[CHORE] Drop go-git support

See https://codeberg.org/forgejo/discussions/issues/164 for the
rationale and discussion of this change.

Everything related to the `go-git` dependency is dropped (Only a single
instance is left in a test file to test for an XSS, it requires crafting
an commit that Git itself refuses to craft). `_gogit` files have
been removed entirely, `go:build: !gogit` is removed, `XXX_nogogit.go` files
either have been renamed or had their code being merged into the
`XXX.go` file.

Makefile

@@ -836,9 +836,6 @@ $(DIST_DIRS):
 .PHONY: release-windows
 release-windows: | $(DIST_DIRS)
 	CGO_CFLAGS="$(CGO_CFLAGS)" $(GO) run $(XGO_PACKAGE) -go $(XGO_VERSION) -buildmode exe -dest $(DIST)/binaries -tags 'osusergo $(TAGS)' -ldflags '-linkmode external -extldflags "-static" $(LDFLAGS)' -targets 'windows/*' -out gitea-$(VERSION) .
-ifeq (,$(findstring gogit,$(TAGS)))
-	CGO_CFLAGS="$(CGO_CFLAGS)" $(GO) run $(XGO_PACKAGE) -go $(XGO_VERSION) -buildmode exe -dest $(DIST)/binaries -tags 'osusergo gogit $(TAGS)' -ldflags '-linkmode external -extldflags "-static" $(LDFLAGS)' -targets 'windows/*' -out gitea-$(VERSION)-gogit .
-endif
 
 .PHONY: release-linux
 release-linux: | $(DIST_DIRS)

go.mod

@@ -31,7 +31,6 @@ require (
 	github.com/dustin/go-humanize v1.0.1
 	github.com/editorconfig/editorconfig-core-go/v2 v2.6.2
 	github.com/emersion/go-imap v1.2.1
-	github.com/emirpasic/gods v1.18.1
 	github.com/felixge/fgprof v0.9.4
 	github.com/fsnotify/fsnotify v1.7.0
 	github.com/gliderlabs/ssh v0.3.7
@@ -42,7 +41,6 @@ require (
 	github.com/go-co-op/gocron v1.37.0
 	github.com/go-enry/go-enry/v2 v2.8.8
 	github.com/go-fed/httpsig v1.1.1-0.20201223112313-55836744818e
-	github.com/go-git/go-billy/v5 v5.5.0
 	github.com/go-git/go-git/v5 v5.11.0
 	github.com/go-ldap/ldap/v3 v3.4.6
 	github.com/go-sql-driver/mysql v1.8.1
@@ -172,6 +170,7 @@ require (
 	github.com/dgryski/go-rendezvous v0.0.0-20200823014737-9f7001d12a5f // indirect
 	github.com/dlclark/regexp2 v1.11.0 // indirect
 	github.com/emersion/go-sasl v0.0.0-20231106173351-e73c9f7bad43 // indirect
+	github.com/emirpasic/gods v1.18.1 // indirect
 	github.com/fatih/color v1.16.0 // indirect
 	github.com/felixge/httpsnoop v1.0.4 // indirect
 	github.com/fxamacker/cbor/v2 v2.5.0 // indirect
@@ -181,6 +180,7 @@ require (
 	github.com/go-faster/city v1.0.1 // indirect
 	github.com/go-faster/errors v0.7.1 // indirect
 	github.com/go-git/gcfg v1.5.1-0.20230307220236-3a3c6141e376 // indirect
+	github.com/go-git/go-billy/v5 v5.5.0 // indirect
 	github.com/go-ini/ini v1.67.0 // indirect
 	github.com/go-openapi/analysis v0.22.2 // indirect
 	github.com/go-openapi/errors v0.21.0 // indirect

modules/git/blob.go

@@ -5,15 +5,119 @@
 package git
 
 import (
+	"bufio"
 	"bytes"
 	"encoding/base64"
 	"io"
 
+	"code.gitea.io/gitea/modules/log"
 	"code.gitea.io/gitea/modules/typesniffer"
 	"code.gitea.io/gitea/modules/util"
 )
 
-// This file contains common functions between the gogit and !gogit variants for git Blobs
+// Blob represents a Git object.
+type Blob struct {
+	ID ObjectID
+
+	gotSize bool
+	size    int64
+	name    string
+	repo    *Repository
+}
+
+// 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)
+
+	_, err := wr.Write([]byte(b.ID.String() + "\n"))
+	if err != nil {
+		cancel()
+		return nil, err
+	}
+	_, _, size, err := ReadBatchLine(rd)
+	if err != nil {
+		cancel()
+		return nil, err
+	}
+	b.gotSize = true
+	b.size = size
+
+	if size < 4096 {
+		bs, err := io.ReadAll(io.LimitReader(rd, size))
+		defer cancel()
+		if err != nil {
+			return nil, err
+		}
+		_, err = rd.Discard(1)
+		return io.NopCloser(bytes.NewReader(bs)), err
+	}
+
+	return &blobReader{
+		rd:     rd,
+		n:      size,
+		cancel: cancel,
+	}, nil
+}
+
+// Size returns the uncompressed size of the blob
+func (b *Blob) Size() int64 {
+	if b.gotSize {
+		return b.size
+	}
+
+	wr, rd, cancel := b.repo.CatFileBatchCheck(b.repo.Ctx)
+	defer cancel()
+	_, 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
+	}
+	_, _, b.size, err = ReadBatchLine(rd)
+	if err != nil {
+		log.Debug("error whilst reading size for %s in %s. Error: %v", b.ID.String(), b.repo.Path, err)
+		return 0
+	}
+
+	b.gotSize = true
+
+	return b.size
+}
+
+type blobReader struct {
+	rd     *bufio.Reader
+	n      int64
+	cancel func()
+}
+
+func (b *blobReader) Read(p []byte) (n int, err error) {
+	if b.n <= 0 {
+		return 0, io.EOF
+	}
+	if int64(len(p)) > b.n {
+		p = p[0:b.n]
+	}
+	n, err = b.rd.Read(p)
+	b.n -= int64(n)
+	return n, err
+}
+
+// Close implements io.Closer
+func (b *blobReader) Close() error {
+	if b.rd == nil {
+		return nil
+	}
+
+	defer b.cancel()
+
+	if err := DiscardFull(b.rd, b.n+1); err != nil {
+		return err
+	}
+
+	b.rd = nil
+
+	return nil
+}
 
 // Name returns name of the tree entry this blob object was created from (or empty string)
 func (b *Blob) Name() string {
@@ -100,3 +204,18 @@ func (b *Blob) GuessContentType() (typesniffer.SniffedType, error) {
 
 	return typesniffer.DetectContentTypeFromReader(r)
 }
+
+// GetBlob finds the blob object in the repository.
+func (repo *Repository) GetBlob(idStr string) (*Blob, error) {
+	id, err := NewIDFromString(idStr)
+	if err != nil {
+		return nil, err
+	}
+	if id.IsZero() {
+		return nil, ErrNotExist{id.String(), ""}
+	}
+	return &Blob{
+		ID:   id,
+		repo: repo,
+	}, nil
+}

modules/git/blob_gogit.go

@@ -1,32 +0,0 @@
-// Copyright 2015 The Gogs Authors. All rights reserved.
-// Copyright 2019 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build gogit
-
-package git
-
-import (
-	"io"
-
-	"github.com/go-git/go-git/v5/plumbing"
-)
-
-// Blob represents a Git object.
-type Blob struct {
-	ID ObjectID
-
-	gogitEncodedObj plumbing.EncodedObject
-	name            string
-}
-
-// 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) {
-	return b.gogitEncodedObj.Reader()
-}
-
-// Size returns the uncompressed size of the blob
-func (b *Blob) Size() int64 {
-	return b.gogitEncodedObj.Size()
-}

modules/git/blob_nogogit.go

@@ -1,118 +0,0 @@
-// Copyright 2020 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build !gogit
-
-package git
-
-import (
-	"bufio"
-	"bytes"
-	"io"
-
-	"code.gitea.io/gitea/modules/log"
-)
-
-// Blob represents a Git object.
-type Blob struct {
-	ID ObjectID
-
-	gotSize bool
-	size    int64
-	name    string
-	repo    *Repository
-}
-
-// 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)
-
-	_, err := wr.Write([]byte(b.ID.String() + "\n"))
-	if err != nil {
-		cancel()
-		return nil, err
-	}
-	_, _, size, err := ReadBatchLine(rd)
-	if err != nil {
-		cancel()
-		return nil, err
-	}
-	b.gotSize = true
-	b.size = size
-
-	if size < 4096 {
-		bs, err := io.ReadAll(io.LimitReader(rd, size))
-		defer cancel()
-		if err != nil {
-			return nil, err
-		}
-		_, err = rd.Discard(1)
-		return io.NopCloser(bytes.NewReader(bs)), err
-	}
-
-	return &blobReader{
-		rd:     rd,
-		n:      size,
-		cancel: cancel,
-	}, nil
-}
-
-// Size returns the uncompressed size of the blob
-func (b *Blob) Size() int64 {
-	if b.gotSize {
-		return b.size
-	}
-
-	wr, rd, cancel := b.repo.CatFileBatchCheck(b.repo.Ctx)
-	defer cancel()
-	_, 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
-	}
-	_, _, b.size, err = ReadBatchLine(rd)
-	if err != nil {
-		log.Debug("error whilst reading size for %s in %s. Error: %v", b.ID.String(), b.repo.Path, err)
-		return 0
-	}
-
-	b.gotSize = true
-
-	return b.size
-}
-
-type blobReader struct {
-	rd     *bufio.Reader
-	n      int64
-	cancel func()
-}
-
-func (b *blobReader) Read(p []byte) (n int, err error) {
-	if b.n <= 0 {
-		return 0, io.EOF
-	}
-	if int64(len(p)) > b.n {
-		p = p[0:b.n]
-	}
-	n, err = b.rd.Read(p)
-	b.n -= int64(n)
-	return n, err
-}
-
-// Close implements io.Closer
-func (b *blobReader) Close() error {
-	if b.rd == nil {
-		return nil
-	}
-
-	defer b.cancel()
-
-	if err := DiscardFull(b.rd, b.n+1); err != nil {
-		return err
-	}
-
-	b.rd = nil
-
-	return nil
-}

modules/git/commit_convert_gogit.go

@@ -1,75 +0,0 @@
-// Copyright 2015 The Gogs Authors. All rights reserved.
-// Copyright 2018 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build gogit
-
-package git
-
-import (
-	"fmt"
-	"strings"
-
-	"github.com/go-git/go-git/v5/plumbing/object"
-)
-
-func convertPGPSignature(c *object.Commit) *ObjectSignature {
-	if c.PGPSignature == "" {
-		return nil
-	}
-
-	var w strings.Builder
-	var err error
-
-	if _, err = fmt.Fprintf(&w, "tree %s\n", c.TreeHash.String()); err != nil {
-		return nil
-	}
-
-	for _, parent := range c.ParentHashes {
-		if _, err = fmt.Fprintf(&w, "parent %s\n", parent.String()); err != nil {
-			return nil
-		}
-	}
-
-	if _, err = fmt.Fprint(&w, "author "); err != nil {
-		return nil
-	}
-
-	if err = c.Author.Encode(&w); err != nil {
-		return nil
-	}
-
-	if _, err = fmt.Fprint(&w, "\ncommitter "); err != nil {
-		return nil
-	}
-
-	if err = c.Committer.Encode(&w); err != nil {
-		return nil
-	}
-
-	if c.Encoding != "" && c.Encoding != "UTF-8" {
-		if _, err = fmt.Fprintf(&w, "\nencoding %s\n", c.Encoding); err != nil {
-			return nil
-		}
-	}
-
-	if _, err = fmt.Fprintf(&w, "\n\n%s", c.Message); err != nil {
-		return nil
-	}
-
-	return &ObjectSignature{
-		Signature: c.PGPSignature,
-		Payload:   w.String(),
-	}
-}
-
-func convertCommit(c *object.Commit) *Commit {
-	return &Commit{
-		ID:            ParseGogitHash(c.Hash),
-		CommitMessage: c.Message,
-		Committer:     &c.Committer,
-		Author:        &c.Author,
-		Signature:     convertPGPSignature(c),
-		Parents:       ParseGogitHashArray(c.ParentHashes),
-	}
-}

modules/git/commit_info.go

@@ -3,9 +3,173 @@
 
 package git
 
+import (
+	"context"
+	"fmt"
+	"io"
+	"path"
+	"sort"
+
+	"code.gitea.io/gitea/modules/log"
+)
+
 // CommitInfo describes the first commit with the provided entry
 type CommitInfo struct {
 	Entry         *TreeEntry
 	Commit        *Commit
 	SubModuleFile *SubModuleFile
 }
+
+// GetCommitsInfo gets information of all commits that are corresponding to these entries
+func (tes Entries) GetCommitsInfo(ctx context.Context, commit *Commit, treePath string) ([]CommitInfo, *Commit, error) {
+	entryPaths := make([]string, len(tes)+1)
+	// Get the commit for the treePath itself
+	entryPaths[0] = ""
+	for i, entry := range tes {
+		entryPaths[i+1] = entry.Name()
+	}
+
+	var err error
+
+	var revs map[string]*Commit
+	if commit.repo.LastCommitCache != nil {
+		var unHitPaths []string
+		revs, unHitPaths, err = getLastCommitForPathsByCache(commit.ID.String(), treePath, entryPaths, commit.repo.LastCommitCache)
+		if err != nil {
+			return nil, nil, err
+		}
+		if len(unHitPaths) > 0 {
+			sort.Strings(unHitPaths)
+			commits, err := GetLastCommitForPaths(ctx, commit, treePath, unHitPaths)
+			if err != nil {
+				return nil, nil, err
+			}
+
+			for pth, found := range commits {
+				revs[pth] = found
+			}
+		}
+	} else {
+		sort.Strings(entryPaths)
+		revs, err = GetLastCommitForPaths(ctx, commit, treePath, entryPaths)
+	}
+	if err != nil {
+		return nil, nil, err
+	}
+
+	commitsInfo := make([]CommitInfo, len(tes))
+	for i, entry := range tes {
+		commitsInfo[i] = CommitInfo{
+			Entry: entry,
+		}
+
+		// Check if we have found a commit for this entry in time
+		if entryCommit, ok := revs[entry.Name()]; ok {
+			commitsInfo[i].Commit = entryCommit
+		} else {
+			log.Debug("missing commit for %s", entry.Name())
+		}
+
+		// If the entry if a submodule add a submodule file for this
+		if entry.IsSubModule() {
+			subModuleURL := ""
+			var fullPath string
+			if len(treePath) > 0 {
+				fullPath = treePath + "/" + entry.Name()
+			} else {
+				fullPath = entry.Name()
+			}
+			if subModule, err := commit.GetSubModule(fullPath); err != nil {
+				return nil, nil, err
+			} else if subModule != nil {
+				subModuleURL = subModule.URL
+			}
+			subModuleFile := NewSubModuleFile(commitsInfo[i].Commit, subModuleURL, entry.ID.String())
+			commitsInfo[i].SubModuleFile = subModuleFile
+		}
+	}
+
+	// Retrieve the commit for the treePath itself (see above). We basically
+	// get it for free during the tree traversal and it's used for listing
+	// pages to display information about newest commit for a given path.
+	var treeCommit *Commit
+	var ok bool
+	if treePath == "" {
+		treeCommit = commit
+	} else if treeCommit, ok = revs[""]; ok {
+		treeCommit.repo = commit.repo
+	}
+	return commitsInfo, treeCommit, nil
+}
+
+func getLastCommitForPathsByCache(commitID, treePath string, paths []string, cache *LastCommitCache) (map[string]*Commit, []string, error) {
+	var unHitEntryPaths []string
+	results := make(map[string]*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
+}
+
+// GetLastCommitForPaths returns last commit information
+func GetLastCommitForPaths(ctx context.Context, commit *Commit, treePath string, paths []string) (map[string]*Commit, error) {
+	// We read backwards from the commit to obtain all of the commits
+	revs, err := WalkGitLog(ctx, commit.repo, commit, treePath, paths...)
+	if err != nil {
+		return nil, err
+	}
+
+	batchStdinWriter, batchReader, cancel := commit.repo.CatFileBatch(ctx)
+	defer cancel()
+
+	commitsMap := map[string]*Commit{}
+	commitsMap[commit.ID.String()] = commit
+
+	commitCommits := map[string]*Commit{}
+	for path, commitID := range revs {
+		c, ok := commitsMap[commitID]
+		if ok {
+			commitCommits[path] = c
+			continue
+		}
+
+		if len(commitID) == 0 {
+			continue
+		}
+
+		_, err := batchStdinWriter.Write([]byte(commitID + "\n"))
+		if err != nil {
+			return nil, err
+		}
+		_, typ, size, err := ReadBatchLine(batchReader)
+		if err != nil {
+			return nil, err
+		}
+		if typ != "commit" {
+			if err := DiscardFull(batchReader, size+1); err != nil {
+				return nil, err
+			}
+			return nil, fmt.Errorf("unexpected type: %s for commit id: %s", typ, commitID)
+		}
+		c, err = CommitFromReader(commit.repo, MustIDFromString(commitID), io.LimitReader(batchReader, size))
+		if err != nil {
+			return nil, err
+		}
+		if _, err := batchReader.Discard(1); err != nil {
+			return nil, err
+		}
+		commitCommits[path] = c
+	}
+
+	return commitCommits, nil
+}

