Refactor the usage of batch catfile (#31754) (#31889)

Backport #31754 by @lunny

When opening a repository, it will call `ensureValidRepository` and also
`CatFileBatch`. But sometimes these will not be used until repository
closed. So it's a waste of CPU to invoke 3 times git command for every
open repository.

This PR removed all of these from `OpenRepository` but only kept
checking whether the folder exists. When a batch is necessary, the
necessary functions will be invoked.

Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>

modules/git/batch.go

@@ -0,0 +1,46 @@
+// Copyright 2024 The Gitea Authors. All rights reserved.
+// SPDX-License-Identifier: MIT
+
+package git
+
+import (
+	"bufio"
+	"context"
+)
+
+type Batch struct {
+	cancel context.CancelFunc
+	Reader *bufio.Reader
+	Writer WriteCloserError
+}
+
+func (repo *Repository) NewBatch(ctx context.Context) (*Batch, error) {
+	// Now because of some insanity with git cat-file not immediately failing if not run in a valid git directory we need to run git rev-parse first!
+	if err := ensureValidGitRepository(ctx, repo.Path); err != nil {
+		return nil, err
+	}
+
+	var batch Batch
+	batch.Writer, batch.Reader, batch.cancel = catFileBatch(ctx, repo.Path)
+	return &batch, nil
+}
+
+func (repo *Repository) NewBatchCheck(ctx context.Context) (*Batch, error) {
+	// Now because of some insanity with git cat-file not immediately failing if not run in a valid git directory we need to run git rev-parse first!
+	if err := ensureValidGitRepository(ctx, repo.Path); err != nil {
+		return nil, err
+	}
+
+	var check Batch
+	check.Writer, check.Reader, check.cancel = catFileBatchCheck(ctx, repo.Path)
+	return &check, nil
+}
+
+func (b *Batch) Close() {
+	if b.cancel != nil {
+		b.cancel()
+		b.Reader = nil
+		b.Writer = nil
+		b.cancel = nil
+	}
+}

modules/git/batch_reader.go

@@ -26,10 +26,10 @@ type WriteCloserError interface {
 	CloseWithError(err error) error
 }
 
-// EnsureValidGitRepository runs git rev-parse in the repository path - thus ensuring that the repository is a valid repository.
+// ensureValidGitRepository runs git rev-parse in the repository path - thus ensuring that the repository is a valid repository.
 // Run before opening git cat-file.
 // This is needed otherwise the git cat-file will hang for invalid repositories.
-func EnsureValidGitRepository(ctx context.Context, repoPath string) error {
+func ensureValidGitRepository(ctx context.Context, repoPath string) error {
 	stderr := strings.Builder{}
 	err := NewCommand(ctx, "rev-parse").
 		SetDescription(fmt.Sprintf("%s rev-parse [repo_path: %s]", GitExecutable, repoPath)).
@@ -43,8 +43,8 @@ func EnsureValidGitRepository(ctx context.Context, repoPath string) error {
 	return nil
 }
 
-// CatFileBatchCheck opens git cat-file --batch-check in the provided repo and returns a stdin pipe, a stdout reader and cancel function
+// catFileBatchCheck opens git cat-file --batch-check in the provided repo and returns a stdin pipe, a stdout reader and cancel function
-func CatFileBatchCheck(ctx context.Context, repoPath string) (WriteCloserError, *bufio.Reader, func()) {
+func catFileBatchCheck(ctx context.Context, repoPath string) (WriteCloserError, *bufio.Reader, func()) {
 	batchStdinReader, batchStdinWriter := io.Pipe()
 	batchStdoutReader, batchStdoutWriter := io.Pipe()
 	ctx, ctxCancel := context.WithCancel(ctx)
@@ -93,8 +93,8 @@ func CatFileBatchCheck(ctx context.Context, repoPath string) (WriteCloserError,
 	return batchStdinWriter, batchReader, cancel
 }
 
