repository.rb 38.5 KB
Newer Older
Robert Speicher's avatar
Robert Speicher committed
1 2 3 4 5 6 7 8 9 10
# Gitlab::Git::Repository is a wrapper around native Rugged::Repository object
require 'tempfile'
require 'forwardable'
require "rubygems/package"

module Gitlab
  module Git
    class Repository
      include Gitlab::Git::Popen

11 12 13 14
      ALLOWED_OBJECT_DIRECTORIES_VARIABLES = %w[
        GIT_OBJECT_DIRECTORY
        GIT_ALTERNATE_OBJECT_DIRECTORIES
      ].freeze
Robert Speicher's avatar
Robert Speicher committed
15 16
      SEARCH_CONTEXT_LINES = 3

17 18 19
      NoRepository = Class.new(StandardError)
      InvalidBlobName = Class.new(StandardError)
      InvalidRef = Class.new(StandardError)
Robert Speicher's avatar
Robert Speicher committed
20 21 22 23 24 25 26 27 28 29

      # Full path to repo
      attr_reader :path

      # Directory name of repo
      attr_reader :name

      # Rugged repo object
      attr_reader :rugged

Jacob Vosmaer's avatar
Jacob Vosmaer committed
30 31
      attr_reader :storage

Robert Speicher's avatar
Robert Speicher committed
32 33
      # 'path' must be the path to a _bare_ git repository, e.g.
      # /path/to/my-repo.git
Jacob Vosmaer's avatar
Jacob Vosmaer committed
34 35
      def initialize(storage, relative_path)
        @storage = storage
36 37
        @relative_path = relative_path

Jacob Vosmaer's avatar
Jacob Vosmaer committed
38
        storage_path = Gitlab.config.repositories.storages[@storage]['path']
39 40
        @path = File.join(storage_path, @relative_path)
        @name = @relative_path.split("/").last
41
        @attributes = Gitlab::Git::Attributes.new(path)
Robert Speicher's avatar
Robert Speicher committed
42 43
      end

Douwe Maan's avatar
Douwe Maan committed
44 45 46 47
      delegate  :empty?,
                :bare?,
                to: :rugged

Robert Speicher's avatar
Robert Speicher committed
48 49
      # Default branch in the repository
      def root_ref
50 51 52 53 54 55 56
        @root_ref ||= gitaly_migrate(:root_ref) do |is_enabled|
          if is_enabled
            gitaly_ref_client.default_branch_name
          else
            discover_default_branch
          end
        end
Robert Speicher's avatar
Robert Speicher committed
57 58 59 60 61 62 63 64
      end

      # Alias to old method for compatibility
      def raw
        rugged
      end

      def rugged
65
        @rugged ||= Rugged::Repository.new(path, alternates: alternate_object_directories)
Robert Speicher's avatar
Robert Speicher committed
66 67 68 69 70 71 72
      rescue Rugged::RepositoryError, Rugged::OSError
        raise NoRepository.new('no repository for such path')
      end

      # Returns an Array of branch names
      # sorted by name ASC
      def branch_names
73 74 75 76 77 78 79
        gitaly_migrate(:branch_names) do |is_enabled|
          if is_enabled
            gitaly_ref_client.branch_names
          else
            branches.map(&:name)
          end
        end
Robert Speicher's avatar
Robert Speicher committed
80 81 82 83 84 85
      end

      # Returns an Array of Branches
      def branches
        rugged.branches.map do |rugged_ref|
          begin
86
            Gitlab::Git::Branch.new(self, rugged_ref.name, rugged_ref.target)
Robert Speicher's avatar
Robert Speicher committed
87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107
          rescue Rugged::ReferenceError
            # Omit invalid branch
          end
        end.compact.sort_by(&:name)
      end

      def reload_rugged
        @rugged = nil
      end

      # Directly find a branch with a simple name (e.g. master)
      #
      # force_reload causes a new Rugged repository to be instantiated
      #
      # This is to work around a bug in libgit2 that causes in-memory refs to
      # be stale/invalid when packed-refs is changed.
      # See https://gitlab.com/gitlab-org/gitlab-ce/issues/15392#note_14538333
      def find_branch(name, force_reload = false)
        reload_rugged if force_reload

        rugged_ref = rugged.branches[name]
108
        Gitlab::Git::Branch.new(self, rugged_ref.name, rugged_ref.target) if rugged_ref
Robert Speicher's avatar
Robert Speicher committed
109 110 111 112
      end

      def local_branches
        rugged.branches.each(:local).map do |branch|
113
          Gitlab::Git::Branch.new(self, branch.name, branch.target)
Robert Speicher's avatar
Robert Speicher committed
114 115 116 117 118
        end
      end

      # Returns the number of valid branches
      def branch_count
119
        gitaly_migrate(:branch_names) do |is_enabled|
120 121 122 123 124 125 126 127 128 129 130 131 132 133 134
          if is_enabled
            gitaly_ref_client.count_branch_names
          else
            rugged.branches.count do |ref|
              begin
                ref.name && ref.target # ensures the branch is valid

                true
              rescue Rugged::ReferenceError
                false
              end
            end
          end
        end
      end
Robert Speicher's avatar
Robert Speicher committed
135