modules/git/commit_info_gogit.go

@@ -1,304 +0,0 @@
-// Copyright 2017 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build gogit
-
-package git
-
-import (
-	"context"
-	"path"
-
-	"github.com/emirpasic/gods/trees/binaryheap"
-	"github.com/go-git/go-git/v5/plumbing"
-	"github.com/go-git/go-git/v5/plumbing/object"
-	cgobject "github.com/go-git/go-git/v5/plumbing/object/commitgraph"
-)
-
-// GetCommitsInfo gets information of all commits that are corresponding to these entries
-func (tes Entries) GetCommitsInfo(ctx context.Context, commit *Commit, treePath string) ([]CommitInfo, *Commit, error) {
-	entryPaths := make([]string, len(tes)+1)
-	// Get the commit for the treePath itself
-	entryPaths[0] = ""
-	for i, entry := range tes {
-		entryPaths[i+1] = entry.Name()
-	}
-
-	commitNodeIndex, commitGraphFile := commit.repo.CommitNodeIndex()
-	if commitGraphFile != nil {
-		defer commitGraphFile.Close()
-	}
-
-	c, err := commitNodeIndex.Get(plumbing.Hash(commit.ID.RawValue()))
-	if err != nil {
-		return nil, nil, err
-	}
-
-	var revs map[string]*Commit
-	if commit.repo.LastCommitCache != nil {
-		var unHitPaths []string
-		revs, unHitPaths, err = getLastCommitForPathsByCache(commit.ID.String(), treePath, entryPaths, commit.repo.LastCommitCache)
-		if err != nil {
-			return nil, nil, err
-		}
-		if len(unHitPaths) > 0 {
-			revs2, err := GetLastCommitForPaths(ctx, commit.repo.LastCommitCache, c, treePath, unHitPaths)
-			if err != nil {
-				return nil, nil, err
-			}
-
-			for k, v := range revs2 {
-				revs[k] = v
-			}
-		}
-	} else {
-		revs, err = GetLastCommitForPaths(ctx, nil, c, treePath, entryPaths)
-	}
-	if err != nil {
-		return nil, nil, err
-	}
-
-	commit.repo.gogitStorage.Close()
-
-	commitsInfo := make([]CommitInfo, len(tes))
-	for i, entry := range tes {
-		commitsInfo[i] = CommitInfo{
-			Entry: entry,
-		}
-
-		// Check if we have found a commit for this entry in time
-		if entryCommit, ok := revs[entry.Name()]; ok {
-			commitsInfo[i].Commit = entryCommit
-		}
-
-		// If the entry if a submodule add a submodule file for this
-		if entry.IsSubModule() {
-			subModuleURL := ""
-			var fullPath string
-			if len(treePath) > 0 {
-				fullPath = treePath + "/" + entry.Name()
-			} else {
-				fullPath = entry.Name()
-			}
-			if subModule, err := commit.GetSubModule(fullPath); err != nil {
-				return nil, nil, err
-			} else if subModule != nil {
-				subModuleURL = subModule.URL
-			}
-			subModuleFile := NewSubModuleFile(commitsInfo[i].Commit, subModuleURL, entry.ID.String())
-			commitsInfo[i].SubModuleFile = subModuleFile
-		}
-	}
-
-	// Retrieve the commit for the treePath itself (see above). We basically
-	// get it for free during the tree traversal and it's used for listing
-	// pages to display information about newest commit for a given path.
-	var treeCommit *Commit
-	var ok bool
-	if treePath == "" {
-		treeCommit = commit
-	} else if treeCommit, ok = revs[""]; ok {
-		treeCommit.repo = commit.repo
-	}
-	return commitsInfo, treeCommit, nil
-}
-
-type commitAndPaths struct {
-	commit cgobject.CommitNode
-	// Paths that are still on the branch represented by commit
-	paths []string
-	// Set of hashes for the paths
-	hashes map[string]plumbing.Hash
-}
-
-func getCommitTree(c cgobject.CommitNode, treePath string) (*object.Tree, error) {
-	tree, err := c.Tree()
-	if err != nil {
-		return nil, err
-	}
-
-	// Optimize deep traversals by focusing only on the specific tree
-	if treePath != "" {
-		tree, err = tree.Tree(treePath)
-		if err != nil {
-			return nil, err
-		}
-	}
-
-	return tree, nil
-}
-
-func getFileHashes(c cgobject.CommitNode, treePath string, paths []string) (map[string]plumbing.Hash, error) {
-	tree, err := getCommitTree(c, treePath)
-	if err == object.ErrDirectoryNotFound {
-		// The whole tree didn't exist, so return empty map
-		return make(map[string]plumbing.Hash), nil
-	}
-	if err != nil {
-		return nil, err
-	}
-
-	hashes := make(map[string]plumbing.Hash)
-	for _, path := range paths {
-		if path != "" {
-			entry, err := tree.FindEntry(path)
-			if err == nil {
-				hashes[path] = entry.Hash
-			}
-		} else {
-			hashes[path] = tree.Hash
-		}
-	}
-
-	return hashes, nil
-}
-
-func getLastCommitForPathsByCache(commitID, treePath string, paths []string, cache *LastCommitCache) (map[string]*Commit, []string, error) {
-	var unHitEntryPaths []string
-	results := make(map[string]*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
-}
-
-// GetLastCommitForPaths returns last commit information
-func GetLastCommitForPaths(ctx context.Context, cache *LastCommitCache, c cgobject.CommitNode, treePath string, paths []string) (map[string]*Commit, error) {
-	refSha := c.ID().String()
-
-	// We do a tree traversal with nodes sorted by commit time
-	heap := binaryheap.NewWith(func(a, b any) int {
-		if a.(*commitAndPaths).commit.CommitTime().Before(b.(*commitAndPaths).commit.CommitTime()) {
-			return 1
-		}
-		return -1
-	})
-
-	resultNodes := make(map[string]cgobject.CommitNode)
-	initialHashes, err := getFileHashes(c, treePath, paths)
-	if err != nil {
-		return nil, err
-	}
-
-	// Start search from the root commit and with full set of paths
-	heap.Push(&commitAndPaths{c, paths, initialHashes})
-heaploop:
-	for {
-		select {
-		case <-ctx.Done():
-			if ctx.Err() == context.DeadlineExceeded {
-				break heaploop
-			}
-			return nil, ctx.Err()
-		default:
-		}
-		cIn, ok := heap.Pop()
-		if !ok {
-			break
-		}
-		current := cIn.(*commitAndPaths)
-
-		// Load the parent commits for the one we are currently examining
-		numParents := current.commit.NumParents()
-		var parents []cgobject.CommitNode
-		for i := 0; i < numParents; i++ {
-			parent, err := current.commit.ParentNode(i)
-			if err != nil {
-				break
-			}
-			parents = append(parents, parent)
-		}
-
-		// Examine the current commit and set of interesting paths
-		pathUnchanged := make([]bool, len(current.paths))
-		parentHashes := make([]map[string]plumbing.Hash, len(parents))
-		for j, parent := range parents {
-			parentHashes[j], err = getFileHashes(parent, treePath, current.paths)
-			if err != nil {
-				break
-			}
-
-			for i, path := range current.paths {
-				if parentHashes[j][path] == current.hashes[path] {
-					pathUnchanged[i] = true
-				}
-			}
-		}
-
-		var remainingPaths []string
-		for i, pth := range current.paths {
-			// The results could already contain some newer change for the same path,
-			// so don't override that and bail out on the file early.
-			if resultNodes[pth] == nil {
-				if pathUnchanged[i] {
-					// The path existed with the same hash in at least one parent so it could
-					// not have been changed in this commit directly.
-					remainingPaths = append(remainingPaths, pth)
-				} else {
-					// There are few possible cases how can we get here:
-					// - The path didn't exist in any parent, so it must have been created by
-					//   this commit.
-					// - The path did exist in the parent commit, but the hash of the file has
-					//   changed.
-					// - We are looking at a merge commit and the hash of the file doesn't
-					//   match any of the hashes being merged. This is more common for directories,
-					//   but it can also happen if a file is changed through conflict resolution.
-					resultNodes[pth] = current.commit
-					if err := cache.Put(refSha, path.Join(treePath, pth), current.commit.ID().String()); err != nil {
-						return nil, err
-					}
-				}
-			}
-		}
-
-		if len(remainingPaths) > 0 {
-			// Add the parent nodes along with remaining paths to the heap for further
-			// processing.
-			for j, parent := range parents {
-				// Combine remainingPath with paths available on the parent branch
-				// and make union of them
-				remainingPathsForParent := make([]string, 0, len(remainingPaths))
-				newRemainingPaths := make([]string, 0, len(remainingPaths))
-				for _, path := range remainingPaths {
-					if parentHashes[j][path] == current.hashes[path] {
-						remainingPathsForParent = append(remainingPathsForParent, path)
-					} else {
-						newRemainingPaths = append(newRemainingPaths, path)
-					}
-				}
-
-				if remainingPathsForParent != nil {
-					heap.Push(&commitAndPaths{parent, remainingPathsForParent, parentHashes[j]})
-				}
-
-				if len(newRemainingPaths) == 0 {
-					break
-				} else {
-					remainingPaths = newRemainingPaths
-				}
-			}
-		}
-	}
-
-	// Post-processing
-	result := make(map[string]*Commit)
-	for path, commitNode := range resultNodes {
-		commit, err := commitNode.Commit()
-		if err != nil {
-			return nil, err
-		}
-		result[path] = convertCommit(commit)
-	}
-
-	return result, nil
-}

modules/git/commit_info_nogogit.go

@@ -1,170 +0,0 @@
-// Copyright 2017 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build !gogit
-
-package git
-
-import (
-	"context"
-	"fmt"
-	"io"
-	"path"
-	"sort"
-
-	"code.gitea.io/gitea/modules/log"
-)
-
-// GetCommitsInfo gets information of all commits that are corresponding to these entries
-func (tes Entries) GetCommitsInfo(ctx context.Context, commit *Commit, treePath string) ([]CommitInfo, *Commit, error) {
-	entryPaths := make([]string, len(tes)+1)
-	// Get the commit for the treePath itself
-	entryPaths[0] = ""
-	for i, entry := range tes {
-		entryPaths[i+1] = entry.Name()
-	}
-
-	var err error
-
-	var revs map[string]*Commit
-	if commit.repo.LastCommitCache != nil {
-		var unHitPaths []string
-		revs, unHitPaths, err = getLastCommitForPathsByCache(commit.ID.String(), treePath, entryPaths, commit.repo.LastCommitCache)
-		if err != nil {
-			return nil, nil, err
-		}
-		if len(unHitPaths) > 0 {
-			sort.Strings(unHitPaths)
-			commits, err := GetLastCommitForPaths(ctx, commit, treePath, unHitPaths)
-			if err != nil {
-				return nil, nil, err
-			}
-
-			for pth, found := range commits {
-				revs[pth] = found
-			}
-		}
-	} else {
-		sort.Strings(entryPaths)
-		revs, err = GetLastCommitForPaths(ctx, commit, treePath, entryPaths)
-	}
-	if err != nil {
-		return nil, nil, err
-	}
-
-	commitsInfo := make([]CommitInfo, len(tes))
-	for i, entry := range tes {
-		commitsInfo[i] = CommitInfo{
-			Entry: entry,
-		}
-
-		// Check if we have found a commit for this entry in time
-		if entryCommit, ok := revs[entry.Name()]; ok {
-			commitsInfo[i].Commit = entryCommit
-		} else {
-			log.Debug("missing commit for %s", entry.Name())
-		}
-
-		// If the entry if a submodule add a submodule file for this
-		if entry.IsSubModule() {
-			subModuleURL := ""
-			var fullPath string
-			if len(treePath) > 0 {
-				fullPath = treePath + "/" + entry.Name()
-			} else {
-				fullPath = entry.Name()
-			}
-			if subModule, err := commit.GetSubModule(fullPath); err != nil {
-				return nil, nil, err
-			} else if subModule != nil {
-				subModuleURL = subModule.URL
-			}
-			subModuleFile := NewSubModuleFile(commitsInfo[i].Commit, subModuleURL, entry.ID.String())
-			commitsInfo[i].SubModuleFile = subModuleFile
-		}
-	}
-
-	// Retrieve the commit for the treePath itself (see above). We basically
-	// get it for free during the tree traversal and it's used for listing
-	// pages to display information about newest commit for a given path.
-	var treeCommit *Commit
-	var ok bool
-	if treePath == "" {
-		treeCommit = commit
-	} else if treeCommit, ok = revs[""]; ok {
-		treeCommit.repo = commit.repo
-	}
-	return commitsInfo, treeCommit, nil
-}
-
-func getLastCommitForPathsByCache(commitID, treePath string, paths []string, cache *LastCommitCache) (map[string]*Commit, []string, error) {
-	var unHitEntryPaths []string
-	results := make(map[string]*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
-}
-
-// GetLastCommitForPaths returns last commit information
-func GetLastCommitForPaths(ctx context.Context, commit *Commit, treePath string, paths []string) (map[string]*Commit, error) {
-	// We read backwards from the commit to obtain all of the commits
-	revs, err := WalkGitLog(ctx, commit.repo, commit, treePath, paths...)
-	if err != nil {
-		return nil, err
-	}
-
-	batchStdinWriter, batchReader, cancel := commit.repo.CatFileBatch(ctx)
-	defer cancel()
-
-	commitsMap := map[string]*Commit{}
-	commitsMap[commit.ID.String()] = commit
-
-	commitCommits := map[string]*Commit{}
-	for path, commitID := range revs {
-		c, ok := commitsMap[commitID]
-		if ok {
-			commitCommits[path] = c
-			continue
-		}
-
-		if len(commitID) == 0 {
-			continue
-		}
-
-		_, err := batchStdinWriter.Write([]byte(commitID + "\n"))
-		if err != nil {
-			return nil, err
-		}
-		_, typ, size, err := ReadBatchLine(batchReader)
-		if err != nil {
-			return nil, err
-		}
-		if typ != "commit" {
-			if err := DiscardFull(batchReader, size+1); err != nil {
-				return nil, err
-			}
-			return nil, fmt.Errorf("unexpected type: %s for commit id: %s", typ, commitID)
-		}
-		c, err = CommitFromReader(commit.repo, MustIDFromString(commitID), io.LimitReader(batchReader, size))
-		if err != nil {
-			return nil, err
-		}
-		if _, err := batchReader.Discard(1); err != nil {
-			return nil, err
-		}
-		commitCommits[path] = c
-	}
-
-	return commitCommits, nil
-}

modules/git/commit_sha256_test.go

@@ -1,8 +1,6 @@
 // Copyright 2023 The Gitea Authors. All rights reserved.
 // SPDX-License-Identifier: MIT
 
-//go:build !gogit
-
 package git
 
 import (

modules/git/git.go

@@ -186,12 +186,12 @@ func InitFull(ctx context.Context) (err error) {
 		globalCommandArgs = append(globalCommandArgs, "-c", "credential.helper=")
 	}
 	SupportProcReceive = CheckGitVersionAtLeast("2.29") == nil
-	SupportHashSha256 = CheckGitVersionAtLeast("2.42") == nil && !isGogit
+	SupportHashSha256 = CheckGitVersionAtLeast("2.42") == nil
 	SupportCheckAttrOnBare = CheckGitVersionAtLeast("2.40") == nil
 	if SupportHashSha256 {
 		SupportedObjectFormats = append(SupportedObjectFormats, Sha256ObjectFormat)
 	} else {
-		log.Warn("sha256 hash support is disabled - requires Git >= 2.42. Gogit is currently unsupported")
+		log.Warn("sha256 hash support is disabled - requires Git >= 2.42")
 	}
 
 	InvertedGitFlushEnv = CheckGitVersionEqual("2.43.1") == nil

modules/git/last_commit_cache.go

@@ -4,6 +4,7 @@
 package git
 
 import (
+	"context"
 	"crypto/sha256"
 	"fmt"
 
@@ -112,3 +113,47 @@ func (c *LastCommitCache) GetCommitByPath(commitID, entryPath string) (*Commit,
 
 	return lastCommit, nil
 }
+
+// CacheCommit will cache the commit from the gitRepository
+func (c *Commit) CacheCommit(ctx context.Context) error {
+	if c.repo.LastCommitCache == nil {
+		return nil
+	}
+	return c.recursiveCache(ctx, &c.Tree, "", 1)
+}
+
+func (c *Commit) recursiveCache(ctx context.Context, tree *Tree, treePath string, level int) error {
+	if level == 0 {
+		return nil
+	}
+
+	entries, err := tree.ListEntries()
+	if err != nil {
+		return err
+	}
+
+	entryPaths := make([]string, len(entries))
+	for i, entry := range entries {
+		entryPaths[i] = entry.Name()
+	}
+
+	_, err = WalkGitLog(ctx, c.repo, c, treePath, entryPaths...)
+	if err != nil {
+		return err
+	}
+
+	for _, treeEntry := range entries {
+		// entryMap won't contain "" therefore skip this.
+		if treeEntry.IsDir() {
+			subTree, err := tree.SubTree(treeEntry.Name())
+			if err != nil {
+				return err
+			}
+			if err := c.recursiveCache(ctx, subTree, treeEntry.Name(), level-1); err != nil {
+				return err
+			}
+		}
+	}
+
+	return nil
+}

modules/git/last_commit_cache_gogit.go

@@ -1,65 +0,0 @@
-// Copyright 2020 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build gogit
-
-package git
-
-import (
-	"context"
-
-	"github.com/go-git/go-git/v5/plumbing"
-	cgobject "github.com/go-git/go-git/v5/plumbing/object/commitgraph"
-)
-
-// CacheCommit will cache the commit from the gitRepository
-func (c *Commit) CacheCommit(ctx context.Context) error {
-	if c.repo.LastCommitCache == nil {
-		return nil
-	}
-	commitNodeIndex, _ := c.repo.CommitNodeIndex()
-
-	index, err := commitNodeIndex.Get(plumbing.Hash(c.ID.RawValue()))
-	if err != nil {
-		return err
-	}
-
-	return c.recursiveCache(ctx, index, &c.Tree, "", 1)
-}
-
-func (c *Commit) recursiveCache(ctx context.Context, index cgobject.CommitNode, tree *Tree, treePath string, 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]*TreeEntry)
-	for i, entry := range entries {
-		entryPaths[i] = entry.Name()
-		entryMap[entry.Name()] = entry
-	}
-
-	commits, err := GetLastCommitForPaths(ctx, c.repo.LastCommitCache, index, treePath, entryPaths)
-	if err != nil {
-		return err
-	}
-
-	for entry := range commits {
-		if entryMap[entry].IsDir() {
-			subTree, err := tree.SubTree(entry)
-			if err != nil {
-				return err
-			}
-			if err := c.recursiveCache(ctx, index, subTree, entry, level-1); err != nil {
-				return err
-			}
-		}
-	}
-
-	return nil
-}

