Cache last commit to accelerate the repository directory page visit (#10069)

* Cache last commit to accelerate the repository directory page visit

* Default use default cache configuration

* add tests for last commit cache

* Simplify last commit cache

* Revert Enabled back

* Change the last commit cache default ttl to 8760h

* Fix test

Author: lunny

custom/conf/app.ini.sample

@@ -43,7 +43,7 @@ DEFAULT_CLOSE_ISSUES_VIA_COMMITS_IN_ANY_BRANCH = false
 ENABLE_PUSH_CREATE_USER = false
 ENABLE_PUSH_CREATE_ORG = false
 ; Comma separated list of globally disabled repo units. Allowed values: repo.issues, repo.ext_issues, repo.pulls, repo.wiki, repo.ext_wiki
-DISABLED_REPO_UNITS = 
+DISABLED_REPO_UNITS =
 ; Comma separated list of default repo units. Allowed values: repo.code, repo.releases, repo.issues, repo.pulls, repo.wiki.
 ; Note: Code and Releases can currently not be deactivated. If you specify default repo units you should still list them for future compatibility.
 ; External wiki and issue tracker can't be enabled by default as it requires additional settings.
@@ -632,6 +632,8 @@ SENDMAIL_PATH = sendmail
 SENDMAIL_ARGS =
 
 [cache]
+; if the cache enabled
+ENABLED = true
 ; Either "memory", "redis", or "memcache", default is "memory"
 ADAPTER = memory
 ; For "memory" only, GC interval in seconds, default is 60
@@ -644,6 +646,16 @@ HOST =
 ; Setting it to 0 disables caching
 ITEM_TTL = 16h
 
+; Last commit cache
+[cache.last_commit]
+; if the cache enabled
+ENABLED = true
+; Time to keep items in cache if not used, default is 8760 hours.
+; Setting it to 0 disables caching
+ITEM_TTL = 8760h
+; Only enable the cache when repository's commits count great than
+COMMITS_COUNT = 1000
+
 [session]
 ; Either "memory", "file", or "redis", default is "memory"
 PROVIDER = memory

docs/content/doc/advanced/config-cheat-sheet.en-us.md

@@ -383,13 +383,20 @@ relation to port exhaustion.
 
 ## Cache (`cache`)
 
+- `ENABLED`: **true**: Enable the cache.
 - `ADAPTER`: **memory**: Cache engine adapter, either `memory`, `redis`, or `memcache`.
 - `INTERVAL`: **60**: Garbage Collection interval (sec), for memory cache only.
 - `HOST`: **\<empty\>**: Connection string for `redis` and `memcache`.
    - Redis: `network=tcp,addr=127.0.0.1:6379,password=macaron,db=0,pool_size=100,idle_timeout=180`
    - Memcache: `127.0.0.1:9090;127.0.0.1:9091`
 - `ITEM_TTL`: **16h**: Time to keep items in cache if not used, Setting it to 0 disables caching.
 
+## Cache - LastCommitCache settings (`cache.last_commit`)
+
+- `ENABLED`: **true**: Enable the cache.
+- `ITEM_TTL`: **8760h**: Time to keep items in cache if not used, Setting it to 0 disables caching.
+- `COMMITS_COUNT`: **1000**: Only enable the cache when repository's commits count great than.
+
 ## Session (`session`)
 
 - `PROVIDER`: **memory**: Session engine provider \[memory, file, redis, mysql, couchbase, memcache, nodb, postgres\].

docs/content/doc/advanced/config-cheat-sheet.zh-cn.md

@@ -148,13 +148,20 @@ menu:
 
 ## Cache (`cache`)
 
+- `ENABLED`: **true**: 是否启用。
 - `ADAPTER`: **memory**: 缓存引擎，可以为 `memory`, `redis` 或 `memcache`。
 - `INTERVAL`: **60**: 只对内存缓存有效，GC间隔，单位秒。
 - `HOST`: **\<empty\>**: 针对redis和memcache有效，主机地址和端口。
     - Redis: `network=tcp,addr=127.0.0.1:6379,password=macaron,db=0,pool_size=100,idle_timeout=180`
     - Memache: `127.0.0.1:9090;127.0.0.1:9091`
 - `ITEM_TTL`: **16h**: 缓存项目失效时间，设置为 0 则禁用缓存。
 
+## Cache - LastCommitCache settings (`cache.last_commit`)
+
+- `ENABLED`: **true**: 是否启用。
+- `ITEM_TTL`: **8760h**: 缓存项目失效时间，设置为 0 则禁用缓存。
+- `COMMITS_COUNT`: **1000**: 仅当仓库的提交数大于时才启用缓存。
+
 ## Session (`session`)
 