136 137
      # Returns the number of valid tags
      def tag_count
138
        gitaly_migrate(:tag_names) do |is_enabled|
139 140 141 142
          if is_enabled
            gitaly_ref_client.count_tag_names
          else
            rugged.tags.count
Robert Speicher's avatar
Robert Speicher committed
143 144 145 146 147 148
          end
        end
      end

      # Returns an Array of tag names
      def tag_names
149 150 151 152 153 154 155
        gitaly_migrate(:tag_names) do |is_enabled|
          if is_enabled
            gitaly_ref_client.tag_names
          else
            rugged.tags.map { |t| t.name }
          end
        end
Robert Speicher's avatar
Robert Speicher committed
156 157 158 159 160 161 162 163 164 165 166 167 168 169 170
      end

      # Returns an Array of Tags
      def tags
        rugged.references.each("refs/tags/*").map do |ref|
          message = nil

          if ref.target.is_a?(Rugged::Tag::Annotation)
            tag_message = ref.target.message

            if tag_message.respond_to?(:chomp)
              message = tag_message.chomp
            end
          end

171
          Gitlab::Git::Tag.new(self, ref.name, ref.target, message)
Robert Speicher's avatar
Robert Speicher committed
172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246
        end.sort_by(&:name)
      end

      # Returns true if the given tag exists
      #
      # name - The name of the tag as a String.
      def tag_exists?(name)
        !!rugged.tags[name]
      end

      # Returns true if the given branch exists
      #
      # name - The name of the branch as a String.
      def branch_exists?(name)
        rugged.branches.exists?(name)

      # If the branch name is invalid (e.g. ".foo") Rugged will raise an error.
      # Whatever code calls this method shouldn't have to deal with that so
      # instead we just return `false` (which is true since a branch doesn't
      # exist when it has an invalid name).
      rescue Rugged::ReferenceError
        false
      end

      # Returns an Array of branch and tag names
      def ref_names
        branch_names + tag_names
      end

      # Deprecated. Will be removed in 5.2
      def heads
        rugged.references.each("refs/heads/*").map do |head|
          Gitlab::Git::Ref.new(self, head.name, head.target)
        end.sort_by(&:name)
      end

      def has_commits?
        !empty?
      end

      def repo_exists?
        !!rugged
      end

      # Discovers the default branch based on the repository's available branches
      #
      # - If no branches are present, returns nil
      # - If one branch is present, returns its name
      # - If two or more branches are present, returns current HEAD or master or first branch
      def discover_default_branch
        names = branch_names

        return if names.empty?

        return names[0] if names.length == 1

        if rugged_head
          extracted_name = Ref.extract_branch_name(rugged_head.name)

          return extracted_name if names.include?(extracted_name)
        end

        if names.include?('master')
          'master'
        else
          names[0]
        end
      end

      def rugged_head
        rugged.head
      rescue Rugged::ReferenceError
        nil
      end

247 248
      def archive_prefix(ref, sha)
        project_name = self.name.chomp('.git')
249
        "#{project_name}-#{ref.tr('/', '-')}-#{sha}"
250 251
      end

Robert Speicher's avatar
Robert Speicher committed
252 253 254 255 256
      def archive_metadata(ref, storage_path, format = "tar.gz")
        ref ||= root_ref
        commit = Gitlab::Git::Commit.find(self, ref)
        return {} if commit.nil?

257
        prefix = archive_prefix(ref, commit.id)
Robert Speicher's avatar
Robert Speicher committed
258 259 260 261 262

        {
          'RepoPath' => path,
          'ArchivePrefix' => prefix,
          'ArchivePath' => archive_file_path(prefix, storage_path, format),
263
          'CommitId' => commit.id
Robert Speicher's avatar
Robert Speicher committed
264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303
        }
      end

      def archive_file_path(name, storage_path, format = "tar.gz")
        # Build file path
        return nil unless name

        extension =
          case format
          when "tar.bz2", "tbz", "tbz2", "tb2", "bz2"
            "tar.bz2"
          when "tar"
            "tar"
          when "zip"
            "zip"
          else
            # everything else should fall back to tar.gz
            "tar.gz"
          end

        file_name = "#{name}.#{extension}"
        File.join(storage_path, self.name, file_name)
      end

      # Return repo size in megabytes
      def size
        size = popen(%w(du -sk), path).first.strip.to_i
        (size.to_f / 1024).round(2)
      end

      # Returns an array of BlobSnippets for files at the specified +ref+ that
      # contain the +query+ string.
      def search_files(query, ref = nil)
        greps = []
        ref ||= root_ref

        populated_index(ref).each do |entry|
          # Discard submodules
          next if submodule?(entry)

304
          blob = Gitlab::Git::Blob.raw(self, entry[:oid])
Robert Speicher's avatar
Robert Speicher committed
305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339

