BigW Consortium Gitlab

repository.rb 29.8 KB
Newer Older
1 2
require 'securerandom'

3
class Repository
4
  include Gitlab::ShellAdapter
5
  include RepositoryMirroring
6

7
  attr_accessor :path_with_namespace, :project
8

9 10
  delegate :ref_name_for_sha, to: :raw_repository

11
  CommitError = Class.new(StandardError)
12
  CreateTreeError = Class.new(StandardError)
13

14 15 16 17 18 19 20 21
  # Methods that cache data from the Git repository.
  #
  # Each entry in this Array should have a corresponding method with the exact
  # same name. The cache key used by those methods must also match method's
  # name.
  #
  # For example, for entry `:readme` there's a method called `readme` which
  # stores its data in the `readme` cache key.
22
  CACHED_METHODS = %i(size commit_count readme contribution_guide
23 24
                      changelog license_blob license_key gitignore koding_yml
                      gitlab_ci_yml branch_names tag_names branch_count
25
                      tag_count avatar exists? empty? root_ref).freeze
26 27 28 29 30 31 32 33 34 35 36 37 38

  # Certain method caches should be refreshed when certain types of files are
  # changed. This Hash maps file types (as returned by Gitlab::FileDetector) to
  # the corresponding methods to call for refreshing caches.
  METHOD_CACHES_FOR_FILE_TYPES = {
    readme: :readme,
    changelog: :changelog,
    license: %i(license_blob license_key),
    contributing: :contribution_guide,
    gitignore: :gitignore,
    koding: :koding_yml,
    gitlab_ci: :gitlab_ci_yml,
    avatar: :avatar
39
  }.freeze
40 41 42 43 44 45 46

  # Wraps around the given method and caches its output in Redis and an instance
  # variable.
  #
  # This only works for methods that do not take any arguments.
  def self.cache_method(name, fallback: nil)
    original = :"_uncached_#{name}"
47

48
    alias_method(original, name)
49

50 51
    define_method(name) do
      cache_method_output(name, fallback: fallback) { __send__(original) }
52
    end
53
  end
54

55
  def initialize(path_with_namespace, project)
56
    @path_with_namespace = path_with_namespace
57
    @project = project
58
  end
59

60 61
  def raw_repository
    return nil unless path_with_namespace
62

63
    @raw_repository ||= initialize_raw_repository
64 65
  end

66
  # Return absolute path to repository
67
  def path_to_repo
68
    @path_to_repo ||= File.expand_path(
69
      File.join(repository_storage_path, path_with_namespace + ".git")
70
    )
71 72
  end

73 74 75 76 77 78 79 80 81 82
  #
  # Git repository can contains some hidden refs like:
  #   /refs/notes/*
  #   /refs/git-as-svn/*
  #   /refs/pulls/*
  # This refs by default not visible in project page and not cloned to client side.
  #
  # This method return true if repository contains some content visible in project page.
  #
  def has_visible_content?
83
    branch_count > 0
84 85
  end

86
  def commit(ref = 'HEAD')
87
    return nil unless exists?
88

89 90 91 92 93 94
    commit =
      if ref.is_a?(Gitlab::Git::Commit)
        ref
      else
        Gitlab::Git::Commit.find(raw_repository, ref)
      end
95

96
    commit = ::Commit.new(commit, @project) if commit
97
    commit
98
  rescue Rugged::OdbError, Rugged::TreeError
99
    nil
100 101
  end

102
  def commits(ref, path: nil, limit: nil, offset: nil, skip_merges: false, after: nil, before: nil)
103
    options = {
104 105 106 107 108
      repo: raw_repository,
      ref: ref,
      path: path,
      limit: limit,
      offset: offset,
109 110
      after: after,
      before: before,
111
      follow: Array(path).length == 1,
112
      skip_merges: skip_merges
113 114 115
    }

    commits = Gitlab::Git::Commit.where(options)
116
    commits = Commit.decorate(commits, @project) if commits.present?
117 118 119
    commits
  end

120 121
  def commits_between(from, to)
    commits = Gitlab::Git::Commit.between(raw_repository, from, to)
122
    commits = Commit.decorate(commits, @project) if commits.present?
123 124 125
    commits
  end

126
  def find_commits_by_message(query, ref = nil, path = nil, limit = 1000, offset = 0)
127 128 129 130
    unless exists? && has_visible_content? && query.present?
      return []
    end

131 132
    ref ||= root_ref

133 134 135 136
    args = %W(
      #{Gitlab.config.git.bin_path} log #{ref} --pretty=%H --skip #{offset}
      --max-count #{limit} --grep=#{query} --regexp-ignore-case
    )
137
    args = args.concat(%W(-- #{path})) if path.present?
138