modules/git/last_commit_cache_nogogit.go

@@ -1,54 +0,0 @@
-// Copyright 2020 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build !gogit
-
-package git
-
-import (
-	"context"
-)
-
-// CacheCommit will cache the commit from the gitRepository
-func (c *Commit) CacheCommit(ctx context.Context) error {
-	if c.repo.LastCommitCache == nil {
-		return nil
-	}
-	return c.recursiveCache(ctx, &c.Tree, "", 1)
-}
-
-func (c *Commit) recursiveCache(ctx context.Context, tree *Tree, treePath string, level int) error {
-	if level == 0 {
-		return nil
-	}
-
-	entries, err := tree.ListEntries()
-	if err != nil {
-		return err
-	}
-
-	entryPaths := make([]string, len(entries))
-	for i, entry := range entries {
-		entryPaths[i] = entry.Name()
-	}
-
-	_, err = WalkGitLog(ctx, c.repo, c, treePath, entryPaths...)
-	if err != nil {
-		return err
-	}
-
-	for _, treeEntry := range entries {
-		// entryMap won't contain "" therefore skip this.
-		if treeEntry.IsDir() {
-			subTree, err := tree.SubTree(treeEntry.Name())
-			if err != nil {
-				return err
-			}
-			if err := c.recursiveCache(ctx, subTree, treeEntry.Name(), level-1); err != nil {
-				return err
-			}
-		}
-	}
-
-	return nil
-}

modules/git/notes.go

@@ -3,6 +3,14 @@
 
 package git
 
+import (
+	"context"
+	"io"
+	"strings"
+
+	"code.gitea.io/gitea/modules/log"
+)
+
 // NotesRef is the git ref where Gitea will look for git-notes data.
 // The value ("refs/notes/commits") is the default ref used by git-notes.
 const NotesRef = "refs/notes/commits"
@@ -12,3 +20,80 @@ type Note struct {
 	Message []byte
 	Commit  *Commit
 }
+
+// GetNote retrieves the git-notes data for a given commit.
+// FIXME: Add LastCommitCache support
+func GetNote(ctx context.Context, repo *Repository, commitID string, note *Note) error {
+	log.Trace("Searching for git note corresponding to the commit %q in the repository %q", commitID, repo.Path)
+	notes, err := repo.GetCommit(NotesRef)
+	if err != nil {
+		if IsErrNotExist(err) {
+			return err
+		}
+		log.Error("Unable to get commit from ref %q. Error: %v", NotesRef, err)
+		return err
+	}
+
+	path := ""
+
+	tree := &notes.Tree
+	log.Trace("Found tree with ID %q while searching for git note corresponding to the commit %q", tree.ID, commitID)
+
+	var entry *TreeEntry
+	originalCommitID := commitID
+	for len(commitID) > 2 {
+		entry, err = tree.GetTreeEntryByPath(commitID)
+		if err == nil {
+			path += commitID
+			break
+		}
+		if IsErrNotExist(err) {
+			tree, err = tree.SubTree(commitID[0:2])
+			path += commitID[0:2] + "/"
+			commitID = commitID[2:]
+		}
+		if err != nil {
+			// Err may have been updated by the SubTree we need to recheck if it's again an ErrNotExist
+			if !IsErrNotExist(err) {
+				log.Error("Unable to find git note corresponding to the commit %q. Error: %v", originalCommitID, err)
+			}
+			return err
+		}
+	}
+
+	blob := entry.Blob()
+	dataRc, err := blob.DataAsync()
+	if err != nil {
+		log.Error("Unable to read blob with ID %q. Error: %v", blob.ID, err)
+		return err
+	}
+	closed := false
+	defer func() {
+		if !closed {
+			_ = dataRc.Close()
+		}
+	}()
+	d, err := io.ReadAll(dataRc)
+	if err != nil {
+		log.Error("Unable to read blob with ID %q. Error: %v", blob.ID, err)
+		return err
+	}
+	_ = dataRc.Close()
+	closed = true
+	note.Message = d
+
+	treePath := ""
+	if idx := strings.LastIndex(path, "/"); idx > -1 {
+		treePath = path[:idx]
+		path = path[idx+1:]
+	}
+
+	lastCommits, err := GetLastCommitForPaths(ctx, notes, treePath, []string{path})
+	if err != nil {
+		log.Error("Unable to get the commit for the path %q. Error: %v", treePath, err)
+		return err
+	}
+	note.Commit = lastCommits[path]
+
+	return nil
+}

modules/git/notes_gogit.go

@@ -1,89 +0,0 @@
-// Copyright 2019 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build gogit
-
-package git
-
-import (
-	"context"
-	"io"
-
-	"code.gitea.io/gitea/modules/log"
-
-	"github.com/go-git/go-git/v5/plumbing"
-	"github.com/go-git/go-git/v5/plumbing/object"
-)
-
-// GetNote retrieves the git-notes data for a given commit.
-// FIXME: Add LastCommitCache support
-func GetNote(ctx context.Context, repo *Repository, commitID string, note *Note) error {
-	log.Trace("Searching for git note corresponding to the commit %q in the repository %q", commitID, repo.Path)
-	notes, err := repo.GetCommit(NotesRef)
-	if err != nil {
-		if IsErrNotExist(err) {
-			return err
-		}
-		log.Error("Unable to get commit from ref %q. Error: %v", NotesRef, err)
-		return err
-	}
-
-	remainingCommitID := commitID
-	path := ""
-	currentTree := notes.Tree.gogitTree
-	log.Trace("Found tree with ID %q while searching for git note corresponding to the commit %q", currentTree.Entries[0].Name, commitID)
-	var file *object.File
-	for len(remainingCommitID) > 2 {
-		file, err = currentTree.File(remainingCommitID)
-		if err == nil {
-			path += remainingCommitID
-			break
-		}
-		if err == object.ErrFileNotFound {
-			currentTree, err = currentTree.Tree(remainingCommitID[0:2])
-			path += remainingCommitID[0:2] + "/"
-			remainingCommitID = remainingCommitID[2:]
-		}
-		if err != nil {
-			if err == object.ErrDirectoryNotFound {
-				return ErrNotExist{ID: remainingCommitID, RelPath: path}
-			}
-			log.Error("Unable to find git note corresponding to the commit %q. Error: %v", commitID, err)
-			return err
-		}
-	}
-
-	blob := file.Blob
-	dataRc, err := blob.Reader()
-	if err != nil {
-		log.Error("Unable to read blob with ID %q. Error: %v", blob.ID, err)
-		return err
-	}
-
-	defer dataRc.Close()
-	d, err := io.ReadAll(dataRc)
-	if err != nil {
-		log.Error("Unable to read blob with ID %q. Error: %v", blob.ID, err)
-		return err
-	}
-	note.Message = d
-
-	commitNodeIndex, commitGraphFile := repo.CommitNodeIndex()
-	if commitGraphFile != nil {
-		defer commitGraphFile.Close()
-	}
-
-	commitNode, err := commitNodeIndex.Get(plumbing.Hash(notes.ID.RawValue()))
-	if err != nil {
-		return err
-	}
-
-	lastCommits, err := GetLastCommitForPaths(ctx, nil, commitNode, "", []string{path})
-	if err != nil {
-		log.Error("Unable to get the commit for the path %q. Error: %v", path, err)
-		return err
-	}
-	note.Commit = lastCommits[path]
-
-	return nil
-}

modules/git/notes_nogogit.go

@@ -1,91 +0,0 @@
-// Copyright 2019 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build !gogit
-
-package git
-
-import (
-	"context"
-	"io"
-	"strings"
-
-	"code.gitea.io/gitea/modules/log"
-)
-
-// GetNote retrieves the git-notes data for a given commit.
-// FIXME: Add LastCommitCache support
-func GetNote(ctx context.Context, repo *Repository, commitID string, note *Note) error {
-	log.Trace("Searching for git note corresponding to the commit %q in the repository %q", commitID, repo.Path)
-	notes, err := repo.GetCommit(NotesRef)
-	if err != nil {
-		if IsErrNotExist(err) {
-			return err
-		}
-		log.Error("Unable to get commit from ref %q. Error: %v", NotesRef, err)
-		return err
-	}
-
-	path := ""
-
-	tree := &notes.Tree
-	log.Trace("Found tree with ID %q while searching for git note corresponding to the commit %q", tree.ID, commitID)
-
-	var entry *TreeEntry
-	originalCommitID := commitID
-	for len(commitID) > 2 {
-		entry, err = tree.GetTreeEntryByPath(commitID)
-		if err == nil {
-			path += commitID
-			break
-		}
-		if IsErrNotExist(err) {
-			tree, err = tree.SubTree(commitID[0:2])
-			path += commitID[0:2] + "/"
-			commitID = commitID[2:]
-		}
-		if err != nil {
-			// Err may have been updated by the SubTree we need to recheck if it's again an ErrNotExist
-			if !IsErrNotExist(err) {
-				log.Error("Unable to find git note corresponding to the commit %q. Error: %v", originalCommitID, err)
-			}
-			return err
-		}
-	}
-
-	blob := entry.Blob()
-	dataRc, err := blob.DataAsync()
-	if err != nil {
-		log.Error("Unable to read blob with ID %q. Error: %v", blob.ID, err)
-		return err
-	}
-	closed := false
-	defer func() {
-		if !closed {
-			_ = dataRc.Close()
-		}
-	}()
-	d, err := io.ReadAll(dataRc)
-	if err != nil {
-		log.Error("Unable to read blob with ID %q. Error: %v", blob.ID, err)
-		return err
-	}
-	_ = dataRc.Close()
-	closed = true
-	note.Message = d
-
-	treePath := ""
-	if idx := strings.LastIndex(path, "/"); idx > -1 {
-		treePath = path[:idx]
-		path = path[idx+1:]
-	}
-
-	lastCommits, err := GetLastCommitForPaths(ctx, notes, treePath, []string{path})
-	if err != nil {
-		log.Error("Unable to get the commit for the path %q. Error: %v", treePath, err)
-		return err
-	}
-	note.Commit = lastCommits[path]
-
-	return nil
-}

modules/git/object_id_gogit.go

@@ -1,30 +0,0 @@
-// Copyright 2023 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-//go:build gogit
-
-package git
-
-import (
-	"github.com/go-git/go-git/v5/plumbing"
-	"github.com/go-git/go-git/v5/plumbing/hash"
-)
-
-func ParseGogitHash(h plumbing.Hash) ObjectID {
-	switch hash.Size {
-	case 20:
-		return Sha1ObjectFormat.MustID(h[:])
-	case 32:
-		return Sha256ObjectFormat.MustID(h[:])
-	}
-
-	return nil
-}
-
-func ParseGogitHashArray(objectIDs []plumbing.Hash) []ObjectID {
-	ret := make([]ObjectID, len(objectIDs))
-	for i, h := range objectIDs {
-		ret[i] = ParseGogitHash(h)
-	}
-
-	return ret
-}

modules/git/parse.go

@@ -1,8 +1,6 @@
 // Copyright 2018 The Gitea Authors. All rights reserved.
 // SPDX-License-Identifier: MIT
 
-//go:build !gogit
-
 package git
 
 import (

modules/git/parse_gogit.go

@@ -1,96 +0,0 @@
-// Copyright 2018 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build gogit
-
-package git
-
-import (
-	"bytes"
-	"fmt"
-	"strconv"
-	"strings"
-
-	"github.com/go-git/go-git/v5/plumbing"
-	"github.com/go-git/go-git/v5/plumbing/filemode"
-	"github.com/go-git/go-git/v5/plumbing/hash"
-	"github.com/go-git/go-git/v5/plumbing/object"
-)
-
-// ParseTreeEntries parses the output of a `git ls-tree -l` command.
-func ParseTreeEntries(data []byte) ([]*TreeEntry, error) {
-	return parseTreeEntries(data, nil)
-}
-
-func parseTreeEntries(data []byte, ptree *Tree) ([]*TreeEntry, error) {
-	entries := make([]*TreeEntry, 0, 10)
-	for pos := 0; pos < len(data); {
-		// expect line to be of the form "<mode> <type> <sha> <space-padded-size>\t<filename>"
-		entry := new(TreeEntry)
-		entry.gogitTreeEntry = &object.TreeEntry{}
-		entry.ptree = ptree
-		if pos+6 > len(data) {
-			return nil, fmt.Errorf("Invalid ls-tree output: %s", string(data))
-		}
-		switch string(data[pos : pos+6]) {
-		case "100644":
-			entry.gogitTreeEntry.Mode = filemode.Regular
-			pos += 12 // skip over "100644 blob "
-		case "100755":
-			entry.gogitTreeEntry.Mode = filemode.Executable
-			pos += 12 // skip over "100755 blob "
-		case "120000":
-			entry.gogitTreeEntry.Mode = filemode.Symlink
-			pos += 12 // skip over "120000 blob "
-		case "160000":
-			entry.gogitTreeEntry.Mode = filemode.Submodule
-			pos += 14 // skip over "160000 object "
-		case "040000":
-			entry.gogitTreeEntry.Mode = filemode.Dir
-			pos += 12 // skip over "040000 tree "
-		default:
-			return nil, fmt.Errorf("unknown type: %v", string(data[pos:pos+6]))
-		}
-
-		// in hex format, not byte format ....
-		if pos+hash.Size*2 > len(data) {
-			return nil, fmt.Errorf("Invalid ls-tree output: %s", string(data))
-		}
-		var err error
-		entry.ID, err = NewIDFromString(string(data[pos : pos+hash.Size*2]))
-		if err != nil {
-			return nil, fmt.Errorf("invalid ls-tree output: %w", err)
-		}
-		entry.gogitTreeEntry.Hash = plumbing.Hash(entry.ID.RawValue())
-		pos += 41 // skip over sha and trailing space
-
-		end := pos + bytes.IndexByte(data[pos:], '\t')
-		if end < pos {
-			return nil, fmt.Errorf("Invalid ls-tree -l output: %s", string(data))
-		}
-		entry.size, _ = strconv.ParseInt(strings.TrimSpace(string(data[pos:end])), 10, 64)
-		entry.sized = true
-
-		pos = end + 1
-
-		end = pos + bytes.IndexByte(data[pos:], '\n')
-		if end < pos {
-			return nil, fmt.Errorf("Invalid ls-tree output: %s", string(data))
-		}
-
-		// In case entry name is surrounded by double quotes(it happens only in git-shell).
-		if data[pos] == '"' {
-			var err error
-			entry.gogitTreeEntry.Name, err = strconv.Unquote(string(data[pos:end]))
-			if err != nil {
-				return nil, fmt.Errorf("Invalid ls-tree output: %w", err)
-			}
-		} else {
-			entry.gogitTreeEntry.Name = string(data[pos:end])
-		}
-
-		pos = end + 1
-		entries = append(entries, entry)
-	}
-	return entries, nil
-}

modules/git/parse_gogit_test.go

@@ -1,79 +0,0 @@
-// Copyright 2018 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build gogit
-
-package git
-
-import (
-	"fmt"
-	"testing"
-
-	"github.com/go-git/go-git/v5/plumbing"
-	"github.com/go-git/go-git/v5/plumbing/filemode"
-	"github.com/go-git/go-git/v5/plumbing/object"
-	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
-)
-
-func TestParseTreeEntries(t *testing.T) {
-	testCases := []struct {
-		Input    string
-		Expected []*TreeEntry
-	}{
-		{
-			Input:    "",
-			Expected: []*TreeEntry{},
-		},
-		{
-			Input: "100644 blob 61ab7345a1a3bbc590068ccae37b8515cfc5843c    1022\texample/file2.txt\n",
-			Expected: []*TreeEntry{
-				{
-					ID: MustIDFromString("61ab7345a1a3bbc590068ccae37b8515cfc5843c"),
-					gogitTreeEntry: &object.TreeEntry{
-						Hash: plumbing.Hash(MustIDFromString("61ab7345a1a3bbc590068ccae37b8515cfc5843c").RawValue()),
-						Name: "example/file2.txt",
-						Mode: filemode.Regular,
-					},
-					size:  1022,
-					sized: true,
-				},
-			},
-		},
-		{
-			Input: "120000 blob 61ab7345a1a3bbc590068ccae37b8515cfc5843c  234131\t\"example/\\n.txt\"\n" +
-				"040000 tree 1d01fb729fb0db5881daaa6030f9f2d3cd3d5ae8       -\texample\n",
-			Expected: []*TreeEntry{
-				{
-					ID: MustIDFromString("61ab7345a1a3bbc590068ccae37b8515cfc5843c"),
-					gogitTreeEntry: &object.TreeEntry{
-						Hash: plumbing.Hash(MustIDFromString("61ab7345a1a3bbc590068ccae37b8515cfc5843c").RawValue()),
-						Name: "example/\n.txt",
-						Mode: filemode.Symlink,
-					},
-					size:  234131,
-					sized: true,
-				},
-				{
-					ID:    MustIDFromString("1d01fb729fb0db5881daaa6030f9f2d3cd3d5ae8"),
-					sized: true,
-					gogitTreeEntry: &object.TreeEntry{
-						Hash: plumbing.Hash(MustIDFromString("1d01fb729fb0db5881daaa6030f9f2d3cd3d5ae8").RawValue()),
-						Name: "example",
-						Mode: filemode.Dir,
-					},
-				},
-			},
-		},
-	}
-
-	for _, testCase := range testCases {
-		entries, err := ParseTreeEntries([]byte(testCase.Input))
-		require.NoError(t, err)
-		if len(entries) > 1 {
-			fmt.Println(testCase.Expected[0].ID)
-			fmt.Println(entries[0].ID)
-		}
-		assert.EqualValues(t, testCase.Expected, entries)
-	}
-}

modules/git/parse_test.go

@@ -1,8 +1,6 @@
 // Copyright 2021 The Gitea Authors. All rights reserved.
 // SPDX-License-Identifier: MIT
 
-//go:build !gogit
-
 package git
 
 import (

modules/git/pipeline/lfs.go

@@ -1,21 +1,42 @@
 // Copyright 2020 The Gitea Authors. All rights reserved.
 // SPDX-License-Identifier: MIT
 
-//go:build !gogit
-
 package pipeline
 
 import (
 	"bufio"
 	"bytes"
+	"fmt"
 	"io"
 	"sort"
 	"strings"
 	"sync"
+	"time"
 
 	"code.gitea.io/gitea/modules/git"
 )
 