-// CatFileBatch opens git cat-file --batch in the provided repo and returns a stdin pipe, a stdout reader and cancel function
+// catFileBatch opens git cat-file --batch in the provided repo and returns a stdin pipe, a stdout reader and cancel function
-func CatFileBatch(ctx context.Context, repoPath string) (WriteCloserError, *bufio.Reader, func()) {
+func catFileBatch(ctx context.Context, repoPath string) (WriteCloserError, *bufio.Reader, func()) {
 	// We often want to feed the commits in order into cat-file --batch, followed by their trees and sub trees as necessary.
 	// so let's create a batch stdin and stdout
 	batchStdinReader, batchStdinWriter := io.Pipe()

modules/git/blob_nogogit.go

@@ -26,9 +26,12 @@ type Blob struct {
 // DataAsync gets a ReadCloser for the contents of a blob without reading it all.
 // Calling the Close function on the result will discard all unread output.
 func (b *Blob) DataAsync() (io.ReadCloser, error) {
-	wr, rd, cancel := b.repo.CatFileBatch(b.repo.Ctx)
+	wr, rd, cancel, err := b.repo.CatFileBatch(b.repo.Ctx)
+	if err != nil {
+		return nil, err
+	}
 
-	_, err := wr.Write([]byte(b.ID.String() + "\n"))
+	_, err = wr.Write([]byte(b.ID.String() + "\n"))
 	if err != nil {
 		cancel()
 		return nil, err
@@ -64,9 +67,13 @@ func (b *Blob) Size() int64 {
 		return b.size
 	}
 
-	wr, rd, cancel := b.repo.CatFileBatchCheck(b.repo.Ctx)
+	wr, rd, cancel, err := b.repo.CatFileBatchCheck(b.repo.Ctx)
+	if err != nil {
+		log.Debug("error whilst reading size for %s in %s. Error: %v", b.ID.String(), b.repo.Path, err)
+		return 0
+	}
 	defer cancel()
-	_, err := wr.Write([]byte(b.ID.String() + "\n"))
+	_, err = wr.Write([]byte(b.ID.String() + "\n"))
 	if err != nil {
 		log.Debug("error whilst reading size for %s in %s. Error: %v", b.ID.String(), b.repo.Path, err)
 		return 0

modules/git/commit_info_nogogit.go

@@ -124,7 +124,10 @@ func GetLastCommitForPaths(ctx context.Context, commit *Commit, treePath string,
 		return nil, err
 	}
 
-	batchStdinWriter, batchReader, cancel := commit.repo.CatFileBatch(ctx)
+	batchStdinWriter, batchReader, cancel, err := commit.repo.CatFileBatch(ctx)
+	if err != nil {
+		return nil, err
+	}
 	defer cancel()
 
 	commitsMap := map[string]*Commit{}

modules/git/pipeline/lfs_nogogit.go

@@ -46,7 +46,10 @@ func FindLFSFile(repo *git.Repository, objectID git.ObjectID) ([]*LFSResult, err
 
 	// Next feed the commits in order into cat-file --batch, followed by their trees and sub trees as necessary.
 	// so let's create a batch stdin and stdout
-	batchStdinWriter, batchReader, cancel := repo.CatFileBatch(repo.Ctx)
+	batchStdinWriter, batchReader, cancel, err := repo.CatFileBatch(repo.Ctx)
+	if err != nil {
+		return nil, err
+	}
 	defer cancel()
 
 	// We'll use a scanner for the revList because it's simpler than a bufio.Reader

modules/git/repo_base_nogogit.go

@@ -25,15 +25,11 @@ type Repository struct {
 
 	gpgSettings *GPGSettings
 
-	batchInUse  bool
+	batchInUse bool
-	batchCancel context.CancelFunc
+	batch      *Batch
-	batchReader *bufio.Reader
-	batchWriter WriteCloserError
 
-	checkInUse  bool
+	checkInUse bool
-	checkCancel context.CancelFunc
+	check      *Batch
-	checkReader *bufio.Reader
-	checkWriter WriteCloserError
 
 	Ctx             context.Context
 	LastCommitCache *LastCommitCache
@@ -55,63 +51,75 @@ func OpenRepository(ctx context.Context, repoPath string) (*Repository, error) {
 		return nil, util.NewNotExistErrorf("no such file or directory")
 	}
 