139 140
    git_log_results = Gitlab::Popen.popen(args, path_to_repo).first.lines
    git_log_results.map { |c| commit(c.chomp) }.compact
141 142
  end

143 144 145 146 147 148 149
  def find_branch(name, fresh_repo: true)
    # Since the Repository object may have in-memory index changes, invalidating the memoized Repository object may
    # cause unintended side effects. Because finding a branch is a read-only operation, we can safely instantiate
    # a new repo here to ensure a consistent state to avoid a libgit2 bug where concurrent access (e.g. via git gc)
    # may cause the branch to "disappear" erroneously or have the wrong SHA.
    #
    # See: https://github.com/libgit2/libgit2/issues/1534 and https://gitlab.com/gitlab-org/gitlab-ce/issues/15392
150
    raw_repo = fresh_repo ? initialize_raw_repository : raw_repository
151 152

    raw_repo.find_branch(name)
153 154 155
  end

  def find_tag(name)
156
    tags.find { |tag| tag.name == name }
157 158
  end

159 160
  def add_branch(user, branch_name, ref)
    newrev = commit(ref).try(:sha)
161

162
    return false unless newrev
163

164
    GitOperationService.new(user, self).add_branch(branch_name, newrev)
165

166
    after_create_branch
167
    find_branch(branch_name)
168 169
  end

170
  def add_tag(user, tag_name, target, message = nil)
171
    newrev = commit(target).try(:id)
172 173
    options = { message: message, tagger: user_to_committer(user) } if message

174 175 176
    return false unless newrev

    GitOperationService.new(user, self).add_tag(tag_name, newrev, options)
177

178
    find_tag(tag_name)
179 180
  end

181
  def rm_branch(user, branch_name)
182
    before_remove_branch
183 184
    branch = find_branch(branch_name)

185
    GitOperationService.new(user, self).rm_branch(branch)
186

187
    after_remove_branch
188
    true
189 190
  end

191
  def rm_tag(user, tag_name)
192
    before_remove_tag
193
    tag = find_tag(tag_name)
194

195 196 197 198
    GitOperationService.new(user, self).rm_tag(tag)

    after_remove_tag
    true
199 200
  end

201 202 203 204
  def ref_names
    branch_names + tag_names
  end

205 206 207 208
  def branch_exists?(branch_name)
    branch_names.include?(branch_name)
  end

209 210
  def ref_exists?(ref)
    rugged.references.exist?(ref)
211 212
  rescue Rugged::ReferenceError
    false
213 214
  end

215 216 217 218
  # Makes sure a commit is kept around when Git garbage collection runs.
  # Git GC will delete commits from the repository that are no longer in any
  # branches or tags, but we want to keep some of these commits around, for
  # example if they have comments or CI builds.
219 220 221 222 223
  def keep_around(sha)
    return unless sha && commit(sha)

    return if kept_around?(sha)

224 225 226 227 228
    # This will still fail if the file is corrupted (e.g. 0 bytes)
    begin
      rugged.references.create(keep_around_ref_name(sha), sha, force: true)
    rescue Rugged::ReferenceError => ex
      Rails.logger.error "Unable to create keep-around reference for repository #{path}: #{ex}"
229 230 231
    rescue Rugged::OSError => ex
      raise unless ex.message =~ /Failed to create locked file/ && ex.message =~ /File exists/
      Rails.logger.error "Unable to create keep-around reference for repository #{path}: #{ex}"
232
    end
233 234 235
  end

  def kept_around?(sha)
236
    ref_exists?(keep_around_ref_name(sha))
237 238
  end

239
  def diverging_commit_counts(branch)
240
    root_ref_hash = raw_repository.rev_parse_target(root_ref).oid
241
    cache.fetch(:"diverging_commit_counts_#{branch.name}") do
242 243
      # Rugged seems to throw a `ReferenceError` when given branch_names rather
      # than SHA-1 hashes
244 245
      number_commits_behind = raw_repository.
        count_commits_between(branch.dereferenced_target.sha, root_ref_hash)
246

247 248
      number_commits_ahead = raw_repository.
        count_commits_between(root_ref_hash, branch.dereferenced_target.sha)
249

250 251 252
      { behind: number_commits_behind, ahead: number_commits_ahead }
    end
  end
253

254 255 256
  def expire_tags_cache
    expire_method_caches(%i(tag_names tag_count))
    @tags = nil
257
  end
258

259 260 261
  def expire_branches_cache
    expire_method_caches(%i(branch_names branch_count))
    @local_branches = nil
262 263
  end

264 265
  def expire_statistics_caches
    expire_method_caches(%i(size commit_count))
266 267
  end

268 269
  def expire_all_method_caches
    expire_method_caches(CACHED_METHODS)
270 271
  end

