bagit 0.3.1 → 0.3.2.pre

data/.gitignore

@@ -0,0 +1,11 @@
+.DS_Store
+*~
+.#*
+\#*#
+*.gem
+tags
+vendor
+Gemfile.lock
+.bundle
+.project
+.idea

data/.rvmrc

@@ -0,0 +1 @@
+rvm use ruby-1.9.3@bagit --create

data/.travis.yml

@@ -0,0 +1,8 @@
+rvm:
+  - "1.9.3"
+  - "1.9.2"
+  - "1.8.7"
+
+branches:
+  only:
+    - master

data/Gemfile

@@ -0,0 +1,3 @@
+source 'https://rubygems.org'
+
+gemspec

data/bagit.gemspec

@@ -1,16 +1,27 @@
+# coding: utf-8
+lib = File.expand_path('../lib', __FILE__)
+$LOAD_PATH.unshift(lib) unless $LOAD_PATH.include?(lib)
+require 'bagit/version'
 
-BAGIT_SPEC = Gem::Specification.new do |spec|
+Gem::Specification.new do |spec|
   spec.name        = "bagit"
-  spec.version     = '0.3.1'
+  spec.version     = BagIt::VERSION
   spec.summary     = "BagIt package generation and validation"
   spec.description = "Ruby Library and Command Line tools for bagit"
   spec.email       = "johnson.tom@gmail.com"
   spec.homepage    = 'http://github.com/tipr/bagit'
   spec.authors     = ["Tom Johnson, Francesco Lazzarino"]
+  spec.license     = "MIT"
 
   spec.add_dependency 'validatable', '~> 1.6'
   spec.add_dependency 'docopt', '~> 0.5.0'
 
-  spec.files = %w(Rakefile README.md LICENSE.txt bagit.gemspec) + Dir["lib/**/*.rb"]
-  spec.executables << 'bagit'
+  spec.add_development_dependency "bundler", "~> 1.3"
+  spec.add_development_dependency 'rake'
+  spec.add_development_dependency 'rspec'
+
+  spec.files         = `git ls-files`.split($/)
+  spec.executables   = spec.files.grep(%r{^bin/}) { |f| File.basename(f) }
+  spec.test_files    = spec.files.grep(%r{^(test|spec|features)/})
+  spec.require_paths = ["lib"]
 end

data/bin/bagit

@@ -36,7 +36,7 @@ DOCOPT
   #             [--bag-info-entry <label> <value>] [-f <file>...] [-t <tagfile>...] BAGPATH
 
 begin
-  opts = Docopt::docopt(doc, version: 'BagIt 0.3.1')
+  opts = Docopt::docopt(doc, version: BagIt::VERSION)
   
   bag = BagIt::Bag.new(opts['BAGPATH'])
 

data/lib/bagit.rb

@@ -4,6 +4,7 @@
 # http://www.cdlib.org/inside/diglib/bagit/bagitspec.html
 
 require 'bagit/bag'
+require 'bagit/version'
 require 'fileutils'
 require 'date'
 

data/lib/bagit/bag.rb

@@ -62,33 +62,33 @@ module BagIt
       FileUtils::mkdir_p File.dirname(path)
 
       if src_path.nil?
-        f = open(path, 'w') { |io| yield io }
+        f = File.open(path, 'w') { |io| yield io }
       else
         f = FileUtils::cp src_path, path
       end
       write_bag_info
       return f
     end
-    
+
     # Remove a bag file
     def remove_file(base_path)
       path = File.join(data_dir, base_path)
       raise "Bag file does not exist: #{base_path}" unless File.exist? path
       FileUtils::rm path
     end
-    
+
     # Retrieve the IO handle for a file in the bag
     def get(base_path)
       path = File.join(data_dir, base_path)
       return nil unless File.exist?(path)
       File.open(path)
     end
-    
+
     # Test if this bag is empty (no files)
     def empty?
       self.bag_files.empty?
     end
-    
+
     # Get all bag file paths relative to the data dir
     def paths
       self.bag_files.collect { |f| f.sub(data_dir + '/', '') }
@@ -103,7 +103,7 @@ module BagIt
       end
       return bytes.to_s + '.' + bag_files.count.to_s
     end
-    
+
     # Remove all empty directory trees from the bag
     def gc!
       Dir.entries(data_dir).each do |f|

data/lib/bagit/info.rb

@@ -4,7 +4,7 @@ module BagIt
 
   module Info
 
-    @@bag_info_headers = {  
+    @@bag_info_headers = {
       :agent => 'Bag-Software-Agent',
       :org => 'Source-Organization',
       :org_addr => 'Organization-Address',
@@ -35,7 +35,7 @@ module BagIt
     end
 
