Prevent RevList failing on non utf8 pathsjej/lfs-rev-list-handles-non-utf-paths-41627

3 files changed, 16 insertions, 2 deletions

diff --git a/changelogs/unreleased/jej-lfs-rev-list-handles-non-utf-paths-41627.yml b/changelogs/unreleased/jej-lfs-rev-list-handles-non-utf-paths-41627.yml
new file mode 100644
index 00000000000..24f18c07ac5
--- /dev/null
+++ b/changelogs/unreleased/jej-lfs-rev-list-handles-non-utf-paths-41627.yml
@@ -0,0 +1,5 @@
+---
+title: Prevent RevList failing on non utf8 paths
+merge_request: 16440
+author:
+type: fixed
diff --git a/lib/gitlab/git/rev_list.rb b/lib/gitlab/git/rev_list.rb
index 4974205b8fd..f8b2e7e0e21 100644
--- a/lib/gitlab/git/rev_list.rb
+++ b/lib/gitlab/git/rev_list.rb
@@ -95,7 +95,7 @@ module Gitlab
         object_output.map do |output_line|
           sha, path = output_line.split(' ', 2)
 
-          next if require_path && path.blank?
+          next if require_path && path.to_s.empty?
 
           sha
         end.reject(&:nil?)
diff --git a/spec/lib/gitlab/git/rev_list_spec.rb b/spec/lib/gitlab/git/rev_list_spec.rb
index eaf74951b0e..90fbef9d248 100644
--- a/spec/lib/gitlab/git/rev_list_spec.rb
+++ b/spec/lib/gitlab/git/rev_list_spec.rb
@@ -39,7 +39,7 @@ describe Gitlab::Git::RevList do
     ]
 
     expect(rev_list).to receive(:popen).with(*params) do |*_, lazy_block:|
-      lazy_block.call(output.split("\n").lazy)
+      lazy_block.call(output.lines.lazy.map(&:chomp))
     end
   end
 
@@ -64,6 +64,15 @@ describe Gitlab::Git::RevList do
       expect(rev_list.new_objects(require_path: true)).to eq(%w[sha2])
     end
 
+    it 'can handle non utf-8 paths' do
+      non_utf_char = [0x89].pack("c*").force_encoding("UTF-8")
+      stub_lazy_popen_rev_list('newrev', '--not', '--all', '--objects', output: "sha2 πå†h/†ø/ƒîlé#{non_utf_char}\nsha1")
+
+      rev_list.new_objects(require_path: true) do |object_ids|
+        expect(object_ids.force).to eq(%w[sha2])
+      end
+    end
+
     it 'can yield a lazy enumerator' do
       stub_lazy_popen_rev_list('newrev', '--not', '--all', '--objects', output: "sha1\nsha2")