 - `PROVIDER`: Session 内容存储方式，可选 `memory`, `file`, `redis` 或 `mysql`。

integrations/repo_test.go

@@ -7,8 +7,10 @@ package integrations
 import (
 	"fmt"
 	"net/http"
+	"path"
 	"strings"
 	"testing"
+	"time"
 
 	"code.gitea.io/gitea/modules/setting"
 
@@ -29,12 +31,71 @@ func TestViewRepo(t *testing.T) {
 	session.MakeRequest(t, req, http.StatusNotFound)
 }
 
-func TestViewRepo2(t *testing.T) {
+func testViewRepo(t *testing.T) {
 	defer prepareTestEnv(t)()
 
 	req := NewRequest(t, "GET", "/user3/repo3")
 	session := loginUser(t, "user2")
-	session.MakeRequest(t, req, http.StatusOK)
+	resp := session.MakeRequest(t, req, http.StatusOK)
+
+	htmlDoc := NewHTMLParser(t, resp.Body)
+	files := htmlDoc.doc.Find("#repo-files-table  > TBODY > TR")
+
+	type file struct {
+		fileName   string
+		commitID   string
+		commitMsg  string
+		commitTime string
+	}
+
+	var items []file
+
+	files.Each(func(i int, s *goquery.Selection) {
+		tds := s.Find("td")
+		var f file
+		tds.Each(func(i int, s *goquery.Selection) {
+			if i == 0 {
+				f.fileName = strings.TrimSpace(s.Text())
+			} else if i == 1 {
+				a := s.Find("a")
+				f.commitMsg = strings.TrimSpace(a.Text())
+				l, _ := a.Attr("href")
+				f.commitID = path.Base(l)
+			}
+		})
+
+		f.commitTime, _ = s.Find("span.time-since").Attr("title")
+		items = append(items, f)
+	})
+
+	assert.EqualValues(t, []file{
+		{
+			fileName:   "doc",
+			commitID:   "2a47ca4b614a9f5a43abbd5ad851a54a616ffee6",
+			commitMsg:  "init project",
+			commitTime: time.Date(2017, time.June, 14, 13, 54, 21, 0, time.UTC).Format(time.RFC1123),
+		},
+		{
+			fileName:   "README.md",
+			commitID:   "2a47ca4b614a9f5a43abbd5ad851a54a616ffee6",
+			commitMsg:  "init project",
+			commitTime: time.Date(2017, time.June, 14, 13, 54, 21, 0, time.UTC).Format(time.RFC1123),
+		},
+	}, items)
+}
+
+func TestViewRepo2(t *testing.T) {
+	// no last commit cache
+	testViewRepo(t)
+
+	// enable last commit cache for all repositories
+	oldCommitsCount := setting.CacheService.LastCommit.CommitsCount
+	setting.CacheService.LastCommit.CommitsCount = 0
+	// first view will not hit the cache
+	testViewRepo(t)
+	// second view will hit the cache
+	testViewRepo(t)
+	setting.CacheService.LastCommit.CommitsCount = oldCommitsCount
 }
 
 func TestViewRepo3(t *testing.T) {

modules/cache/cache.go

@@ -16,20 +16,28 @@ import (
 	_ "gitea.com/macaron/cache/redis"
 )
 
-var conn mc.Cache
+var (
+	conn mc.Cache
+)
+
+func newCache(cacheConfig setting.Cache) (mc.Cache, error) {
+	return mc.NewCacher(cacheConfig.Adapter, mc.Options{
+		Adapter:       cacheConfig.Adapter,
+		AdapterConfig: cacheConfig.Conn,
+		Interval:      cacheConfig.Interval,
+	})
+}
 
 // NewContext start cache service
 func NewContext() error {
-	if setting.CacheService == nil || conn != nil {
-		return nil
+	var err error
+
+	if conn == nil && setting.CacheService.Enabled {
+		if conn, err = newCache(setting.CacheService.Cache); err != nil {
+			return err
+		}
 	}
 
-	var err error
-	conn, err = mc.NewCacher(setting.CacheService.Adapter, mc.Options{
-		Adapter:       setting.CacheService.Adapter,
-		AdapterConfig: setting.CacheService.Conn,
-		Interval:      setting.CacheService.Interval,
-	})
 	return err
 }
 