     def write_bag_info(hash={})
-      hash = bag_info.merge(hash) 
+      hash = bag_info.merge(hash)
       hash[@@bag_info_headers[:agent]] = "BagIt Ruby Gem (http://bagit.rubyforge.org)" if hash[@@bag_info_headers[:agent]].nil?
       hash[@@bag_info_headers[:date]] = Date.today.strftime('%Y-%m-%d') if hash[@@bag_info_headers[:date]].nil?
       hash[@@bag_info_headers[:oxum]] = payload_oxum
@@ -49,7 +49,7 @@ module BagIt
     def bagit
       read_info_file bagit_txt_file
     end
-    
+
     def write_bagit(hash)
       write_info_file bagit_txt_file, hash
     end
@@ -62,18 +62,18 @@ module BagIt
     protected
 
     def read_info_file(file)
-      
-      open(file) do |io|
-        
+
+      File.open(file) do |io|
+
         entries = io.read.split /\n(?=[^\s])/
-        
+
         entries.inject({}) do |hash, line|
           name, value = line.chomp.split /\s*:\s*/, 2
           hash.merge({name => value})
         end
-        
+
       end
-      
+
     end
 
     def write_info_file(file, hash)
@@ -82,25 +82,25 @@ module BagIt
         a = hash.keys.grep(/#{key}/i)
         acc + (a.size > 1 ? a : [])
       end
-      
+
       raise "Multiple labels (#{dups.to_a.join ', '}) in #{file}" unless dups.empty?
-      
-      open(file, 'w') do |io|
-        
+
+      File.open(file, 'w') do |io|
+
         hash.each do |name, value|
           simple_entry = "#{name}: #{value.gsub /\s+/, ' '}"
-          
+
           entry = if simple_entry.length > 79
                     simple_entry.wrap(77).indent(2)
                   else
                     simple_entry
                   end
-          
+
           io.puts entry
         end
-        
+
       end
-      
+
     end
 
   end

data/lib/bagit/manifest.rb

@@ -10,7 +10,7 @@ module BagIt
     # All tag files that are bag manifest files (manifest-[algorithm].txt)
     def manifest_files
       files = Dir[File.join(@bag_dir, '*')].select { |f|
-        File.file? f and File.basename(f) =~ /^manifest-.*.txt/ 
+        File.file? f and File.basename(f) =~ /^manifest-.*.txt/
       }
       files
     end
@@ -32,19 +32,19 @@ module BagIt
 
         # sha1
         sha1 = Digest::SHA1.file f
-        open(manifest_file(:sha1), 'a') { |io| io.puts "#{sha1} #{rel_path}" }
+        File.open(manifest_file(:sha1), 'a') { |io| io.puts "#{sha1} #{rel_path}" }
 
         # md5
         md5 = Digest::MD5.file f
-        open(manifest_file(:md5), 'a') { |io| io.puts "#{md5} #{rel_path}" }
+        File.open(manifest_file(:md5), 'a') { |io| io.puts "#{md5} #{rel_path}" }
       end
-
+      tagmanifest!
     end
 
     # All tag files that are bag manifest files (tagmanifest-[algorithm].txt)
     def tagmanifest_files
       files = Dir[File.join(@bag_dir, '*')].select { |f|
-        File.file? f and File.basename(f) =~ /^tagmanifest-.*.txt/ 
+        File.file? f and File.basename(f) =~ /^tagmanifest-.*.txt/
       }
       files
     end
@@ -57,12 +57,12 @@ module BagIt
     # Generate manifest files for all the tag files (except the tag
     # manifest files)
     def tagmanifest!(tags=nil)
-      
+
       tags = tag_files if tags == nil
 
       # nuke all the existing tagmanifest files
       tagmanifest_files.each { |f| FileUtils::rm f }
-      
+
       # ensure presence of manfiest files
       manifest_files.each do |manifest|
         tags << manifest unless tags.include?(manifest)
@@ -81,36 +81,36 @@ module BagIt
 
     def add_tag_file(path, src_path=nil)
 
-      f = File.join(@bag_dir, path) 
+      f = File.join(@bag_dir, path)
       raise "Tag file already in manifest: #{path}" if tag_files.include?(f)
-      
+
       if not File.exist? f
         FileUtils::mkdir_p File.dirname(f)
 
         # write file
         if src_path.nil?
-          open(f, 'w') { |io| yield io }
+          File.open(f, 'w') { |io| yield io }
         else
           FileUtils::cp src_path, f
         end
         # this adds the manifest and bag info files on initial creation
-        # it must only run when the manifest doesn't already exist or it will 
+        # it must only run when the manifest doesn't already exist or it will
         # infinitely recall add_tag_file. Better way of doing this?
-        tagmanifest! 
-      elsif not src_path.nil? 
+        tagmanifest!
+      elsif not src_path.nil?
         raise "Tag file already exists, will not overwrite: #{path}\n Use add_tag_file(path) to add an existing tag file."
       end
 
-      data = open(f) { |io| io.read }
+      data = File.open(f) { |io| io.read }
       rel_path = Pathname.new(f).relative_path_from(Pathname.new(bag_dir)).to_s
 
       # sha1
       sha1 = Digest::SHA1.hexdigest data
-      open(tagmanifest_file(:sha1), 'a') { |io| io.puts "#{sha1} #{rel_path}" }
+      File.open(tagmanifest_file(:sha1), 'a') { |io| io.puts "#{sha1} #{rel_path}" }
 