-	// Now because of some insanity with git cat-file not immediately failing if not run in a valid git directory we need to run git rev-parse first!
+	return &Repository{
-	if err := EnsureValidGitRepository(ctx, repoPath); err != nil {
-		return nil, err
-	}
-
-	repo := &Repository{
 		Path:     repoPath,
 		tagCache: newObjectCache(),
 		Ctx:      ctx,
-	}
+	}, nil
-
-	repo.batchWriter, repo.batchReader, repo.batchCancel = CatFileBatch(ctx, repoPath)
-	repo.checkWriter, repo.checkReader, repo.checkCancel = CatFileBatchCheck(ctx, repoPath)
-
-	return repo, nil
 }
 
 // CatFileBatch obtains a CatFileBatch for this repository
-func (repo *Repository) CatFileBatch(ctx context.Context) (WriteCloserError, *bufio.Reader, func()) {
+func (repo *Repository) CatFileBatch(ctx context.Context) (WriteCloserError, *bufio.Reader, func(), error) {
-	if repo.batchCancel == nil || repo.batchInUse {
+	if repo.batch == nil {
-		log.Debug("Opening temporary cat file batch for: %s", repo.Path)
+		var err error
-		return CatFileBatch(ctx, repo.Path)
+		repo.batch, err = repo.NewBatch(ctx)
+		if err != nil {
+			return nil, nil, nil, err
+		}
 	}
-	repo.batchInUse = true
+
-	return repo.batchWriter, repo.batchReader, func() {
+	if !repo.batchInUse {
-		repo.batchInUse = false
+		repo.batchInUse = true
+		return repo.batch.Writer, repo.batch.Reader, func() {
+			repo.batchInUse = false
+		}, nil
 	}
+
+	log.Debug("Opening temporary cat file batch for: %s", repo.Path)
+	tempBatch, err := repo.NewBatch(ctx)
+	if err != nil {
+		return nil, nil, nil, err
+	}
+	return tempBatch.Writer, tempBatch.Reader, tempBatch.Close, nil
 }
 
 // CatFileBatchCheck obtains a CatFileBatchCheck for this repository
-func (repo *Repository) CatFileBatchCheck(ctx context.Context) (WriteCloserError, *bufio.Reader, func()) {
+func (repo *Repository) CatFileBatchCheck(ctx context.Context) (WriteCloserError, *bufio.Reader, func(), error) {
-	if repo.checkCancel == nil || repo.checkInUse {
+	if repo.check == nil {
-		log.Debug("Opening temporary cat file batch-check for: %s", repo.Path)
+		var err error
-		return CatFileBatchCheck(ctx, repo.Path)
+		repo.check, err = repo.NewBatchCheck(ctx)
+		if err != nil {
+			return nil, nil, nil, err
+		}
 	}
-	repo.checkInUse = true
+
-	return repo.checkWriter, repo.checkReader, func() {
+	if !repo.checkInUse {
-		repo.checkInUse = false
+		repo.checkInUse = true
+		return repo.check.Writer, repo.check.Reader, func() {
+			repo.checkInUse = false
+		}, nil
 	}
+
+	log.Debug("Opening temporary cat file batch-check for: %s", repo.Path)
+	tempBatchCheck, err := repo.NewBatchCheck(ctx)
+	if err != nil {
+		return nil, nil, nil, err
+	}
+	return tempBatchCheck.Writer, tempBatchCheck.Reader, tempBatchCheck.Close, nil
 }
 
 func (repo *Repository) Close() error {
 	if repo == nil {
 		return nil
 	}
-	if repo.batchCancel != nil {
+	if repo.batch != nil {
-		repo.batchCancel()
+		repo.batch.Close()
-		repo.batchReader = nil
+		repo.batch = nil
-		repo.batchWriter = nil
-		repo.batchCancel = nil
 		repo.batchInUse = false
 	}
-	if repo.checkCancel != nil {
+	if repo.check != nil {
-		repo.checkCancel()
+		repo.check.Close()
-		repo.checkCancel = nil
+		repo.check = nil
-		repo.checkReader = nil
-		repo.checkWriter = nil
 		repo.checkInUse = false
 	}
 	repo.LastCommitCache = nil

modules/git/repo_branch_nogogit.go

@@ -22,9 +22,13 @@ func (repo *Repository) IsObjectExist(name string) bool {
 		return false
 	}
 
