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.

dump.go 15KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635
  1. // Copyright 2020 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 migrations
  5. import (
  6. "context"
  7. "fmt"
  8. "io"
  9. "net/http"
  10. "net/url"
  11. "os"
  12. "path"
  13. "path/filepath"
  14. "strconv"
  15. "strings"
  16. "time"
  17. user_model "code.gitea.io/gitea/models/user"
  18. "code.gitea.io/gitea/modules/git"
  19. "code.gitea.io/gitea/modules/log"
  20. base "code.gitea.io/gitea/modules/migration"
  21. "code.gitea.io/gitea/modules/repository"
  22. "code.gitea.io/gitea/modules/structs"
  23. "gopkg.in/yaml.v2"
  24. )
  25. var _ base.Uploader = &RepositoryDumper{}
  26. // RepositoryDumper implements an Uploader to the local directory
  27. type RepositoryDumper struct {
  28. ctx context.Context
  29. baseDir string
  30. repoOwner string
  31. repoName string
  32. opts base.MigrateOptions
  33. milestoneFile *os.File
  34. labelFile *os.File
  35. releaseFile *os.File
  36. issueFile *os.File
  37. commentFiles map[int64]*os.File
  38. pullrequestFile *os.File
  39. reviewFiles map[int64]*os.File
  40. gitRepo *git.Repository
  41. prHeadCache map[string]struct{}
  42. }
  43. // NewRepositoryDumper creates an gitea Uploader
  44. func NewRepositoryDumper(ctx context.Context, baseDir, repoOwner, repoName string, opts base.MigrateOptions) (*RepositoryDumper, error) {
  45. baseDir = filepath.Join(baseDir, repoOwner, repoName)
  46. if err := os.MkdirAll(baseDir, os.ModePerm); err != nil {
  47. return nil, err
  48. }
  49. return &RepositoryDumper{
  50. ctx: ctx,
  51. opts: opts,
  52. baseDir: baseDir,
  53. repoOwner: repoOwner,
  54. repoName: repoName,
  55. prHeadCache: make(map[string]struct{}),
  56. commentFiles: make(map[int64]*os.File),
  57. reviewFiles: make(map[int64]*os.File),
  58. }, nil
  59. }
  60. // MaxBatchInsertSize returns the table's max batch insert size
  61. func (g *RepositoryDumper) MaxBatchInsertSize(tp string) int {
  62. return 1000
  63. }
  64. func (g *RepositoryDumper) gitPath() string {
  65. return filepath.Join(g.baseDir, "git")
  66. }
  67. func (g *RepositoryDumper) wikiPath() string {
  68. return filepath.Join(g.baseDir, "wiki")
  69. }
  70. func (g *RepositoryDumper) commentDir() string {
  71. return filepath.Join(g.baseDir, "comments")
  72. }
  73. func (g *RepositoryDumper) reviewDir() string {
  74. return filepath.Join(g.baseDir, "reviews")
  75. }
  76. func (g *RepositoryDumper) setURLToken(remoteAddr string) (string, error) {
  77. if len(g.opts.AuthToken) > 0 || len(g.opts.AuthUsername) > 0 {
  78. u, err := url.Parse(remoteAddr)
  79. if err != nil {
  80. return "", err
  81. }
  82. u.User = url.UserPassword(g.opts.AuthUsername, g.opts.AuthPassword)
  83. if len(g.opts.AuthToken) > 0 {
  84. u.User = url.UserPassword("oauth2", g.opts.AuthToken)
  85. }
  86. remoteAddr = u.String()
  87. }
  88. return remoteAddr, nil
  89. }
  90. // CreateRepo creates a repository
  91. func (g *RepositoryDumper) CreateRepo(repo *base.Repository, opts base.MigrateOptions) error {
  92. f, err := os.Create(filepath.Join(g.baseDir, "repo.yml"))
  93. if err != nil {
  94. return err
  95. }
  96. defer f.Close()
  97. bs, err := yaml.Marshal(map[string]interface{}{
  98. "name": repo.Name,
  99. "owner": repo.Owner,
  100. "description": repo.Description,
  101. "clone_addr": opts.CloneAddr,
  102. "original_url": repo.OriginalURL,
  103. "is_private": opts.Private,
  104. "service_type": opts.GitServiceType,
  105. "wiki": opts.Wiki,
  106. "issues": opts.Issues,
  107. "milestones": opts.Milestones,
  108. "labels": opts.Labels,
  109. "releases": opts.Releases,
