Cache last commit when pushing for big repository (#10109)

* Cache last commit when pushing for big repository

* Fix bug

* detect force push

* Refactor cache push

* Finish cache last commit info when push

* Some improvements

* Fix lint

* Remove unused changes

* Move pull request test before cache

* Fix test mysql

Co-authored-by: techknowlogick <techknowlogick@gitea.io>
Co-authored-by: Lauris BH <lauris@nix.lv>

modules/git/commit_info.go

@@ -40,7 +40,7 @@ func (tes Entries) GetCommitsInfo(commit *Commit, treePath string, cache LastCom
 			return nil, nil, err
 		}
 		if len(unHitPaths) > 0 {
-			revs2, err := getLastCommitForPaths(c, treePath, unHitPaths)
+			revs2, err := GetLastCommitForPaths(c, treePath, unHitPaths)
 			if err != nil {
 				return nil, nil, err
 			}
@@ -53,7 +53,7 @@ func (tes Entries) GetCommitsInfo(commit *Commit, treePath string, cache LastCom
 			}
 		}
 	} else {
-		revs, err = getLastCommitForPaths(c, treePath, entryPaths)
+		revs, err = GetLastCommitForPaths(c, treePath, entryPaths)
 	}
 	if err != nil {
 		return nil, nil, err
@@ -170,7 +170,8 @@ func getLastCommitForPathsByCache(commitID, treePath string, paths []string, cac
 	return results, unHitEntryPaths, nil
 }
 
-func getLastCommitForPaths(c cgobject.CommitNode, treePath string, paths []string) (map[string]*object.Commit, error) {
+// GetLastCommitForPaths returns last commit information
+func GetLastCommitForPaths(c cgobject.CommitNode, treePath string, paths []string) (map[string]*object.Commit, error) {
 	// We do a tree traversal with nodes sorted by commit time
 	heap := binaryheap.NewWith(func(a, b interface{}) int {
 		if a.(*commitAndPaths).commit.CommitTime().Before(b.(*commitAndPaths).commit.CommitTime()) {

modules/git/notes.go

@@ -70,7 +70,7 @@ func GetNote(repo *Repository, commitID string, note *Note) error {
 		return err
 	}
 
-	lastCommits, err := getLastCommitForPaths(commitNode, "", []string{path})
+	lastCommits, err := GetLastCommitForPaths(commitNode, "", []string{path})
 	if err != nil {
 		return err
 	}

modules/repository/cache.go

@@ -0,0 +1,97 @@
+// Copyright 2020 The Gitea Authors. All rights reserved.
+// Use of this source code is governed by a MIT-style
+// license that can be found in the LICENSE file.
+
+package repository
+
+import (
+	"path"
+	"strings"
+
+	"code.gitea.io/gitea/models"
+	"code.gitea.io/gitea/modules/cache"
+	"code.gitea.io/gitea/modules/git"
+	"code.gitea.io/gitea/modules/setting"
+
+	cgobject "github.com/go-git/go-git/v5/plumbing/object/commitgraph"
+)
+
+func recusiveCache(gitRepo *git.Repository, c cgobject.CommitNode, tree *git.Tree, treePath string, ca *cache.LastCommitCache, level int) error {
+	if level == 0 {
+		return nil
+	}
+
+	entries, err := tree.ListEntries()
+	if err != nil {
+		return err
+	}
+
+	entryPaths := make([]string, len(entries))
+	entryMap := make(map[string]*git.TreeEntry)
+	for i, entry := range entries {
+		entryPaths[i] = entry.Name()
+		entryMap[entry.Name()] = entry
+	}
+
+	commits, err := git.GetLastCommitForPaths(c, treePath, entryPaths)
+	if err != nil {
+		return err
+	}
+
+	for entry, cm := range commits {
+		if err := ca.Put(c.ID().String(), path.Join(treePath, entry), cm.ID().String()); err != nil {
+			return err
+		}
+		if entryMap[entry].IsDir() {
+			subTree, err := tree.SubTree(entry)
+			if err != nil {
+				return err
+			}
+			if err := recusiveCache(gitRepo, c, subTree, entry, ca, level-1); err != nil {
+				return err
+			}
+		}
+	}
+
+	return nil
+}
+
+func getRefName(fullRefName string) string {
+	if strings.HasPrefix(fullRefName, git.TagPrefix) {
+		return fullRefName[len(git.TagPrefix):]
+	} else if strings.HasPrefix(fullRefName, git.BranchPrefix) {
+		return fullRefName[len(git.BranchPrefix):]
+	}
+	return ""
+}
+
+// CacheRef cachhe last commit information of the branch or the tag
+func CacheRef(repo *models.Repository, gitRepo *git.Repository, fullRefName string) error {
+	if !setting.CacheService.LastCommit.Enabled {
+		return nil
+	}
+
+	commit, err := gitRepo.GetCommit(fullRefName)
+	if err != nil {
+		return err
+	}
+
+	commitsCount, err := cache.GetInt64(repo.GetCommitsCountCacheKey(getRefName(fullRefName), true), commit.CommitsCount)
+	if err != nil {
+		return err
+	}
+	if commitsCount < setting.CacheService.LastCommit.CommitsCount {
+		return nil
+	}
+
+	commitNodeIndex, _ := gitRepo.CommitNodeIndex()
+
+	c, err := commitNodeIndex.Get(commit.ID)
+	if err != nil {
+		return err
+	}
+
+	ca := cache.NewLastCommitCache(repo.FullName(), gitRepo, int64(setting.CacheService.LastCommit.TTL.Seconds()))
+
+	return recusiveCache(gitRepo, c, &commit.Tree, "", ca, 1)
+}