-	wr, rd, cancel := repo.CatFileBatchCheck(repo.Ctx)
+	wr, rd, cancel, err := repo.CatFileBatchCheck(repo.Ctx)
+	if err != nil {
+		log.Debug("Error writing to CatFileBatchCheck %v", err)
+		return false
+	}
 	defer cancel()
-	_, err := wr.Write([]byte(name + "\n"))
+	_, err = wr.Write([]byte(name + "\n"))
 	if err != nil {
 		log.Debug("Error writing to CatFileBatchCheck %v", err)
 		return false
@@ -39,9 +43,13 @@ func (repo *Repository) IsReferenceExist(name string) bool {
 		return false
 	}
 
-	wr, rd, cancel := repo.CatFileBatchCheck(repo.Ctx)
+	wr, rd, cancel, err := repo.CatFileBatchCheck(repo.Ctx)
+	if err != nil {
+		log.Debug("Error writing to CatFileBatchCheck %v", err)
+		return false
+	}
 	defer cancel()
-	_, err := wr.Write([]byte(name + "\n"))
+	_, err = wr.Write([]byte(name + "\n"))
 	if err != nil {
 		log.Debug("Error writing to CatFileBatchCheck %v", err)
 		return false

modules/git/repo_commit_nogogit.go

@@ -33,9 +33,12 @@ func (repo *Repository) ResolveReference(name string) (string, error) {
 
 // GetRefCommitID returns the last commit ID string of given reference (branch or tag).
 func (repo *Repository) GetRefCommitID(name string) (string, error) {
-	wr, rd, cancel := repo.CatFileBatchCheck(repo.Ctx)
+	wr, rd, cancel, err := repo.CatFileBatchCheck(repo.Ctx)
+	if err != nil {
+		return "", err
+	}
 	defer cancel()
-	_, err := wr.Write([]byte(name + "\n"))
+	_, err = wr.Write([]byte(name + "\n"))
 	if err != nil {
 		return "", err
 	}
@@ -61,12 +64,19 @@ func (repo *Repository) RemoveReference(name string) error {
 
 // IsCommitExist returns true if given commit exists in current repository.
 func (repo *Repository) IsCommitExist(name string) bool {
+	if err := ensureValidGitRepository(repo.Ctx, repo.Path); err != nil {
+		log.Error("IsCommitExist: %v", err)
+		return false
+	}
 	_, _, err := NewCommand(repo.Ctx, "cat-file", "-e").AddDynamicArguments(name).RunStdString(&RunOpts{Dir: repo.Path})
 	return err == nil
 }
 
 func (repo *Repository) getCommit(id ObjectID) (*Commit, error) {
-	wr, rd, cancel := repo.CatFileBatch(repo.Ctx)
+	wr, rd, cancel, err := repo.CatFileBatch(repo.Ctx)
+	if err != nil {
+		return nil, err
+	}
 	defer cancel()
 
 	_, _ = wr.Write([]byte(id.String() + "\n"))
@@ -143,7 +153,10 @@ func (repo *Repository) ConvertToGitID(commitID string) (ObjectID, error) {
 		}
 	}
 
-	wr, rd, cancel := repo.CatFileBatchCheck(repo.Ctx)
+	wr, rd, cancel, err := repo.CatFileBatchCheck(repo.Ctx)
+	if err != nil {
+		return nil, err
+	}
 	defer cancel()
 	_, err = wr.Write([]byte(commitID + "\n"))
 	if err != nil {

modules/git/repo_language_stats_nogogit.go

@@ -20,7 +20,10 @@ import (
 func (repo *Repository) GetLanguageStats(commitID string) (map[string]int64, error) {
 	// We will feed the commit IDs in order into cat-file --batch, followed by blobs as necessary.
 	// so let's create a batch stdin and stdout
-	batchStdinWriter, batchReader, cancel := repo.CatFileBatch(repo.Ctx)
+	batchStdinWriter, batchReader, cancel, err := repo.CatFileBatch(repo.Ctx)
+	if err != nil {
+		return nil, err
+	}
 	defer cancel()
 
 	writeID := func(id string) error {

modules/git/repo_tag_nogogit.go

@@ -31,9 +31,12 @@ func (repo *Repository) GetTags(skip, limit int) (tags []string, err error) {
 