  110. "comments": opts.Comments,
  111. "pulls": opts.PullRequests,
  112. "assets": opts.ReleaseAssets,
  113. })
  114. if err != nil {
  115. return err
  116. }
  117. if _, err := f.Write(bs); err != nil {
  118. return err
  119. }
  120. repoPath := g.gitPath()
  121. if err := os.MkdirAll(repoPath, os.ModePerm); err != nil {
  122. return err
  123. }
  124. migrateTimeout := 2 * time.Hour
  125. remoteAddr, err := g.setURLToken(repo.CloneURL)
  126. if err != nil {
  127. return err
  128. }
  129. err = git.Clone(g.ctx, remoteAddr, repoPath, git.CloneRepoOptions{
  130. Mirror: true,
  131. Quiet: true,
  132. Timeout: migrateTimeout,
  133. })
  134. if err != nil {
  135. return fmt.Errorf("Clone: %v", err)
  136. }
  137. if opts.Wiki {
  138. wikiPath := g.wikiPath()
  139. wikiRemotePath := repository.WikiRemoteURL(g.ctx, remoteAddr)
  140. if len(wikiRemotePath) > 0 {
  141. if err := os.MkdirAll(wikiPath, os.ModePerm); err != nil {
  142. return fmt.Errorf("Failed to remove %s: %v", wikiPath, err)
  143. }
  144. if err := git.Clone(g.ctx, wikiRemotePath, wikiPath, git.CloneRepoOptions{
  145. Mirror: true,
  146. Quiet: true,
  147. Timeout: migrateTimeout,
  148. Branch: "master",
  149. }); err != nil {
  150. log.Warn("Clone wiki: %v", err)
  151. if err := os.RemoveAll(wikiPath); err != nil {
  152. return fmt.Errorf("Failed to remove %s: %v", wikiPath, err)
  153. }
  154. }
  155. }
  156. }
  157. g.gitRepo, err = git.OpenRepositoryCtx(g.ctx, g.gitPath())
  158. return err
  159. }
  160. // Close closes this uploader
  161. func (g *RepositoryDumper) Close() {
  162. if g.gitRepo != nil {
  163. g.gitRepo.Close()
  164. }
  165. if g.milestoneFile != nil {
  166. g.milestoneFile.Close()
  167. }
  168. if g.labelFile != nil {
  169. g.labelFile.Close()
  170. }
  171. if g.releaseFile != nil {
  172. g.releaseFile.Close()
  173. }
  174. if g.issueFile != nil {
  175. g.issueFile.Close()
  176. }
  177. for _, f := range g.commentFiles {
  178. f.Close()
  179. }
  180. if g.pullrequestFile != nil {
  181. g.pullrequestFile.Close()
  182. }
  183. for _, f := range g.reviewFiles {
  184. f.Close()
  185. }
  186. }
  187. // CreateTopics creates topics
  188. func (g *RepositoryDumper) CreateTopics(topics ...string) error {
  189. f, err := os.Create(filepath.Join(g.baseDir, "topic.yml"))
  190. if err != nil {
  191. return err
  192. }
  193. defer f.Close()
  194. bs, err := yaml.Marshal(map[string]interface{}{
  195. "topics": topics,
  196. })
  197. if err != nil {
  198. return err
  199. }
  200. if _, err := f.Write(bs); err != nil {
  201. return err
  202. }
  203. return nil
  204. }
  205. // CreateMilestones creates milestones
  206. func (g *RepositoryDumper) CreateMilestones(milestones ...*base.Milestone) error {
  207. var err error
  208. if g.milestoneFile == nil {
  209. g.milestoneFile, err = os.Create(filepath.Join(g.baseDir, "milestone.yml"))
  210. if err != nil {
  211. return err
  212. }
  213. }
  214. bs, err := yaml.Marshal(milestones)
  215. if err != nil {
  216. return err
  217. }
  218. if _, err := g.milestoneFile.Write(bs); err != nil {
  219. return err
  220. }
  221. return nil
  222. }
  223. // CreateLabels creates labels
  224. func (g *RepositoryDumper) CreateLabels(labels ...*base.Label) error {
  225. var err error
  226. if g.labelFile == nil {
  227. g.labelFile, err = os.Create(filepath.Join(g.baseDir, "label.yml"))
  228. if err != nil {
  229. return err
  230. }
  231. }
  232. bs, err := yaml.Marshal(labels)
  233. if err != nil {
  234. return err
  235. }
  236. if _, err := g.labelFile.Write(bs); err != nil {
  237. return err
  238. }
  239. return nil
  240. }