       # md5
       md5 = Digest::MD5.hexdigest data
-      open(tagmanifest_file(:md5), 'a') { |io| io.puts "#{md5} #{rel_path}" }
+      File.open(tagmanifest_file(:md5), 'a') { |io| io.puts "#{md5} #{rel_path}" }
       tag_files
     end
 
@@ -146,11 +146,11 @@ module BagIt
 
         # check it, an unknown algorithm is always true
         unless algo == :unknown
-          lines = open(mf) { |io| io.readlines }
+          lines = File.open(mf) { |io| io.readlines }
 
           lines.all? do |line|
             manifested_digest, path = line.chomp.split /\s+/, 2
-            actual_digest = open(File.join(@bag_dir, path)) { |io| algo.hexdigest io.read }
+            actual_digest = File.open(File.join(@bag_dir, path)) { |io| algo.hexdigest io.read }
             actual_digest == manifested_digest
           end
 

data/lib/bagit/valid.rb

@@ -4,8 +4,8 @@ module BagIt
 
   class Bag
     include Validatable
-    validates_true_for :consistency, :logic => Proc.new { complete? }
-    validates_true_for :completeness, :logic => Proc.new { consistent? }
+    validates_true_for :consistency, :logic => Proc.new { consistent? }
+    validates_true_for :completeness, :logic => Proc.new { complete? }
   end
 
   module Validity
@@ -24,7 +24,7 @@ module BagIt
       tag_empty_manifests.each do |file|
         errors.add :completeness, "#{file} is a manifested tag but not present"
       end
-      
+
       errors.on(:completeness).nil?
     end
 
@@ -42,7 +42,7 @@ module BagIt
                  :unknown
                end
         # Check every file in the manifest
-        open(mf) do |io|
+        File.open(mf) do |io|
           io.each_line do |line|
             expected, path = line.chomp.split /\s+/, 2
             file = File.join(bag_dir, path)
@@ -55,7 +55,7 @@ module BagIt
           end
         end
       end
-      
+
 
       errors.on(:consistency).nil?
     end
@@ -93,7 +93,7 @@ module BagIt
 
       manifest_files.inject([]) do |acc, mf|
 
-        files = open(mf) do |io|
+        files = File.open(mf) do |io|
 
           io.readlines.map do |line|
             digest, path = line.chomp.split /\s+/, 2
@@ -109,7 +109,7 @@ module BagIt
     # Returns a list of all files in the tag manifest files
     def tag_manifested_files
       tagmanifest_files.inject([]) do |acc, mf|
-        files = open(mf) do |io|
+        files = File.open(mf) do |io|
           io.readlines.map do |line|
             digest, path = line.chomp.split /\s+/, 2
             path