 // GetTagType gets the type of the tag, either commit (simple) or tag (annotated)
 func (repo *Repository) GetTagType(id ObjectID) (string, error) {
-	wr, rd, cancel := repo.CatFileBatchCheck(repo.Ctx)
+	wr, rd, cancel, err := repo.CatFileBatchCheck(repo.Ctx)
+	if err != nil {
+		return "", err
+	}
 	defer cancel()
-	_, err := wr.Write([]byte(id.String() + "\n"))
+	_, err = wr.Write([]byte(id.String() + "\n"))
 	if err != nil {
 		return "", err
 	}
@@ -89,7 +92,10 @@ func (repo *Repository) getTag(tagID ObjectID, name string) (*Tag, error) {
 	}
 
 	// The tag is an annotated tag with a message.
-	wr, rd, cancel := repo.CatFileBatch(repo.Ctx)
+	wr, rd, cancel, err := repo.CatFileBatch(repo.Ctx)
+	if err != nil {
+		return nil, err
+	}
 	defer cancel()
 
 	if _, err := wr.Write([]byte(tagID.String() + "\n")); err != nil {

modules/git/repo_tree_nogogit.go

@@ -10,7 +10,10 @@ import (
 )
 
 func (repo *Repository) getTree(id ObjectID) (*Tree, error) {
-	wr, rd, cancel := repo.CatFileBatch(repo.Ctx)
+	wr, rd, cancel, err := repo.CatFileBatch(repo.Ctx)
+	if err != nil {
+		return nil, err
+	}
 	defer cancel()
 
 	_, _ = wr.Write([]byte(id.String() + "\n"))

modules/git/tree_entry_nogogit.go

@@ -42,9 +42,13 @@ func (te *TreeEntry) Size() int64 {
 		return te.size
 	}
 
-	wr, rd, cancel := te.ptree.repo.CatFileBatchCheck(te.ptree.repo.Ctx)
+	wr, rd, cancel, err := te.ptree.repo.CatFileBatchCheck(te.ptree.repo.Ctx)
+	if err != nil {
+		log.Debug("error whilst reading size for %s in %s. Error: %v", te.ID.String(), te.ptree.repo.Path, err)
+		return 0
+	}
 	defer cancel()
-	_, err := wr.Write([]byte(te.ID.String() + "\n"))
+	_, err = wr.Write([]byte(te.ID.String() + "\n"))
 	if err != nil {
 		log.Debug("error whilst reading size for %s in %s. Error: %v", te.ID.String(), te.ptree.repo.Path, err)
 		return 0

modules/git/tree_nogogit.go

@@ -33,7 +33,10 @@ func (t *Tree) ListEntries() (Entries, error) {
 	}
 
 	if t.repo != nil {
-		wr, rd, cancel := t.repo.CatFileBatch(t.repo.Ctx)
+		wr, rd, cancel, err := t.repo.CatFileBatch(t.repo.Ctx)
+		if err != nil {
+			return nil, err
+		}
 		defer cancel()
 
 		_, _ = wr.Write([]byte(t.ID.String() + "\n"))