          # Skip binary files
          next if blob.data.encoding == Encoding::ASCII_8BIT

          blob.load_all_data!(self)
          greps += build_greps(blob.data, query, ref, entry[:path])
        end

        greps
      end

      # Use the Rugged Walker API to build an array of commits.
      #
      # Usage.
      #   repo.log(
      #     ref: 'master',
      #     path: 'app/models',
      #     limit: 10,
      #     offset: 5,
      #     after: Time.new(2016, 4, 21, 14, 32, 10)
      #   )
      #
      def log(options)
        default_options = {
          limit: 10,
          offset: 0,
          path: nil,
          follow: false,
          skip_merges: false,
          disable_walk: false,
          after: nil,
          before: nil
        }

        options = default_options.merge(options)
340
        options[:limit] ||= 0
Robert Speicher's avatar
Robert Speicher committed
341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367
        options[:offset] ||= 0
        actual_ref = options[:ref] || root_ref
        begin
          sha = sha_from_ref(actual_ref)
        rescue Rugged::OdbError, Rugged::InvalidError, Rugged::ReferenceError
          # Return an empty array if the ref wasn't found
          return []
        end

        if log_using_shell?(options)
          log_by_shell(sha, options)
        else
          log_by_walk(sha, options)
        end
      end

      def log_using_shell?(options)
        options[:path].present? ||
          options[:disable_walk] ||
          options[:skip_merges] ||
          options[:after] ||
          options[:before]
      end

      def log_by_walk(sha, options)
        walk_options = {
          show: sha,
368
          sort: Rugged::SORT_NONE,
Robert Speicher's avatar
Robert Speicher committed
369 370 371 372 373 374 375
          limit: options[:limit],
          offset: options[:offset]
        }
        Rugged::Walker.walk(rugged, walk_options).to_a
      end

      def log_by_shell(sha, options)
376 377 378 379 380 381 382 383 384
        limit = options[:limit].to_i
        offset = options[:offset].to_i
        use_follow_flag = options[:follow] && options[:path].present?

        # We will perform the offset in Ruby because --follow doesn't play well with --skip.
        # See: https://gitlab.com/gitlab-org/gitlab-ce/issues/3574#note_3040520
        offset_in_ruby = use_follow_flag && options[:offset].present?
        limit += offset if offset_in_ruby

385 386 387 388 389 390 391 392 393
        cmd = %W[#{Gitlab.config.git.bin_path} --git-dir=#{path} log]
        cmd << "--max-count=#{limit}"
        cmd << '--format=%H'
        cmd << "--skip=#{offset}" unless offset_in_ruby
        cmd << '--follow' if use_follow_flag
        cmd << '--no-merges' if options[:skip_merges]
        cmd << "--after=#{options[:after].iso8601}" if options[:after]
        cmd << "--before=#{options[:before].iso8601}" if options[:before]
        cmd << sha
394 395 396 397 398 399

        # :path can be a string or an array of strings
        if options[:path].present?
          cmd << '--'
          cmd += Array(options[:path])
        end
Robert Speicher's avatar
Robert Speicher committed
400

401 402
        raw_output = IO.popen(cmd) { |io| io.read }
        lines = offset_in_ruby ? raw_output.lines.drop(offset) : raw_output.lines
Robert Speicher's avatar
Robert Speicher committed
403

404
        lines.map! { |c| Rugged::Commit.new(rugged, c.strip) }
Robert Speicher's avatar
Robert Speicher committed
405 406
      end

407
      def count_commits(options)
408 409 410 411 412 413 414 415
        cmd = %W[#{Gitlab.config.git.bin_path} --git-dir=#{path} rev-list]
        cmd << "--after=#{options[:after].iso8601}" if options[:after]
        cmd << "--before=#{options[:before].iso8601}" if options[:before]
        cmd += %W[--count #{options[:ref]}]
        cmd += %W[-- #{options[:path]}] if options[:path].present?

        raw_output = IO.popen(cmd) { |io| io.read }

416 417 418
        raw_output.to_i
      end

Robert Speicher's avatar
Robert Speicher committed
419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434
      def sha_from_ref(ref)
        rev_parse_target(ref).oid
      end

      # Return the object that +revspec+ points to.  If +revspec+ is an
      # annotated tag, then return the tag's target instead.
      def rev_parse_target(revspec)
        obj = rugged.rev_parse(revspec)
        Ref.dereference_object(obj)
      end

      # Return a collection of Rugged::Commits between the two revspec arguments.
      # See http://git-scm.com/docs/git-rev-parse.html#_specifying_revisions for
      # a detailed list of valid arguments.
      def commits_between(from, to)
        walker = Rugged::Walker.new(rugged)
435
        walker.sorting(Rugged::SORT_NONE | Rugged::SORT_REVERSE)
Robert Speicher's avatar
Robert Speicher committed
436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458

        sha_from = sha_from_ref(from)
        sha_to = sha_from_ref(to)

        walker.push(sha_to)
        walker.hide(sha_from)

        commits = walker.to_a
        walker.reset

        commits
      end

      # Counts the amount of commits between `from` and `to`.
      def count_commits_between(from, to)
        commits_between(from, to).size
      end

      # Returns the SHA of the most recent common ancestor of +from+ and +to+
      def merge_base_commit(from, to)
        rugged.merge_base(from, to)
      end

