Merge branch '24704-restore-download-repository-path' into 'master'

Restore "download repository path" feature, now with fixes and a feature flag

Closes gitlab-workhorse#218 and #24704

See merge request gitlab-org/gitlab-ce!27275

2 files changed, 64 insertions, 8 deletions

diff --git a/lib/gitlab/git/repository.rb b/lib/gitlab/git/repository.rb
index be9e926728c..a22e3c4b9dd 100644
--- a/lib/gitlab/git/repository.rb
+++ b/lib/gitlab/git/repository.rb
@@ -231,12 +231,12 @@ module Gitlab
         end
       end
 
-      def archive_metadata(ref, storage_path, project_path, format = "tar.gz", append_sha:)
+      def archive_metadata(ref, storage_path, project_path, format = "tar.gz", append_sha:, path: nil)
         ref ||= root_ref
         commit = Gitlab::Git::Commit.find(self, ref)
         return {} if commit.nil?
 
-        prefix = archive_prefix(ref, commit.id, project_path, append_sha: append_sha)
+        prefix = archive_prefix(ref, commit.id, project_path, append_sha: append_sha, path: path)
 
         {
           'ArchivePrefix' => prefix,
@@ -248,13 +248,14 @@ module Gitlab
 
       # This is both the filename of the archive (missing the extension) and the
       # name of the top-level member of the archive under which all files go
-      def archive_prefix(ref, sha, project_path, append_sha:)
+      def archive_prefix(ref, sha, project_path, append_sha:, path:)
         append_sha = (ref != sha) if append_sha.nil?
 
         formatted_ref = ref.tr('/', '-')
 
         prefix_segments = [project_path, formatted_ref]
         prefix_segments << sha if append_sha
+        prefix_segments << path.tr('/', '-').gsub(%r{^/|/$}, '') if path
 
         prefix_segments.join('-')
       end
diff --git a/lib/gitlab/workhorse.rb b/lib/gitlab/workhorse.rb
index 0c2acac3d1e..46a7b5b982a 100644
--- a/lib/gitlab/workhorse.rb
+++ b/lib/gitlab/workhorse.rb
@@ -63,16 +63,34 @@ module Gitlab
         ]
       end
 
-      def send_git_archive(repository, ref:, format:, append_sha:)
+      def send_git_archive(repository, ref:, format:, append_sha:, path: nil)
+        path_enabled = Feature.enabled?(:git_archive_path, default_enabled: true)
+        path = nil unless path_enabled
+
         format ||= 'tar.gz'
         format = format.downcase
-        params = repository.archive_metadata(ref, Gitlab.config.gitlab.repository_downloads_path, format, append_sha: append_sha)
-        raise "Repository or ref not found" if params.empty?
 
-        params['GitalyServer'] = gitaly_server_hash(repository)
+        metadata = repository.archive_metadata(
+          ref,
+          Gitlab.config.gitlab.repository_downloads_path,
+          format,
+          append_sha: append_sha,
+          path: path
+        )
 
-        # If present DisableCache must be a Boolean. Otherwise workhorse ignores it.
+        raise "Repository or ref not found" if metadata.empty?
+
+        params =
+          if path_enabled
+            send_git_archive_params(repository, metadata, path, archive_format(format))
+          else
+            metadata
+          end
+
+        # If present, DisableCache must be a Boolean. Otherwise
+        # workhorse ignores it.
         params['DisableCache'] = true if git_archive_cache_disabled?
+        params['GitalyServer'] = gitaly_server_hash(repository)
 
         [
           SEND_DATA_HEADER,
@@ -216,10 +234,19 @@ module Gitlab
 
       protected
 
+      # This is the outermost encoding of a senddata: header. It is safe for
+      # inclusion in HTTP response headers
       def encode(hash)
         Base64.urlsafe_encode64(JSON.dump(hash))
       end
 
+      # This is for encoding individual fields inside the senddata JSON that
+      # contain binary data. In workhorse, the corresponding struct field should
+      # be type []byte
+      def encode_binary(binary)
+        Base64.encode64(binary)
+      end
+
       def gitaly_server_hash(repository)
         {
           address: Gitlab::GitalyClient.address(repository.project.repository_storage),
@@ -238,6 +265,34 @@ module Gitlab
       def git_archive_cache_disabled?
         ENV['WORKHORSE_ARCHIVE_CACHE_DISABLED'].present? || Feature.enabled?(:workhorse_archive_cache_disabled)
       end
+
+      def archive_format(format)
+        case format
+        when "tar.bz2", "tbz", "tbz2", "tb2", "bz2"
+          Gitaly::GetArchiveRequest::Format::TAR_BZ2
+        when "tar"
+          Gitaly::GetArchiveRequest::Format::TAR
+        when "zip"
+          Gitaly::GetArchiveRequest::Format::ZIP
+        else
+          Gitaly::GetArchiveRequest::Format::TAR_GZ
+        end
+      end
+
+      def send_git_archive_params(repository, metadata, path, format)
+        {
+          'ArchivePath' => metadata['ArchivePath'],
+          'GetArchiveRequest' => encode_binary(
+            Gitaly::GetArchiveRequest.new(
+              repository: repository.gitaly_repository,
+              commit_id: metadata['CommitId'],
+              prefix: metadata['ArchivePrefix'],
+              format: format,
+              path: path.presence || ""
+            ).to_proto
+          )
+        }
+      end
     end
   end
 end