modules/indexer/code/bleve/bleve.go

@@ -16,10 +16,10 @@ import (
 	"code.gitea.io/gitea/modules/analyze"
 	"code.gitea.io/gitea/modules/charset"
 	"code.gitea.io/gitea/modules/git"
+	"code.gitea.io/gitea/modules/gitrepo"
 	"code.gitea.io/gitea/modules/indexer/code/internal"
 	indexer_internal "code.gitea.io/gitea/modules/indexer/internal"
 	inner_bleve "code.gitea.io/gitea/modules/indexer/internal/bleve"
-	"code.gitea.io/gitea/modules/log"
 	"code.gitea.io/gitea/modules/setting"
 	"code.gitea.io/gitea/modules/timeutil"
 	"code.gitea.io/gitea/modules/typesniffer"
@@ -189,21 +189,23 @@ func (b *Indexer) addDelete(filename string, repo *repo_model.Repository, batch
 func (b *Indexer) Index(ctx context.Context, repo *repo_model.Repository, sha string, changes *internal.RepoChanges) error {
 	batch := inner_bleve.NewFlushingBatch(b.inner.Indexer, maxBatchSize)
 	if len(changes.Updates) > 0 {
-		// Now because of some insanity with git cat-file not immediately failing if not run in a valid git directory we need to run git rev-parse first!
+		r, err := gitrepo.OpenRepository(ctx, repo)
-		if err := git.EnsureValidGitRepository(ctx, repo.RepoPath()); err != nil {
+		if err != nil {
-			log.Error("Unable to open git repo: %s for %-v: %v", repo.RepoPath(), repo, err)
 			return err
 		}
-
+		defer r.Close()
-		batchWriter, batchReader, cancel := git.CatFileBatch(ctx, repo.RepoPath())
+		gitBatch, err := r.NewBatch(ctx)
-		defer cancel()
+		if err != nil {
+			return err
+		}
+		defer gitBatch.Close()
 
 		for _, update := range changes.Updates {
-			if err := b.addUpdate(ctx, batchWriter, batchReader, sha, update, repo, batch); err != nil {
+			if err := b.addUpdate(ctx, gitBatch.Writer, gitBatch.Reader, sha, update, repo, batch); err != nil {
 				return err
 			}
 		}
-		cancel()
+		gitBatch.Close()
 	}
 	for _, filename := range changes.RemovedFilenames {
 		if err := b.addDelete(filename, repo, batch); err != nil {

modules/indexer/code/elasticsearch/elasticsearch.go

@@ -15,11 +15,11 @@ import (
 	"code.gitea.io/gitea/modules/analyze"
 	"code.gitea.io/gitea/modules/charset"
 	"code.gitea.io/gitea/modules/git"
+	"code.gitea.io/gitea/modules/gitrepo"
 	"code.gitea.io/gitea/modules/indexer/code/internal"
 	indexer_internal "code.gitea.io/gitea/modules/indexer/internal"
 	inner_elasticsearch "code.gitea.io/gitea/modules/indexer/internal/elasticsearch"
 	"code.gitea.io/gitea/modules/json"
-	"code.gitea.io/gitea/modules/log"
 	"code.gitea.io/gitea/modules/setting"
 	"code.gitea.io/gitea/modules/timeutil"
 	"code.gitea.io/gitea/modules/typesniffer"
@@ -154,17 +154,19 @@ func (b *Indexer) addDelete(filename string, repo *repo_model.Repository) elasti
 func (b *Indexer) Index(ctx context.Context, repo *repo_model.Repository, sha string, changes *internal.RepoChanges) error {
 	reqs := make([]elastic.BulkableRequest, 0)
 	if len(changes.Updates) > 0 {
-		// Now because of some insanity with git cat-file not immediately failing if not run in a valid git directory we need to run git rev-parse first!
+		r, err := gitrepo.OpenRepository(ctx, repo)
-		if err := git.EnsureValidGitRepository(ctx, repo.RepoPath()); err != nil {
+		if err != nil {
-			log.Error("Unable to open git repo: %s for %-v: %v", repo.RepoPath(), repo, err)
 			return err
 		}
-
+		defer r.Close()
-		batchWriter, batchReader, cancel := git.CatFileBatch(ctx, repo.RepoPath())
+		batch, err := r.NewBatch(ctx)
-		defer cancel()
+		if err != nil {
+			return err
+		}
+		defer batch.Close()
 
 		for _, update := range changes.Updates {
-			updateReqs, err := b.addUpdate(ctx, batchWriter, batchReader, sha, update, repo)
+			updateReqs, err := b.addUpdate(ctx, batch.Writer, batch.Reader, sha, update, repo)
 			if err != nil {
 				return err
 			}
@@ -172,7 +174,7 @@ func (b *Indexer) Index(ctx context.Context, repo *repo_model.Repository, sha st
 				reqs = append(reqs, updateReqs...)
 			}
 		}
-		cancel()
+		batch.Close()
 	}
 
 	for _, filename := range changes.RemovedFilenames {