459 460
      # Returns true is +from+ is direct ancestor to +to+, otherwise false
      def is_ancestor?(from, to)
461
        gitaly_commit_client.is_ancestor(from, to)
462 463
      end

Robert Speicher's avatar
Robert Speicher committed
464 465 466 467 468
      # Return an array of Diff objects that represent the diff
      # between +from+ and +to+.  See Diff::filter_diff_options for the allowed
      # diff options.  The +options+ hash can also include :break_rewrites to
      # split larger rewrites into delete/add pairs.
      def diff(from, to, options = {}, *paths)
469
        Gitlab::Git::DiffCollection.new(diff_patches(from, to, options, *paths), options)
Robert Speicher's avatar
Robert Speicher committed
470 471
      end

472 473
      # Returns a RefName for a given SHA
      def ref_name_for_sha(ref_path, sha)
474 475
        # NOTE: This feature is intentionally disabled until
        # https://gitlab.com/gitlab-org/gitaly/issues/180 is resolved
476
        # gitaly_migrate(:find_ref_name) do |is_enabled|
477 478 479 480 481 482 483 484 485 486
        #   if is_enabled
        #     gitaly_ref_client.find_ref_name(sha, ref_path)
        #   else
        args = %W(#{Gitlab.config.git.bin_path} for-each-ref --count=1 #{ref_path} --contains #{sha})

        # Not found -> ["", 0]
        # Found -> ["b8d95eb4969eefacb0a58f6a28f6803f8070e7b9 commit\trefs/environments/production/77\n", 0]
        Gitlab::Popen.popen(args, @path).first.split.last
        #   end
        # end
487 488
      end

Robert Speicher's avatar
Robert Speicher committed
489 490 491 492 493 494 495 496 497 498 499 500 501 502 503
      # Returns commits collection
      #
      # Ex.
      #   repo.find_commits(
      #     ref: 'master',
      #     max_count: 10,
      #     skip: 5,
      #     order: :date
      #   )
      #
      #   +options+ is a Hash of optional arguments to git
      #     :ref is the ref from which to begin (SHA1 or name)
      #     :contains is the commit contained by the refs from which to begin (SHA1 or name)
      #     :max_count is the maximum number of commits to fetch
      #     :skip is the number of commits to skip
504 505 506
      #     :order is the commits order and allowed value is :none (default), :date,
      #        :topo, or any combination of them (in an array). Commit ordering types
      #        are documented here:
507
      #        http://www.rubydoc.info/github/libgit2/rugged/Rugged#SORT_NONE-constant)
Robert Speicher's avatar
Robert Speicher committed
508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534
      #
      def find_commits(options = {})
        actual_options = options.dup

        allowed_options = [:ref, :max_count, :skip, :contains, :order]

        actual_options.keep_if do |key|
          allowed_options.include?(key)
        end

        default_options = { skip: 0 }
        actual_options = default_options.merge(actual_options)

        walker = Rugged::Walker.new(rugged)

        if actual_options[:ref]
          walker.push(rugged.rev_parse_oid(actual_options[:ref]))
        elsif actual_options[:contains]
          branches_contains(actual_options[:contains]).each do |branch|
            walker.push(branch.target_id)
          end
        else
          rugged.references.each("refs/heads/*") do |ref|
            walker.push(ref.target_id)
          end
        end

535 536
        sort_type = rugged_sort_type(actual_options[:order])
        walker.sorting(sort_type)
Robert Speicher's avatar
Robert Speicher committed
537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655

        commits = []
        offset = actual_options[:skip]
        limit = actual_options[:max_count]
        walker.each(offset: offset, limit: limit) do |commit|
          gitlab_commit = Gitlab::Git::Commit.decorate(commit)
          commits.push(gitlab_commit)
        end

        walker.reset

        commits
      rescue Rugged::OdbError
        []
      end

      # Returns branch names collection that contains the special commit(SHA1
      # or name)
      #
      # Ex.
      #   repo.branch_names_contains('master')
      #
      def branch_names_contains(commit)
        branches_contains(commit).map { |c| c.name }
      end

      # Returns branch collection that contains the special commit(SHA1 or name)
      #
      # Ex.
      #   repo.branch_names_contains('master')
      #
      def branches_contains(commit)
        commit_obj = rugged.rev_parse(commit)
        parent = commit_obj.parents.first unless commit_obj.parents.empty?

        walker = Rugged::Walker.new(rugged)

        rugged.branches.select do |branch|
          walker.push(branch.target_id)
          walker.hide(parent) if parent
          result = walker.any? { |c| c.oid == commit_obj.oid }
          walker.reset

          result
        end
      end

      # Get refs hash which key is SHA1
      # and value is a Rugged::Reference
      def refs_hash
        # Initialize only when first call
        if @refs_hash.nil?
          @refs_hash = Hash.new { |h, k| h[k] = [] }

          rugged.references.each do |r|
            # Symbolic/remote references may not have an OID; skip over them
            target_oid = r.target.try(:oid)
            if target_oid
              sha = rev_parse_target(target_oid).oid
              @refs_hash[sha] << r
            end
          end
        end
        @refs_hash
      end