272 273 274 275 276 277 278 279 280
  # Expires the caches of a specific set of methods
  def expire_method_caches(methods)
    methods.each do |key|
      cache.expire(key)

      ivar = cache_instance_variable_name(key)

      remove_instance_variable(ivar) if instance_variable_defined?(ivar)
    end
281 282
  end

283 284 285 286 287 288 289 290 291 292 293 294 295 296 297
  def expire_avatar_cache
    expire_method_caches(%i(avatar))
  end

  # Refreshes the method caches of this repository.
  #
  # types - An Array of file types (e.g. `:readme`) used to refresh extra
  #         caches.
  def refresh_method_caches(types)
    to_refresh = []

    types.each do |type|
      methods = METHOD_CACHES_FOR_FILE_TYPES[type.to_sym]

      to_refresh.concat(Array(methods)) if methods
298
    end
299

300
    expire_method_caches(to_refresh)
301

302
    to_refresh.each { |method| send(method) }
303
  end
304

305 306 307 308 309 310 311
  def expire_branch_cache(branch_name = nil)
    # When we push to the root branch we have to flush the cache for all other
    # branches as their statistics are based on the commits relative to the
    # root branch.
    if !branch_name || branch_name == root_ref
      branches.each do |branch|
        cache.expire(:"diverging_commit_counts_#{branch.name}")
312
        cache.expire(:"commit_count_#{branch.name}")
313 314 315 316 317
      end
    # In case a commit is pushed to a non-root branch we only have to flush the
    # cache for said branch.
    else
      cache.expire(:"diverging_commit_counts_#{branch_name}")
318
      cache.expire(:"commit_count_#{branch_name}")
319
    end
Dmitriy Zaporozhets committed
320 321
  end

322
  def expire_root_ref_cache
323
    expire_method_caches(%i(root_ref))
324 325
  end

326 327
  # Expires the cache(s) used to determine if a repository is empty or not.
  def expire_emptiness_caches
328
    return unless empty?
329

330
    expire_method_caches(%i(empty?))
331 332
  end

333 334 335 336
  def lookup_cache
    @lookup_cache ||= {}
  end

337
  def expire_exists_cache
338
    expire_method_caches(%i(exists?))
339 340
  end

341 342 343 344 345 346 347
  # expire cache that doesn't depend on repository data (when expiring)
  def expire_content_cache
    expire_tags_cache
    expire_branches_cache
    expire_root_ref_cache
    expire_emptiness_caches
    expire_exists_cache
348
    expire_statistics_caches
349 350
  end

351 352 353
  # Runs code after a repository has been created.
  def after_create
    expire_exists_cache
354 355
    expire_root_ref_cache
    expire_emptiness_caches
356 357

    repository_event(:create_repository)
358 359
  end

360 361
  # Runs code just before a repository is deleted.
  def before_delete
362
    expire_exists_cache
363 364
    expire_all_method_caches
    expire_branch_cache if exists?
365
    expire_content_cache
366 367

    repository_event(:remove_repository)
368 369 370 371 372 373 374
  end

  # Runs code just before the HEAD of a repository is changed.
  def before_change_head
    # Cached divergent commit counts are based on repository head
    expire_branch_cache
    expire_root_ref_cache
375 376

    repository_event(:change_default_branch)
377 378
  end

379 380
  # Runs code before pushing (= creating or removing) a tag.
  def before_push_tag
381 382
    expire_statistics_caches
    expire_emptiness_caches
383
    expire_tags_cache
384 385

    repository_event(:push_tag)
386 387 388 389 390
  end

  # Runs code before removing a tag.
  def before_remove_tag
    expire_tags_cache
391
    expire_statistics_caches
392 393

    repository_event(:remove_tag)
394 395
  end

396 397 398 399 400
  # Runs code after removing a tag.
  def after_remove_tag
    expire_tags_cache
  end

401 402 403 404 405
  # Runs code after the HEAD of a repository is changed.
  def after_change_head
    expire_method_caches(METHOD_CACHES_FOR_FILE_TYPES.keys)
  end

406 407
  # Runs code after a repository has been forked/imported.
  def after_import
408
    expire_content_cache
409 410 411
  end

  # Runs code after a new commit has been pushed.
412 413 414
  def after_push_commit(branch_name)
    expire_statistics_caches
    expire_branch_cache(branch_name)
415 416

    repository_event(:push_commit, branch: branch_name)
417 418 419 420
  end

  # Runs code after a new branch has been created.
  def after_create_branch
421
    expire_branches_cache
422 423

    repository_event(:push_branch)
424 425
  end

426 427 428
  # Runs code before removing an existing branch.
  def before_remove_branch
    expire_branches_cache
429 430

    repository_event(:remove_branch)
431 432
  end

433 434
  # Runs code after an existing branch has been removed.
  def after_remove_branch
435
    expire_branches_cache
436 437
  end

