revert to using a simple representation

lib/api/entities.rb

@@ -295,7 +295,7 @@ module API
       expose :new_file?, as: :new_file
       expose :renamed_file?, as: :renamed_file
       expose :deleted_file?, as: :deleted_file
-      expose :diff
+      expose :json_safe_diff, as: :diff
     end
 
     class ProtectedRefAccess < Grape::Entity

lib/gitlab/encoding_helper.rb

@@ -24,7 +24,7 @@ module Gitlab
 
       # return message if message type is binary
       detect = CharlockHolmes::EncodingDetector.detect(message)
-      return message.force_encoding("BINARY") if binary?(message, detect)
+      return message.force_encoding("BINARY") if all_binary?(message, detect)
 
       if detect && detect[:encoding] && detect[:confidence] > ENCODING_CONFIDENCE_THRESHOLD
         # force detected encoding if we have sufficient confidence.
@@ -34,14 +34,21 @@ module Gitlab
       # encode and clean the bad chars
       message.replace clean(message)
     rescue => e
-      byebug
       encoding = detect ? detect[:encoding] : "unknown"
       "--broken encoding: #{encoding}"
     end
 
-    def binary?(message, detect=nil)
-      detect ||= CharlockHolmes::EncodingDetector.detect(message)
-      detect && detect[:type] == :binary && detect[:confidence] == 100
+    def all_binary?(data, detect=nil)
+      detect ||= CharlockHolmes::EncodingDetector.detect(data)
+      detect && detect[:type] == :binary
+    end
+
+    def libgit2_binary?(data)
+      # EncodingDetector checks the first 1024 * 1024 bytes for NUL byte, libgit2 checks
+      # only the first 8000 (https://github.com/libgit2/libgit2/blob/2ed855a9e8f9af211e7274021c2264e600c0f86b/src/filter.h#L15),
+      # which is what we use below to keep a consistent behavior.
+      detect = CharlockHolmes::EncodingDetector.new(8000).detect(data)
+      all_binary?(data, detect)
     end
 
     def encode_utf8(message)

lib/gitlab/git/blob.rb

@@ -42,14 +42,6 @@ module Gitlab
           end
         end
 
-        def binary?(data)
-          # EncodingDetector checks the first 1024 * 1024 bytes for NUL byte, libgit2 checks
-          # only the first 8000 (https://github.com/libgit2/libgit2/blob/2ed855a9e8f9af211e7274021c2264e600c0f86b/src/filter.h#L15),
-          # which is what we use below to keep a consistent behavior.
-          detect = CharlockHolmes::EncodingDetector.new(8000).detect(data)
-          detect && detect[:type] == :binary
-        end
-
         # Returns an array of Blob instances, specified in blob_references as
         # [[commit_sha, path], [commit_sha, path], ...]. If blob_size_limit < 0 then the
         # full blob contents are returned. If blob_size_limit >= 0 then each blob will
@@ -169,6 +161,10 @@ module Gitlab
             end
           end
         end
+
+        def binary?(data)
+          EncodingHelper.libgit2_binary?(data)
+        end
       end
 
       def initialize(options)

lib/gitlab/git/diff.rb

@@ -197,6 +197,13 @@ module Gitlab
         @collapsed = true
       end
 
+      def json_safe_diff
+        return @diff unless all_binary?(@diff)
+
+        # the diff is binary, let's make a message for it
+        Diff::binary_message(@old_path, @new_path)
+      end
+
       private
 
       def init_from_rugged(rugged)
@@ -221,14 +228,7 @@ module Gitlab
         # binary we're not going to display anything so we skip the size check.
         return if !patch.delta.binary? && prune_large_patch(patch)
 
-        diff = strip_diff_headers(patch.to_s)
-        @diff = if binary?(diff)
-                  # the diff is binary, let's make a message for it
-                  Diff::binary_message(patch.delta.old_file[:path],
-                                       patch.delta.new_file[:path])
-                else
-                  encode!(diff)
-                end
+        @diff = encode!(strip_diff_headers(patch.to_s))
       end
 
       def init_from_hash(hash)