modules/cache/last_commit.go

@@ -0,0 +1,64 @@
+// 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 cache
+
+import (
+	"fmt"
+
+	"code.gitea.io/gitea/modules/git"
+	"code.gitea.io/gitea/modules/log"
+
+	mc "gitea.com/macaron/cache"
+	"gopkg.in/src-d/go-git.v4/plumbing/object"
+)
+
+// LastCommitCache represents a cache to store last commit
+type LastCommitCache struct {
+	repoPath    string
+	ttl         int64
+	repo        *git.Repository
+	commitCache map[string]*object.Commit
+	mc.Cache
+}
+
+// NewLastCommitCache creates a new last commit cache for repo
+func NewLastCommitCache(repoPath string, gitRepo *git.Repository, ttl int64) *LastCommitCache {
+	return &LastCommitCache{
+		repoPath:    repoPath,
+		repo:        gitRepo,
+		commitCache: make(map[string]*object.Commit),
+		ttl:         ttl,
+		Cache:       conn,
+	}
+}
+
+// Get get the last commit information by commit id and entry path
+func (c LastCommitCache) Get(ref, entryPath string) (*object.Commit, error) {
+	v := c.Cache.Get(fmt.Sprintf("last_commit:%s:%s:%s", c.repoPath, ref, entryPath))
+	if vs, ok := v.(string); ok {
+		log.Trace("LastCommitCache hit level 1: [%s:%s:%s]", ref, entryPath, vs)
+		if commit, ok := c.commitCache[vs]; ok {
+			log.Trace("LastCommitCache hit level 2: [%s:%s:%s]", ref, entryPath, vs)
+			return commit, nil
+		}
+		id, err := c.repo.ConvertToSHA1(vs)
+		if err != nil {
+			return nil, err
+		}
+		commit, err := c.repo.GoGitRepo().CommitObject(id)
+		if err != nil {
+			return nil, err
+		}
+		c.commitCache[vs] = commit
+		return commit, nil
+	}
+	return nil, nil
+}
+
+// Put put the last commit id with commit and entry path
+func (c LastCommitCache) Put(ref, entryPath, commitID string) error {
+	log.Trace("LastCommitCache save: [%s:%s:%s]", ref, entryPath, commitID)
+	return c.Cache.Put(fmt.Sprintf("last_commit:%s:%s:%s", c.repoPath, ref, entryPath), commitID, c.ttl)
+}

modules/git/cache.go

@@ -4,8 +4,10 @@
 
 package git
 
+import "gopkg.in/src-d/go-git.v4/plumbing/object"
+
 // LastCommitCache cache
 type LastCommitCache interface {
-	Get(repoPath, ref, entryPath string) (*Commit, error)
-	Put(repoPath, ref, entryPath string, commit *Commit) error
+	Get(ref, entryPath string) (*object.Commit, error)
+	Put(ref, entryPath, commitID string) error
 }

modules/git/commit_info.go

@@ -5,6 +5,8 @@
 package git
 
 import (
+	"path"
+
 	"github.com/emirpasic/gods/trees/binaryheap"
 	"gopkg.in/src-d/go-git.v4/plumbing"
 	"gopkg.in/src-d/go-git.v4/plumbing/object"
@@ -30,7 +32,29 @@ func (tes Entries) GetCommitsInfo(commit *Commit, treePath string, cache LastCom
 		return nil, nil, err
 	}
 
-	revs, err := getLastCommitForPaths(c, treePath, entryPaths)
+	var revs map[string]*object.Commit
+	if cache != nil {
+		var unHitPaths []string
+		revs, unHitPaths, err = getLastCommitForPathsByCache(commit.ID.String(), treePath, entryPaths, cache)
+		if err != nil {
+			return nil, nil, err
+		}
+		if len(unHitPaths) > 0 {
+			revs2, err := getLastCommitForPaths(c, treePath, unHitPaths)
+			if err != nil {
+				return nil, nil, err
+			}
+
+			for k, v := range revs2 {
+				if err := cache.Put(commit.ID.String(), path.Join(treePath, k), v.ID().String()); err != nil {
+					return nil, nil, err
+				}
+				revs[k] = v
+			}
+		}
+	} else {
+		revs, err = getLastCommitForPaths(c, treePath, entryPaths)
+	}
 	if err != nil {
 		return nil, nil, err
 	}
@@ -127,6 +151,25 @@ func getFileHashes(c cgobject.CommitNode, treePath string, paths []string) (map[
 	return hashes, nil
 }
 
+func getLastCommitForPathsByCache(commitID, treePath string, paths []string, cache LastCommitCache) (map[string]*object.Commit, []string, error) {
+	var unHitEntryPaths []string
+	var results = make(map[string]*object.Commit)
+	for _, p := range paths {
+		lastCommit, err := cache.Get(commitID, path.Join(treePath, p))
+		if err != nil {
+			return nil, nil, err
+		}
+		if lastCommit != nil {
+			results[p] = lastCommit
+			continue
+		}
+
+		unHitEntryPaths = append(unHitEntryPaths, p)
+	}
+
+	return results, unHitEntryPaths, nil
+}
+
 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 {