      # Lookup for rugged object by oid or ref name
      def lookup(oid_or_ref_name)
        rugged.rev_parse(oid_or_ref_name)
      end

      # Return hash with submodules info for this repository
      #
      # Ex.
      #   {
      #     "rack"  => {
      #       "id" => "c67be4624545b4263184c4a0e8f887efd0a66320",
      #       "path" => "rack",
      #       "url" => "git://github.com/chneukirchen/rack.git"
      #     },
      #     "encoding" => {
      #       "id" => ....
      #     }
      #   }
      #
      def submodules(ref)
        commit = rev_parse_target(ref)
        return {} unless commit

        begin
          content = blob_content(commit, ".gitmodules")
        rescue InvalidBlobName
          return {}
        end

        parse_gitmodules(commit, content)
      end

      # Return total commits count accessible from passed ref
      def commit_count(ref)
        walker = Rugged::Walker.new(rugged)
        walker.sorting(Rugged::SORT_TOPO | Rugged::SORT_REVERSE)
        oid = rugged.rev_parse_oid(ref)
        walker.push(oid)
        walker.count
      end

      # Sets HEAD to the commit specified by +ref+; +ref+ can be a branch or
      # tag name or a commit SHA.  Valid +reset_type+ values are:
      #
      #  [:soft]
      #    the head will be moved to the commit.
      #  [:mixed]
      #    will trigger a +:soft+ reset, plus the index will be replaced
      #    with the content of the commit tree.
      #  [:hard]
      #    will trigger a +:mixed+ reset and the working directory will be
      #    replaced with the content of the index. (Untracked and ignored files
      #    will be left alone)
Douwe Maan's avatar
Douwe Maan committed
656
      delegate :reset, to: :rugged
Robert Speicher's avatar
Robert Speicher committed
657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841

      # Mimic the `git clean` command and recursively delete untracked files.
      # Valid keys that can be passed in the +options+ hash are:
      #
      # :d - Remove untracked directories
      # :f - Remove untracked directories that are managed by a different
      #      repository
      # :x - Remove ignored files
      #
      # The value in +options+ must evaluate to true for an option to take
      # effect.
      #
      # Examples:
      #
      #   repo.clean(d: true, f: true) # Enable the -d and -f options
      #
      #   repo.clean(d: false, x: true) # -x is enabled, -d is not
      def clean(options = {})
        strategies = [:remove_untracked]
        strategies.push(:force) if options[:f]
        strategies.push(:remove_ignored) if options[:x]

        # TODO: implement this method
      end

      # Check out the specified ref. Valid options are:
      #
      #  :b - Create a new branch at +start_point+ and set HEAD to the new
      #       branch.
      #
      #  * These options are passed to the Rugged::Repository#checkout method:
      #
      #  :progress ::
      #    A callback that will be executed for checkout progress notifications.
      #    Up to 3 parameters are passed on each execution:
      #
      #    - The path to the last updated file (or +nil+ on the very first
      #      invocation).
      #    - The number of completed checkout steps.
      #    - The number of total checkout steps to be performed.
      #
      #  :notify ::
      #    A callback that will be executed for each checkout notification
      #    types specified with +:notify_flags+. Up to 5 parameters are passed
      #    on each execution:
      #
      #    - An array containing the +:notify_flags+ that caused the callback
      #      execution.
      #    - The path of the current file.
      #    - A hash describing the baseline blob (or +nil+ if it does not
      #      exist).
      #    - A hash describing the target blob (or +nil+ if it does not exist).
      #    - A hash describing the workdir blob (or +nil+ if it does not
      #      exist).
      #
      #  :strategy ::
      #    A single symbol or an array of symbols representing the strategies
      #    to use when performing the checkout. Possible values are:
      #
      #    :none ::
      #      Perform a dry run (default).
      #
      #    :safe ::
      #      Allow safe updates that cannot overwrite uncommitted data.
      #
      #    :safe_create ::
      #      Allow safe updates plus creation of missing files.
      #
      #    :force ::
      #      Allow all updates to force working directory to look like index.
      #
      #    :allow_conflicts ::
      #      Allow checkout to make safe updates even if conflicts are found.
      #
      #    :remove_untracked ::
      #      Remove untracked files not in index (that are not ignored).
      #
      #    :remove_ignored ::
      #      Remove ignored files not in index.
      #
      #    :update_only ::
      #      Only update existing files, don't create new ones.
      #
      #    :dont_update_index ::
      #      Normally checkout updates index entries as it goes; this stops
      #      that.
      #
      #    :no_refresh ::
      #      Don't refresh index/config/etc before doing checkout.
      #
      #    :disable_pathspec_match ::
      #      Treat pathspec as simple list of exact match file paths.
      #
      #    :skip_locked_directories ::
      #      Ignore directories in use, they will be left empty.
      #
      #    :skip_unmerged ::
      #      Allow checkout to skip unmerged files (NOT IMPLEMENTED).
      #
      #    :use_ours ::
      #      For unmerged files, checkout stage 2 from index (NOT IMPLEMENTED).
      #
      #    :use_theirs ::
      #      For unmerged files, checkout stage 3 from index (NOT IMPLEMENTED).
      #
      #    :update_submodules ::
      #      Recursively checkout submodules with same options (NOT
      #      IMPLEMENTED).
      #
      #    :update_submodules_if_changed ::
      #      Recursively checkout submodules if HEAD moved in super repo (NOT
      #      IMPLEMENTED).
      #
      #  :disable_filters ::
      #    If +true+, filters like CRLF line conversion will be disabled.
      #
      #  :dir_mode ::
      #    Mode for newly created directories. Default: +0755+.
      #
      #  :file_mode ::
      #    Mode for newly created files. Default: +0755+ or +0644+.
      #
      #  :file_open_flags ::
      #    Mode for opening files. Default:
      #    <code>IO::CREAT | IO::TRUNC | IO::WRONLY</code>.
      #
      #  :notify_flags ::
      #    A single symbol or an array of symbols representing the cases in
      #    which the +:notify+ callback should be invoked. Possible values are:
      #
      #    :none ::
      #      Do not invoke the +:notify+ callback (default).
      #
      #    :conflict ::
      #      Invoke the callback for conflicting paths.
      #
      #    :dirty ::
      #      Invoke the callback for "dirty" files, i.e. those that do not need
      #      an update but no longer match the baseline.
      #
      #    :updated ::
      #      Invoke the callback for any file that was changed.
      #
      #    :untracked ::
      #      Invoke the callback for untracked files.
      #
      #    :ignored ::
      #      Invoke the callback for ignored files.
      #
      #    :all ::
      #      Invoke the callback for all these cases.
      #
      #  :paths ::
      #    A glob string or an array of glob strings specifying which paths
      #    should be taken into account for the checkout operation. +nil+ will
      #    match all files.  Default: +nil+.
      #
      #  :baseline ::
      #    A Rugged::Tree that represents the current, expected contents of the
      #    workdir.  Default: +HEAD+.
      #
      #  :target_directory ::
      #    A path to an alternative workdir directory in which the checkout
      #    should be performed.
      def checkout(ref, options = {}, start_point = "HEAD")
        if options[:b]
          rugged.branches.create(ref, start_point)
          options.delete(:b)
        end
        default_options = { strategy: [:recreate_missing, :safe] }
        rugged.checkout(ref, default_options.merge(options))
      end

