git-health-check 0.0.2 → 0.0.3

data/Rakefile

@@ -1 +1 @@
-require "bundler/gem_tasks"
+require 'bundler/gem_tasks'

data/git-health-check.gemspec

@@ -4,20 +4,20 @@ $LOAD_PATH.unshift(lib) unless $LOAD_PATH.include?(lib)
 require 'git-health-check/version'
 
 Gem::Specification.new do |gem|
-  gem.name = "git-health-check"
+  gem.name = 'git-health-check'
   gem.version = GitHealthCheck::VERSION
-  gem.authors = ["Ben Snape"]
-  gem.email = ["bsnape@gmail.com"]
+  gem.authors = ['Ben Snape']
+  gem.email = ['bsnape@gmail.com']
   gem.description = %q{Git Health Check}
   gem.summary = %q{Git Health Check}
-  gem.homepage = "http://wwww.bensnape.com"
+  gem.homepage = 'http://wwww.bensnape.com'
 
   gem.files = `git ls-files`.split($/)
   gem.executables = gem.files.grep(%r{^bin/}).map { |f| File.basename(f) }
   gem.test_files = gem.files.grep(%r{^(test|spec|features)/})
-  gem.require_paths = ["lib"]
+  gem.require_paths = ['lib']
 
-  gem.add_dependency("ruport")
-  gem.add_development_dependency("rake")
+  gem.add_dependency('ruport')
+  gem.add_development_dependency('rake')
 
 end

data/lib/git-health-check.rb

@@ -1,4 +1,4 @@
-require "git-health-check/version"
+require 'git-health-check/version'
 require 'git-health-check/history'
 require 'git-health-check/working_copy'
 require 'git-health-check/report'

data/lib/git-health-check/cli/git_health_check_command.rb

@@ -12,18 +12,18 @@ module GitHealthCheck
       end
 
       def execute(view)
-        history = GitHealthCheck::History.new(@repository, 'HEAD', @threshold)
+        history = GitHealthCheck::History.new(@repository, @threshold)
         working_copy = GitHealthCheck::WorkingCopy.new @repository
         packfile = GitHealthCheck::Packfile.new(@repository)
         packfile.packfile_stats
 
-        working_copy_output = working_copy.fast_find_in_working_copy
+        working_copy_output = working_copy.find_in_working_copy
         history_output = history.search
 
-        working_copy_report = Table("object sha", "size (kB)", "compressed (kB)", "path")
-        history_report = Table("object sha", "size (MB)", "path", "commit details", "author")
+        working_copy_report = Table('object sha', 'size (MB)', 'path')
+        history_report = Table('object sha', 'size (MB)', 'path', 'commit details', 'author')
 
-        working_copy_output.each { |sha, (size, csize, path)| working_copy_report << [sha, size, csize, path] }
+        working_copy_output.each { |sha, size, path| working_copy_report << [sha, size, path] }
         history_output.each { |sha, size, path, where, who| history_report << [sha, size, path, where, who] }
 
         report = GitHealthCheck::Report.new(working_copy_report.to_html, history_report.to_html, packfile)

data/lib/git-health-check/cli/options.rb

@@ -29,13 +29,13 @@ EOB
       def set_parser_options
         @parser.banner = banner
 
-        @parser.separator "Options:"
+        @parser.separator 'Options:'
 
-        @parser.on("-v", "--version", "Displays the gem version") { @command_class = VersionCommand }
+        @parser.on('-v', '--version', 'Displays the gem version') { @command_class = VersionCommand }
 
-        @parser.on("-h", "--help", "Displays this help message") { @command_class = HelpCommand }
+        @parser.on('-h', '--help', 'Displays this help message') { @command_class = HelpCommand }
 