+// LFSResult represents commits found using a provided pointer file hash
+type LFSResult struct {
+	Name           string
+	SHA            string
+	Summary        string
+	When           time.Time
+	ParentHashes   []git.ObjectID
+	BranchName     string
+	FullCommitName string
+}
+
+type lfsResultSlice []*LFSResult
+
+func (a lfsResultSlice) Len() int           { return len(a) }
+func (a lfsResultSlice) Swap(i, j int)      { a[i], a[j] = a[j], a[i] }
+func (a lfsResultSlice) Less(i, j int) bool { return a[j].When.After(a[i].When) }
+
+func lfsError(msg string, err error) error {
+	return fmt.Errorf("LFS error occurred, %s: err: %w", msg, err)
+}
+
 // FindLFSFile finds commits that contain a provided pointer file hash
 func FindLFSFile(repo *git.Repository, objectID git.ObjectID) ([]*LFSResult, error) {
 	resultsMap := map[string]*LFSResult{}

modules/git/pipeline/lfs_common.go

@@ -1,32 +0,0 @@
-// Copyright 2024 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-package pipeline
-
-import (
-	"fmt"
-	"time"
-
-	"code.gitea.io/gitea/modules/git"
-)
-
-// LFSResult represents commits found using a provided pointer file hash
-type LFSResult struct {
-	Name           string
-	SHA            string
-	Summary        string
-	When           time.Time
-	ParentHashes   []git.ObjectID
-	BranchName     string
-	FullCommitName string
-}
-
-type lfsResultSlice []*LFSResult
-
-func (a lfsResultSlice) Len() int           { return len(a) }
-func (a lfsResultSlice) Swap(i, j int)      { a[i], a[j] = a[j], a[i] }
-func (a lfsResultSlice) Less(i, j int) bool { return a[j].When.After(a[i].When) }
-
-func lfsError(msg string, err error) error {
-	return fmt.Errorf("LFS error occurred, %s: err: %w", msg, err)
-}

modules/git/pipeline/lfs_gogit.go

@@ -1,146 +0,0 @@
-// Copyright 2020 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build gogit
-
-package pipeline
-
-import (
-	"bufio"
-	"io"
-	"sort"
-	"strings"
-	"sync"
-
-	"code.gitea.io/gitea/modules/git"
-
-	gogit "github.com/go-git/go-git/v5"
-	"github.com/go-git/go-git/v5/plumbing"
-	"github.com/go-git/go-git/v5/plumbing/object"
-)
-
-// FindLFSFile finds commits that contain a provided pointer file hash
-func FindLFSFile(repo *git.Repository, objectID git.ObjectID) ([]*LFSResult, error) {
-	resultsMap := map[string]*LFSResult{}
-	results := make([]*LFSResult, 0)
-
-	basePath := repo.Path
-	gogitRepo := repo.GoGitRepo()
-
-	commitsIter, err := gogitRepo.Log(&gogit.LogOptions{
-		Order: gogit.LogOrderCommitterTime,
-		All:   true,
-	})
-	if err != nil {
-		return nil, lfsError("failed to get GoGit CommitsIter", err)
-	}
-
-	err = commitsIter.ForEach(func(gitCommit *object.Commit) error {
-		tree, err := gitCommit.Tree()
-		if err != nil {
-			return err
-		}
-		treeWalker := object.NewTreeWalker(tree, true, nil)
-		defer treeWalker.Close()
-		for {
-			name, entry, err := treeWalker.Next()
-			if err == io.EOF {
-				break
-			}
-			if entry.Hash == plumbing.Hash(objectID.RawValue()) {
-				parents := make([]git.ObjectID, len(gitCommit.ParentHashes))
-				for i, parentCommitID := range gitCommit.ParentHashes {
-					parents[i] = git.ParseGogitHash(parentCommitID)
-				}
-
-				result := LFSResult{
-					Name:         name,
-					SHA:          gitCommit.Hash.String(),
-					Summary:      strings.Split(strings.TrimSpace(gitCommit.Message), "\n")[0],
-					When:         gitCommit.Author.When,
-					ParentHashes: parents,
-				}
-				resultsMap[gitCommit.Hash.String()+":"+name] = &result
-			}
-		}
-		return nil
-	})
-	if err != nil && err != io.EOF {
-		return nil, lfsError("failure in CommitIter.ForEach", err)
-	}
-
-	for _, result := range resultsMap {
-		hasParent := false
-		for _, parentHash := range result.ParentHashes {
-			if _, hasParent = resultsMap[parentHash.String()+":"+result.Name]; hasParent {
-				break
-			}
-		}
-		if !hasParent {
-			results = append(results, result)
-		}
-	}
-
-	sort.Sort(lfsResultSlice(results))
-
-	// Should really use a go-git function here but name-rev is not completed and recapitulating it is not simple
-	shasToNameReader, shasToNameWriter := io.Pipe()
-	nameRevStdinReader, nameRevStdinWriter := io.Pipe()
-	errChan := make(chan error, 1)
-	wg := sync.WaitGroup{}
-	wg.Add(3)
-
-	go func() {
-		defer wg.Done()
-		scanner := bufio.NewScanner(nameRevStdinReader)
-		i := 0
-		for scanner.Scan() {
-			line := scanner.Text()
-			if len(line) == 0 {
-				continue
-			}
-			result := results[i]
-			result.FullCommitName = line
-			result.BranchName = strings.Split(line, "~")[0]
-			i++
-		}
-	}()
-	go NameRevStdin(repo.Ctx, shasToNameReader, nameRevStdinWriter, &wg, basePath)
-	go func() {
-		defer wg.Done()
-		defer shasToNameWriter.Close()
-		for _, result := range results {
-			i := 0
-			if i < len(result.SHA) {
-				n, err := shasToNameWriter.Write([]byte(result.SHA)[i:])
-				if err != nil {
-					errChan <- err
-					break
-				}
-				i += n
-			}
-			n := 0
-			for n < 1 {
-				n, err = shasToNameWriter.Write([]byte{'\n'})
-				if err != nil {
-					errChan <- err
-					break
-				}
-
-			}
-
-		}
-	}()
-
-	wg.Wait()
-
-	select {
-	case err, has := <-errChan:
-		if has {
-			return nil, lfsError("unable to obtain name for LFS files", err)
-		}
-	default:
-	}
-
-	return results, nil
-}

modules/git/repo_base.go

@@ -1,6 +1,116 @@
-// Copyright 2021 The Gitea Authors. All rights reserved.
+// Copyright 2015 The Gogs Authors. All rights reserved.
+// Copyright 2017 The Gitea Authors. All rights reserved.
 // SPDX-License-Identifier: MIT
 
 package git
 
-var isGogit bool
+import (
+	"bufio"
+	"context"
+	"errors"
+	"path/filepath"
+
+	"code.gitea.io/gitea/modules/log"
+)
+
+// Repository represents a Git repository.
+type Repository struct {
+	Path string
+
+	tagCache *ObjectCache
+
+	gpgSettings *GPGSettings
+
+	batchInUse  bool
+	batchCancel context.CancelFunc
+	batchReader *bufio.Reader
+	batchWriter WriteCloserError
+
+	checkInUse  bool
+	checkCancel context.CancelFunc
+	checkReader *bufio.Reader
+	checkWriter WriteCloserError
+
+	Ctx             context.Context
+	LastCommitCache *LastCommitCache
+
+	objectFormat ObjectFormat
+}
+
+// openRepositoryWithDefaultContext opens the repository at the given path with DefaultContext.
+func openRepositoryWithDefaultContext(repoPath string) (*Repository, error) {
+	return OpenRepository(DefaultContext, repoPath)
+}
+
+// OpenRepository opens the repository at the given path with the provided context.
+func OpenRepository(ctx context.Context, repoPath string) (*Repository, error) {
+	repoPath, err := filepath.Abs(repoPath)
+	if err != nil {
+		return nil, err
+	} else if !isDir(repoPath) {
+		return nil, errors.New("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!
+	if err := EnsureValidGitRepository(ctx, repoPath); err != nil {
+		return nil, err
+	}
+
+	repo := &Repository{
+		Path:     repoPath,
+		tagCache: newObjectCache(),
+		Ctx:      ctx,
+	}
+
+	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()) {
+	if repo.batchCancel == nil || repo.batchInUse {
+		log.Debug("Opening temporary cat file batch for: %s", repo.Path)
+		return CatFileBatch(ctx, repo.Path)
+	}
+	repo.batchInUse = true
+	return repo.batchWriter, repo.batchReader, func() {
+		repo.batchInUse = false
+	}
+}
+
+// CatFileBatchCheck obtains a CatFileBatchCheck for this repository
+func (repo *Repository) CatFileBatchCheck(ctx context.Context) (WriteCloserError, *bufio.Reader, func()) {
+	if repo.checkCancel == nil || repo.checkInUse {
+		log.Debug("Opening temporary cat file batch-check for: %s", repo.Path)
+		return CatFileBatchCheck(ctx, repo.Path)
+	}
+	repo.checkInUse = true
+	return repo.checkWriter, repo.checkReader, func() {
+		repo.checkInUse = false
+	}
+}
+
+func (repo *Repository) Close() error {
+	if repo == nil {
+		return nil
+	}
+	if repo.batchCancel != nil {
+		repo.batchCancel()
+		repo.batchReader = nil
+		repo.batchWriter = nil
+		repo.batchCancel = nil
+		repo.batchInUse = false
+	}
+	if repo.checkCancel != nil {
+		repo.checkCancel()
+		repo.checkCancel = nil
+		repo.checkReader = nil
+		repo.checkWriter = nil
+		repo.checkInUse = false
+	}
+	repo.LastCommitCache = nil
+	repo.tagCache = nil
+	return nil
+}

modules/git/repo_base_gogit.go

@@ -1,107 +0,0 @@
-// Copyright 2015 The Gogs Authors. All rights reserved.
-// Copyright 2017 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build gogit
-
-package git
-
-import (
-	"context"
-	"errors"
-	"path/filepath"
-
-	gitealog "code.gitea.io/gitea/modules/log"
-	"code.gitea.io/gitea/modules/setting"
-
-	"github.com/go-git/go-billy/v5"
-	"github.com/go-git/go-billy/v5/osfs"
-	gogit "github.com/go-git/go-git/v5"
-	"github.com/go-git/go-git/v5/plumbing"
-	"github.com/go-git/go-git/v5/plumbing/cache"
-	"github.com/go-git/go-git/v5/storage/filesystem"
-)
-
-func init() {
-	isGogit = true
-}
-
-// Repository represents a Git repository.
-type Repository struct {
-	Path string
-
-	tagCache *ObjectCache
-
-	gogitRepo    *gogit.Repository
-	gogitStorage *filesystem.Storage
-	gpgSettings  *GPGSettings
-
-	Ctx             context.Context
-	LastCommitCache *LastCommitCache
-	objectFormat    ObjectFormat
-}
-
-// openRepositoryWithDefaultContext opens the repository at the given path with DefaultContext.
-func openRepositoryWithDefaultContext(repoPath string) (*Repository, error) {
-	return OpenRepository(DefaultContext, repoPath)
-}
-
-// OpenRepository opens the repository at the given path within the context.Context
-func OpenRepository(ctx context.Context, repoPath string) (*Repository, error) {
-	repoPath, err := filepath.Abs(repoPath)
-	if err != nil {
-		return nil, err
-	} else if !isDir(repoPath) {
-		return nil, errors.New("no such file or directory")
-	}
-
-	fs := osfs.New(repoPath)
-	_, err = fs.Stat(".git")
-	if err == nil {
-		fs, err = fs.Chroot(".git")
-		if err != nil {
-			return nil, err
-		}
-	}
-	// the "clone --shared" repo doesn't work well with go-git AlternativeFS, https://github.com/go-git/go-git/issues/1006
-	// so use "/" for AlternatesFS, I guess it is the same behavior as current nogogit (no limitation or check for the "objects/info/alternates" paths), trust the "clone" command executed by the server.
-	var altFs billy.Filesystem
-	if setting.IsWindows {
-		altFs = osfs.New(filepath.VolumeName(setting.RepoRootPath) + "\\") // TODO: does it really work for Windows? Need some time to check.
-	} else {
-		altFs = osfs.New("/")
-	}
-	storage := filesystem.NewStorageWithOptions(fs, cache.NewObjectLRUDefault(), filesystem.Options{KeepDescriptors: true, LargeObjectThreshold: setting.Git.LargeObjectThreshold, AlternatesFS: altFs})
-	gogitRepo, err := gogit.Open(storage, fs)
-	if err != nil {
-		return nil, err
-	}
-
-	return &Repository{
-		Path:         repoPath,
-		gogitRepo:    gogitRepo,
-		gogitStorage: storage,
-		tagCache:     newObjectCache(),
-		Ctx:          ctx,
-		objectFormat: ParseGogitHash(plumbing.ZeroHash).Type(),
-	}, nil
-}
-
-// Close this repository, in particular close the underlying gogitStorage if this is not nil
-func (repo *Repository) Close() error {
-	if repo == nil || repo.gogitStorage == nil {
-		return nil
-	}
-	if err := repo.gogitStorage.Close(); err != nil {
-		gitealog.Error("Error closing storage: %v", err)
-	}
-	repo.gogitStorage = nil
-	repo.LastCommitCache = nil
-	repo.tagCache = nil
-	return nil
-}
-
-// GoGitRepo gets the go-git repo representation
-func (repo *Repository) GoGitRepo() *gogit.Repository {
-	return repo.gogitRepo
-}

modules/git/repo_base_nogogit.go

@@ -1,122 +0,0 @@
-// Copyright 2015 The Gogs Authors. All rights reserved.
-// Copyright 2017 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build !gogit
-
-package git
-
-import (
-	"bufio"
-	"context"
-	"errors"
-	"path/filepath"
-
-	"code.gitea.io/gitea/modules/log"
-)
-
-func init() {
-	isGogit = false
-}
-
-// Repository represents a Git repository.
-type Repository struct {
-	Path string
-
-	tagCache *ObjectCache
-
-	gpgSettings *GPGSettings
-
-	batchInUse  bool
-	batchCancel context.CancelFunc
-	batchReader *bufio.Reader
-	batchWriter WriteCloserError
-
-	checkInUse  bool
-	checkCancel context.CancelFunc
-	checkReader *bufio.Reader
-	checkWriter WriteCloserError
-
-	Ctx             context.Context
-	LastCommitCache *LastCommitCache
-
-	objectFormat ObjectFormat
-}
-
-// openRepositoryWithDefaultContext opens the repository at the given path with DefaultContext.
-func openRepositoryWithDefaultContext(repoPath string) (*Repository, error) {
-	return OpenRepository(DefaultContext, repoPath)
-}
-
-// OpenRepository opens the repository at the given path with the provided context.
-func OpenRepository(ctx context.Context, repoPath string) (*Repository, error) {
-	repoPath, err := filepath.Abs(repoPath)
-	if err != nil {
-		return nil, err
-	} else if !isDir(repoPath) {
-		return nil, errors.New("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!
-	if err := EnsureValidGitRepository(ctx, repoPath); err != nil {
-		return nil, err
-	}
-
-	repo := &Repository{
-		Path:     repoPath,
-		tagCache: newObjectCache(),
-		Ctx:      ctx,
-	}
-
-	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()) {
-	if repo.batchCancel == nil || repo.batchInUse {
-		log.Debug("Opening temporary cat file batch for: %s", repo.Path)
-		return CatFileBatch(ctx, repo.Path)
-	}
-	repo.batchInUse = true
-	return repo.batchWriter, repo.batchReader, func() {
-		repo.batchInUse = false
-	}
-}
-
-// CatFileBatchCheck obtains a CatFileBatchCheck for this repository
-func (repo *Repository) CatFileBatchCheck(ctx context.Context) (WriteCloserError, *bufio.Reader, func()) {
-	if repo.checkCancel == nil || repo.checkInUse {
-		log.Debug("Opening temporary cat file batch-check for: %s", repo.Path)
-		return CatFileBatchCheck(ctx, repo.Path)
-	}
-	repo.checkInUse = true
-	return repo.checkWriter, repo.checkReader, func() {
-		repo.checkInUse = false
-	}
-}
-
-func (repo *Repository) Close() error {
-	if repo == nil {
-		return nil
-	}
-	if repo.batchCancel != nil {
-		repo.batchCancel()
-		repo.batchReader = nil
-		repo.batchWriter = nil
-		repo.batchCancel = nil
-		repo.batchInUse = false
-	}
-	if repo.checkCancel != nil {
-		repo.checkCancel()
-		repo.checkCancel = nil
-		repo.checkReader = nil
-		repo.checkWriter = nil
-		repo.checkInUse = false
-	}
-	repo.LastCommitCache = nil
-	repo.tagCache = nil
-	return nil
-}

modules/git/repo_blob.go

@@ -1,13 +0,0 @@
-// Copyright 2020 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-package git
-
-// GetBlob finds the blob object in the repository.
-func (repo *Repository) GetBlob(idStr string) (*Blob, error) {
-	id, err := NewIDFromString(idStr)
-	if err != nil {
-		return nil, err
-	}
-	return repo.getBlob(id)
-}

modules/git/repo_blob_gogit.go

@@ -1,22 +0,0 @@
-// Copyright 2018 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build gogit
-
-package git
-
-import (
-	"github.com/go-git/go-git/v5/plumbing"
-)
-
-func (repo *Repository) getBlob(id ObjectID) (*Blob, error) {
-	encodedObj, err := repo.gogitRepo.Storer.EncodedObject(plumbing.AnyObject, plumbing.Hash(id.RawValue()))
-	if err != nil {
-		return nil, ErrNotExist{id.String(), ""}
-	}
-
-	return &Blob{
-		ID:              id,
-		gogitEncodedObj: encodedObj,
-	}, nil
-}

modules/git/repo_blob_nogogit.go

@@ -1,16 +0,0 @@
-// Copyright 2020 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build !gogit
-
-package git
-
-func (repo *Repository) getBlob(id ObjectID) (*Blob, error) {
-	if id.IsZero() {
-		return nil, ErrNotExist{id.String(), ""}
-	}
-	return &Blob{
-		ID:   id,
-		repo: repo,
-	}, nil
-}

modules/git/repo_branch.go

@@ -5,10 +5,15 @@
 package git
 
 import (
+	"bufio"
+	"bytes"
 	"context"
 	"errors"
 	"fmt"
+	"io"
 	"strings"
+
+	"code.gitea.io/gitea/modules/log"
 )
 
