Add latest changes from gitlab-org/security/gitlab@14-0-stable-ee

2 files changed, 11 insertions, 2 deletions

diff --git a/lib/gitlab/diff/file.rb b/lib/gitlab/diff/file.rb
index dcd4bbdabf5..35581952f4a 100644
--- a/lib/gitlab/diff/file.rb
+++ b/lib/gitlab/diff/file.rb
@@ -250,7 +250,7 @@ module Gitlab
       end
 
       def diffable?
-        repository.attributes(file_path).fetch('diff') { true }
+        diffable_by_attribute? && !text_with_binary_notice?
       end
 
       def binary_in_repo?
@@ -366,6 +366,15 @@ module Gitlab
 
       private
 
+      def diffable_by_attribute?
+        repository.attributes(file_path).fetch('diff') { true }
+      end
+
+      # NOTE: Files with unsupported encodings (e.g. UTF-16) are treated as binary by git, but they are recognized as text files during encoding detection. These files have `Binary files a/filename and b/filename differ' as their raw diff content which cannot be used. We need to handle this special case and avoid displaying incorrect diff.
+      def text_with_binary_notice?
+        text? && has_binary_notice?
+      end
+
       def fetch_blob(sha, path)
         return unless sha
 
diff --git a/lib/gitlab/diff/parser.rb b/lib/gitlab/diff/parser.rb
index 4a47e4b80b6..adb711ca89f 100644
--- a/lib/gitlab/diff/parser.rb
+++ b/lib/gitlab/diff/parser.rb
@@ -6,7 +6,7 @@ module Gitlab
       include Enumerable
 
       def parse(lines, diff_file: nil)
-        return [] if lines.blank?
+        return [] if lines.blank? || Git::Diff.has_binary_notice?(lines.first)
 
         @lines = lines
         line_obj_index = 0