  241. // CreateReleases creates releases
  242. func (g *RepositoryDumper) CreateReleases(releases ...*base.Release) error {
  243. if g.opts.ReleaseAssets {
  244. for _, release := range releases {
  245. attachDir := filepath.Join("release_assets", release.TagName)
  246. if err := os.MkdirAll(filepath.Join(g.baseDir, attachDir), os.ModePerm); err != nil {
  247. return err
  248. }
  249. for _, asset := range release.Assets {
  250. attachLocalPath := filepath.Join(attachDir, asset.Name)
  251. // download attachment
  252. err := func(attachPath string) error {
  253. var rc io.ReadCloser
  254. var err error
  255. if asset.DownloadURL == nil {
  256. rc, err = asset.DownloadFunc()
  257. if err != nil {
  258. return err
  259. }
  260. } else {
  261. resp, err := http.Get(*asset.DownloadURL)
  262. if err != nil {
  263. return err
  264. }
  265. rc = resp.Body
  266. }
  267. defer rc.Close()
  268. fw, err := os.Create(attachPath)
  269. if err != nil {
  270. return fmt.Errorf("Create: %v", err)
  271. }
  272. defer fw.Close()
  273. _, err = io.Copy(fw, rc)
  274. return err
  275. }(filepath.Join(g.baseDir, attachLocalPath))
  276. if err != nil {
  277. return err
  278. }
  279. asset.DownloadURL = &attachLocalPath // to save the filepath on the yml file, change the source
  280. }
  281. }
  282. }
  283. var err error
  284. if g.releaseFile == nil {
  285. g.releaseFile, err = os.Create(filepath.Join(g.baseDir, "release.yml"))
  286. if err != nil {
  287. return err
  288. }
  289. }
  290. bs, err := yaml.Marshal(releases)
  291. if err != nil {
  292. return err
  293. }
  294. if _, err := g.releaseFile.Write(bs); err != nil {
  295. return err
  296. }
  297. return nil
  298. }
  299. // SyncTags syncs releases with tags in the database
  300. func (g *RepositoryDumper) SyncTags() error {
  301. return nil
  302. }
  303. // CreateIssues creates issues
  304. func (g *RepositoryDumper) CreateIssues(issues ...*base.Issue) error {
  305. var err error
  306. if g.issueFile == nil {
  307. g.issueFile, err = os.Create(filepath.Join(g.baseDir, "issue.yml"))
  308. if err != nil {
  309. return err
  310. }
  311. }
  312. bs, err := yaml.Marshal(issues)
  313. if err != nil {
  314. return err
  315. }
  316. if _, err := g.issueFile.Write(bs); err != nil {
  317. return err
  318. }
  319. return nil
  320. }
  321. func (g *RepositoryDumper) createItems(dir string, itemFiles map[int64]*os.File, itemsMap map[int64][]interface{}) error {
  322. if err := os.MkdirAll(dir, os.ModePerm); err != nil {
  323. return err
  324. }
  325. for number, items := range itemsMap {
  326. var err error
  327. itemFile := itemFiles[number]
  328. if itemFile == nil {
  329. itemFile, err = os.Create(filepath.Join(dir, fmt.Sprintf("%d.yml", number)))
  330. if err != nil {
  331. return err
  332. }
  333. itemFiles[number] = itemFile
  334. }
  335. bs, err := yaml.Marshal(items)
  336. if err != nil {
  337. return err
  338. }
  339. if _, err := itemFile.Write(bs); err != nil {
  340. return err
  341. }
  342. }
  343. return nil
  344. }
  345. // CreateComments creates comments of issues
  346. func (g *RepositoryDumper) CreateComments(comments ...*base.Comment) error {
  347. commentsMap := make(map[int64][]interface{}, len(comments))
  348. for _, comment := range comments {
  349. commentsMap[comment.IssueIndex] = append(commentsMap[comment.IssueIndex], comment)
  350. }
  351. return g.createItems(g.commentDir(), g.commentFiles, commentsMap)
  352. }
  353. // CreatePullRequests creates pull requests
  354. func (g *RepositoryDumper) CreatePullRequests(prs ...*base.PullRequest) error {
  355. for _, pr := range prs {
  356. // download patch file