 // BranchPrefix base dir of the branch information file store on git
@@ -157,3 +162,180 @@ func (repo *Repository) RenameBranch(from, to string) error {
 	_, _, err := NewCommand(repo.Ctx, "branch", "-m").AddDynamicArguments(from, to).RunStdString(&RunOpts{Dir: repo.Path})
 	return err
 }
+
+// IsObjectExist returns true if given reference exists in the repository.
+func (repo *Repository) IsObjectExist(name string) bool {
+	if name == "" {
+		return false
+	}
+
+	wr, rd, cancel := repo.CatFileBatchCheck(repo.Ctx)
+	defer cancel()
+	_, err := wr.Write([]byte(name + "\n"))
+	if err != nil {
+		log.Debug("Error writing to CatFileBatchCheck %v", err)
+		return false
+	}
+	sha, _, _, err := ReadBatchLine(rd)
+	return err == nil && bytes.HasPrefix(sha, []byte(strings.TrimSpace(name)))
+}
+
+// IsReferenceExist returns true if given reference exists in the repository.
+func (repo *Repository) IsReferenceExist(name string) bool {
+	if name == "" {
+		return false
+	}
+
+	wr, rd, cancel := repo.CatFileBatchCheck(repo.Ctx)
+	defer cancel()
+	_, err := wr.Write([]byte(name + "\n"))
+	if err != nil {
+		log.Debug("Error writing to CatFileBatchCheck %v", err)
+		return false
+	}
+	_, _, _, err = ReadBatchLine(rd)
+	return err == nil
+}
+
+// IsBranchExist returns true if given branch exists in current repository.
+func (repo *Repository) IsBranchExist(name string) bool {
+	if repo == nil || name == "" {
+		return false
+	}
+
+	return repo.IsReferenceExist(BranchPrefix + name)
+}
+
+// GetBranchNames returns branches from the repository, skipping "skip" initial branches and
+// returning at most "limit" branches, or all branches if "limit" is 0.
+func (repo *Repository) GetBranchNames(skip, limit int) ([]string, int, error) {
+	return callShowRef(repo.Ctx, repo.Path, BranchPrefix, TrustedCmdArgs{BranchPrefix, "--sort=-committerdate"}, skip, limit)
+}
+
+// WalkReferences walks all the references from the repository
+// refType should be empty, ObjectTag or ObjectBranch. All other values are equivalent to empty.
+func (repo *Repository) WalkReferences(refType ObjectType, skip, limit int, walkfn func(sha1, refname string) error) (int, error) {
+	var args TrustedCmdArgs
+	switch refType {
+	case ObjectTag:
+		args = TrustedCmdArgs{TagPrefix, "--sort=-taggerdate"}
+	case ObjectBranch:
+		args = TrustedCmdArgs{BranchPrefix, "--sort=-committerdate"}
+	}
+
+	return WalkShowRef(repo.Ctx, repo.Path, args, skip, limit, walkfn)
+}
+
+// callShowRef return refs, if limit = 0 it will not limit
+func callShowRef(ctx context.Context, repoPath, trimPrefix string, extraArgs TrustedCmdArgs, skip, limit int) (branchNames []string, countAll int, err error) {
+	countAll, err = WalkShowRef(ctx, repoPath, extraArgs, skip, limit, func(_, branchName string) error {
+		branchName = strings.TrimPrefix(branchName, trimPrefix)
+		branchNames = append(branchNames, branchName)
+
+		return nil
+	})
+	return branchNames, countAll, err
+}
+
+func WalkShowRef(ctx context.Context, repoPath string, extraArgs TrustedCmdArgs, skip, limit int, walkfn func(sha1, refname string) error) (countAll int, err error) {
+	stdoutReader, stdoutWriter := io.Pipe()
+	defer func() {
+		_ = stdoutReader.Close()
+		_ = stdoutWriter.Close()
+	}()
+
+	go func() {
+		stderrBuilder := &strings.Builder{}
+		args := TrustedCmdArgs{"for-each-ref", "--format=%(objectname) %(refname)"}
+		args = append(args, extraArgs...)
+		err := NewCommand(ctx, args...).Run(&RunOpts{
+			Dir:    repoPath,
+			Stdout: stdoutWriter,
+			Stderr: stderrBuilder,
+		})
+		if err != nil {
+			if stderrBuilder.Len() == 0 {
+				_ = stdoutWriter.Close()
+				return
+			}
+			_ = stdoutWriter.CloseWithError(ConcatenateError(err, stderrBuilder.String()))
+		} else {
+			_ = stdoutWriter.Close()
+		}
+	}()
+
+	i := 0
+	bufReader := bufio.NewReader(stdoutReader)
+	for i < skip {
+		_, isPrefix, err := bufReader.ReadLine()
+		if err == io.EOF {
+			return i, nil
+		}
+		if err != nil {
+			return 0, err
+		}
+		if !isPrefix {
+			i++
+		}
+	}
+	for limit == 0 || i < skip+limit {
+		// The output of show-ref is simply a list:
+		// <sha> SP <ref> LF
+		sha, err := bufReader.ReadString(' ')
+		if err == io.EOF {
+			return i, nil
+		}
+		if err != nil {
+			return 0, err
+		}
+
+		branchName, err := bufReader.ReadString('\n')
+		if err == io.EOF {
+			// This shouldn't happen... but we'll tolerate it for the sake of peace
+			return i, nil
+		}
+		if err != nil {
+			return i, err
+		}
+
+		if len(branchName) > 0 {
+			branchName = branchName[:len(branchName)-1]
+		}
+
+		if len(sha) > 0 {
+			sha = sha[:len(sha)-1]
+		}
+
+		err = walkfn(sha, branchName)
+		if err != nil {
+			return i, err
+		}
+		i++
+	}
+	// count all refs
+	for limit != 0 {
+		_, isPrefix, err := bufReader.ReadLine()
+		if err == io.EOF {
+			return i, nil
+		}
+		if err != nil {
+			return 0, err
+		}
+		if !isPrefix {
+			i++
+		}
+	}
+	return i, nil
+}
+
+// GetRefsBySha returns all references filtered with prefix that belong to a sha commit hash
+func (repo *Repository) GetRefsBySha(sha, prefix string) ([]string, error) {
+	var revList []string
+	_, err := WalkShowRef(repo.Ctx, repo.Path, nil, 0, 0, func(walkSha, refname string) error {
+		if walkSha == sha && strings.HasPrefix(refname, prefix) {
+			revList = append(revList, refname)
+		}
+		return nil
+	})
+	return revList, err
+}

modules/git/repo_branch_gogit.go

@@ -1,147 +0,0 @@
-// Copyright 2015 The Gogs Authors. All rights reserved.
-// Copyright 2018 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build gogit
-
-package git
-
-import (
-	"sort"
-	"strings"
-
-	"github.com/go-git/go-git/v5/plumbing"
-	"github.com/go-git/go-git/v5/plumbing/storer"
-)
-
-// IsObjectExist returns true if given reference exists in the repository.
-func (repo *Repository) IsObjectExist(name string) bool {
-	if name == "" {
-		return false
-	}
-
-	_, err := repo.gogitRepo.ResolveRevision(plumbing.Revision(name))
-
-	return err == nil
-}
-
-// IsReferenceExist returns true if given reference exists in the repository.
-func (repo *Repository) IsReferenceExist(name string) bool {
-	if name == "" {
-		return false
-	}
-
-	reference, err := repo.gogitRepo.Reference(plumbing.ReferenceName(name), true)
-	if err != nil {
-		return false
-	}
-	return reference.Type() != plumbing.InvalidReference
-}
-
-// IsBranchExist returns true if given branch exists in current repository.
-func (repo *Repository) IsBranchExist(name string) bool {
-	if name == "" {
-		return false
-	}
-	reference, err := repo.gogitRepo.Reference(plumbing.ReferenceName(BranchPrefix+name), true)
-	if err != nil {
-		return false
-	}
-	return reference.Type() != plumbing.InvalidReference
-}
-
-// GetBranches returns branches from the repository, skipping "skip" initial branches and
-// returning at most "limit" branches, or all branches if "limit" is 0.
-// Branches are returned with sort of `-commiterdate` as the nogogit
-// implementation. This requires full fetch, sort and then the
-// skip/limit applies later as gogit returns in undefined order.
-func (repo *Repository) GetBranchNames(skip, limit int) ([]string, int, error) {
-	type BranchData struct {
-		name          string
-		committerDate int64
-	}
-	var branchData []BranchData
-
-	branchIter, err := repo.gogitRepo.Branches()
-	if err != nil {
-		return nil, 0, err
-	}
-
-	_ = branchIter.ForEach(func(branch *plumbing.Reference) error {
-		obj, err := repo.gogitRepo.CommitObject(branch.Hash())
-		if err != nil {
-			// skip branch if can't find commit
-			return nil
-		}
-
-		branchData = append(branchData, BranchData{strings.TrimPrefix(branch.Name().String(), BranchPrefix), obj.Committer.When.Unix()})
-		return nil
-	})
-
-	sort.Slice(branchData, func(i, j int) bool {
-		return !(branchData[i].committerDate < branchData[j].committerDate)
-	})
-
-	var branchNames []string
-	maxPos := len(branchData)
-	if limit > 0 {
-		maxPos = min(skip+limit, maxPos)
-	}
-	for i := skip; i < maxPos; i++ {
-		branchNames = append(branchNames, branchData[i].name)
-	}
-
-	return branchNames, len(branchData), nil
-}
-
-// WalkReferences walks all the references from the repository
-func (repo *Repository) WalkReferences(arg ObjectType, skip, limit int, walkfn func(sha1, refname string) error) (int, error) {
-	i := 0
-	var iter storer.ReferenceIter
-	var err error
-	switch arg {
-	case ObjectTag:
-		iter, err = repo.gogitRepo.Tags()
-	case ObjectBranch:
-		iter, err = repo.gogitRepo.Branches()
-	default:
-		iter, err = repo.gogitRepo.References()
-	}
-	if err != nil {
-		return i, err
-	}
-	defer iter.Close()
-
-	err = iter.ForEach(func(ref *plumbing.Reference) error {
-		if i < skip {
-			i++
-			return nil
-		}
-		err := walkfn(ref.Hash().String(), string(ref.Name()))
-		i++
-		if err != nil {
-			return err
-		}
-		if limit != 0 && i >= skip+limit {
-			return storer.ErrStop
-		}
-		return nil
-	})
-	return i, err
-}
-
-// GetRefsBySha returns all references filtered with prefix that belong to a sha commit hash
-func (repo *Repository) GetRefsBySha(sha, prefix string) ([]string, error) {
-	var revList []string
-	iter, err := repo.gogitRepo.References()
-	if err != nil {
-		return nil, err
-	}
-	err = iter.ForEach(func(ref *plumbing.Reference) error {
-		if ref.Hash().String() == sha && strings.HasPrefix(string(ref.Name()), prefix) {
-			revList = append(revList, string(ref.Name()))
-		}
-		return nil
-	})
-	return revList, err
-}

modules/git/repo_branch_nogogit.go

@@ -1,194 +0,0 @@
-// Copyright 2015 The Gogs Authors. All rights reserved.
-// Copyright 2018 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build !gogit
-
-package git
-
-import (
-	"bufio"
-	"bytes"
-	"context"
-	"io"
-	"strings"
-
-	"code.gitea.io/gitea/modules/log"
-)
-
-// IsObjectExist returns true if given reference exists in the repository.
-func (repo *Repository) IsObjectExist(name string) bool {
-	if name == "" {
-		return false
-	}
-
-	wr, rd, cancel := repo.CatFileBatchCheck(repo.Ctx)
-	defer cancel()
-	_, err := wr.Write([]byte(name + "\n"))
-	if err != nil {
-		log.Debug("Error writing to CatFileBatchCheck %v", err)
-		return false
-	}
-	sha, _, _, err := ReadBatchLine(rd)
-	return err == nil && bytes.HasPrefix(sha, []byte(strings.TrimSpace(name)))
-}
-
-// IsReferenceExist returns true if given reference exists in the repository.
-func (repo *Repository) IsReferenceExist(name string) bool {
-	if name == "" {
-		return false
-	}
-
-	wr, rd, cancel := repo.CatFileBatchCheck(repo.Ctx)
-	defer cancel()
-	_, err := wr.Write([]byte(name + "\n"))
-	if err != nil {
-		log.Debug("Error writing to CatFileBatchCheck %v", err)
-		return false
-	}
-	_, _, _, err = ReadBatchLine(rd)
-	return err == nil
-}
-
-// IsBranchExist returns true if given branch exists in current repository.
-func (repo *Repository) IsBranchExist(name string) bool {
-	if repo == nil || name == "" {
-		return false
-	}
-
-	return repo.IsReferenceExist(BranchPrefix + name)
-}
-
-// GetBranchNames returns branches from the repository, skipping "skip" initial branches and
-// returning at most "limit" branches, or all branches if "limit" is 0.
-func (repo *Repository) GetBranchNames(skip, limit int) ([]string, int, error) {
-	return callShowRef(repo.Ctx, repo.Path, BranchPrefix, TrustedCmdArgs{BranchPrefix, "--sort=-committerdate"}, skip, limit)
-}
-
-// WalkReferences walks all the references from the repository
-// refType should be empty, ObjectTag or ObjectBranch. All other values are equivalent to empty.
-func (repo *Repository) WalkReferences(refType ObjectType, skip, limit int, walkfn func(sha1, refname string) error) (int, error) {
-	var args TrustedCmdArgs
-	switch refType {
-	case ObjectTag:
-		args = TrustedCmdArgs{TagPrefix, "--sort=-taggerdate"}
-	case ObjectBranch:
-		args = TrustedCmdArgs{BranchPrefix, "--sort=-committerdate"}
-	}
-
-	return WalkShowRef(repo.Ctx, repo.Path, args, skip, limit, walkfn)
-}
-
-// callShowRef return refs, if limit = 0 it will not limit
-func callShowRef(ctx context.Context, repoPath, trimPrefix string, extraArgs TrustedCmdArgs, skip, limit int) (branchNames []string, countAll int, err error) {
-	countAll, err = WalkShowRef(ctx, repoPath, extraArgs, skip, limit, func(_, branchName string) error {
-		branchName = strings.TrimPrefix(branchName, trimPrefix)
-		branchNames = append(branchNames, branchName)
-
-		return nil
-	})
-	return branchNames, countAll, err
-}
-
-func WalkShowRef(ctx context.Context, repoPath string, extraArgs TrustedCmdArgs, skip, limit int, walkfn func(sha1, refname string) error) (countAll int, err error) {
-	stdoutReader, stdoutWriter := io.Pipe()
-	defer func() {
-		_ = stdoutReader.Close()
-		_ = stdoutWriter.Close()
-	}()
-
-	go func() {
-		stderrBuilder := &strings.Builder{}
-		args := TrustedCmdArgs{"for-each-ref", "--format=%(objectname) %(refname)"}
-		args = append(args, extraArgs...)
-		err := NewCommand(ctx, args...).Run(&RunOpts{
-			Dir:    repoPath,
-			Stdout: stdoutWriter,
-			Stderr: stderrBuilder,
-		})
-		if err != nil {
-			if stderrBuilder.Len() == 0 {
-				_ = stdoutWriter.Close()
-				return
-			}
-			_ = stdoutWriter.CloseWithError(ConcatenateError(err, stderrBuilder.String()))
-		} else {
-			_ = stdoutWriter.Close()
-		}
-	}()
-
-	i := 0
-	bufReader := bufio.NewReader(stdoutReader)
-	for i < skip {
-		_, isPrefix, err := bufReader.ReadLine()
-		if err == io.EOF {
-			return i, nil
-		}
-		if err != nil {
-			return 0, err
-		}
-		if !isPrefix {
-			i++
-		}
-	}
-	for limit == 0 || i < skip+limit {
-		// The output of show-ref is simply a list:
-		// <sha> SP <ref> LF
-		sha, err := bufReader.ReadString(' ')
-		if err == io.EOF {
-			return i, nil
-		}
-		if err != nil {
-			return 0, err
-		}
-
-		branchName, err := bufReader.ReadString('\n')
-		if err == io.EOF {
-			// This shouldn't happen... but we'll tolerate it for the sake of peace
-			return i, nil
-		}
-		if err != nil {
-			return i, err
-		}
-
-		if len(branchName) > 0 {
-			branchName = branchName[:len(branchName)-1]
-		}
-
-		if len(sha) > 0 {
-			sha = sha[:len(sha)-1]
-		}
-
-		err = walkfn(sha, branchName)
-		if err != nil {
-			return i, err
-		}
-		i++
-	}
-	// count all refs
-	for limit != 0 {
-		_, isPrefix, err := bufReader.ReadLine()
-		if err == io.EOF {
-			return i, nil
-		}
-		if err != nil {
-			return 0, err
-		}
-		if !isPrefix {
-			i++
-		}
-	}
-	return i, nil
-}
-
-// GetRefsBySha returns all references filtered with prefix that belong to a sha commit hash
-func (repo *Repository) GetRefsBySha(sha, prefix string) ([]string, error) {
-	var revList []string
-	_, err := WalkShowRef(repo.Ctx, repo.Path, nil, 0, 0, func(walkSha, refname string) error {
-		if walkSha == sha && strings.HasPrefix(refname, prefix) {
-			revList = append(revList, refname)
-		}
-		return nil
-	})
-	return revList, err
-}

modules/git/repo_commit.go

@@ -5,12 +5,15 @@
 package git
 
 import (
+	"bufio"
 	"bytes"
+	"errors"
 	"io"
 	"strconv"
 	"strings"
 
 	"code.gitea.io/gitea/modules/cache"
+	"code.gitea.io/gitea/modules/log"
 	"code.gitea.io/gitea/modules/setting"
 )
 
