You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

git.go 4.6KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172
  1. // Copyright 2019 The Gitea Authors. All rights reserved.
  2. // Use of this source code is governed by a MIT-style
  3. // license that can be found in the LICENSE file.
  4. package code
  5. import (
  6. "strconv"
  7. "strings"
  8. "code.gitea.io/gitea/models"
  9. "code.gitea.io/gitea/modules/git"
  10. "code.gitea.io/gitea/modules/log"
  11. "code.gitea.io/gitea/modules/setting"
  12. )
  13. type fileUpdate struct {
  14. Filename string
  15. BlobSha string
  16. }
  17. // repoChanges changes (file additions/updates/removals) to a repo
  18. type repoChanges struct {
  19. Updates []fileUpdate
  20. RemovedFilenames []string
  21. }
  22. func getDefaultBranchSha(repo *models.Repository) (string, error) {
  23. stdout, err := git.NewCommand("show-ref", "-s", git.BranchPrefix+repo.DefaultBranch).RunInDir(repo.RepoPath())
  24. if err != nil {
  25. return "", err
  26. }
  27. return strings.TrimSpace(stdout), nil
  28. }
  29. // getRepoChanges returns changes to repo since last indexer update
  30. func getRepoChanges(repo *models.Repository, revision string) (*repoChanges, error) {
  31. if err := repo.GetIndexerStatus(); err != nil {
  32. return nil, err
  33. }
  34. if len(repo.IndexerStatus.CommitSha) == 0 {
  35. return genesisChanges(repo, revision)
  36. }
  37. return nonGenesisChanges(repo, revision)
  38. }
  39. func isIndexable(entry *git.TreeEntry) bool {
  40. if !entry.IsRegular() && !entry.IsExecutable() {
  41. return false
  42. }
  43. name := strings.ToLower(entry.Name())
  44. for _, g := range setting.Indexer.ExcludePatterns {
  45. if g.Match(name) {
  46. return false
  47. }
  48. }
  49. for _, g := range setting.Indexer.IncludePatterns {
  50. if g.Match(name) {
  51. return true
  52. }
  53. }
  54. return len(setting.Indexer.IncludePatterns) == 0
  55. }
  56. // parseGitLsTreeOutput parses the output of a `git ls-tree -r --full-name` command
  57. func parseGitLsTreeOutput(stdout []byte) ([]fileUpdate, error) {
  58. entries, err := git.ParseTreeEntries(stdout)
  59. if err != nil {
  60. return nil, err
  61. }
  62. var idxCount = 0
  63. updates := make([]fileUpdate, len(entries))
  64. for _, entry := range entries {
  65. if isIndexable(entry) {
  66. updates[idxCount] = fileUpdate{
  67. Filename: entry.Name(),
  68. BlobSha: entry.ID.String(),
  69. }
  70. idxCount++
  71. }
  72. }
  73. return updates[:idxCount], nil
  74. }
  75. // genesisChanges get changes to add repo to the indexer for the first time
  76. func genesisChanges(repo *models.Repository, revision string) (*repoChanges, error) {
  77. var changes repoChanges
  78. stdout, err := git.NewCommand("ls-tree", "--full-tree", "-r", revision).
  79. RunInDirBytes(repo.RepoPath())
  80. if err != nil {
  81. return nil, err
  82. }
  83. changes.Updates, err = parseGitLsTreeOutput(stdout)
  84. return &changes, err
  85. }
  86. // nonGenesisChanges get changes since the previous indexer update
  87. func nonGenesisChanges(repo *models.Repository, revision string) (*repoChanges, error) {
  88. diffCmd := git.NewCommand("diff", "--name-status",
  89. repo.IndexerStatus.CommitSha, revision)
  90. stdout, err := diffCmd.RunInDir(repo.RepoPath())
  91. if err != nil {
  92. // previous commit sha may have been removed by a force push, so
  93. // try rebuilding from scratch
  94. log.Warn("git diff: %v", err)
  95. if err = indexer.Delete(repo.ID); err != nil {
  96. return nil, err
  97. }
  98. return genesisChanges(repo, revision)
  99. }
  100. var changes repoChanges
  101. updatedFilenames := make([]string, 0, 10)
  102. for _, line := range strings.Split(stdout, "\n") {
  103. line = strings.TrimSpace(line)
  104. if len(line) == 0 {
  105. continue
  106. }
  107. fields := strings.Split(line, "\t")
  108. if len(fields) < 2 {
  109. log.Warn("Unparseable output for diff --name-status: `%s`)", line)
  110. continue
  111. }
  112. filename := fields[1]
  113. if len(filename) == 0 {
  114. continue
  115. } else if filename[0] == '"' {
  116. filename, err = strconv.Unquote(filename)
  117. if err != nil {
  118. return nil, err
  119. }
  120. }
  121. switch status := fields[0][0]; status {
  122. case 'M', 'A':
  123. updatedFilenames = append(updatedFilenames, filename)
  124. case 'D':
  125. changes.RemovedFilenames = append(changes.RemovedFilenames, filename)
  126. case 'R', 'C':
  127. if len(fields) < 3 {
  128. log.Warn("Unparseable output for diff --name-status: `%s`)", line)
  129. continue
  130. }
  131. dest := fields[2]
  132. if len(dest) == 0 {
  133. log.Warn("Unparseable output for diff --name-status: `%s`)", line)
  134. continue
  135. }
  136. if dest[0] == '"' {
  137. dest, err = strconv.Unquote(dest)
  138. if err != nil {
  139. return nil, err
  140. }
  141. }
  142. if status == 'R' {
  143. changes.RemovedFilenames = append(changes.RemovedFilenames, filename)
  144. }
  145. updatedFilenames = append(updatedFilenames, dest)
  146. default:
  147. log.Warn("Unrecognized status: %c (line=%s)", status, line)
  148. }
  149. }
  150. cmd := git.NewCommand("ls-tree", "--full-tree", revision, "--")
  151. cmd.AddArguments(updatedFilenames...)
  152. lsTreeStdout, err := cmd.RunInDirBytes(repo.RepoPath())
  153. if err != nil {
  154. return nil, err
  155. }
  156. changes.Updates, err = parseGitLsTreeOutput(lsTreeStdout)
  157. return &changes, err
  158. }