  357. err := func() error {
  358. u, err := g.setURLToken(pr.PatchURL)
  359. if err != nil {
  360. return err
  361. }
  362. resp, err := http.Get(u)
  363. if err != nil {
  364. return err
  365. }
  366. defer resp.Body.Close()
  367. pullDir := filepath.Join(g.gitPath(), "pulls")
  368. if err = os.MkdirAll(pullDir, os.ModePerm); err != nil {
  369. return err
  370. }
  371. fPath := filepath.Join(pullDir, fmt.Sprintf("%d.patch", pr.Number))
  372. f, err := os.Create(fPath)
  373. if err != nil {
  374. return err
  375. }
  376. defer f.Close()
  377. if _, err = io.Copy(f, resp.Body); err != nil {
  378. return err
  379. }
  380. pr.PatchURL = "git/pulls/" + fmt.Sprintf("%d.patch", pr.Number)
  381. return nil
  382. }()
  383. if err != nil {
  384. return err
  385. }
  386. // set head information
  387. pullHead := filepath.Join(g.gitPath(), "refs", "pull", fmt.Sprintf("%d", pr.Number))
  388. if err := os.MkdirAll(pullHead, os.ModePerm); err != nil {
  389. return err
  390. }
  391. p, err := os.Create(filepath.Join(pullHead, "head"))
  392. if err != nil {
  393. return err
  394. }
  395. _, err = p.WriteString(pr.Head.SHA)
  396. p.Close()
  397. if err != nil {
  398. return err
  399. }
  400. if pr.IsForkPullRequest() && pr.State != "closed" {
  401. if pr.Head.OwnerName != "" {
  402. remote := pr.Head.OwnerName
  403. _, ok := g.prHeadCache[remote]
  404. if !ok {
  405. // git remote add
  406. // TODO: how to handle private CloneURL?
  407. err := g.gitRepo.AddRemote(remote, pr.Head.CloneURL, true)
  408. if err != nil {
  409. log.Error("AddRemote failed: %s", err)
  410. } else {
  411. g.prHeadCache[remote] = struct{}{}
  412. ok = true
  413. }
  414. }
  415. if ok {
  416. _, err = git.NewCommandContext(g.ctx, "fetch", remote, pr.Head.Ref).RunInDir(g.gitPath())
  417. if err != nil {
  418. log.Error("Fetch branch from %s failed: %v", pr.Head.CloneURL, err)
  419. } else {
  420. // a new branch name with <original_owner_name/original_branchname> will be created to as new head branch
  421. ref := path.Join(pr.Head.OwnerName, pr.Head.Ref)
  422. headBranch := filepath.Join(g.gitPath(), "refs", "heads", ref)
  423. if err := os.MkdirAll(filepath.Dir(headBranch), os.ModePerm); err != nil {
  424. return err
  425. }
  426. b, err := os.Create(headBranch)
  427. if err != nil {
  428. return err
  429. }
  430. _, err = b.WriteString(pr.Head.SHA)
  431. b.Close()
  432. if err != nil {
  433. return err
  434. }
  435. pr.Head.Ref = ref
  436. }
  437. }
  438. }
  439. }
  440. // whatever it's a forked repo PR, we have to change head info as the same as the base info
  441. pr.Head.OwnerName = pr.Base.OwnerName
  442. pr.Head.RepoName = pr.Base.RepoName
  443. }
  444. var err error
  445. if g.pullrequestFile == nil {
  446. if err := os.MkdirAll(g.baseDir, os.ModePerm); err != nil {
  447. return err
  448. }
  449. g.pullrequestFile, err = os.Create(filepath.Join(g.baseDir, "pull_request.yml"))
  450. if err != nil {
  451. return err
  452. }
  453. }
  454. bs, err := yaml.Marshal(prs)
  455. if err != nil {
  456. return err
  457. }
  458. if _, err := g.pullrequestFile.Write(bs); err != nil {
  459. return err
  460. }
  461. return nil
  462. }
  463. // CreateReviews create pull request reviews
  464. func (g *RepositoryDumper) CreateReviews(reviews ...*base.Review) error {
  465. reviewsMap := make(map[int64][]interface{}, len(reviews))
  466. for _, review := range reviews {
  467. reviewsMap[review.IssueIndex] = append(reviewsMap[review.IssueIndex], review)
  468. }
  469. return g.createItems(g.reviewDir(), g.reviewFiles, reviewsMap)
  470. }
  471. // Rollback when migrating failed, this will rollback all the changes.