@@ -513,3 +516,149 @@ func (repo *Repository) AddLastCommitCache(cacheKey, fullName, sha string) error
 	}
 	return nil
 }
+
+// ResolveReference resolves a name to a reference
+func (repo *Repository) ResolveReference(name string) (string, error) {
+	stdout, _, err := NewCommand(repo.Ctx, "show-ref", "--hash").AddDynamicArguments(name).RunStdString(&RunOpts{Dir: repo.Path})
+	if err != nil {
+		if strings.Contains(err.Error(), "not a valid ref") {
+			return "", ErrNotExist{name, ""}
+		}
+		return "", err
+	}
+	stdout = strings.TrimSpace(stdout)
+	if stdout == "" {
+		return "", ErrNotExist{name, ""}
+	}
+
+	return stdout, nil
+}
+
+// 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)
+	defer cancel()
+	_, err := wr.Write([]byte(name + "\n"))
+	if err != nil {
+		return "", err
+	}
+	shaBs, _, _, err := ReadBatchLine(rd)
+	if IsErrNotExist(err) {
+		return "", ErrNotExist{name, ""}
+	}
+
+	return string(shaBs), nil
+}
+
+// SetReference sets the commit ID string of given reference (e.g. branch or tag).
+func (repo *Repository) SetReference(name, commitID string) error {
+	_, _, err := NewCommand(repo.Ctx, "update-ref").AddDynamicArguments(name, commitID).RunStdString(&RunOpts{Dir: repo.Path})
+	return err
+}
+
+// RemoveReference removes the given reference (e.g. branch or tag).
+func (repo *Repository) RemoveReference(name string) error {
+	_, _, err := NewCommand(repo.Ctx, "update-ref", "--no-deref", "-d").AddDynamicArguments(name).RunStdString(&RunOpts{Dir: repo.Path})
+	return err
+}
+
+// IsCommitExist returns true if given commit exists in current repository.
+func (repo *Repository) IsCommitExist(name string) bool {
+	_, _, 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)
+	defer cancel()
+
+	_, _ = wr.Write([]byte(id.String() + "\n"))
+
+	return repo.getCommitFromBatchReader(rd, id)
+}
+
+func (repo *Repository) getCommitFromBatchReader(rd *bufio.Reader, id ObjectID) (*Commit, error) {
+	_, typ, size, err := ReadBatchLine(rd)
+	if err != nil {
+		if errors.Is(err, io.EOF) || IsErrNotExist(err) {
+			return nil, ErrNotExist{ID: id.String()}
+		}
+		return nil, err
+	}
+
+	switch typ {
+	case "missing":
+		return nil, ErrNotExist{ID: id.String()}
+	case "tag":
+		// then we need to parse the tag
+		// and load the commit
+		data, err := io.ReadAll(io.LimitReader(rd, size))
+		if err != nil {
+			return nil, err
+		}
+		_, err = rd.Discard(1)
+		if err != nil {
+			return nil, err
+		}
+		tag, err := parseTagData(id.Type(), data)
+		if err != nil {
+			return nil, err
+		}
+
+		commit, err := tag.Commit(repo)
+		if err != nil {
+			return nil, err
+		}
+
+		return commit, nil
+	case "commit":
+		commit, err := CommitFromReader(repo, id, io.LimitReader(rd, size))
+		if err != nil {
+			return nil, err
+		}
+		_, err = rd.Discard(1)
+		if err != nil {
+			return nil, err
+		}
+
+		return commit, nil
+	default:
+		log.Debug("Unknown typ: %s", typ)
+		if err := DiscardFull(rd, size+1); err != nil {
+			return nil, err
+		}
+		return nil, ErrNotExist{
+			ID: id.String(),
+		}
+	}
+}
+
+// ConvertToGitID returns a GitHash object from a potential ID string
+func (repo *Repository) ConvertToGitID(commitID string) (ObjectID, error) {
+	objectFormat, err := repo.GetObjectFormat()
+	if err != nil {
+		return nil, err
+	}
+	if len(commitID) == objectFormat.FullLength() && objectFormat.IsValid(commitID) {
+		ID, err := NewIDFromString(commitID)
+		if err == nil {
+			return ID, nil
+		}
+	}
+
+	wr, rd, cancel := repo.CatFileBatchCheck(repo.Ctx)
+	defer cancel()
+	_, err = wr.Write([]byte(commitID + "\n"))
+	if err != nil {
+		return nil, err
+	}
+	sha, _, _, err := ReadBatchLine(rd)
+	if err != nil {
+		if IsErrNotExist(err) {
+			return nil, ErrNotExist{commitID, ""}
+		}
+		return nil, err
+	}
+
+	return MustIDFromString(string(sha)), nil
+}

modules/git/repo_commit_gogit.go

@@ -1,111 +0,0 @@
-// Copyright 2015 The Gogs Authors. All rights reserved.
-// Copyright 2019 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build gogit
-
-package git
-
-import (
-	"strings"
-
-	"github.com/go-git/go-git/v5/plumbing"
-	"github.com/go-git/go-git/v5/plumbing/hash"
-	"github.com/go-git/go-git/v5/plumbing/object"
-)
-
-// GetRefCommitID returns the last commit ID string of given reference (branch or tag).
-func (repo *Repository) GetRefCommitID(name string) (string, error) {
-	ref, err := repo.gogitRepo.Reference(plumbing.ReferenceName(name), true)
-	if err != nil {
-		if err == plumbing.ErrReferenceNotFound {
-			return "", ErrNotExist{
-				ID: name,
-			}
-		}
-		return "", err
-	}
-
-	return ref.Hash().String(), nil
-}
-
-// SetReference sets the commit ID string of given reference (e.g. branch or tag).
-func (repo *Repository) SetReference(name, commitID string) error {
-	return repo.gogitRepo.Storer.SetReference(plumbing.NewReferenceFromStrings(name, commitID))
-}
-
-// RemoveReference removes the given reference (e.g. branch or tag).
-func (repo *Repository) RemoveReference(name string) error {
-	return repo.gogitRepo.Storer.RemoveReference(plumbing.ReferenceName(name))
-}
-
-// ConvertToHash returns a Hash object from a potential ID string
-func (repo *Repository) ConvertToGitID(commitID string) (ObjectID, error) {
-	objectFormat, err := repo.GetObjectFormat()
-	if err != nil {
-		return nil, err
-	}
-	if len(commitID) == hash.HexSize && objectFormat.IsValid(commitID) {
-		ID, err := NewIDFromString(commitID)
-		if err == nil {
-			return ID, nil
-		}
-	}
-
-	actualCommitID, _, err := NewCommand(repo.Ctx, "rev-parse", "--verify").AddDynamicArguments(commitID).RunStdString(&RunOpts{Dir: repo.Path})
-	actualCommitID = strings.TrimSpace(actualCommitID)
-	if err != nil {
-		if strings.Contains(err.Error(), "unknown revision or path") ||
-			strings.Contains(err.Error(), "fatal: Needed a single revision") {
-			return objectFormat.EmptyObjectID(), ErrNotExist{commitID, ""}
-		}
-		return objectFormat.EmptyObjectID(), err
-	}
-
-	return NewIDFromString(actualCommitID)
-}
-
-// IsCommitExist returns true if given commit exists in current repository.
-func (repo *Repository) IsCommitExist(name string) bool {
-	hash, err := repo.ConvertToGitID(name)
-	if err != nil {
-		return false
-	}
-	_, err = repo.gogitRepo.CommitObject(plumbing.Hash(hash.RawValue()))
-	return err == nil
-}
-
-func (repo *Repository) getCommit(id ObjectID) (*Commit, error) {
-	var tagObject *object.Tag
-
-	commitID := plumbing.Hash(id.RawValue())
-	gogitCommit, err := repo.gogitRepo.CommitObject(commitID)
-	if err == plumbing.ErrObjectNotFound {
-		tagObject, err = repo.gogitRepo.TagObject(commitID)
-		if err == plumbing.ErrObjectNotFound {
-			return nil, ErrNotExist{
-				ID: id.String(),
-			}
-		}
-		if err == nil {
-			gogitCommit, err = repo.gogitRepo.CommitObject(tagObject.Target)
-		}
-		// if we get a plumbing.ErrObjectNotFound here then the repository is broken and it should be 500
-	}
-	if err != nil {
-		return nil, err
-	}
-
-	commit := convertCommit(gogitCommit)
-	commit.repo = repo
-
-	tree, err := gogitCommit.Tree()
-	if err != nil {
-		return nil, err
-	}
-
-	commit.Tree.ID = ParseGogitHash(tree.Hash)
-	commit.Tree.gogitTree = tree
-
-	return commit, nil
-}

modules/git/repo_commit_nogogit.go

@@ -1,161 +0,0 @@
-// Copyright 2020 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build !gogit
-
-package git
-
-import (
-	"bufio"
-	"errors"
-	"io"
-	"strings"
-
-	"code.gitea.io/gitea/modules/log"
-)
-
-// ResolveReference resolves a name to a reference
-func (repo *Repository) ResolveReference(name string) (string, error) {
-	stdout, _, err := NewCommand(repo.Ctx, "show-ref", "--hash").AddDynamicArguments(name).RunStdString(&RunOpts{Dir: repo.Path})
-	if err != nil {
-		if strings.Contains(err.Error(), "not a valid ref") {
-			return "", ErrNotExist{name, ""}
-		}
-		return "", err
-	}
-	stdout = strings.TrimSpace(stdout)
-	if stdout == "" {
-		return "", ErrNotExist{name, ""}
-	}
-
-	return stdout, nil
-}
-
-// 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)
-	defer cancel()
-	_, err := wr.Write([]byte(name + "\n"))
-	if err != nil {
-		return "", err
-	}
-	shaBs, _, _, err := ReadBatchLine(rd)
-	if IsErrNotExist(err) {
-		return "", ErrNotExist{name, ""}
-	}
-
-	return string(shaBs), nil
-}
-
-// SetReference sets the commit ID string of given reference (e.g. branch or tag).
-func (repo *Repository) SetReference(name, commitID string) error {
-	_, _, err := NewCommand(repo.Ctx, "update-ref").AddDynamicArguments(name, commitID).RunStdString(&RunOpts{Dir: repo.Path})
-	return err
-}
-
-// RemoveReference removes the given reference (e.g. branch or tag).
-func (repo *Repository) RemoveReference(name string) error {
-	_, _, err := NewCommand(repo.Ctx, "update-ref", "--no-deref", "-d").AddDynamicArguments(name).RunStdString(&RunOpts{Dir: repo.Path})
-	return err
-}
-
-// IsCommitExist returns true if given commit exists in current repository.
-func (repo *Repository) IsCommitExist(name string) bool {
-	_, _, 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)
-	defer cancel()
-
-	_, _ = wr.Write([]byte(id.String() + "\n"))
-
-	return repo.getCommitFromBatchReader(rd, id)
-}
-
-func (repo *Repository) getCommitFromBatchReader(rd *bufio.Reader, id ObjectID) (*Commit, error) {
-	_, typ, size, err := ReadBatchLine(rd)
-	if err != nil {
-		if errors.Is(err, io.EOF) || IsErrNotExist(err) {
-			return nil, ErrNotExist{ID: id.String()}
-		}
-		return nil, err
-	}
-
-	switch typ {
-	case "missing":
-		return nil, ErrNotExist{ID: id.String()}
-	case "tag":
-		// then we need to parse the tag
-		// and load the commit
-		data, err := io.ReadAll(io.LimitReader(rd, size))
-		if err != nil {
-			return nil, err
-		}
-		_, err = rd.Discard(1)
-		if err != nil {
-			return nil, err
-		}
-		tag, err := parseTagData(id.Type(), data)
-		if err != nil {
-			return nil, err
-		}
-
-		commit, err := tag.Commit(repo)
-		if err != nil {
-			return nil, err
-		}
-
-		return commit, nil
-	case "commit":
-		commit, err := CommitFromReader(repo, id, io.LimitReader(rd, size))
-		if err != nil {
-			return nil, err
-		}
-		_, err = rd.Discard(1)
-		if err != nil {
-			return nil, err
-		}
-
-		return commit, nil
-	default:
-		log.Debug("Unknown typ: %s", typ)
-		if err := DiscardFull(rd, size+1); err != nil {
-			return nil, err
-		}
-		return nil, ErrNotExist{
-			ID: id.String(),
-		}
-	}
-}
-
-// ConvertToGitID returns a GitHash object from a potential ID string
-func (repo *Repository) ConvertToGitID(commitID string) (ObjectID, error) {
-	objectFormat, err := repo.GetObjectFormat()
-	if err != nil {
-		return nil, err
-	}
-	if len(commitID) == objectFormat.FullLength() && objectFormat.IsValid(commitID) {
-		ID, err := NewIDFromString(commitID)
-		if err == nil {
-			return ID, nil
-		}
-	}
-
-	wr, rd, cancel := repo.CatFileBatchCheck(repo.Ctx)
-	defer cancel()
-	_, err = wr.Write([]byte(commitID + "\n"))
-	if err != nil {
-		return nil, err
-	}
-	sha, _, _, err := ReadBatchLine(rd)
-	if err != nil {
-		if IsErrNotExist(err) {
-			return nil, ErrNotExist{commitID, ""}
-		}
-		return nil, err
-	}
-
-	return MustIDFromString(string(sha)), nil
-}

modules/git/repo_commitgraph_gogit.go

@@ -1,37 +0,0 @@
-// Copyright 2019 The Gitea Authors.
-// All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build gogit
-
-package git
-
-import (
-	"os"
-	"path"
-
-	gitealog "code.gitea.io/gitea/modules/log"
-
-	commitgraph "github.com/go-git/go-git/v5/plumbing/format/commitgraph/v2"
-	cgobject "github.com/go-git/go-git/v5/plumbing/object/commitgraph"
-)
-
-// CommitNodeIndex returns the index for walking commit graph
-func (r *Repository) CommitNodeIndex() (cgobject.CommitNodeIndex, *os.File) {
-	indexPath := path.Join(r.Path, "objects", "info", "commit-graph")
-
-	file, err := os.Open(indexPath)
-	if err == nil {
-		var index commitgraph.Index
-		index, err = commitgraph.OpenFileIndex(file)
-		if err == nil {
-			return cgobject.NewGraphCommitNodeIndex(index, r.gogitRepo.Storer), file
-		}
-	}
-
-	if !os.IsNotExist(err) {
-		gitealog.Warn("Unable to read commit-graph for %s: %v", r.Path, err)
-	}
-
-	return cgobject.NewObjectCommitNodeIndex(r.gogitRepo.Storer), nil
-}

modules/git/repo_language_stats.go

@@ -4,8 +4,17 @@
 package git
 
 import (
+	"bytes"
+	"cmp"
+	"io"
 	"strings"
 	"unicode"
+
+	"code.gitea.io/gitea/modules/analyze"
+	"code.gitea.io/gitea/modules/log"
+	"code.gitea.io/gitea/modules/optional"
+
+	"github.com/go-enry/go-enry/v2"
 )
 
 const (
@@ -46,3 +55,194 @@ func mergeLanguageStats(stats map[string]int64) map[string]int64 {
 	}
 	return res
 }
+
+// GetLanguageStats calculates language stats for git repository at specified commit
+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)
+	defer cancel()
+
+	writeID := func(id string) error {
+		_, err := batchStdinWriter.Write([]byte(id + "\n"))
+		return err
+	}
+
+	if err := writeID(commitID); err != nil {
+		return nil, err
+	}
+	shaBytes, typ, size, err := ReadBatchLine(batchReader)
+	if typ != "commit" {
+		log.Debug("Unable to get commit for: %s. Err: %v", commitID, err)
+		return nil, ErrNotExist{commitID, ""}
+	}
+
+	sha, err := NewIDFromString(string(shaBytes))
+	if err != nil {
+		log.Debug("Unable to get commit for: %s. Err: %v", commitID, err)
+		return nil, ErrNotExist{commitID, ""}
+	}
+
+	commit, err := CommitFromReader(repo, sha, io.LimitReader(batchReader, size))
+	if err != nil {
+		log.Debug("Unable to get commit for: %s. Err: %v", commitID, err)
+		return nil, err
+	}
+	if _, err = batchReader.Discard(1); err != nil {
+		return nil, err
+	}
+
+	tree := commit.Tree
+
+	entries, err := tree.ListEntriesRecursiveWithSize()
+	if err != nil {
+		return nil, err
+	}
+
+	checker, err := repo.GitAttributeChecker(commitID, LinguistAttributes...)
+	if err != nil {
+		return nil, err
+	}
+	defer checker.Close()
+
+	contentBuf := bytes.Buffer{}
+	var content []byte
+
+	// sizes contains the current calculated size of all files by language
+	sizes := make(map[string]int64)
+	// by default we will only count the sizes of programming languages or markup languages
+	// unless they are explicitly set using linguist-language
+	includedLanguage := map[string]bool{}
+	// or if there's only one language in the repository
+	firstExcludedLanguage := ""
+	firstExcludedLanguageSize := int64(0)
+
+	isTrue := func(v optional.Option[bool]) bool {
+		return v.ValueOrDefault(false)
+	}
+	isFalse := func(v optional.Option[bool]) bool {
+		return !v.ValueOrDefault(true)
+	}
+
+	for _, f := range entries {
+		select {
+		case <-repo.Ctx.Done():
+			return sizes, repo.Ctx.Err()
+		default:
+		}
+
+		contentBuf.Reset()
+		content = contentBuf.Bytes()
+
+		if f.Size() == 0 {
+			continue
+		}
+
+		isVendored := optional.None[bool]()
+		isGenerated := optional.None[bool]()
+		isDocumentation := optional.None[bool]()
+		isDetectable := optional.None[bool]()
+
+		attrs, err := checker.CheckPath(f.Name())
+		if err == nil {
+			isVendored = attrs["linguist-vendored"].Bool()
+			isGenerated = attrs["linguist-generated"].Bool()
+			isDocumentation = attrs["linguist-documentation"].Bool()
+			isDetectable = attrs["linguist-detectable"].Bool()
+			if language := cmp.Or(
+				attrs["linguist-language"].String(),
+				attrs["gitlab-language"].Prefix(),
+			); language != "" {
+				// group languages, such as Pug -> HTML; SCSS -> CSS
+				group := enry.GetLanguageGroup(language)
+				if len(group) != 0 {
+					language = group
+				}
+
+				// this language will always be added to the size
+				sizes[language] += f.Size()
+				continue
+			}
+		}
+
+		if isFalse(isDetectable) || isTrue(isVendored) || isTrue(isDocumentation) ||
+			(!isFalse(isVendored) && analyze.IsVendor(f.Name())) ||
+			enry.IsDotFile(f.Name()) ||
+			enry.IsConfiguration(f.Name()) ||
+			(!isFalse(isDocumentation) && enry.IsDocumentation(f.Name())) {
+			continue
+		}
+
+		// If content can not be read or file is too big just do detection by filename
+
+		if f.Size() <= bigFileSize {
+			if err := writeID(f.ID.String()); err != nil {
+				return nil, err
+			}
+			_, _, size, err := ReadBatchLine(batchReader)
+			if err != nil {
+				log.Debug("Error reading blob: %s Err: %v", f.ID.String(), err)
+				return nil, err
+			}
+
+			sizeToRead := size
+			discard := int64(1)
+			if size > fileSizeLimit {
+				sizeToRead = fileSizeLimit
+				discard = size - fileSizeLimit + 1
+			}
+
+			_, err = contentBuf.ReadFrom(io.LimitReader(batchReader, sizeToRead))
+			if err != nil {
+				return nil, err
+			}
+			content = contentBuf.Bytes()
+			if err := DiscardFull(batchReader, discard); err != nil {
+				return nil, err
+			}
+		}
+		if !isTrue(isGenerated) && enry.IsGenerated(f.Name(), content) {
+			continue
+		}
+
+		// FIXME: Why can't we split this and the IsGenerated tests to avoid reading the blob unless absolutely necessary?
+		// - eg. do the all the detection tests using filename first before reading content.
+		language := analyze.GetCodeLanguage(f.Name(), content)
+		if language == "" {
+			continue
+		}
+
+		// group languages, such as Pug -> HTML; SCSS -> CSS
+		group := enry.GetLanguageGroup(language)
+		if group != "" {
+			language = group
+		}
+
+		included, checked := includedLanguage[language]
+		langType := enry.GetLanguageType(language)
+		if !checked {
+			included = langType == enry.Programming || langType == enry.Markup
+			if !included && (isTrue(isDetectable) || (langType == enry.Prose && isFalse(isDocumentation))) {
+				included = true
+			}
+			includedLanguage[language] = included
+		}
+		if included {
+			sizes[language] += f.Size()
+		} else if len(sizes) == 0 && (firstExcludedLanguage == "" || firstExcludedLanguage == language) {
+			// Only consider Programming or Markup languages as fallback
+			if !(langType == enry.Programming || langType == enry.Markup) {
+				continue
+			}
+			firstExcludedLanguage = language
+			firstExcludedLanguageSize += f.Size()
+		}
+	}
+
+	// If there are no included languages add the first excluded language
+	if len(sizes) == 0 && firstExcludedLanguage != "" {
+		sizes[firstExcludedLanguage] = firstExcludedLanguageSize
+	}
+
+	return mergeLanguageStats(sizes), nil
+}