data/lib/bagit/version.rb

@@ -0,0 +1,3 @@
+module BagIt
+  VERSION = "0.3.2.pre"
+end

data/spec/bagit_spec.rb

@@ -0,0 +1,160 @@
+require 'spec_helper'
+
+# based on v0.96 http://www.cdlib.org/inside/diglib/bagit/bagitspec.html
+describe BagIt::Bag do
+  describe 'empty bag' do
+    before(:each) do
+      @sandbox = Sandbox.new
+      # make the bag
+      @bag_path = File.join @sandbox.to_s, 'the_bag'
+      @bag = BagIt::Bag.new @bag_path
+    end
+
+    after(:each) do
+      @sandbox.cleanup!
+    end
+
+    it "should be empty" do
+      @bag.should be_empty
+    end
+  end
+
+
+  describe 'bag with files' do
+    before(:each) do
+      @sandbox = Sandbox.new
+
+      # make the bag
+      @bag_path = File.join @sandbox.to_s, 'the_bag'
+      @bag = BagIt::Bag.new @bag_path
+
+      # add some files
+      File.open('/dev/urandom') do |rio|
+        10.times do |n|
+          @bag.add_file("file-#{n}") { |io| io.write rio.read(16) }
+        end
+      end
+    end
+
+    after(:each) do
+      @sandbox.cleanup!
+    end
+
+    it "should be a directory" do
+      File.directory?(@bag_path).should be_true
+    end
+
+    it "should not be empty" do
+      @bag.should_not be_empty
+    end
+
+    it "should have a sub-directory called data" do
+      data_path = File.join @bag_path, 'data'
+      File.directory?(data_path).should be_true
+    end
+
+    describe "#add_file" do
+      it "should allow addition of files via io" do
+        @bag.add_file("foo") { |io| io.puts 'all alone' }
+        File.join(@bag_path, "data", "foo").should exist_on_fs
+      end
+
+      it "should allow addition of files via copy" do
+        src_path = File.join @sandbox.to_s, 'somefile'
+        File.open(src_path, 'w') { |io| io.puts "something" }
+        @bag.add_file("foo", src_path) { |io| io.puts 'all alone' }
+        File.join(@bag_path, "data", "foo").should exist_on_fs
+      end
+
+      it "should allow addition of files with deep paths" do
+        @bag.add_file("deep/dir/structure/file") { |io| io.puts 'all alone' }
+        File.join(@bag_path, "data", "deep/dir/structure/file").should exist_on_fs
+      end
+
+      it "should not allow overwriting of files" do
+        lambda { @bag.add_file("file-0") { |io| io.puts 'overwrite!' } }.should raise_error
+      end
+
+      it "should update payload oxum" do
+        oxum_count = @bag.bag_info["Payload-Oxum"].split('.')[1].to_i
+        @bag.add_file("foo") { |io| io.puts 'all alone' }
+        @bag.bag_info["Payload-Oxum"].split('.')[1].to_i.should == oxum_count + 1
+      end
+    end
+
+    describe "#remove_file" do
+      it "should raise an error when deleing non existant files" do
+        lambda { @bag.remove_file("file-x") }.should raise_error
+      end
+    end
+
+    describe "#get" do
+      describe "file not in bag" do
+        it "should return nil" do
+          @bag.get('foobar').should be_nil
+        end
+      end
+
+      describe "file in bag" do
+        before do
+          @contents = 'all alone'
+          @bag.add_file("foo") { |io| io << 'all alone' }
+          @file = @bag.get("foo")
+        end
+
+        it "should return an IO object for the given path" do
+          @file.should be_a_kind_of(IO)
+        end
+
+        it "should have the same content as the file added" do
+          @file.read.should == @contents
+        end
+
+        it "should accept an optional leading slash or ./" do
+          @bag.get("/foo").read.should == @contents
+          @bag.get("./foo").read.should == @contents
+        end
+      end
+    end
+
+    describe "#paths" do
+      before do
+        @paths = @bag.paths
+      end
+
+      it "should return a non-empty Array of Strings" do
+        @paths.should be_a_kind_of(Array)
+        @paths.should_not be_empty
+        @paths.each do |p|
+          p.should be_a_kind_of(String)
+        end
+      end
+
+      it "should return relative paths to all files in the data directory" do
+        @paths.should =~ (0..9).collect { |x| "file-#{x}" }
+      end
+    end
+
+    describe "#payload-oxum" do
+      it "should return a valid oxum" do
+        @bag.payload_oxum.should =~ /^[0-9]+\.[0-9]+$/
+      end
+
+      it "should accurately specify the number of payload files" do
+        @bag.add_tag_file('non-payload') { |f| f.puts "I shouldn't count in the oxum" }
+        @bag.payload_oxum.split('.')[1] == @bag.bag_files.count
+      end
+    end
+
+    describe "#gc!" do
+      it "should clean up empty directories" do
+        f = File.join "1", "2", "3", "file"
+        @bag.add_file(f) { |io| io.puts 'all alone' }
+        @bag.remove_file f
+        File.exist?(File.dirname(File.join(@bag_path, 'data', f))).should be_true
+        @bag.gc!
+        File.exist?(File.dirname(File.join(@bag_path, 'data', f))).should be_false
+      end
+    end
+  end
+end

data/spec/fetch_spec.rb

@@ -0,0 +1,56 @@
+require 'spec_helper'
+
+describe "fetch.txt" do
+
+  before(:each) do
+
+    @sandbox = Sandbox.new
+
+    # make the bag
+    @bag_path = File.join @sandbox.to_s, 'the_bag'
+    @bag = BagIt::Bag.new @bag_path
+
+    # add some files
+    File.open('/dev/urandom') do |rio|
+
+      10.times do |n|
+        @bag.add_file("file-#{n}") { |io| io.write rio.read(16) }
+      end
+
+    end
+
+  end
+
+  after(:each) do
+    @sandbox.cleanup!
+  end
+
+  before(:each) do
+    @bag.add_remote_file('http://www.gnu.org/graphics/heckert_gnu.small.png', 'gnu.png', 6322,
+                         '390c0a30976f899cbdf951eab5cce60fe9743ac9',
+                         'a3bd7ab2442028bb91b51d9f6722ec98')
+
+    path = File.join @bag_path, 'fetch.txt'
+    @lines = File.open(path) { |io| io.readlines }
+  end
+
+  it "should not be empty" do
+    @lines.should_not be_empty
+  end
+
+  it "should only contain lines of the format URL LENGTH FILENAME" do
+    @lines.each { |line| line.chomp.should =~ /^[^\s]+\s+(\d+|\-)\s+[^\s]+$/ }
+  end
+
+  it "should contain manifested files" do
+    path = File.join @bag_path, 'manifest-sha1.txt'
+    data = File.open(path) { |io| io.read }
+    data.should include('gnu.png')
+  end
+
+  it "should be gone when fetch is complete" do
+    @bag.fetch!
+    File.exist?(File.join(@bag_path, 'fetch.txt')).should_not be_true
+  end
+
+end

