb', 21]]
     end
+
+    it "interprets a single entry as zero churn" do
+      calculator.stub(:ruby_files_changed_in_scm) {
+        [
+          ['lib/eddies.rb', 4],
+        ]
+      }
+      calculator.compute_mean = false
+      calculator.changes_by_ruby_file.should =~ [ ['lib/eddies.rb', 0] ]
+    end
 
     it "groups and takes the mean of churns, excluding the last" do
       calculator.compute_mean = true
@@ -94,6 +104,21 @@
     end
   end
 
+  describe "::calculate_mean_of_churn" do
+    it "handles zero sample size" do
+      calculator.calculate_mean_of_churn(8,0).should == 8
+    end
+
+    it "returns original churn for sample size = 1"  do
+      calculator.calculate_mean_of_churn(8,1).should == 8
+    end
+
+    it "returns churn divided by sample size" do
+      calculator.calculate_mean_of_churn(25,3).should == 8
+    end
+
+  end
+
   context "Full stack tests" do
     context "when one ruby file is given" do
       context "with two log entries for file" do

spec/turbulence/command_line_interface_spec.rb

@@ -14,7 +14,7 @@
     end
     it "bundles the files" do
       cli.generate_bundle
-      Dir.glob('turbulence/*').should eq(["turbulence/cc.js", "turbulence/highcharts.js", "turbulence/jquery.min.js", "turbulence/turbulence.html"])
+      Dir.glob('turbulence/*').sort.should eq(["turbulence/cc.js", "turbulence/highcharts.js", "turbulence/jquery.min.js", "turbulence/turbulence.html"])
     end
   end
   describe "command line options" do

spec/turbulence/scatter_splot_generator_spec.rb

@@ -5,28 +5,80 @@
     it "generates JavaScript" do
       generator = Turbulence::ScatterPlotGenerator.new(
         "foo.rb" => {
-          Turbulence::Calculators::Churn => 1,
-          Turbulence::Calculators::Complexity => 2
-        }
+        Turbulence::Calculators::Churn => 1,
+        Turbulence::Calculators::Complexity => 2
+      }
       )
       generator.to_js.should =~ /var directorySeries/ 
-      generator.to_js.should =~ /\"filename\"\:\"foo.rb\"/ 
-      generator.to_js.should =~ /\"x\":1/ 
-      generator.to_js.should =~ /\"y\":2/ 
+        generator.to_js.should =~ /\"filename\"\:\"foo.rb\"/ 
+        generator.to_js.should =~ /\"x\":1/ 
+        generator.to_js.should =~ /\"y\":2/ 
     end  
   end
 
   context "with a missing Metric" do
     it "generates JavaScript" do
       generator = Turbulence::ScatterPlotGenerator.new(
         "foo.rb" => {
-          Turbulence::Calculators::Churn => 1
-        }
+        Turbulence::Calculators::Churn => 1
+      }
       )
       generator.to_js.should == 'var directorySeries = {};'
     end  
   end
 
+  describe "#clean_metrics_from_missing_data" do 
+    let(:spg) {Turbulence::ScatterPlotGenerator.new({})}
+
+    it "removes entries with missing churn" do 
+      spg.stub(:metrics_hash).and_return("foo.rb" => { 
+        Turbulence::Calculators::Complexity => 88.3})
+        spg.clean_metrics_from_missing_data.should == {}
+    end
+
+    it "removes entries with missing complexity" do
+      spg.stub(:metrics_hash).and_return("foo.rb" => { 
+        Turbulence::Calculators::Churn => 1})
+        spg.clean_metrics_from_missing_data.should == {}
+    end
+
+    it "keeps entries with churn and complexity present" do
+      spg.stub(:metrics_hash).and_return("foo.rb" => { 
+        Turbulence::Calculators::Churn => 1,
+        Turbulence::Calculators::Complexity => 88.3})
+        spg.clean_metrics_from_missing_data.should_not == {}
+    end
+  end
+
+  describe "#grouped_by_directory" do
+    let(:spg) {Turbulence::ScatterPlotGenerator.new("lib/foo/foo.rb" => {
+      Turbulence::Calculators::Churn => 1},
+      "lib/bar.rb" => {
+      Turbulence::Calculators::Churn => 2} )}
+
+      it "uses \".\" to denote flat hierarchy" do
+        spg.stub(:metrics_hash).and_return("foo.rb" => {
+          Turbulence::Calculators::Churn => 1
+        })
+        spg.grouped_by_directory.should ==  {"." => [["foo.rb", Turbulence::Calculators::Churn => 1]]}
+      end
+
+      it "takes full path into account" do
+        spg.grouped_by_directory.should ==  {"lib/foo" => [["lib/foo/foo.rb", Turbulence::Calculators::Churn => 1]],
+          "lib" => [["lib/bar.rb", Turbulence::Calculators::Churn => 2]]}
+      end
+  end
+
+  describe "#file_metrics_for_directory" do
+    let(:spg) {Turbulence::ScatterPlotGenerator.new({})}
+    it "assigns :filename, :x, :y" do
+      spg.file_metrics_for_directory("lib/foo/foo.rb" => {
+        Turbulence::Calculators::Churn => 1,
+        Turbulence::Calculators::Complexity => 88.2}).should == [{:filename => "lib/foo/foo.rb",
+        :x => 1, :y => 88.2}]
+    end
+  end
+
   describe Turbulence::FileNameMangler do
     subject { Turbulence::FileNameMangler.new }
     it "anonymizes a string" do