modules/git/repo_language_stats_gogit.go

@@ -1,194 +0,0 @@
-// Copyright 2020 The Gitea Authors. All rights reserved.
-// Copyright 2024 The Forgejo Authors c/o Codeberg e.V.. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build gogit
-
-package git
-
-import (
-	"bytes"
-	"io"
-	"strings"
-
-	"code.gitea.io/gitea/modules/analyze"
-	"code.gitea.io/gitea/modules/optional"
-
-	"github.com/go-enry/go-enry/v2"
-	"github.com/go-git/go-git/v5"
-	"github.com/go-git/go-git/v5/plumbing"
-	"github.com/go-git/go-git/v5/plumbing/object"
-)
-
-// GetLanguageStats calculates language stats for git repository at specified commit
-func (repo *Repository) GetLanguageStats(commitID string) (map[string]int64, error) {
-	r, err := git.PlainOpen(repo.Path)
-	if err != nil {
-		return nil, err
-	}
-
-	rev, err := r.ResolveRevision(plumbing.Revision(commitID))
-	if err != nil {
-		return nil, err
-	}
-
-	commit, err := r.CommitObject(*rev)
-	if err != nil {
-		return nil, err
-	}
-
-	tree, err := commit.Tree()
-	if err != nil {
-		return nil, err
-	}
-
-	checker, deferable := repo.CheckAttributeReader(commitID)
-	defer deferable()
-
-	// sizes contains the current calculated size of all files by language
-	sizes := make(map[string]int64)
-	// by default we will only count the sizes of programming languages or markup languages
-	// unless they are explicitly set using linguist-language
-	includedLanguage := map[string]bool{}
-	// or if there's only one language in the repository
-	firstExcludedLanguage := ""
-	firstExcludedLanguageSize := int64(0)
-
-	isTrue := func(v optional.Option[bool]) bool {
-		return v.ValueOrDefault(false)
-	}
-	isFalse := func(v optional.Option[bool]) bool {
-		return !v.ValueOrDefault(true)
-	}
-
-	err = tree.Files().ForEach(func(f *object.File) error {
-		if f.Size == 0 {
-			return nil
-		}
-
-		isVendored := optional.None[bool]()
-		isGenerated := optional.None[bool]()
-		isDocumentation := optional.None[bool]()
-		isDetectable := optional.None[bool]()
-
-		if checker != nil {
-			attrs, err := checker.CheckPath(f.Name)
-			if err == nil {
-				isVendored = attributeToBool(attrs, "linguist-vendored")
-				isGenerated = attributeToBool(attrs, "linguist-generated")
-				isDocumentation = attributeToBool(attrs, "linguist-documentation")
-				isDetectable = attributeToBool(attrs, "linguist-detectable")
-				if language, has := attrs["linguist-language"]; has && language != "unspecified" && language != "" {
-					// group languages, such as Pug -> HTML; SCSS -> CSS
-					group := enry.GetLanguageGroup(language)
-					if len(group) != 0 {
-						language = group
-					}
-
-					// this language will always be added to the size
-					sizes[language] += f.Size
-					return nil
-				} else if language, has := attrs["gitlab-language"]; has && language != "unspecified" && language != "" {
-					// strip off a ? if present
-					if idx := strings.IndexByte(language, '?'); idx >= 0 {
-						language = language[:idx]
-					}
-					if len(language) != 0 {
-						// group languages, such as Pug -> HTML; SCSS -> CSS
-						group := enry.GetLanguageGroup(language)
-						if len(group) != 0 {
-							language = group
-						}
-
-						// this language will always be added to the size
-						sizes[language] += f.Size
-						return nil
-					}
-				}
-			}
-		}
-
-		if isFalse(isDetectable) || isTrue(isVendored) || isTrue(isDocumentation) ||
-			(!isFalse(isVendored) && analyze.IsVendor(f.Name)) ||
-			enry.IsDotFile(f.Name) ||
-			enry.IsConfiguration(f.Name) ||
-			(!isFalse(isDocumentation) && enry.IsDocumentation(f.Name)) {
-			return nil
-		}
-
-		// If content can not be read or file is too big just do detection by filename
-		var content []byte
-		if f.Size <= bigFileSize {
-			content, _ = readFile(f, fileSizeLimit)
-		}
-		if !isTrue(isGenerated) && enry.IsGenerated(f.Name, content) {
-			return nil
-		}
-
-		// TODO: Use .gitattributes file for linguist overrides
-		language := analyze.GetCodeLanguage(f.Name, content)
-		if language == enry.OtherLanguage || language == "" {
-			return nil
-		}
-
-		// group languages, such as Pug -> HTML; SCSS -> CSS
-		group := enry.GetLanguageGroup(language)
-		if group != "" {
-			language = group
-		}
-
-		included, checked := includedLanguage[language]
-		langType := enry.GetLanguageType(language)
-		if !checked {
-			included = langType == enry.Programming || langType == enry.Markup
-			if !included && (isTrue(isDetectable) || (langType == enry.Prose && isFalse(isDocumentation))) {
-				included = true
-			}
-			includedLanguage[language] = included
-		}
-		if included {
-			sizes[language] += f.Size
-		} else if len(sizes) == 0 && (firstExcludedLanguage == "" || firstExcludedLanguage == language) {
-			// Only consider Programming or Markup languages as fallback
-			if !(langType == enry.Programming || langType == enry.Markup) {
-				return nil
-			}
-
-			firstExcludedLanguage = language
-			firstExcludedLanguageSize += f.Size
-		}
-
-		return nil
-	})
-	if err != nil {
-		return nil, err
-	}
-
-	// If there are no included languages add the first excluded language
-	if len(sizes) == 0 && firstExcludedLanguage != "" {
-		sizes[firstExcludedLanguage] = firstExcludedLanguageSize
-	}
-
-	return mergeLanguageStats(sizes), nil
-}
-
-func readFile(f *object.File, limit int64) ([]byte, error) {
-	r, err := f.Reader()
-	if err != nil {
-		return nil, err
-	}
-	defer r.Close()
-
-	if limit <= 0 {
-		return io.ReadAll(r)
-	}
-
-	size := f.Size
-	if limit > 0 && size > limit {
-		size = limit
-	}
-	buf := bytes.NewBuffer(nil)
-	buf.Grow(int(size))
-	_, err = io.Copy(buf, io.LimitReader(r, limit))
-	return buf.Bytes(), err
-}

modules/git/repo_language_stats_nogogit.go

@@ -1,210 +0,0 @@
-// Copyright 2020 The Gitea Authors. All rights reserved.
-// Copyright 2024 The Forgejo Authors c/o Codeberg e.V.. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build !gogit
-
-package git
-
-import (
-	"bytes"
-	"cmp"
-	"io"
-
-	"code.gitea.io/gitea/modules/analyze"
-	"code.gitea.io/gitea/modules/log"
-	"code.gitea.io/gitea/modules/optional"
-
-	"github.com/go-enry/go-enry/v2"
-)
-
-// GetLanguageStats calculates language stats for git repository at specified commit
-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)
-	defer cancel()
-
-	writeID := func(id string) error {
-		_, err := batchStdinWriter.Write([]byte(id + "\n"))
-		return err
-	}
-
-	if err := writeID(commitID); err != nil {
-		return nil, err
-	}
-	shaBytes, typ, size, err := ReadBatchLine(batchReader)
-	if typ != "commit" {
-		log.Debug("Unable to get commit for: %s. Err: %v", commitID, err)
-		return nil, ErrNotExist{commitID, ""}
-	}
-
-	sha, err := NewIDFromString(string(shaBytes))
-	if err != nil {
-		log.Debug("Unable to get commit for: %s. Err: %v", commitID, err)
-		return nil, ErrNotExist{commitID, ""}
-	}
-
-	commit, err := CommitFromReader(repo, sha, io.LimitReader(batchReader, size))
-	if err != nil {
-		log.Debug("Unable to get commit for: %s. Err: %v", commitID, err)
-		return nil, err
-	}
-	if _, err = batchReader.Discard(1); err != nil {
-		return nil, err
-	}
-
-	tree := commit.Tree
-
-	entries, err := tree.ListEntriesRecursiveWithSize()
-	if err != nil {
-		return nil, err
-	}
-
-	checker, err := repo.GitAttributeChecker(commitID, LinguistAttributes...)
-	if err != nil {
-		return nil, err
-	}
-	defer checker.Close()
-
-	contentBuf := bytes.Buffer{}
-	var content []byte
-
-	// sizes contains the current calculated size of all files by language
-	sizes := make(map[string]int64)
-	// by default we will only count the sizes of programming languages or markup languages
-	// unless they are explicitly set using linguist-language
-	includedLanguage := map[string]bool{}
-	// or if there's only one language in the repository
-	firstExcludedLanguage := ""
-	firstExcludedLanguageSize := int64(0)
-
-	isTrue := func(v optional.Option[bool]) bool {
-		return v.ValueOrDefault(false)
-	}
-	isFalse := func(v optional.Option[bool]) bool {
-		return !v.ValueOrDefault(true)
-	}
-
-	for _, f := range entries {
-		select {
-		case <-repo.Ctx.Done():
-			return sizes, repo.Ctx.Err()
-		default:
-		}
-
-		contentBuf.Reset()
-		content = contentBuf.Bytes()
-
-		if f.Size() == 0 {
-			continue
-		}
-
-		isVendored := optional.None[bool]()
-		isGenerated := optional.None[bool]()
-		isDocumentation := optional.None[bool]()
-		isDetectable := optional.None[bool]()
-
-		attrs, err := checker.CheckPath(f.Name())
-		if err == nil {
-			isVendored = attrs["linguist-vendored"].Bool()
-			isGenerated = attrs["linguist-generated"].Bool()
-			isDocumentation = attrs["linguist-documentation"].Bool()
-			isDetectable = attrs["linguist-detectable"].Bool()
-			if language := cmp.Or(
-				attrs["linguist-language"].String(),
-				attrs["gitlab-language"].Prefix(),
-			); language != "" {
-				// group languages, such as Pug -> HTML; SCSS -> CSS
-				group := enry.GetLanguageGroup(language)
-				if len(group) != 0 {
-					language = group
-				}
-
-				// this language will always be added to the size
-				sizes[language] += f.Size()
-				continue
-			}
-		}
-
-		if isFalse(isDetectable) || isTrue(isVendored) || isTrue(isDocumentation) ||
-			(!isFalse(isVendored) && analyze.IsVendor(f.Name())) ||
-			enry.IsDotFile(f.Name()) ||
-			enry.IsConfiguration(f.Name()) ||
-			(!isFalse(isDocumentation) && enry.IsDocumentation(f.Name())) {
-			continue
-		}
-
-		// If content can not be read or file is too big just do detection by filename
-
-		if f.Size() <= bigFileSize {
-			if err := writeID(f.ID.String()); err != nil {
-				return nil, err
-			}
-			_, _, size, err := ReadBatchLine(batchReader)
-			if err != nil {
-				log.Debug("Error reading blob: %s Err: %v", f.ID.String(), err)
-				return nil, err
-			}
-
-			sizeToRead := size
-			discard := int64(1)
-			if size > fileSizeLimit {
-				sizeToRead = fileSizeLimit
-				discard = size - fileSizeLimit + 1
-			}
-
-			_, err = contentBuf.ReadFrom(io.LimitReader(batchReader, sizeToRead))
-			if err != nil {
-				return nil, err
-			}
-			content = contentBuf.Bytes()
-			if err := DiscardFull(batchReader, discard); err != nil {
-				return nil, err
-			}
-		}
-		if !isTrue(isGenerated) && enry.IsGenerated(f.Name(), content) {
-			continue
-		}
-
-		// FIXME: Why can't we split this and the IsGenerated tests to avoid reading the blob unless absolutely necessary?
-		// - eg. do the all the detection tests using filename first before reading content.
-		language := analyze.GetCodeLanguage(f.Name(), content)
-		if language == "" {
-			continue
-		}
-
-		// group languages, such as Pug -> HTML; SCSS -> CSS
-		group := enry.GetLanguageGroup(language)
-		if group != "" {
-			language = group
-		}
-
-		included, checked := includedLanguage[language]
-		langType := enry.GetLanguageType(language)
-		if !checked {
-			included = langType == enry.Programming || langType == enry.Markup
-			if !included && (isTrue(isDetectable) || (langType == enry.Prose && isFalse(isDocumentation))) {
-				included = true
-			}
-			includedLanguage[language] = included
-		}
-		if included {
-			sizes[language] += f.Size()
-		} else if len(sizes) == 0 && (firstExcludedLanguage == "" || firstExcludedLanguage == language) {
-			// Only consider Programming or Markup languages as fallback
-			if !(langType == enry.Programming || langType == enry.Markup) {
-				continue
-			}
-			firstExcludedLanguage = language
-			firstExcludedLanguageSize += f.Size()
-		}
-	}
-
-	// If there are no included languages add the first excluded language
-	if len(sizes) == 0 && firstExcludedLanguage != "" {
-		sizes[firstExcludedLanguage] = firstExcludedLanguageSize
-	}
-
-	return mergeLanguageStats(sizes), nil
-}

modules/git/repo_language_stats_test.go

@@ -1,8 +1,6 @@
 // Copyright 2020 The Gitea Authors. All rights reserved.
 // SPDX-License-Identifier: MIT
 
-//go:build !gogit
-
 package git
 
 import (

modules/git/repo_ref.go

@@ -4,8 +4,10 @@
 package git
 
 import (
+	"bufio"
 	"context"
 	"fmt"
+	"io"
 	"strings"
 
 	"code.gitea.io/gitea/modules/util"
@@ -78,3 +80,78 @@ func (repo *Repository) ExpandRef(ref string) (string, error) {
 	}
 	return "", fmt.Errorf("could not expand reference '%s'", ref)
 }
+
+// GetRefsFiltered returns all references of the repository that matches patterm exactly or starting with.
+func (repo *Repository) GetRefsFiltered(pattern string) ([]*Reference, error) {
+	stdoutReader, stdoutWriter := io.Pipe()
+	defer func() {
+		_ = stdoutReader.Close()
+		_ = stdoutWriter.Close()
+	}()
+
+	go func() {
+		stderrBuilder := &strings.Builder{}
+		err := NewCommand(repo.Ctx, "for-each-ref").Run(&RunOpts{
+			Dir:    repo.Path,
+			Stdout: stdoutWriter,
+			Stderr: stderrBuilder,
+		})
+		if err != nil {
+			_ = stdoutWriter.CloseWithError(ConcatenateError(err, stderrBuilder.String()))
+		} else {
+			_ = stdoutWriter.Close()
+		}
+	}()
+
+	refs := make([]*Reference, 0)
+	bufReader := bufio.NewReader(stdoutReader)
+	for {
+		// The output of for-each-ref is simply a list:
+		// <sha> SP <type> TAB <ref> LF
+		sha, err := bufReader.ReadString(' ')
+		if err == io.EOF {
+			break
+		}
+		if err != nil {
+			return nil, err
+		}
+		sha = sha[:len(sha)-1]
+
+		typ, err := bufReader.ReadString('\t')
+		if err == io.EOF {
+			// This should not happen, but we'll tolerate it
+			break
+		}
+		if err != nil {
+			return nil, err
+		}
+		typ = typ[:len(typ)-1]
+
+		refName, err := bufReader.ReadString('\n')
+		if err == io.EOF {
+			// This should not happen, but we'll tolerate it
+			break
+		}
+		if err != nil {
+			return nil, err
+		}
+		refName = refName[:len(refName)-1]
+
+		// refName cannot be HEAD but can be remotes or stash
+		if strings.HasPrefix(refName, RemotePrefix) || refName == "/refs/stash" {
+			continue
+		}
+
+		if pattern == "" || strings.HasPrefix(refName, pattern) {
+			r := &Reference{
+				Name:   refName,
+				Object: MustIDFromString(sha),
+				Type:   typ,
+				repo:   repo,
+			}
+			refs = append(refs, r)
+		}
+	}
+
+	return refs, nil
+}

modules/git/repo_ref_gogit.go

@@ -1,51 +0,0 @@
-// Copyright 2018 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build gogit
-
-package git
-
-import (
-	"strings"
-
-	"github.com/go-git/go-git/v5"
-	"github.com/go-git/go-git/v5/plumbing"
-)
-
-// GetRefsFiltered returns all references of the repository that matches patterm exactly or starting with.
-func (repo *Repository) GetRefsFiltered(pattern string) ([]*Reference, error) {
-	r, err := git.PlainOpen(repo.Path)
-	if err != nil {
-		return nil, err
-	}
-
-	refsIter, err := r.References()
-	if err != nil {
-		return nil, err
-	}
-	refs := make([]*Reference, 0)
-	if err = refsIter.ForEach(func(ref *plumbing.Reference) error {
-		if ref.Name() != plumbing.HEAD && !ref.Name().IsRemote() &&
-			(pattern == "" || strings.HasPrefix(ref.Name().String(), pattern)) {
-			refType := string(ObjectCommit)
-			if ref.Name().IsTag() {
-				// tags can be of type `commit` (lightweight) or `tag` (annotated)
-				if tagType, _ := repo.GetTagType(ParseGogitHash(ref.Hash())); err == nil {
-					refType = tagType
-				}
-			}
-			r := &Reference{
-				Name:   ref.Name().String(),
-				Object: ParseGogitHash(ref.Hash()),
-				Type:   refType,
-				repo:   repo,
-			}
-			refs = append(refs, r)
-		}
-		return nil
-	}); err != nil {
-		return nil, err
-	}
-
-	return refs, nil
-}