data/spec/manifest_spec.rb

@@ -0,0 +1,144 @@
+require 'spec_helper'
+
+describe "BagIt Manifests" do
+
+  before(:each) do
+
+    @sandbox = Sandbox.new
+
+    # make the bag
+    @bag_path = File.join @sandbox.to_s, 'the_bag'
+    @bag = BagIt::Bag.new @bag_path
+
+    # add some files
+    File.open('/dev/urandom') do |rio|
+
+      10.times do |n|
+        @bag.add_file("file-#{n}") { |io| io.write rio.read(16) }
+        @bag.add_tag_file("tag-#{n}") { |io| io.write rio.read(16) }
+      end
+
+    end
+
+  end
+
+  after(:each) do
+    @sandbox.cleanup!
+  end
+
+  shared_examples_for "a manifest file" do
+
+    before do
+      pattern = File.join @bag_path, '*manifest-*.txt'
+      @manifest_files = Dir.glob pattern
+    end
+
+    it "should have valid algorithm in the name (at least md5 or sha1)" do
+      algorithms = @manifest_files.map { |mf| mf =~ /manifest-(.*).txt$/; $1 }
+      algorithms.each { |a| a.should be_in('md5', 'sha1') }
+    end
+
+    it "should not be an empty file" do
+      @manifest_files.each { |mf| File.size(mf).should_not == 0 }
+    end
+
+    it "should only contain lines of the format CHECKSUM FILENAME" do
+      @manifest_files.each do |file|
+        File.open(file) do |io|
+          io.each_line { |line| line.chomp.should =~ /^[a-fA-F0-9]+\s+[^\s].+$/ }
+        end
+      end
+    end
+
+    it "should validate after adding a file and remanifesting" do
+      @bag.add_file('newfile.txt') { |io| io.puts("new file to remanifest") }
+      @bag.manifest!
+      @bag.should be_valid
+    end
+
+  end
+
+  describe "bag manifest files" do
+
+    before do
+      @bag.manifest!
+    end
+
+    it_behaves_like "a manifest file"
+
+    it "should have a manifest file" do
+      @bag.manifest_files.should_not be_empty
+    end
+
+    it "should only contain bag files" do
+      @bag.manifest_files.each do |mf|
+        File.open(mf) do |io|
+          io.each_line do |line|
+            line.chomp.should =~ /^[a-f0-9]+\s+data\/[^\s].+$/
+          end
+        end
+      end
+    end
+
+  end
+
+  describe "tag manifest files" do
+
+    before do
+      @bag.add_tag_file("test-tag") { |f| f.puts "all alone" }
+    end
+
+    it_should_behave_like "a manifest file"
+
+    it "should have a tag manifest file" do
+      @bag.tagmanifest_files.should_not be_empty
+    end
+    it "should only contain tag files" do
+      @bag.tagmanifest_files.each do |mf|
+        File.open(mf) do |io|
+          io.each_line do |line|
+            line.chomp.should =~ /^[a-fA-F0-9]+\s+(?!data\/)[^\s].+$/
+          end
+        end
+      end
+    end
+    it "should contain manifest and bag info files" do
+      @bag.tagmanifest_files.each do |mf|
+        File.open(mf).read.should include(File.basename(@bag.bag_info_txt_file))
+        File.open(mf).read.should include(File.basename(@bag.bagit_txt_file))
+        @bag.manifest_files.each do |man|
+          File.open(mf).read.should include(man)
+        end
+      end
+    end
+    it "should not contain the untracked tag file" do
+      @bag.tagmanifest_files.each do |mf|
+        File.open(mf) do |io|
+          io.read.should_not include "tag-notrack"
+        end
+      end
+    end
+    describe "removing tracked files" do
+      before(:each) do
+        @bag.remove_tag_file "tag-1"
+        @bag.delete_tag_file "tag-2"
+      end
+      it "should still have the untracked tag file on the file system" do
+        File.join(@bag_path, "tag-1").should exist_on_fs
+      end
+      it "should not have the deleted tag file on the file system" do
+        File.join(@bag_path, "tag-2").should_not exist_on_fs
+      end
+      it "should not have the removed or deleted tag files in the manifest" do
+        @bag.tagmanifest_files.each do |mf|
+          File.open(mf) do |io|
+            io.read.should_not include "tag-1"
+            io.read.should_not include "tag-2"
+          end
+        end
+      end
+    end
+  end
+
+
+end

data/spec/spec_helper.rb

@@ -0,0 +1,33 @@
+require 'rubygems'
+require 'bundler'
+Bundler.require(:default, :test)
+
+require File.expand_path('./util/bagit_matchers', File.dirname(__FILE__))
+
+RSpec.configure do |config|
+  config.include(BagitMatchers)
+end
+
+$:.unshift File.expand_path('../lib', File.dirname(__FILE__))
+require 'bagit'
+
+require 'tempfile'
+
+class Sandbox
+
+  def initialize
+    tf = Tempfile.open 'sandbox'
+    @path = tf.path
+    tf.close!
+    FileUtils::mkdir @path
+  end
+
+  def cleanup!
+    FileUtils::rm_rf @path
+  end
+
+  def to_s
+    @path
+  end
+
+end