      # Delete the specified branch from the repository
      def delete_branch(branch_name)
        rugged.branches.delete(branch_name)
      end

      # Create a new branch named **ref+ based on **stat_point+, HEAD by default
      #
      # Examples:
      #   create_branch("feature")
      #   create_branch("other-feature", "master")
      def create_branch(ref, start_point = "HEAD")
        rugged_ref = rugged.branches.create(ref, start_point)
842
        Gitlab::Git::Branch.new(self, rugged_ref.name, rugged_ref.target)
Robert Speicher's avatar
Robert Speicher committed
843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910
      rescue Rugged::ReferenceError => e
        raise InvalidRef.new("Branch #{ref} already exists") if e.to_s =~ /'refs\/heads\/#{ref}'/
        raise InvalidRef.new("Invalid reference #{start_point}")
      end

      # Return an array of this repository's remote names
      def remote_names
        rugged.remotes.each_name.to_a
      end

      # Delete the specified remote from this repository.
      def remote_delete(remote_name)
        rugged.remotes.delete(remote_name)
      end

      # Add a new remote to this repository.  Returns a Rugged::Remote object
      def remote_add(remote_name, url)
        rugged.remotes.create(remote_name, url)
      end

      # Update the specified remote using the values in the +options+ hash
      #
      # Example
      # repo.update_remote("origin", url: "path/to/repo")
      def remote_update(remote_name, options = {})
        # TODO: Implement other remote options
        rugged.remotes.set_url(remote_name, options[:url]) if options[:url]
      end

      # Fetch the specified remote
      def fetch(remote_name)
        rugged.remotes[remote_name].fetch
      end

      # Push +*refspecs+ to the remote identified by +remote_name+.
      def push(remote_name, *refspecs)
        rugged.remotes[remote_name].push(refspecs)
      end

      AUTOCRLF_VALUES = {
        "true" => true,
        "false" => false,
        "input" => :input
      }.freeze

      def autocrlf
        AUTOCRLF_VALUES[rugged.config['core.autocrlf']]
      end

      def autocrlf=(value)
        rugged.config['core.autocrlf'] = AUTOCRLF_VALUES.invert[value]
      end

      # Returns result like "git ls-files" , recursive and full file path
      #
      # Ex.
      #   repo.ls_files('master')
      #
      def ls_files(ref)
        actual_ref = ref || root_ref

        begin
          sha_from_ref(actual_ref)
        rescue Rugged::OdbError, Rugged::InvalidError, Rugged::ReferenceError
          # Return an empty array if the ref wasn't found
          return []
        end

911
        cmd = %W(#{Gitlab.config.git.bin_path} --git-dir=#{path} ls-tree)
Robert Speicher's avatar
Robert Speicher committed
912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968
        cmd += %w(-r)
        cmd += %w(--full-tree)
        cmd += %w(--full-name)
        cmd += %W(-- #{actual_ref})

        raw_output = IO.popen(cmd, &:read).split("\n").map do |f|
          stuff, path = f.split("\t")
          _mode, type, _sha = stuff.split(" ")
          path if type == "blob"
          # Contain only blob type
        end

        raw_output.compact
      end

      def copy_gitattributes(ref)
        begin
          commit = lookup(ref)
        rescue Rugged::ReferenceError
          raise InvalidRef.new("Ref #{ref} is invalid")
        end