438
  def method_missing(m, *args, &block)
439 440 441 442 443 444
    if m == :lookup && !block_given?
      lookup_cache[m] ||= {}
      lookup_cache[m][args.join(":")] ||= raw_repository.send(m, *args, &block)
    else
      raw_repository.send(m, *args, &block)
    end
445 446
  end

447 448
  def respond_to_missing?(method, include_private = false)
    raw_repository.respond_to?(method, include_private) || super
449
  end
450 451

  def blob_at(sha, path)
452
    unless Gitlab::Git.blank_ref?(sha)
453
      Blob.decorate(Gitlab::Git::Blob.find(self, sha, path))
454
    end
Douwe Maan committed
455 456
  rescue Gitlab::Git::Repository::NoRepository
    nil
457
  end
458

459 460 461 462
  def blob_by_oid(oid)
    Gitlab::Git::Blob.raw(self, oid)
  end

463 464 465 466 467 468 469 470 471 472 473 474 475 476 477
  def root_ref
    if raw_repository
      raw_repository.root_ref
    else
      # When the repo does not exist we raise this error so no data is cached.
      raise Rugged::ReferenceError
    end
  end
  cache_method :root_ref

  def exists?
    refs_directory_exists?
  end
  cache_method :exists?

Douwe Maan committed
478
  delegate :empty?, to: :raw_repository
479 480 481 482 483 484 485 486 487 488 489 490 491
  cache_method :empty?

  # The size of this repository in megabytes.
  def size
    exists? ? raw_repository.size : 0.0
  end
  cache_method :size, fallback: 0.0

  def commit_count
    root_ref ? raw_repository.commit_count(root_ref) : 0
  end
  cache_method :commit_count, fallback: 0

492
  def commit_count_for_ref(ref)
493
    return 0 unless exists?
494

495 496 497 498
    begin
      cache.fetch(:"commit_count_#{ref}") { raw_repository.commit_count(ref) }
    rescue Rugged::ReferenceError
      0
499 500 501
    end
  end

502
  delegate :branch_names, to: :raw_repository
503 504
  cache_method :branch_names, fallback: []

Douwe Maan committed
505
  delegate :tag_names, to: :raw_repository
506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524
  cache_method :tag_names, fallback: []

  def branch_count
    branches.size
  end
  cache_method :branch_count, fallback: 0

  def tag_count
    raw_repository.rugged.tags.count
  end
  cache_method :tag_count, fallback: 0

  def avatar
    if tree = file_on_head(:avatar)
      tree.path
    end
  end
  cache_method :avatar

525
  def readme
526 527 528
    if head = tree(:head)
      head.readme
    end
529
  end
530
  cache_method :readme
531

532
  def contribution_guide
533
    file_on_head(:contributing)
534
  end
535
  cache_method :contribution_guide
536 537

  def changelog
538
    file_on_head(:changelog)
539
  end
540
  cache_method :changelog
541

542
  def license_blob
543
    file_on_head(:license)
544
  end
545
  cache_method :license_blob
546

547
  def license_key
548
    return unless exists?
549

550
    Licensee.license(path).try(:key)
551
  end
552
  cache_method :license_key
553

554
  def gitignore
555
    file_on_head(:gitignore)
556
  end
557
  cache_method :gitignore
558

559
  def koding_yml
560
    file_on_head(:koding)
561
  end
562
  cache_method :koding_yml
563

564
  def gitlab_ci_yml
565
    file_on_head(:gitlab_ci)
566
  end
567
  cache_method :gitlab_ci_yml
568

569
  def head_commit
570 571 572 573
    @head_commit ||= commit(self.root_ref)
  end

  def head_tree
574 575 576
    if head_commit
      @head_tree ||= Tree.new(self, head_commit.sha, nil)
    end
577 578
  end

579
  def tree(sha = :head, path = nil, recursive: false)
580
    if sha == :head
581 582
      return unless head_commit

583 584 585 586 587
      if path.nil?
        return head_tree
      else
        sha = head_commit.sha
      end
588 589
    end

590
    Tree.new(self, sha, path, recursive: recursive)
591
  end
592 593

  def blob_at_branch(branch_name, path)
594
    last_commit = commit(branch_name)
595

596 597 598 599 600
    if last_commit
      blob_at(last_commit.sha, path)
    else
      nil
    end
601
  end
602 603 604 605 606 607 608 609