  472. func (g *RepositoryDumper) Rollback() error {
  473. g.Close()
  474. return os.RemoveAll(g.baseDir)
  475. }
  476. // Finish when migrating succeed, this will update something.
  477. func (g *RepositoryDumper) Finish() error {
  478. return nil
  479. }
  480. // DumpRepository dump repository according MigrateOptions to a local directory
  481. func DumpRepository(ctx context.Context, baseDir, ownerName string, opts base.MigrateOptions) error {
  482. downloader, err := newDownloader(ctx, ownerName, opts)
  483. if err != nil {
  484. return err
  485. }
  486. uploader, err := NewRepositoryDumper(ctx, baseDir, ownerName, opts.RepoName, opts)
  487. if err != nil {
  488. return err
  489. }
  490. if err := migrateRepository(downloader, uploader, opts, nil); err != nil {
  491. if err1 := uploader.Rollback(); err1 != nil {
  492. log.Error("rollback failed: %v", err1)
  493. }
  494. return err
  495. }
  496. return nil
  497. }
  498. func updateOptionsUnits(opts *base.MigrateOptions, units []string) {
  499. if len(units) == 0 {
  500. opts.Wiki = true
  501. opts.Issues = true
  502. opts.Milestones = true
  503. opts.Labels = true
  504. opts.Releases = true
  505. opts.Comments = true
  506. opts.PullRequests = true
  507. opts.ReleaseAssets = true
  508. } else {
  509. for _, unit := range units {
  510. switch strings.ToLower(unit) {
  511. case "wiki":
  512. opts.Wiki = true
  513. case "issues":
  514. opts.Issues = true
  515. case "milestones":
  516. opts.Milestones = true
  517. case "labels":
  518. opts.Labels = true
  519. case "releases":
  520. opts.Releases = true
  521. case "release_assets":
  522. opts.ReleaseAssets = true
  523. case "comments":
  524. opts.Comments = true
  525. case "pull_requests":
  526. opts.PullRequests = true
  527. }
  528. }
  529. }
  530. }
  531. // RestoreRepository restore a repository from the disk directory
  532. func RestoreRepository(ctx context.Context, baseDir, ownerName, repoName string, units []string, validation bool) error {
  533. doer, err := user_model.GetAdminUser()
  534. if err != nil {
  535. return err
  536. }
  537. uploader := NewGiteaLocalUploader(ctx, doer, ownerName, repoName)
  538. downloader, err := NewRepositoryRestorer(ctx, baseDir, ownerName, repoName, validation)
  539. if err != nil {
  540. return err
  541. }
  542. opts, err := downloader.getRepoOptions()
  543. if err != nil {
  544. return err
  545. }
  546. tp, _ := strconv.Atoi(opts["service_type"])
  547. migrateOpts := base.MigrateOptions{
  548. GitServiceType: structs.GitServiceType(tp),
  549. }
  550. updateOptionsUnits(&migrateOpts, units)
  551. if err = migrateRepository(downloader, uploader, migrateOpts, nil); err != nil {
  552. if err1 := uploader.Rollback(); err1 != nil {
  553. log.Error("rollback failed: %v", err1)
  554. }
  555. return err
  556. }
  557. return updateMigrationPosterIDByGitService(ctx, structs.GitServiceType(tp))
  558. }