        # Create the paths
        info_dir_path = File.join(path, 'info')
        info_attributes_path = File.join(info_dir_path, 'attributes')

        begin
          # Retrieve the contents of the blob
          gitattributes_content = blob_content(commit, '.gitattributes')
        rescue InvalidBlobName
          # No .gitattributes found. Should now remove any info/attributes and return
          File.delete(info_attributes_path) if File.exist?(info_attributes_path)
          return
        end

        # Create the info directory if needed
        Dir.mkdir(info_dir_path) unless File.directory?(info_dir_path)

        # Write the contents of the .gitattributes file to info/attributes
        # Use binary mode to prevent Rails from converting ASCII-8BIT to UTF-8
        File.open(info_attributes_path, "wb") do |file|
          file.write(gitattributes_content)
        end
      end

      # Checks if the blob should be diffable according to its attributes
      def diffable?(blob)
        attributes(blob.path).fetch('diff') { blob.text? }
      end

      # Returns the Git attributes for the given file path.
      #
      # See `Gitlab::Git::Attributes` for more information.
      def attributes(path)
        @attributes.attributes(path)
      end

969
      def gitaly_repository
Jacob Vosmaer's avatar
Jacob Vosmaer committed
970
        Gitlab::GitalyClient::Util.repository(@storage, @relative_path)
971 972
      end

Robert Speicher's avatar
Robert Speicher committed
973 974
      private

975 976 977 978
      def alternate_object_directories
        Gitlab::Git::Env.all.values_at(*ALLOWED_OBJECT_DIRECTORIES_VARIABLES).compact
      end

Robert Speicher's avatar
Robert Speicher committed
979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197
      # Get the content of a blob for a given commit.  If the blob is a commit
      # (for submodules) then return the blob's OID.
      def blob_content(commit, blob_name)
        blob_entry = tree_entry(commit, blob_name)

        unless blob_entry
          raise InvalidBlobName.new("Invalid blob name: #{blob_name}")
        end

        case blob_entry[:type]
        when :commit
          blob_entry[:oid]
        when :tree
          raise InvalidBlobName.new("#{blob_name} is a tree, not a blob")
        when :blob
          rugged.lookup(blob_entry[:oid]).content
        end
      end

      # Parses the contents of a .gitmodules file and returns a hash of
      # submodule information.
      def parse_gitmodules(commit, content)
        results = {}

        current = ""
        content.split("\n").each do |txt|
          if txt =~ /^\s*\[/
            current = txt.match(/(?<=").*(?=")/)[0]
            results[current] = {}
          else
            next unless results[current]
            match_data = txt.match(/(\w+)\s*=\s*(.*)/)
            next unless match_data
            target = match_data[2].chomp
            results[current][match_data[1]] = target

            if match_data[1] == "path"
              begin
                results[current]["id"] = blob_content(commit, target)
              rescue InvalidBlobName
                results.delete(current)
              end
            end
          end
        end

        results
      end

      # Returns true if +commit+ introduced changes to +path+, using commit
      # trees to make that determination.  Uses the history simplification
      # rules that `git log` uses by default, where a commit is omitted if it
      # is TREESAME to any parent.
      #
      # If the +follow+ option is true and the file specified by +path+ was
      # renamed, then the path value is set to the old path.
      def commit_touches_path?(commit, path, follow, walker)
        entry = tree_entry(commit, path)

        if commit.parents.empty?
          # This is the root commit, return true if it has +path+ in its tree
          return !entry.nil?
        end

        num_treesame = 0
        commit.parents.each do |parent|
          parent_entry = tree_entry(parent, path)

          # Only follow the first TREESAME parent for merge commits
          if num_treesame > 0
            walker.hide(parent)
            next
          end

          if entry.nil? && parent_entry.nil?
            num_treesame += 1
          elsif entry && parent_entry && entry[:oid] == parent_entry[:oid]
            num_treesame += 1
          end
        end

        case num_treesame
        when 0
          detect_rename(commit, commit.parents.first, path) if follow
          true
        else false
        end
      end

      # Find the entry for +path+ in the tree for +commit+
      def tree_entry(commit, path)
        pathname = Pathname.new(path)
        first = true
        tmp_entry = nil

        pathname.each_filename do |dir|
          if first
            tmp_entry = commit.tree[dir]
            first = false
          elsif tmp_entry.nil?
            return nil
          else
            tmp_entry = rugged.lookup(tmp_entry[:oid])
            return nil unless tmp_entry.type == :tree
            tmp_entry = tmp_entry[dir]
          end
        end

        tmp_entry
      end

      # Compare +commit+ and +parent+ for +path+.  If +path+ is a file and was
      # renamed in +commit+, then set +path+ to the old filename.
      def detect_rename(commit, parent, path)
        diff = parent.diff(commit, paths: [path], disable_pathspec_match: true)