data/spec/tag_info_spec.rb

@@ -0,0 +1,132 @@
+require 'spec_helper'
+
+describe "Tag Info Files" do
+
+  before(:each) do
+
+    @sandbox = Sandbox.new
+
+    # make the bag
+    @bag_path = File.join @sandbox.to_s, 'the_bag'
+    @bag = BagIt::Bag.new @bag_path
+
+    # add some files
+    File.open('/dev/urandom') do |rio|
+      10.times do |n|
+        @bag.add_file("file-#{n}") { |io| io.write rio.read(16) }
+      end
+    end
+
+  end
+
+  after(:each) do
+    @sandbox.cleanup!
+  end
+
+  describe "bagit.txt" do
+
+    before do
+      path = File.join @bag_path, 'bagit.txt'
+      @lines = File.open(path) { |io| io.readlines }
+    end
+
+    it "should create a file bagit.txt on bag initialization" do
+      File.join(@bag_path, 'bagit.txt').should exist_on_fs
+    end
+
+    it "should have exactly two lines" do
+      @lines.size.should == 2
+    end
+
+    it "should have a bagit version" do
+      a = @lines.select { |line| line.chomp =~ /BagIt-Version:\s*\d+\.\d+/ }
+      a.should_not be_empty
+    end
+
+    it "should have a tag file encoding" do
+      a = @lines.select { |line| line.chomp =~ /Tag-File-Character-Encoding:\s*.+/ }
+      a.should_not be_empty
+    end
+
+  end
+
+  describe "bag-info.txt" do
+
+    before(:each) do
+      path = File.join @bag_path, 'bag-info.txt'
+      @lines = File.open(path) { |io| io.readlines }
+    end
+
+    it "should not be empty" do
+      @lines.should_not be_empty
+    end
+
+    it "should contain lines of the format LABEL: VALUE (like an email header)" do
+      @lines.each { |line| line.chomp.should =~ /^[^\s]+\s*:\s+.*$/ }
+    end
+
+    it "should be case insensitive with respect to LABELs" do
+      path = File.join @bag_path, 'bag-info.txt'
+      lambda { @bag.write_bag_info 'foo' => 'lowercase', 'Foo' => 'capital' }.should raise_error(/Multiple labels/)
+    end
+
+    it "should fold long VALUEs" do
+      longline = <<LOREM
+Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do
+  eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enimad
+  minim veniam, quis nostrud exercitation ullamco laboris nisi ut
+  aliquip ex ea commodo consequat. Duis aute irure dolor in
+  reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla
+  pariatur. Excepteur sint occaecat cupidatat non proident, sunt in
+  culpa qui officia deserunt mollit anim id est laborum.
+LOREM
+      @bag.write_bag_info 'Lorem' => longline
+      @bag.bag_info.keys.length.should == 4 # this isn't a great test. Changed it from 1 to 4 because unrelated changes caused failure.
+    end
+
+    it "should specify a bag software agent" do
+      @bag.bag_info.keys.should include("Bag-Software-Agent")
+    end
+
+    it "should contain a valid bagging date" do
+      @bag.bag_info.keys.should include("Bagging-Date")
+      @bag.bag_info["Bagging-Date"] =~ /^^[0-9]{4}-[0-9]{2}-[0-9]{2}$/
+    end
+
+    it "should contain a payload oxum" do
+      @bag.bag_info.keys.should include("Payload-Oxum")
+    end
+    it "should not override any previous values" do
+      path = File.join @bag_path, 'bag-info.txt'
+      @bag.write_bag_info 'Bag-Software-Agent' => 'Some Other Agent'
+      @bag.write_bag_info 'Source-Organization' => 'Awesome Inc.'
+      @bag.write_bag_info 'Bagging-Date' => '1901-01-01'
+      @bag.write_bag_info
+      contents = File.open(path).read
+      contents.should include "Some Other Agent"
+      contents.should include "Awesome Inc."
+      contents.should include "1901-01-01"
+    end
+    it "should override previous tags when they collide with new ones" do
+      path = File.join @bag_path, 'bag-info.txt'
+      @bag.write_bag_info 'Source-Organization' => 'Awesome Inc.'
+      @bag.write_bag_info 'Source-Organization' => 'Awesome LLC.'
+      contents = File.open(path).read
+      contents.should include "Awesome LLC."
+      contents.should_not include "Awesome Inc."
+    end
+    it "should contain values passed to bag" do
+      hash = {"Bag-Software-Agent" => "rspec",
+        "Bagging-Date" => "2012-11-21",
+        "Contact-Name" => "Willis Corto",
+        "Some-Tag" => "Some Value"
+      }
+      bag_with_info = BagIt::Bag.new(@bag_path + '2', hash)
+      hash.each do |key, value|
+        bag_with_info.bag_info[key].should == value
+      end
+    end
+
+  end
+
+end

