Refs: #4273@20h; Reverse Engeneeiring RubyZip fixed ZipService

        - Replzed Lazy Enum over get_next_entry and group by with explicit looping and yielding to an instance level enumerator
          that yields streams per subdir (wich represents a referential).

        - Fixtures and Specs that use a REAL usecase.

        Next: Get rid of metadata kludge (should fix failing specs)

4 files changed, 59 insertions, 73 deletions

diff --git a/spec/services/zip_service/regression_4273_spec.rb b/spec/services/zip_service/regression_4273_spec.rb
new file mode 100644
index 000000000..4fe0f6539
--- /dev/null
+++ b/spec/services/zip_service/regression_4273_spec.rb
@@ -0,0 +1,59 @@
+RSpec.describe ZipService do
+  describe 'Regression Issue # 4273 https://projects.af83.io/issues/4273' do
+    let( :zip_service ){ described_class }
+    let( :unzipper ){ zip_service.new(zip_data) }
+    let( :zip_data ){ File.read zip_file }
+
+    context 'real test data' do
+      let( :subdir_names ){ %w<OFFRE_TRANSDEV_20170301122517 OFFRE_TRANSDEV_20170301122519>  }
+      let( :expected_chksums ){
+        checksum_trees( subdir_names.map{ |sn| subdir_file(sn, prefix: 'source_') } )
+      }
+
+      let( :zip_file ){ fixtures_path 'OFFRE_TRANSDEV_2017030112251.zip' }
+      #
+      # Remove potential test artefacts
+      before do
+        subdir_names.each do | subdir_name |
+          File.unlink( subdir_file subdir_name, suffix: '.zip' ) rescue nil
+          Dir.unlink( subdir_file subdir_name ) rescue nil
+        end
+      end
+
+      it "yields the correct content" do
+        subdir_contents = {}
+        # Write ZipService Streams to files and inflate them to file system
+        unzipper.subdirs.each do | subdir |
+          File.open(subdir_file( subdir.name, suffix: '.zip' ), 'wb'){ |f| f.write subdir.stream.string }
+          unzip_subdir subdir
+        end
+        # Represent the inflated file_system as a checksum tree
+        actual_checksums = 
+          checksum_trees( subdir_names.map{ |sn| subdir_file(sn, prefix: 'target/') } )
+        expect( actual_checksums ).to eq( expected_chksums )
+      end
+
+    end
+
+  end
+
+  def checksum_trees *dirs
+    dirs.flatten.inject({},&method(:checksum_tree))
+  end
+  def checksum_tree repr, dir
+    Dir.glob("#{dir}/**/*").each do |file|
+      if !File.directory?(file)
+        repr.merge!( File.basename(file) => %x{cksum #{file}}.split.first ){ |_, ov, nv| Array(ov) << nv }
+      end
+    end
+    repr
+  end
+
+  def subdir_file( subdir, prefix: 'target_', suffix: '' )
+    fixtures_path("#{prefix}#{subdir}#{suffix}")
+  end
+
+  def unzip_subdir subdir
+    %x{unzip -oqq #{subdir_file subdir.name, suffix: '.zip'} -d #{fixture_path}/target}
+  end
+end
diff --git a/spec/services/zip_service/zip_entry_data_spec.rb b/spec/services/zip_service/zip_entry_data_spec.rb
deleted file mode 100644
index 2a7226eb4..000000000
--- a/spec/services/zip_service/zip_entry_data_spec.rb
+++ /dev/null
@@ -1,32 +0,0 @@
-RSpec.describe ZipService do
-
-  subject{ described_class.new(read_fixture('multiple_references_import.zip')) }
-
-  it 'can group all entries' do
-    expect( subject.entry_groups.keys ).to eq(%w{ref1 ref2})
-  end
-
-  context 'creates correct zip data for each subdir' do
-    it 'e.g. reference1' do
-      reference1_stream = subject.entry_group_streams['ref1']
-      control_stream = Zip::InputStream.open( reference1_stream )
-      control_entries = described_class.entries(control_stream)
-      expect( control_entries.map{ |e| [e.name, e.get_input_stream.read]}.force ).to eq([
-        ["multiref/ref1/", ""],
-        ["multiref/ref1/datum-1", "multi-ref1-datum1\n"],
-        ["multiref/ref1/datum-2", "multi-ref1-datum2\n"]
-      ])
-    end
-    it 'e.g. reference2' do
-      reference2_stream = subject.entry_group_streams['ref2']
-      control_stream = Zip::InputStream.open( reference2_stream )
-      control_entries = described_class.entries(control_stream)
-      expect( control_entries.map{ |e| [e.name, e.get_input_stream.read]}.force ).to eq([
-        ["multiref/ref2/", ""],
-        ["multiref/ref2/datum-1", "multi-ref2-datum1\n"],
-        ["multiref/ref2/datum-2", "multi-ref2-datum2\n"]
-      ])
-    end
-  end
-
-end
diff --git a/spec/services/zip_service/zip_entry_dirs_spec.rb b/spec/services/zip_service/zip_entry_dirs_spec.rb
deleted file mode 100644
index 8ca1b0f1a..000000000
--- a/spec/services/zip_service/zip_entry_dirs_spec.rb
+++ /dev/null
@@ -1,33 +0,0 @@
-RSpec.describe ZipService do
-
-  let( :zip_service ){ described_class }
-
-  let( :zip_data ){ File.read zip_file }
-
-  shared_examples_for 'a correct zip entry reader' do
-    it 'gets all entries of the zip file' do
-      expect( zip_service.new(zip_data).entry_groups.keys ).to eq(expected)
-    end
-  end
-
-  context 'single entry' do
-    let( :zip_file ){ fixtures_path 'multiple_references_import.zip' }
-    let( :expected ){ %w{ref1 ref2} }
-
-    it_behaves_like 'a correct zip entry reader'
-  end
-
-  context 'more entries' do
-    let( :zip_file ){ fixtures_path 'single_reference_import.zip' }
-    let( :expected ){ %w{ref} }
-
-    it_behaves_like 'a correct zip entry reader'
-  end
-  
-  context 'illegal file' do
-    let( :zip_file ){ fixtures_path 'nozip.zip' }
-    let( :expected ){ [] }
-
-    it_behaves_like 'a correct zip entry reader'
-  end
-end
diff --git a/spec/services/zip_service/zip_output_streams_spec.rb b/spec/services/zip_service/zip_output_streams_spec.rb
deleted file mode 100644
index 742f9b996..000000000
--- a/spec/services/zip_service/zip_output_streams_spec.rb
+++ /dev/null
@@ -1,8 +0,0 @@
-RSpec.describe ZipService do
-  
-  subject{ described_class.new(read_fixture('multiple_references_import.zip')) }
-
-  it "exposes its size" do
-    expect( subject.entry_group_streams.size ).to eq(2)
-  end
-end