Fix binary encoding error on MR diffs

3 files changed, 16 insertions, 1 deletions

diff --git a/changelogs/unreleased/fix-encoding-binary-issue.yml b/changelogs/unreleased/fix-encoding-binary-issue.yml
new file mode 100644
index 00000000000..ac9aff64a88
--- /dev/null
+++ b/changelogs/unreleased/fix-encoding-binary-issue.yml
@@ -0,0 +1,4 @@
+---
+title: Fix binary encoding error on MR diffs
+merge_request: 11929
+author:
diff --git a/lib/gitlab/encoding_helper.rb b/lib/gitlab/encoding_helper.rb
index dbe28e6bb93..781f9c56a42 100644
--- a/lib/gitlab/encoding_helper.rb
+++ b/lib/gitlab/encoding_helper.rb
@@ -38,7 +38,7 @@ module Gitlab
 
     def encode_utf8(message)
       detect = CharlockHolmes::EncodingDetector.detect(message)
-      if detect
+      if detect && detect[:encoding]
         begin
           CharlockHolmes::Converter.convert(message, detect[:encoding], 'UTF-8')
         rescue ArgumentError => e
diff --git a/spec/models/merge_request_diff_spec.rb b/spec/models/merge_request_diff_spec.rb
index 0a10ee01506..ed9fde57bf7 100644
--- a/spec/models/merge_request_diff_spec.rb
+++ b/spec/models/merge_request_diff_spec.rb
@@ -139,4 +139,15 @@ describe MergeRequestDiff, models: true do
       expect(subject.commits_count).to eq 2
     end
   end
+
+  describe '#utf8_st_diffs' do
+    it 'does not raise error when a hash value is in binary' do
+      subject.st_diffs = [
+        { diff: "\0" },
+        { diff: "\x05\x00\x68\x65\x6c\x6c\x6f" }
+      ]
+
+      expect { subject.utf8_st_diffs }.not_to raise_error
+    end
+  end
 end