data/spec/tag_spec.rb

@@ -0,0 +1,61 @@
+require 'spec_helper'
+
+describe "Tag Specs" do
+
+  before(:each) do
+
+    @sandbox = Sandbox.new
+
+    # make the bag
+    @bag_path = File.join @sandbox.to_s, 'the_bag'
+    @bag = BagIt::Bag.new @bag_path
+
+    # add some files
+    File.open('/dev/urandom') do |rio|
+
+      10.times do |n|
+        @bag.add_file("file-#{n}") { |io| io.write rio.read(16) }
+        @bag.add_tag_file("tag-#{n}") { |io| io.write rio.read(16)}
+      end
+
+    end
+
+  end
+
+  after(:each) do
+    @sandbox.cleanup!
+  end
+  describe "#add_tag_file" do
+    it "should allow addition of tag files via io" do
+      @bag.add_tag_file("foo") { |io| io.puts 'all alone' }
+      File.join(@bag_path, "foo").should exist_on_fs
+    end
+    it "should allow addition of bag files within directories using io" do
+      @bag.add_tag_file("fedora/foo") { |io| io.puts 'all alone' }
+      File.join(@bag_path, "fedora","foo").should exist_on_fs
+    end
+    it "should allow addition of deep tag files" do
+      @bag.add_tag_file("fedora/foo/newfoo/deep") {|io| io.puts "woah that's deep"}
+      File.join(@bag_path,"fedora","foo","newfoo","deep").should exist_on_fs
+    end
+    it "should not allow overwriting of tag files" do
+      lambda { @bag.add_tag_file("tag-0") { |io| io.puts 'overwrite!' } }.should raise_error
+    end
+    it "should allow addition of tag files via copy" do
+      src_path = File.join @sandbox.to_s, 'somefile'
+      File.open(src_path, 'w') { |io| io.puts "something" }
+      @bag.add_tag_file("foo", src_path) { |io| io.puts 'all alone' }
+      File.join(@bag_path, "foo").should exist_on_fs
+    end
+  end
+  describe "#remove_tag_file" do
+    it "should raise an error when removing non existant files" do
+      lambda { @bag.remove_tag_file("file-x") }.should raise_error
+    end
+  end
+  describe "#delete_tag_file" do
+    it "should raise an error when deleting non existant tag files" do
+      lambda { @bag.delete_tag_file("file-x") }.should raise_error
+  end
+  end
+end

data/spec/util/bagit_matchers.rb

@@ -0,0 +1,49 @@
+module BagitMatchers
+
+  class BeIn
+
+    def initialize(*expected_collection)
+      @expected = expected_collection
+    end
+
+    def matches?(target)
+      @target = target
+      @expected.include? @target
+    end
+
+    def failure_message
+      "expected <#{@target}> to be in collection <#{@expected}>"
+    end
+
+    def negative_failure_message
+      "expected <#{@target}> to not be in collection <#{@expected}>"
+    end
+
+  end
+
+  def be_in(*expected_collection)
+    BeIn.new(*expected_collection)
+  end
+
+  class ExistOnFS
+
+    def matches?(target)
+      @target = target
+      File.exist? target
+    end
+
+    def failure_message
+      "expected <#{@target}> to exist, but it doesn't"
+    end
+
+    def negative_failure_message
+      "expected <#{@target}> to not exist but it does"
+    end
+
+  end
+
+  def exist_on_fs
+    ExistOnFS.new
+  end
+
+end