services/repository/push.go

@@ -30,7 +30,7 @@ type PushUpdateOptions struct {
 	PusherName   string
 	RepoUserName string
 	RepoName     string
-	RefFullName  string
+	RefFullName  string // branch, tag or other name to push
 	OldCommitID  string
 	NewCommitID  string
 }
@@ -95,11 +95,36 @@ func (opts PushUpdateOptions) BranchName() string {
 	return opts.RefFullName[len(git.BranchPrefix):]
 }
 
+// RefName returns simple name for ref
+func (opts PushUpdateOptions) RefName() string {
+	if strings.HasPrefix(opts.RefFullName, git.TagPrefix) {
+		return opts.RefFullName[len(git.TagPrefix):]
+	} else if strings.HasPrefix(opts.RefFullName, git.BranchPrefix) {
+		return opts.RefFullName[len(git.BranchPrefix):]
+	}
+	return ""
+}
+
 // RepoFullName returns repo full name
 func (opts PushUpdateOptions) RepoFullName() string {
 	return opts.RepoUserName + "/" + opts.RepoName
 }
 
+// isForcePush detect if a push is a force push
+func isForcePush(repoPath string, opts *PushUpdateOptions) (bool, error) {
+	if !opts.IsUpdateBranch() {
+		return false, nil
+	}
+
+	output, err := git.NewCommand("rev-list", "--max-count=1", opts.OldCommitID, "^"+opts.NewCommitID).RunInDir(repoPath)
+	if err != nil {
+		return false, err
+	} else if len(output) > 0 {
+		return true, nil
+	}
+	return false, nil
+}
+
 // pushQueue represents a queue to handle update pull request tests
 var pushQueue queue.Queue
 
@@ -184,7 +209,6 @@ func pushUpdates(optsList []*PushUpdateOptions) error {
 			if opts.IsDelRef() {
 				delTags = append(delTags, tagName)
 			} else { // is new tag
-				cache.Remove(repo.GetCommitsCountCacheKey(tagName, true))
 				addTags = append(addTags, tagName)
 			}
 		} else if opts.IsBranch() { // If is branch reference
@@ -197,8 +221,8 @@ func pushUpdates(optsList []*PushUpdateOptions) error {
 
 			branch := opts.BranchName()
 			if !opts.IsDelRef() {
-				// Clear cache for branch commit count
-				cache.Remove(repo.GetCommitsCountCacheKey(opts.BranchName(), true))
+				log.Trace("TriggerTask '%s/%s' by %s", repo.Name, branch, pusher.Name)
+				go pull_service.AddTestPullRequestTask(pusher, repo.ID, branch, true, opts.OldCommitID, opts.NewCommitID)
 
 				newCommit, err := gitRepo.GetCommit(opts.NewCommitID)
 				if err != nil {
@@ -217,6 +241,20 @@ func pushUpdates(optsList []*PushUpdateOptions) error {
 					if err != nil {
 						return fmt.Errorf("newCommit.CommitsBeforeUntil: %v", err)
 					}
+
+					isForce, err := isForcePush(repo.RepoPath(), opts)
+					if err != nil {
+						log.Error("isForcePush %s/%s failed: %v", repo.ID, branch, err)
+					}
+
+					if isForce {
+						log.Trace("Push %s is a force push", opts.NewCommitID)
+
+						cache.Remove(repo.GetCommitsCountCacheKey(opts.RefName(), true))
+					} else {
+						// TODO: increment update the commit count cache but not remove
+						cache.Remove(repo.GetCommitsCountCacheKey(opts.RefName(), true))
+					}
 				}
 
 				commits = repo_module.ListToPushCommits(l)
@@ -225,9 +263,10 @@ func pushUpdates(optsList []*PushUpdateOptions) error {
 					log.Error("models.RemoveDeletedBranch %s/%s failed: %v", repo.ID, branch, err)
 				}
 
-				log.Trace("TriggerTask '%s/%s' by %s", repo.Name, branch, pusher.Name)
-
-				go pull_service.AddTestPullRequestTask(pusher, repo.ID, branch, true, opts.OldCommitID, opts.NewCommitID)
+				// Cache for big repository
+				if err := repo_module.CacheRef(repo, gitRepo, opts.RefFullName); err != nil {
+					log.Error("repo_module.CacheRef %s/%s failed: %v", repo.ID, branch, err)
+				}
 			} else if err = pull_service.CloseBranchPulls(pusher, repo.ID, branch); err != nil {
 				// close all related pulls
 				log.Error("close related pull request failed: %v", err)