  # Returns url for submodule
  #
  # Ex.
  #   @repository.submodule_url_for('master', 'rack')
  #   # => git@localhost:rack.git
  #
  def submodule_url_for(ref, path)
610
    if submodules(ref).any?
611 612 613 614 615 616 617
      submodule = submodules(ref)[path]

      if submodule
        submodule['url']
      end
    end
  end
618 619

  def last_commit_for_path(sha, path)
Hiroyuki Sato committed
620
    sha = last_commit_id_for_path(sha, path)
621
    commit(sha)
622
  end
623

Hiroyuki Sato committed
624 625
  def last_commit_id_for_path(sha, path)
    key = path.blank? ? "last_commit_id_for_path:#{sha}" : "last_commit_id_for_path:#{sha}:#{Digest::SHA1.hexdigest(path)}"
Hiroyuki Sato committed
626

627
    cache.fetch(key) do
Hiroyuki Sato committed
628 629
      args = %W(#{Gitlab.config.git.bin_path} rev-list --max-count=1 #{sha} -- #{path})
      Gitlab::Popen.popen(args, path_to_repo).first.strip
630 631 632
    end
  end

633
  def next_branch(name, opts = {})
634 635 636
    branch_ids = self.branch_names.map do |n|
      next 1 if n == name
      result = n.match(/\A#{name}-([0-9]+)\z/)
637 638 639
      result[1].to_i if result
    end.compact

640
    highest_branch_id = branch_ids.max || 0
641

642 643 644
    return name if opts[:mild] && 0 == highest_branch_id

    "#{name}-#{highest_branch_id + 1}"
645 646
  end

647
  # Remove archives older than 2 hours
648 649
  def branches_sorted_by(value)
    case value
650 651
    when 'name'
      branches.sort_by(&:name)
652
    when 'updated_desc'
653
      branches.sort do |a, b|
654
        commit(b.dereferenced_target).committed_date <=> commit(a.dereferenced_target).committed_date
655
      end
656
    when 'updated_asc'
657
      branches.sort do |a, b|
658
        commit(a.dereferenced_target).committed_date <=> commit(b.dereferenced_target).committed_date
659 660 661 662 663
      end
    else
      branches
    end
  end
664

665 666 667
  def tags_sorted_by(value)
    case value
    when 'name'
668
      VersionSorter.rsort(tags) { |tag| tag.name }
669 670 671 672 673 674 675 676 677
    when 'updated_desc'
      tags_sorted_by_committed_date.reverse
    when 'updated_asc'
      tags_sorted_by_committed_date
    else
      tags
    end
  end

678
  def contributors
679
    commits = self.commits(nil, limit: 2000, offset: 0, skip_merges: true)
680

681
    commits.group_by(&:author_email).map do |email, commits|
682 683
      contributor = Gitlab::Contributor.new
      contributor.email = email
684

685
      commits.each do |commit|
686
        if contributor.name.blank?
687
          contributor.name = commit.author_name
688 689
        end

690
        contributor.commits += 1
691 692
      end

693 694
      contributor
    end
695
  end
696

697 698
  def refs_contains_sha(ref_type, sha)
    args = %W(#{Gitlab.config.git.bin_path} #{ref_type} --contains #{sha})
699 700 701 702 703 704 705 706 707 708 709 710 711 712
    names = Gitlab::Popen.popen(args, path_to_repo).first

    if names.respond_to?(:split)
      names = names.split("\n").map(&:strip)

      names.each do |name|
        name.slice! '* '
      end

      names
    else
      []
    end
  end
713

714 715 716
  def branch_names_contains(sha)
    refs_contains_sha('branch', sha)
  end
717

718 719
  def tag_names_contains(sha)
    refs_contains_sha('tag', sha)
720
  end
721

722
  def local_branches
723
    @local_branches ||= raw_repository.local_branches
724 725
  end

726 727
  alias_method :branches, :local_branches

728 729 730 731
  def tags
    @tags ||= raw_repository.tags
  end

Douwe Maan committed
732 733 734
  def create_dir(user, path, **options)
    options[:user] = user
    options[:actions] = [{ action: :create_dir, file_path: path }]
735

Douwe Maan committed
736
    multi_action(**options)
Stan Hu committed
737
  end
738

Douwe Maan committed
739 740 741
  def create_file(user, path, content, **options)
    options[:user] = user
    options[:actions] = [{ action: :create, file_path: path, content: content }]
742

Douwe Maan committed
743
    multi_action(**options)
744 745
  end

Douwe Maan committed
746 747 748 749 750 751 752 753
  def update_file(user, path, content, **options)
    previous_path = options.delete(:previous_path)
    action = previous_path && previous_path != path ? :move : :update

    options[:user] = user
    options[:actions] = [{ action: action, file_path: path, previous_path: previous_path, content: content }]

    multi_action(**options)
754 755
  end

Douwe Maan committed
756 757 758
  def delete_file(user, path, **options)
    options[:user] = user
    options[:actions] = [{ action: :delete, file_path: path }]
759