-        @parser.on("-t", "--threshold THRESHOLD", Float, "Specify history size threshold in MB (default 0.5)") do |n|
+        @parser.on('-t', '--threshold THRESHOLD', Float, 'Specify history size threshold in MB (default 0.5)') do |n|
           @threshold = n
         end
       end

data/lib/git-health-check/git_lib.rb

@@ -38,5 +38,21 @@ module GitHealthCheck
       `git count-objects -v`
     end
 
+    def get_largest_files(number=10)
+      `git ls-files -z | xargs -0 ls -l | sort -nrk5 | head -n #{number}`
+    end
+
+    def get_object_sha_from_path(path)
+      `git ls-files -s #{path} | cut -d ' ' -f 2`
+    end
+
+    def get_revision_list(head='HEAD')
+      `git rev-list #{head}`
+    end
+
+    def get_treeish_contents(treeish)
+      `git ls-tree -zrl #{treeish}`
+    end
+
   end
 end

data/lib/git-health-check/history.rb

@@ -5,36 +5,36 @@ module GitHealthCheck
 
   class History
 
-    MEGABYTE = 1000 ** 2
+    MEGABYTE = 1024 ** 2
 
-    def initialize(repository, head = 'HEAD', threshold = 0.1)
-      @head = head
+    def initialize(repository, threshold)
+      @repository = repository
       @bytes_threshold = threshold.to_f * MEGABYTE
       Dir.chdir repository
       @git_lib = GitHealthCheck::GitLib.new repository
     end
 
     def search
-
+      revision_list = @git_lib.get_revision_list.split "\n"
       big_files = {}
 
-      # list commit objects in chronological order
-      IO.popen("git rev-list #@head", 'r') do |rev_list|
-        rev_list.each_line do |commit|
-          # list contents of the tree object
-          `git ls-tree -zrl #{commit.chomp!}`.split("\0").each do |object|
-            bits, type, sha, size, path = object.split(/\s+/, 5)
-            size = size.to_i
-            big_files[sha] = [path, size, commit] if size >= @bytes_threshold
-          end
+      revision_list.each do |commit|
+        @git_lib.get_treeish_contents(commit).split("\0").each do |object|
+          bits, type, sha, size, path = object.split
+          next if File.exist?("#@repository/#{path}")
+          size = size.to_f
+          big_files[path] = [sha, size, commit] if size >= @bytes_threshold
         end
       end
 
-      big_files.map do |sha, (path, size, commit_sha)|
+      big_files = big_files.map do |path, (sha, size, commit_sha)|
         where = @git_lib.get_commit_details commit_sha
         who = @git_lib.get_commit_author commit_sha
-        [sha, size.to_f / MEGABYTE, path, where, who]
+        [sha, (size / MEGABYTE).round(2), path, where, who]
       end
+
+      big_files.sort_by! { |a| [a[1], a[2]] }.reverse!
+
     end
 
   end

data/lib/git-health-check/report.rb

@@ -8,7 +8,7 @@ module GitHealthCheck
       @working_copy = working_copy
       @history = history
       @packfile = packfile
-      @report_directory = Dir.pwd + "/healthcheck"
+      @report_directory = Dir.pwd + '/healthcheck'
       @repository = Dir.pwd
     end
 
@@ -17,7 +17,7 @@ module GitHealthCheck
     end
 
     def get_template
-      File.read(File.dirname(__FILE__) + "/report/report.erb")
+      File.read(File.dirname(__FILE__) + '/report/report.erb')
     end
 
     def create
@@ -26,7 +26,7 @@ module GitHealthCheck
 
       Dir.mkdir @report_directory unless File.directory? @report_directory
 
-      File.open(@report_directory + "/report.html", "w+") do |f|
+      File.open(@report_directory + '/report.html', 'w+') do |f|
         f.write output
       end
     end

data/lib/git-health-check/report/report.erb

@@ -39,14 +39,15 @@
 
       h3 {
           font-size: 16px;
+          padding: 0 0 0 20px;
       }
 
       #history {
-          padding: 0 0 0 20px;
+          padding: 0 0 0 30px;
       }
 
       #working-copy {
-          padding: 0 0 0 20px;
+          padding: 0 0 0 30px;
       }
   </style>
 </head>
@@ -64,11 +65,11 @@
     <li><strong>Total packfile size: </strong>
       <% size = @packfile.size_of_pack[0].to_f %>
       <% if size >= 1048576 %>
-          <%= size / 1048576 %> GB
+          <%= (size / 1048576).round 2 %> GB
       <% elsif size >= 1024 %>
-          <%= size / 1024 %> MB
+          <%= (size / 1024).round 2 %> MB
       <% else %>
-          <%= size %> kB
+          <%= size.round 2 %> kB
       <% end %>
     </li>
   </ul>
@@ -78,24 +79,41 @@
 