        # If +path+ is a filename, not a directory, then we should only have
        # one delta.  We don't need to follow renames for directories.
        return nil if diff.each_delta.count > 1

        delta = diff.each_delta.first
        if delta.added?
          full_diff = parent.diff(commit)
          full_diff.find_similar!

          full_diff.each_delta do |full_delta|
            if full_delta.renamed? && path == full_delta.new_file[:path]
              # Look for the old path in ancestors
              path.replace(full_delta.old_file[:path])
            end
          end
        end
      end

      # Returns true if the index entry has the special file mode that denotes
      # a submodule.
      def submodule?(index_entry)
        index_entry[:mode] == 57344
      end

      # Return a Rugged::Index that has read from the tree at +ref_name+
      def populated_index(ref_name)
        commit = rev_parse_target(ref_name)
        index = rugged.index
        index.read_tree(commit.tree)
        index
      end

      # Return an array of BlobSnippets for lines in +file_contents+ that match
      # +query+
      def build_greps(file_contents, query, ref, filename)
        # The file_contents string is potentially huge so we make sure to loop
        # through it one line at a time. This gives Ruby the chance to GC lines
        # we are not interested in.
        #
        # We need to do a little extra work because we are not looking for just
        # the lines that matches the query, but also for the context
        # (surrounding lines). We will use Enumerable#each_cons to efficiently
        # loop through the lines while keeping surrounding lines on hand.
        #
        # First, we turn "foo\nbar\nbaz" into
        # [
        #  [nil, -3], [nil, -2], [nil, -1],
        #  ['foo', 0], ['bar', 1], ['baz', 3],
        #  [nil, 4], [nil, 5], [nil, 6]
        # ]
        lines_with_index = Enumerator.new do |yielder|
          # Yield fake 'before' lines for the first line of file_contents
          (-SEARCH_CONTEXT_LINES..-1).each do |i|
            yielder.yield [nil, i]
          end

          # Yield the actual file contents
          count = 0
          file_contents.each_line do |line|
            line.chomp!
            yielder.yield [line, count]
            count += 1
          end

          # Yield fake 'after' lines for the last line of file_contents
          (count + 1..count + SEARCH_CONTEXT_LINES).each do |i|
            yielder.yield [nil, i]
          end
        end

        greps = []

        # Loop through consecutive blocks of lines with indexes
        lines_with_index.each_cons(2 * SEARCH_CONTEXT_LINES + 1) do |line_block|
          # Get the 'middle' line and index from the block
          line, _ = line_block[SEARCH_CONTEXT_LINES]

          next unless line && line.match(/#{Regexp.escape(query)}/i)

          # Yay, 'line' contains a match!
          # Get an array with just the context lines (no indexes)
          match_with_context = line_block.map(&:first)
          # Remove 'nil' lines in case we are close to the first or last line
          match_with_context.compact!

          # Get the line number (1-indexed) of the first context line
          first_context_line_number = line_block[0][1] + 1

          greps << Gitlab::Git::BlobSnippet.new(
            ref,
            match_with_context,
            first_context_line_number,
            filename
          )
        end

        greps
      end

      # Return the Rugged patches for the diff between +from+ and +to+.
      def diff_patches(from, to, options = {}, *paths)
        options ||= {}
        break_rewrites = options[:break_rewrites]
1198
        actual_options = Gitlab::Git::Diff.filter_diff_options(options.merge(paths: paths))
Robert Speicher's avatar
Robert Speicher committed
1199 1200 1201 1202 1203

        diff = rugged.diff(from, to, actual_options)
        diff.find_similar!(break_rewrites: break_rewrites)
        diff.each_patch
      end
1204 1205

      def gitaly_ref_client
1206
        @gitaly_ref_client ||= Gitlab::GitalyClient::Ref.new(self)
1207
      end
1208

1209 1210 1211 1212
      def gitaly_commit_client
        @gitaly_commit_client ||= Gitlab::GitalyClient::Commit.new(self)
      end

1213 1214 1215 1216 1217 1218 1219 1220
      def gitaly_migrate(method, &block)
        Gitlab::GitalyClient.migrate(method, &block)
      rescue GRPC::NotFound => e
        raise NoRepository.new(e)
      rescue GRPC::BadStatus => e
        raise CommandError.new(e)
      end

1221 1222 1223 1224 1225
      # Returns the `Rugged` sorting type constant for one or more given
      # sort types. Valid keys are `:none`, `:topo`, and `:date`, or an array
      # containing more than one of them. `:date` uses a combination of date and
      # topological sorting to closer mimic git's native ordering.
      def rugged_sort_type(sort_type)
1226 1227 1228
        @rugged_sort_types ||= {
          none: Rugged::SORT_NONE,
          topo: Rugged::SORT_TOPO,
1229
          date: Rugged::SORT_DATE | Rugged::SORT_TOPO
1230 1231
        }

1232
        @rugged_sort_types.fetch(sort_type, Rugged::SORT_NONE)
1233
      end
Robert Speicher's avatar
Robert Speicher committed
1234 1235 1236
    end
  end
end