Douwe Maan committed
760
    multi_action(**options)
761 762
  end

763
  # rubocop:disable Metrics/ParameterLists
764
  def multi_action(
765
    user:, branch_name:, message:, actions:,
766
    author_email: nil, author_name: nil,
767
    start_branch_name: nil, start_project: project)
Douwe Maan committed
768

769
    GitOperationService.new(user, self).with_branch(
770
      branch_name,
771 772
      start_branch_name: start_branch_name,
      start_project: start_project) do |start_commit|
773

774
      index = Gitlab::Git::Index.new(raw_repository)
Marc Siegfriedt committed
775

776 777 778 779 780 781 782 783
      if start_commit
        index.read_tree(start_commit.raw_commit.tree)
        parents = [start_commit.sha]
      else
        parents = []
      end

      actions.each do |options|
Douwe Maan committed
784
        index.public_send(options.delete(:action), options)
Marc Siegfriedt committed
785 786 787
      end

      options = {
788
        tree: index.write_tree,
Marc Siegfriedt committed
789 790 791 792 793 794 795 796
        message: message,
        parents: parents
      }
      options.merge!(get_committer_and_author(user, email: author_email, name: author_name))

      Rugged::Commit.create(rugged, options)
    end
  end
797
  # rubocop:enable Metrics/ParameterLists
Marc Siegfriedt committed
798

799 800
  def get_committer_and_author(user, email: nil, name: nil)
    committer = user_to_committer(user)
801
    author = Gitlab::Git.committer_hash(email: email, name: name) || committer
802

803
    {
804 805
      author: author,
      committer: committer
806 807 808
    }
  end

809
  def user_to_committer(user)
810
    Gitlab::Git.committer_hash(email: user.email, name: user.name)
811 812
  end

813 814 815 816 817 818 819 820 821 822 823
  def can_be_merged?(source_sha, target_branch)
    our_commit = rugged.branches[target_branch].target
    their_commit = rugged.lookup(source_sha)

    if our_commit && their_commit
      !rugged.merge_commits(our_commit, their_commit).conflicts?
    else
      false
    end
  end

824
  def merge(user, source, merge_request, options = {})
825
    GitOperationService.new(user, self).with_branch(
826 827
      merge_request.target_branch) do |start_commit|
      our_commit = start_commit.sha
828
      their_commit = source
829

830 831
      raise 'Invalid merge target' unless our_commit
      raise 'Invalid merge source' unless their_commit
832

833 834
      merge_index = rugged.merge_commits(our_commit, their_commit)
      break if merge_index.conflicts?
835

836 837 838 839
      actual_options = options.merge(
        parents: [our_commit, their_commit],
        tree: merge_index.write_tree(rugged),
      )
840

841 842 843
      commit_id = Rugged::Commit.create(rugged, actual_options)
      merge_request.update(in_progress_merge_commit_sha: commit_id)
      commit_id
844
    end
845 846
  rescue Repository::CommitError # when merge_index.conflicts?
    false
847 848
  end

849
  def revert(
850
    user, commit, branch_name,
851
    start_branch_name: nil, start_project: project)
852
    GitOperationService.new(user, self).with_branch(
853
      branch_name,
854 855
      start_branch_name: start_branch_name,
      start_project: start_project) do |start_commit|
856

857 858
      revert_tree_id = check_revert_content(commit, start_commit.sha)
      unless revert_tree_id
859
        raise Repository::CreateTreeError.new('Failed to revert commit')
860 861
      end

862
      committer = user_to_committer(user)
863

Lin Jen-Shin committed
864
      Rugged::Commit.create(rugged,
865
        message: commit.revert_message(user),
866 867
        author: committer,
        committer: committer,
868
        tree: revert_tree_id,
869
        parents: [start_commit.sha])
870
    end
871 872
  end

873
  def cherry_pick(
874
    user, commit, branch_name,
875
    start_branch_name: nil, start_project: project)
876
    GitOperationService.new(user, self).with_branch(
877
      branch_name,
878 879
      start_branch_name: start_branch_name,
      start_project: start_project) do |start_commit|
880

881 882
      cherry_pick_tree_id = check_cherry_pick_content(commit, start_commit.sha)
      unless cherry_pick_tree_id
883
        raise Repository::CreateTreeError.new('Failed to cherry-pick commit')
884 885
      end

886
      committer = user_to_committer(user)
887

Lin Jen-Shin committed
888
      Rugged::Commit.create(rugged,
889 890 891 892 893 894 895 896
        message: commit.message,
        author: {
          email: commit.author_email,
          name: commit.author_name,
          time: commit.authored_date
        },
        committer: committer,
        tree: cherry_pick_tree_id,
897
        parents: [start_commit.sha])
898 899 900
    end
  end