data/spec/validation_spec.rb

@@ -0,0 +1,115 @@
+require 'spec_helper'
+
+describe "a valid bag" do
+
+  before(:each) do
+
+    @sandbox = Sandbox.new
+
+    # make the bag
+    @bag_path = File.join @sandbox.to_s, 'the_bag'
+    @bag = BagIt::Bag.new @bag_path
+
+    # add some files
+    File.open('/dev/urandom') do |rio|
+
+      10.times do |n|
+        @bag.add_file("file-#{n}") { |io| io.write rio.read(16) }
+        @bag.add_tag_file("tag-#{n}") { |io| io.write rio.read(16) }
+      end
+
+    end
+
+    @bag.manifest!
+  end
+
+  after(:each) do
+    @sandbox.cleanup!
+  end
+
+  it "should validate with no errors" do
+    @bag.should be_valid
+  end
+
+  it "should not be lewd (some file is not covered by the manifest)" do
+    # add a file into the bag through the back door
+    File.open(File.join(@bag.data_dir, 'not-manifested'), 'w') do |io|
+      io.puts 'nothing to see here, move along'
+    end
+
+    @bag.validate_only('true_for/completeness')
+    @bag.errors.on(:completeness).should_not be_empty
+    @bag.should_not be_valid
+  end
+
+  it "should not be prude (the manifest covers files that do not exist)" do
+    # add a file and then remove it through the back door
+    @bag.add_file("file-k") { |io| io.puts 'time to go' }
+    @bag.manifest!
+
+    FileUtils::rm File.join(@bag.bag_dir, 'data', 'file-k')
+
+    @bag.validate_only('true_for/completeness')
+    @bag.errors.on(:completeness).should_not be_empty
+    @bag.should_not be_valid
+  end
+
+  it "should be consistent (fixity)" do
+    # tweak a file through the back door
+    File.open(@bag.bag_files[0], 'a') { |io| io.puts 'oops!' }
+
+    @bag.validate_only('true_for/consistency')
+    @bag.errors.on(:consistency).should_not be_empty
+    @bag.should_not be_valid
+  end
+
+  it "should calculate sha1 correctly for a big file" do
+    @bag.add_file 'big-data-file' do |fh|
+      count = 0
+      while count < 1024 * 512 do
+        fh.write "1" * 1024
+        count += 1
+      end
+    end
+    @bag.manifest!
+    sha1_manifest = File.join @bag_path, 'manifest-sha1.txt'
+    checksums = {}
+    File.open(sha1_manifest).each_line do |line|
+        fixity, path = line.split(' ')
+        checksums[path] = fixity
+    end
+    expected = checksums['data/big-data-file']
+    expected.should == '12be64c30968bb90136ee695dc58f4b2276968c6'
+  end
+
+  it "should validate by oxum when needed" do
+    @bag.valid_oxum?.should == true
+  end
+
+  it "should validate false by oxum when file count is incorrect" do
+    # tweak oxum through backdoor
+    File.open(@bag.bag_info_txt_file, 'a') { |f| f.write "Payload-Oxum: " + @bag.bag_info["Payload-Oxum"].split('.')[0] + '.0' }
+    @bag.valid_oxum?.should == false
+  end
+
+  it "should validate false by oxum when octetstream size is incorrect" do
+    # tweak oxum through backdoor
+    File.open(@bag.bag_info_txt_file, 'a') { |f| f.write "Payload-Oxum: 1." + @bag.bag_info["Payload-Oxum"].split('.')[1] }
+    @bag.valid_oxum?.should == false
+  end
+
+  describe "tag manifest validation" do
+    it "should be invalid if listed tag file does not exist" do
+      # add a file and then remove it through the back door
+      @bag.add_tag_file("tag-k") { |io| io.puts 'time to go' }
+      @bag.tagmanifest!
+
+      FileUtils::rm File.join(@bag.bag_dir, 'tag-k')
+
+      # @bag.should_not be_valid
+      @bag.should_not be_valid
+      @bag.errors.on(:completeness).should_not be_empty
+    end
+  end
+
+end

metadata

@@ -1,15 +1,15 @@
 --- !ruby/object:Gem::Specification
 name: bagit
 version: !ruby/object:Gem::Version
-  version: 0.3.1
-  prerelease: 
+  version: 0.3.2.pre
+  prerelease: 6
 platform: ruby
 authors:
 - Tom Johnson, Francesco Lazzarino
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-04-28 00:00:00.000000000 Z
+date: 2013-08-07 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: validatable
@@ -43,6 +43,54 @@ dependencies:
     - - ~>
       - !ruby/object:Gem::Version
         version: 0.5.0
+- !ruby/object:Gem::Dependency
+  name: bundler
+  requirement: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: '1.3'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: '1.3'
+- !ruby/object:Gem::Dependency
+  name: rake
+  requirement: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: rspec
+  requirement: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
 description: Ruby Library and Command Line tools for bagit
 email: johnson.tom@gmail.com
 executables:
@@ -50,21 +98,35 @@ executables:
 extensions: []
 extra_rdoc_files: []
 files:
-- Rakefile
-- README.md
+- .gitignore
+- .rvmrc
+- .travis.yml
+- Gemfile
 - LICENSE.txt
+- README.md
+- Rakefile
 - bagit.gemspec
+- bin/bagit
 - lib/bagit.rb
-- lib/bagit/string.rb
+- lib/bagit/bag.rb
 - lib/bagit/fetch.rb
-- lib/bagit/manifest.rb
-- lib/bagit/info.rb
 - lib/bagit/file.rb
+- lib/bagit/info.rb
+- lib/bagit/manifest.rb
+- lib/bagit/string.rb
 - lib/bagit/valid.rb
-- lib/bagit/bag.rb
-- bin/bagit
+- lib/bagit/version.rb
+- spec/bagit_spec.rb
+- spec/fetch_spec.rb
+- spec/manifest_spec.rb
+- spec/spec_helper.rb
+- spec/tag_info_spec.rb
+- spec/tag_spec.rb
+- spec/util/bagit_matchers.rb
+- spec/validation_spec.rb
 homepage: http://github.com/tipr/bagit
-licenses: []
+licenses:
+- MIT
 post_install_message: 
 rdoc_options: []
 require_paths:
@@ -78,13 +140,21 @@ required_ruby_version: !ruby/object:Gem::Requirement
 required_rubygems_version: !ruby/object:Gem::Requirement
   none: false
   requirements:
-  - - ! '>='
+  - - ! '>'
     - !ruby/object:Gem::Version
-      version: '0'
+      version: 1.3.1
 requirements: []
 rubyforge_project: 
 rubygems_version: 1.8.25
 signing_key: 
 specification_version: 3
 summary: BagIt package generation and validation
-test_files: []
+test_files:
+- spec/bagit_spec.rb
+- spec/fetch_spec.rb
+- spec/manifest_spec.rb
+- spec/spec_helper.rb
+- spec/tag_info_spec.rb
+- spec/tag_spec.rb
+- spec/util/bagit_matchers.rb
+- spec/validation_spec.rb