Merge branch '277162-regexp-support-for-search_files_by_name' into 'master'

Implement wildcard support for file searching in repository

See merge request gitlab-org/gitlab!55871

app/models/repository.rb

@@ -995,6 +995,12 @@ class Repository
     raw_repository.search_files_by_name(query, ref)
   end
 
+  def search_files_by_wildcard_path(path, ref = 'HEAD')
+    # We need to use RE2 to match Gitaly's regexp engine
+    regexp_string = RE2::Regexp.escape(path).gsub('\*', '.*?')
+    raw_repository.search_files_by_regexp("^#{regexp_string}$", ref)
+  end
+
   def copy_gitattributes(ref)
     actual_ref = ref || root_ref
     begin

lib/gitlab/git/repository.rb

@@ -1017,6 +1017,10 @@ module Gitlab
         gitaly_repository_client.search_files_by_name(ref, safe_query)
       end
 
+      def search_files_by_regexp(filter, ref = 'HEAD')
+        gitaly_repository_client.search_files_by_regexp(ref, filter)
+      end
+
       def find_commits_by_message(query, ref, path, limit, offset)
         wrapped_gitaly_errors do
           gitaly_commit_client

lib/gitlab/gitaly_client/repository_service.rb

@@ -339,6 +339,11 @@ module Gitlab
         search_results_from_response(response, options)
       end
 
+      def search_files_by_regexp(ref, filter)
+        request = Gitaly::SearchFilesByNameRequest.new(repository: @gitaly_repo, ref: ref, query: '.', filter: filter)
+        GitalyClient.call(@storage, :repository_service, :search_files_by_name, request, timeout: GitalyClient.fast_timeout).flat_map(&:files)
+      end
+
       def disconnect_alternates
         request = Gitaly::DisconnectGitAlternatesRequest.new(
           repository: @gitaly_repo

spec/lib/gitlab/git/repository_spec.rb

@@ -564,6 +564,41 @@ RSpec.describe Gitlab::Git::Repository, :seed_helper do
     end
   end
 
+  describe '#search_files_by_regexp' do
+    let(:ref) { 'master' }
+
+    subject(:result) { mutable_repository.search_files_by_regexp(filter, ref) }
+
+    context 'when sending a valid regexp' do
+      let(:filter) { 'files\/.*\/.*\.rb' }
+
+      it 'returns matched files' do
+        expect(result).to contain_exactly('files/links/regex.rb',
+                                          'files/ruby/popen.rb',
+                                          'files/ruby/regex.rb',
+                                          'files/ruby/version_info.rb')
+      end
+    end
+
+    context 'when sending an ivalid regexp' do
+      let(:filter) { '*.rb' }
+
+      it 'raises error' do
+        expect { result }.to raise_error(GRPC::InvalidArgument,
+                                         /missing argument to repetition operator: `*`/)
+      end
+    end
+
+    context "when the ref doesn't exist" do
+      let(:filter) { 'files\/.*\/.*\.rb' }
+      let(:ref) { 'non-existing-branch' }
+
+      it 'returns an empty array' do
+        expect(result).to eq([])
+      end
+    end
+  end
+
   describe '#find_remote_root_ref' do
     it 'gets the remote root ref from GitalyClient' do
       expect_any_instance_of(Gitlab::GitalyClient::RemoteService)

spec/lib/gitlab/gitaly_client/repository_service_spec.rb

@@ -246,6 +246,21 @@ RSpec.describe Gitlab::GitalyClient::RepositoryService do
     end
   end
 
+  describe '#search_files_by_regexp' do
+    subject(:result) { client.search_files_by_regexp('master', '.*') }
+
+    before do
+      expect_any_instance_of(Gitaly::RepositoryService::Stub)
+        .to receive(:search_files_by_name)
+        .with(gitaly_request_with_path(storage_name, relative_path), kind_of(Hash))
+        .and_return([double(files: ['file1.txt']), double(files: ['file2.txt'])])
+    end
+
+    it 'sends a search_files_by_name message and returns a flatten array' do
+      expect(result).to contain_exactly('file1.txt', 'file2.txt')
+    end
+  end
+
   describe '#disconnect_alternates' do
     let(:project) { create(:project, :repository) }
     let(:repository) { project.repository }

spec/models/repository_spec.rb

@@ -977,6 +977,57 @@ RSpec.describe Repository do
     end
   end
 
+  describe '#search_files_by_wildcard_path' do
+    let(:ref) { 'master' }
+
+    subject(:result) { repository.search_files_by_wildcard_path(path, ref) }
+
+    context 'when specifying a normal path' do
+      let(:path) { 'files/images/logo-black.png' }
+
+      it 'returns the path' do
+        expect(result).to eq(['files/images/logo-black.png'])
+      end
+    end
+
+    context 'when specifying a path with wildcard' do
+      let(:path) { 'files/*/*.png' }
+
+      it 'returns all files matching the path' do
+        expect(result).to contain_exactly('files/images/logo-black.png',
+                                          'files/images/logo-white.png')
+      end
+    end
+
+    context 'when specifying an extension with wildcard' do
+      let(:path) { '*.rb' }
+
+      it 'returns all files matching the extension' do
+        expect(result).to contain_exactly('encoding/russian.rb',
+                                          'files/ruby/popen.rb',
+                                          'files/ruby/regex.rb',
+                                          'files/ruby/version_info.rb')
+      end
+    end
+
+    context 'when sending regexp' do
+      let(:path) { '.*\.rb' }
+
+      it 'ignores the regexp and returns an empty array' do
+        expect(result).to eq([])
+      end
+    end
+
+    context 'when sending another ref' do
+      let(:path) { 'files' }
+      let(:ref) { 'other-branch' }
+
+      it 'returns an empty array' do
+        expect(result).to eq([])
+      end
+    end
+  end
+
   describe '#async_remove_remote' do
     before do
       masterrev = repository.find_branch('master').dereferenced_target