   <h2>Repository Size</h2>
 
+  <h3>Working Copy</h3>
+
   <div id="working-copy">
 
-    <h3>Working Copy</h3>
+    <h4>Description</h4>
+
+    <p>This metric inspects your repository's working copy using
+      <a href="http://www.kernel.org/pub/software/scm/git/docs/git-ls-files.html">git ls-files</a> to identify the
+      <em>n</em> (default 10) largest files - and consequently the <strong>largest objects</strong>.</p>
 
-    <p>This metric inspects your repository's <a href="http://git-scm.com/book/en/Git-Internals-Packfiles">packfile</a>
-      contents via its index to identify the <em>n</em> (default 10) largest objects - and consequently the
-      <strong>largest files</strong> - in your working copy.</p> <br>
+    <p>In some cases the same object hash will be displayed for multiple paths. <strong>This is not a bug.</strong>
+      The way git stores <a href="http://git-scm.com/book/en/Git-Internals-Git-Objects">blob objects</a> - i.e. bytes on
+      the filesystem that could be anything (e.g. a text file, source code, an image) - means that if you have
+      <em>exactly</em> the same file in multiple places in your repository - perhaps a static test file - then the
+      report will correctly output <em>all the paths on the filesystem where the object is referenced</em>.</p>
+
+    <h4>Statistics</h4>
 
     <p><%= @working_copy %></p>
 
   </div>
 
+  <br>
+
+  <h3>History</h3>
+
   <div id="history">
-    <h3>History</h3>
+
+    <h4>Description</h4>
 
     <p>This metric thoroughly inspects your repository's history and identifies the largest files over a configurable
-      threshold (defaults at 0.5 MB) that have been committed in the past <strong>but are
-      no longer part of the working copy</strong>.</p> <br>
+      threshold (defaults at 0.5 MB) that have been committed in the past <strong>but are no longer part of the working
+      copy</strong>.</p>
+
+    <h4>Statistics</h4>
 
     <p><%= @history %></p>
   </div>

data/lib/git-health-check/version.rb

@@ -1,3 +1,3 @@
 module GitHealthCheck
-  VERSION = "0.0.2"
+  VERSION = '0.0.3'
 end

data/lib/git-health-check/working_copy.rb

@@ -7,34 +7,23 @@ module GitHealthCheck
       @git_lib = GitHealthCheck::GitLib.new repository
     end
 
-    def fast_find_in_working_copy(number=10)
-      raise 'no packs found' if @git_lib.get_pack_number == 0
-      objects = `git verify-pack -v .git/objects/pack/pack-*.idx | grep -E 'tree|blob' | sort -k4nr`
-
-      objects = objects.split("\n")
-
-      test = {}
-
-      objects.each do |object|
-        break if test.size == number
-        sha = object.match(/^\w+/)[0]
-        #path = `git rev-list --all --objects | grep #{sha} | cut -f 2 -d ' '`.chomp
-        path = `git rev-list --all --objects | grep #{sha}`.match(/\w+\s(.*)/)[1]
-        # an empty path means you need to do a garbage collection
-        next unless File.exist?("#@repository/#{path}")
-
-        sha, type, size, size_in_pack, offset = object.split
-
-        # convert from byte to kilobyte
-        size = (size.to_f / 1024).round 2
-        size_in_pack = (size_in_pack.to_f / 1024).round 2
-
-        test[sha] = [size, size_in_pack, path]
+    def find_in_working_copy(number=10)
+      largest_files = @git_lib.get_largest_files number
+      largest_files = largest_files.split "\n"
+
+      files = []
+
+      largest_files.each do |file|
+        split = file.split
+        size = split[4]
+        path = split[8]
+        sha = @git_lib.get_object_sha_from_path path
+        size = (size.to_f * 9.53674e-7).round 2 # bytes to MB
+        files << [sha, size, path] # no hash in case of blob-reuse in various paths
       end
 
-      test
+      files
     end
 
-
   end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: git-health-check
 version: !ruby/object:Gem::Version
-  version: 0.0.2
+  version: 0.0.3
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-01-15 00:00:00.000000000 Z
+date: 2013-02-08 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: ruport