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>

3 files changed, 102 insertions, 4 deletions

diff --git a/modules/git/commit_info.go b/modules/git/commit_info.go
index f4ba5aceb7..e03ea00fc6 100644
--- a/modules/git/commit_info.go
+++ b/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()) {
diff --git a/modules/git/notes.go b/modules/git/notes.go
index 8184a568fa..ba19fa4893 100644
--- a/modules/git/notes.go
+++ b/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
 	}
diff --git a/modules/repository/cache.go b/modules/repository/cache.go
new file mode 100644
index 0000000000..508e5bec0b
--- /dev/null
+++ b/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)
+}