901 902
  def resolve_conflicts(user, branch_name, params)
    GitOperationService.new(user, self).with_branch(branch_name) do
903 904 905 906 907 908
      committer = user_to_committer(user)

      Rugged::Commit.create(rugged, params.merge(author: committer, committer: committer))
    end
  end

909 910
  def check_revert_content(target_commit, source_sha)
    args = [target_commit.sha, source_sha]
911
    args << { mainline: 1 } if target_commit.merge_commit?
912 913 914 915 916 917 918 919 920 921

    revert_index = rugged.revert_commit(*args)
    return false if revert_index.conflicts?

    tree_id = revert_index.write_tree(rugged)
    return false unless diff_exists?(source_sha, tree_id)

    tree_id
  end

922 923
  def check_cherry_pick_content(target_commit, source_sha)
    args = [target_commit.sha, source_sha]
924
    args << 1 if target_commit.merge_commit?
925 926 927 928 929 930 931 932 933 934

    cherry_pick_index = rugged.cherrypick_commit(*args)
    return false if cherry_pick_index.conflicts?

    tree_id = cherry_pick_index.write_tree(rugged)
    return false unless diff_exists?(source_sha, tree_id)

    tree_id
  end

935 936
  def diff_exists?(sha1, sha2)
    rugged.diff(sha1, sha2).size > 0
937 938
  end

939 940 941 942 943
  def merged_to_root_ref?(branch_name)
    branch_commit = commit(branch_name)
    root_ref_commit = commit(root_ref)

    if branch_commit
944 945
      same_head = branch_commit.id == root_ref_commit.id
      !same_head && is_ancestor?(branch_commit.id, root_ref_commit.id)
946 947 948 949 950
    else
      nil
    end
  end

951
  def merge_base(first_commit_id, second_commit_id)
952 953
    first_commit_id = commit(first_commit_id).try(:id) || first_commit_id
    second_commit_id = commit(second_commit_id).try(:id) || second_commit_id
954
    rugged.merge_base(first_commit_id, second_commit_id)
Douwe Maan committed
955 956
  rescue Rugged::ReferenceError
    nil
957 958
  end

959
  def is_ancestor?(ancestor_id, descendant_id)
960 961 962 963 964 965 966 967 968
    # NOTE: This feature is intentionally disabled until
    # https://gitlab.com/gitlab-org/gitlab-ce/issues/30586 is resolved
    # Gitlab::GitalyClient.migrate(:is_ancestor) do |is_enabled|
    #   if is_enabled
    #     raw_repository.is_ancestor?(ancestor_id, descendant_id)
    #   else
    merge_base_commit(ancestor_id, descendant_id) == ancestor_id
    #   end
    # end
969 970
  end

971 972 973 974 975 976
  def empty_repo?
    !exists? || !has_visible_content?
  end

  def search_files_by_content(query, ref)
    return [] if empty_repo? || query.blank?
Valery Sizov committed
977

978
    offset = 2
979
    args = %W(#{Gitlab.config.git.bin_path} grep -i -I -n --before-context #{offset} --after-context #{offset} -E -e #{Regexp.escape(query)} #{ref || root_ref})
980 981 982
    Gitlab::Popen.popen(args, path_to_repo).first.scrub.split(/^--$/)
  end

983 984 985 986 987 988 989
  def search_files_by_name(query, ref)
    return [] if empty_repo? || query.blank?

    args = %W(#{Gitlab.config.git.bin_path} ls-tree --full-tree -r #{ref || root_ref} --name-status | #{Regexp.escape(query)})
    Gitlab::Popen.popen(args, path_to_repo).first.lines.map(&:strip)
  end

990
  def with_repo_branch_commit(start_repository, start_branch_name)
991 992
    return yield(nil) if start_repository.empty_repo?

993
    branch_name_or_sha =
994 995
      if start_repository == self
        start_branch_name
996 997
      else
        tmp_ref = "refs/tmp/#{SecureRandom.hex}/head"
998

999
        fetch_ref(
1000 1001
          start_repository.path_to_repo,
          "#{Gitlab::Git::BRANCH_REF_PREFIX}#{start_branch_name}",
1002 1003 1004
          tmp_ref
        )

1005
        start_repository.commit(start_branch_name).sha
1006
      end
1007

1008
    yield(commit(branch_name_or_sha))
1009 1010

  ensure
1011
    rugged.references.delete(tmp_ref) if tmp_ref
1012 1013
  end

1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028
  def add_remote(name, url)
    raw_repository.remote_add(name, url)
  rescue Rugged::ConfigError
    raw_repository.remote_update(name, url: url)
  end

  def remove_remote(name)
    raw_repository.remote_delete(name)
    true
  rescue Rugged::ConfigError
    false
  end

  def fetch_remote(remote, forced: false, no_tags: false)
    gitlab_shell.fetch_remote(repository_storage_path, path_with_namespace, remote, forced: forced, no_tags: no_tags)
1029 1030
  end

1031
  def fetch_ref(source_path, source_ref, target_ref)
1032
    args = %W(#{Gitlab.config.git.bin_path} fetch --no-tags -f #{source_path} #{source_ref}:#{target_ref})
1033 1034 1035
    Gitlab::Popen.popen(args, path_to_repo)
  end

1036 1037 1038 1039
  def create_ref(ref, ref_path)
    fetch_ref(path_to_repo, ref, ref_path)
  end

1040 1041 1042 1043 1044
  def ls_files(ref)
    actual_ref = ref || root_ref
    raw_repository.ls_files(actual_ref)
  end

1045 1046 1047 1048
  def gitattribute(path, name)
    raw_repository.attributes(path)[name]
  end

1049 1050 1051 1052 1053 1054 1055 1056 1057 1058
  def copy_gitattributes(ref)
    actual_ref = ref || root_ref
    begin
      raw_repository.copy_gitattributes(actual_ref)
      true
    rescue Gitlab::Git::Repository::InvalidRef
      false
    end
  end

1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071
  # Caches the supplied block both in a cache and in an instance variable.
  #
  # The cache key and instance variable are named the same way as the value of
  # the `key` argument.
  #
  # This method will return `nil` if the corresponding instance variable is also
  # set to `nil`. This ensures we don't keep yielding the block when it returns
  # `nil`.
  #
  # key - The name of the key to cache the data in.
  # fallback - A value to fall back to in the event of a Git error.
  def cache_method_output(key, fallback: nil, &block)
    ivar = cache_instance_variable_name(key)
1072

1073 1074 1075 1076 1077 1078 1079 1080 1081
    if instance_variable_defined?(ivar)
      instance_variable_get(ivar)
    else
      begin
        instance_variable_set(ivar, cache.fetch(key, &block))
      rescue Rugged::ReferenceError, Gitlab::Git::Repository::NoRepository
        # if e.g. HEAD or the entire repository doesn't exist we want to
        # gracefully handle this and not cache anything.
        fallback
1082 1083 1084 1085
      end
    end
  end

1086 1087 1088
  def cache_instance_variable_name(key)
    :"@#{key.to_s.tr('?!', '')}"
  end
1089

1090 1091 1092 1093
  def file_on_head(type)
    if head = tree(:head)
      head.blobs.find do |file|
        Gitlab::FileDetector.type_of(file.name) == type
1094 1095
      end
    end
1096
  end
1097

Douwe Maan committed
1098 1099 1100 1101 1102 1103 1104 1105 1106
  def route_map_for(sha)
    blob_data_at(sha, '.gitlab/route-map.yml')
  end

  def gitlab_ci_yml_for(sha)
    blob_data_at(sha, '.gitlab-ci.yml')
  end

  private
1107

Douwe Maan committed
1108 1109
  def blob_data_at(sha, path)
    blob = blob_at(sha, path)
1110
    return unless blob
1111

1112 1113 1114 1115
    blob.load_all_data!(self)
    blob.data
  end

1116 1117 1118 1119
  def refs_directory_exists?
    return false unless path_with_namespace

    File.exist?(File.join(path_to_repo, 'refs'))
1120
  end
1121

1122
  def cache
1123
    @cache ||= RepositoryCache.new(path_with_namespace, @project.id)
1124
  end
1125 1126

  def tags_sorted_by_committed_date
1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138
    tags.sort_by do |tag|
      # Annotated tags can point to any object (e.g. a blob), but generally
      # tags point to a commit. If we don't have a commit, then just default
      # to putting the tag at the end of the list.
      target = tag.dereferenced_target

      if target
        target.committed_date
      else
        Time.now
      end
    end
1139
  end
1140 1141 1142 1143

  def keep_around_ref_name(sha)
    "refs/keep-around/#{sha}"
  end
1144 1145 1146 1147

  def repository_event(event, tags = {})
    Gitlab::Metrics.add_event(event, { path: path_with_namespace }.merge(tags))
  end
1148 1149 1150 1151

  def repository_storage_path
    @project.repository_storage_path
  end
1152

1153
  delegate :gitaly_channel, :gitaly_repository, to: :raw_repository
1154

1155
  def initialize_raw_repository
1156 1157
    Gitlab::Git::Repository.new(project.repository_storage, path_with_namespace + '.git')
  end
1158
end