modules/git/repo_ref_nogogit.go

@@ -1,87 +0,0 @@
-// Copyright 2020 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build !gogit
-
-package git
-
-import (
-	"bufio"
-	"io"
-	"strings"
-)
-
-// GetRefsFiltered returns all references of the repository that matches patterm exactly or starting with.
-func (repo *Repository) GetRefsFiltered(pattern string) ([]*Reference, error) {
-	stdoutReader, stdoutWriter := io.Pipe()
-	defer func() {
-		_ = stdoutReader.Close()
-		_ = stdoutWriter.Close()
-	}()
-
-	go func() {
-		stderrBuilder := &strings.Builder{}
-		err := NewCommand(repo.Ctx, "for-each-ref").Run(&RunOpts{
-			Dir:    repo.Path,
-			Stdout: stdoutWriter,
-			Stderr: stderrBuilder,
-		})
-		if err != nil {
-			_ = stdoutWriter.CloseWithError(ConcatenateError(err, stderrBuilder.String()))
-		} else {
-			_ = stdoutWriter.Close()
-		}
-	}()
-
-	refs := make([]*Reference, 0)
-	bufReader := bufio.NewReader(stdoutReader)
-	for {
-		// The output of for-each-ref is simply a list:
-		// <sha> SP <type> TAB <ref> LF
-		sha, err := bufReader.ReadString(' ')
-		if err == io.EOF {
-			break
-		}
-		if err != nil {
-			return nil, err
-		}
-		sha = sha[:len(sha)-1]
-
-		typ, err := bufReader.ReadString('\t')
-		if err == io.EOF {
-			// This should not happen, but we'll tolerate it
-			break
-		}
-		if err != nil {
-			return nil, err
-		}
-		typ = typ[:len(typ)-1]
-
-		refName, err := bufReader.ReadString('\n')
-		if err == io.EOF {
-			// This should not happen, but we'll tolerate it
-			break
-		}
-		if err != nil {
-			return nil, err
-		}
-		refName = refName[:len(refName)-1]
-
-		// refName cannot be HEAD but can be remotes or stash
-		if strings.HasPrefix(refName, RemotePrefix) || refName == "/refs/stash" {
-			continue
-		}
-
-		if pattern == "" || strings.HasPrefix(refName, pattern) {
-			r := &Reference{
-				Name:   refName,
-				Object: MustIDFromString(sha),
-				Type:   typ,
-				repo:   repo,
-			}
-			refs = append(refs, r)
-		}
-	}
-
-	return refs, nil
-}

modules/git/repo_tag.go

@@ -6,11 +6,13 @@ package git
 
 import (
 	"context"
+	"errors"
 	"fmt"
 	"io"
 	"strings"
 
 	"code.gitea.io/gitea/modules/git/foreachref"
+	"code.gitea.io/gitea/modules/log"
 	"code.gitea.io/gitea/modules/util"
 )
 
@@ -236,3 +238,123 @@ func (repo *Repository) GetAnnotatedTag(sha string) (*Tag, error) {
 	}
 	return tag, nil
 }
+
+// IsTagExist returns true if given tag exists in the repository.
+func (repo *Repository) IsTagExist(name string) bool {
+	if repo == nil || name == "" {
+		return false
+	}
+
+	return repo.IsReferenceExist(TagPrefix + name)
+}
+
+// GetTags returns all tags of the repository.
+// returning at most limit tags, or all if limit is 0.
+func (repo *Repository) GetTags(skip, limit int) (tags []string, err error) {
+	tags, _, err = callShowRef(repo.Ctx, repo.Path, TagPrefix, TrustedCmdArgs{TagPrefix, "--sort=-taggerdate"}, skip, limit)
+	return tags, err
+}
+
+// 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)
+	defer cancel()
+	_, err := wr.Write([]byte(id.String() + "\n"))
+	if err != nil {
+		return "", err
+	}
+	_, typ, _, err := ReadBatchLine(rd)
+	if IsErrNotExist(err) {
+		return "", ErrNotExist{ID: id.String()}
+	}
+	return typ, nil
+}
+
+func (repo *Repository) getTag(tagID ObjectID, name string) (*Tag, error) {
+	t, ok := repo.tagCache.Get(tagID.String())
+	if ok {
+		log.Debug("Hit cache: %s", tagID)
+		tagClone := *t.(*Tag)
+		tagClone.Name = name // This is necessary because lightweight tags may have same id
+		return &tagClone, nil
+	}
+
+	tp, err := repo.GetTagType(tagID)
+	if err != nil {
+		return nil, err
+	}
+
+	// Get the commit ID and tag ID (may be different for annotated tag) for the returned tag object
+	commitIDStr, err := repo.GetTagCommitID(name)
+	if err != nil {
+		// every tag should have a commit ID so return all errors
+		return nil, err
+	}
+	commitID, err := NewIDFromString(commitIDStr)
+	if err != nil {
+		return nil, err
+	}
+
+	// If type is "commit, the tag is a lightweight tag
+	if ObjectType(tp) == ObjectCommit {
+		commit, err := repo.GetCommit(commitIDStr)
+		if err != nil {
+			return nil, err
+		}
+		tag := &Tag{
+			Name:    name,
+			ID:      tagID,
+			Object:  commitID,
+			Type:    tp,
+			Tagger:  commit.Committer,
+			Message: commit.Message(),
+		}
+
+		repo.tagCache.Set(tagID.String(), tag)
+		return tag, nil
+	}
+
+	// The tag is an annotated tag with a message.
+	wr, rd, cancel := repo.CatFileBatch(repo.Ctx)
+	defer cancel()
+
+	if _, err := wr.Write([]byte(tagID.String() + "\n")); err != nil {
+		return nil, err
+	}
+	_, typ, size, err := ReadBatchLine(rd)
+	if err != nil {
+		if errors.Is(err, io.EOF) || IsErrNotExist(err) {
+			return nil, ErrNotExist{ID: tagID.String()}
+		}
+		return nil, err
+	}
+	if typ != "tag" {
+		if err := DiscardFull(rd, size+1); err != nil {
+			return nil, err
+		}
+		return nil, ErrNotExist{ID: tagID.String()}
+	}
+
+	// then we need to parse the tag
+	// and load the commit
+	data, err := io.ReadAll(io.LimitReader(rd, size))
+	if err != nil {
+		return nil, err
+	}
+	_, err = rd.Discard(1)
+	if err != nil {
+		return nil, err
+	}
+
+	tag, err := parseTagData(tagID.Type(), data)
+	if err != nil {
+		return nil, err
+	}
+
+	tag.Name = name
+	tag.ID = tagID
+	tag.Type = tp
+
+	repo.tagCache.Set(tagID.String(), tag)
+	return tag, nil
+}

modules/git/repo_tag_gogit.go

@@ -1,135 +0,0 @@
-// Copyright 2015 The Gogs Authors. All rights reserved.
-// Copyright 2019 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build gogit
-
-package git
-
-import (
-	"strings"
-
-	"code.gitea.io/gitea/modules/log"
-
-	"github.com/go-git/go-git/v5/plumbing"
-)
-
-// IsTagExist returns true if given tag exists in the repository.
-func (repo *Repository) IsTagExist(name string) bool {
-	_, err := repo.gogitRepo.Reference(plumbing.ReferenceName(TagPrefix+name), true)
-	return err == nil
-}
-
-// GetTags returns all tags of the repository.
-// returning at most limit tags, or all if limit is 0.
-func (repo *Repository) GetTags(skip, limit int) ([]string, error) {
-	var tagNames []string
-
-	tags, err := repo.gogitRepo.Tags()
-	if err != nil {
-		return nil, err
-	}
-
-	_ = tags.ForEach(func(tag *plumbing.Reference) error {
-		tagNames = append(tagNames, strings.TrimPrefix(tag.Name().String(), TagPrefix))
-		return nil
-	})
-
-	// Reverse order
-	for i := 0; i < len(tagNames)/2; i++ {
-		j := len(tagNames) - i - 1
-		tagNames[i], tagNames[j] = tagNames[j], tagNames[i]
-	}
-
-	// since we have to reverse order we can paginate only afterwards
-	if len(tagNames) < skip {
-		tagNames = []string{}
-	} else {
-		tagNames = tagNames[skip:]
-	}
-	if limit != 0 && len(tagNames) > limit {
-		tagNames = tagNames[:limit]
-	}
-
-	return tagNames, nil
-}
-
-// GetTagType gets the type of the tag, either commit (simple) or tag (annotated)
-func (repo *Repository) GetTagType(id ObjectID) (string, error) {
-	// Get tag type
-	obj, err := repo.gogitRepo.Object(plumbing.AnyObject, plumbing.Hash(id.RawValue()))
-	if err != nil {
-		if err == plumbing.ErrReferenceNotFound {
-			return "", &ErrNotExist{ID: id.String()}
-		}
-		return "", err
-	}
-
-	return obj.Type().String(), nil
-}
-
-func (repo *Repository) getTag(tagID ObjectID, name string) (*Tag, error) {
-	t, ok := repo.tagCache.Get(tagID.String())
-	if ok {
-		log.Debug("Hit cache: %s", tagID)
-		tagClone := *t.(*Tag)
-		tagClone.Name = name // This is necessary because lightweight tags may have same id
-		return &tagClone, nil
-	}
-
-	tp, err := repo.GetTagType(tagID)
-	if err != nil {
-		return nil, err
-	}
-
-	// Get the commit ID and tag ID (may be different for annotated tag) for the returned tag object
-	commitIDStr, err := repo.GetTagCommitID(name)
-	if err != nil {
-		// every tag should have a commit ID so return all errors
-		return nil, err
-	}
-	commitID, err := NewIDFromString(commitIDStr)
-	if err != nil {
-		return nil, err
-	}
-
-	// If type is "commit, the tag is a lightweight tag
-	if ObjectType(tp) == ObjectCommit {
-		commit, err := repo.GetCommit(commitIDStr)
-		if err != nil {
-			return nil, err
-		}
-		tag := &Tag{
-			Name:    name,
-			ID:      tagID,
-			Object:  commitID,
-			Type:    tp,
-			Tagger:  commit.Committer,
-			Message: commit.Message(),
-		}
-
-		repo.tagCache.Set(tagID.String(), tag)
-		return tag, nil
-	}
-
-	gogitTag, err := repo.gogitRepo.TagObject(plumbing.Hash(tagID.RawValue()))
-	if err != nil {
-		if err == plumbing.ErrReferenceNotFound {
-			return nil, &ErrNotExist{ID: tagID.String()}
-		}
-
-		return nil, err
-	}
-
-	tag := &Tag{
-		Name:    name,
-		ID:      tagID,
-		Object:  commitID.Type().MustID(gogitTag.Target[:]),
-		Type:    tp,
-		Tagger:  &gogitTag.Tagger,
-		Message: gogitTag.Message,
-	}
-
-	repo.tagCache.Set(tagID.String(), tag)
-	return tag, nil
-}

modules/git/repo_tag_nogogit.go

@@ -1,134 +0,0 @@
-// Copyright 2015 The Gogs Authors. All rights reserved.
-// Copyright 2019 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build !gogit
-
-package git
-
-import (
-	"errors"
-	"io"
-
-	"code.gitea.io/gitea/modules/log"
-)
-
-// IsTagExist returns true if given tag exists in the repository.
-func (repo *Repository) IsTagExist(name string) bool {
-	if repo == nil || name == "" {
-		return false
-	}
-
-	return repo.IsReferenceExist(TagPrefix + name)
-}
-
-// GetTags returns all tags of the repository.
-// returning at most limit tags, or all if limit is 0.
-func (repo *Repository) GetTags(skip, limit int) (tags []string, err error) {
-	tags, _, err = callShowRef(repo.Ctx, repo.Path, TagPrefix, TrustedCmdArgs{TagPrefix, "--sort=-taggerdate"}, skip, limit)
-	return tags, err
-}
-
-// 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)
-	defer cancel()
-	_, err := wr.Write([]byte(id.String() + "\n"))
-	if err != nil {
-		return "", err
-	}
-	_, typ, _, err := ReadBatchLine(rd)
-	if IsErrNotExist(err) {
-		return "", ErrNotExist{ID: id.String()}
-	}
-	return typ, nil
-}
-
-func (repo *Repository) getTag(tagID ObjectID, name string) (*Tag, error) {
-	t, ok := repo.tagCache.Get(tagID.String())
-	if ok {
-		log.Debug("Hit cache: %s", tagID)
-		tagClone := *t.(*Tag)
-		tagClone.Name = name // This is necessary because lightweight tags may have same id
-		return &tagClone, nil
-	}
-
-	tp, err := repo.GetTagType(tagID)
-	if err != nil {
-		return nil, err
-	}
-
-	// Get the commit ID and tag ID (may be different for annotated tag) for the returned tag object
-	commitIDStr, err := repo.GetTagCommitID(name)
-	if err != nil {
-		// every tag should have a commit ID so return all errors
-		return nil, err
-	}
-	commitID, err := NewIDFromString(commitIDStr)
-	if err != nil {
-		return nil, err
-	}
-
-	// If type is "commit, the tag is a lightweight tag
-	if ObjectType(tp) == ObjectCommit {
-		commit, err := repo.GetCommit(commitIDStr)
-		if err != nil {
-			return nil, err
-		}
-		tag := &Tag{
-			Name:    name,
-			ID:      tagID,
-			Object:  commitID,
-			Type:    tp,
-			Tagger:  commit.Committer,
-			Message: commit.Message(),
-		}
-
-		repo.tagCache.Set(tagID.String(), tag)
-		return tag, nil
-	}
-
-	// The tag is an annotated tag with a message.
-	wr, rd, cancel := repo.CatFileBatch(repo.Ctx)
-	defer cancel()
-
-	if _, err := wr.Write([]byte(tagID.String() + "\n")); err != nil {
-		return nil, err
-	}
-	_, typ, size, err := ReadBatchLine(rd)
-	if err != nil {
-		if errors.Is(err, io.EOF) || IsErrNotExist(err) {
-			return nil, ErrNotExist{ID: tagID.String()}
-		}
-		return nil, err
-	}
-	if typ != "tag" {
-		if err := DiscardFull(rd, size+1); err != nil {
-			return nil, err
-		}
-		return nil, ErrNotExist{ID: tagID.String()}
-	}
-
-	// then we need to parse the tag
-	// and load the commit
-	data, err := io.ReadAll(io.LimitReader(rd, size))
-	if err != nil {
-		return nil, err
-	}
-	_, err = rd.Discard(1)
-	if err != nil {
-		return nil, err
-	}
-
-	tag, err := parseTagData(tagID.Type(), data)
-	if err != nil {
-		return nil, err
-	}
-
-	tag.Name = name
-	tag.ID = tagID
-	tag.Type = tp
-
-	repo.tagCache.Set(tagID.String(), tag)
-	return tag, nil
-}

modules/git/repo_tree.go

@@ -6,6 +6,7 @@ package git
 
 import (
 	"bytes"
+	"io"
 	"os"
 	"strings"
 	"time"
@@ -65,3 +66,88 @@ func (repo *Repository) CommitTree(author, committer *Signature, tree *Tree, opt
 	}
 	return NewIDFromString(strings.TrimSpace(stdout.String()))
 }
+
+func (repo *Repository) getTree(id ObjectID) (*Tree, error) {
+	wr, rd, cancel := repo.CatFileBatch(repo.Ctx)
+	defer cancel()
+
+	_, _ = wr.Write([]byte(id.String() + "\n"))
+
+	// ignore the SHA
+	_, typ, size, err := ReadBatchLine(rd)
+	if err != nil {
+		return nil, err
+	}
+
+	switch typ {
+	case "tag":
+		resolvedID := id
+		data, err := io.ReadAll(io.LimitReader(rd, size))
+		if err != nil {
+			return nil, err
+		}
+		tag, err := parseTagData(id.Type(), data)
+		if err != nil {
+			return nil, err
+		}
+		commit, err := tag.Commit(repo)
+		if err != nil {
+			return nil, err
+		}
+		commit.Tree.ResolvedID = resolvedID
+		return &commit.Tree, nil
+	case "commit":
+		commit, err := CommitFromReader(repo, id, io.LimitReader(rd, size))
+		if err != nil {
+			return nil, err
+		}
+		if _, err := rd.Discard(1); err != nil {
+			return nil, err
+		}
+		commit.Tree.ResolvedID = commit.ID
+		return &commit.Tree, nil
+	case "tree":
+		tree := NewTree(repo, id)
+		tree.ResolvedID = id
+		objectFormat, err := repo.GetObjectFormat()
+		if err != nil {
+			return nil, err
+		}
+		tree.entries, err = catBatchParseTreeEntries(objectFormat, tree, rd, size)
+		if err != nil {
+			return nil, err
+		}
+		tree.entriesParsed = true
+		return tree, nil
+	default:
+		if err := DiscardFull(rd, size+1); err != nil {
+			return nil, err
+		}
+		return nil, ErrNotExist{
+			ID: id.String(),
+		}
+	}
+}
+
+// GetTree find the tree object in the repository.
+func (repo *Repository) GetTree(idStr string) (*Tree, error) {
+	objectFormat, err := repo.GetObjectFormat()
+	if err != nil {
+		return nil, err
+	}
+	if len(idStr) != objectFormat.FullLength() {
+		res, err := repo.GetRefCommitID(idStr)
+		if err != nil {
+			return nil, err
+		}
+		if len(res) > 0 {
+			idStr = res
+		}
+	}
+	id, err := NewIDFromString(idStr)
+	if err != nil {
+		return nil, err
+	}
+
+	return repo.getTree(id)
+}

modules/git/repo_tree_gogit.go

@@ -1,53 +0,0 @@
-// Copyright 2015 The Gogs Authors. All rights reserved.
-// Copyright 2019 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build gogit
-
-package git
-
-import "github.com/go-git/go-git/v5/plumbing"
-
-func (repo *Repository) getTree(id ObjectID) (*Tree, error) {
-	gogitTree, err := repo.gogitRepo.TreeObject(plumbing.Hash(id.RawValue()))
-	if err != nil {
-		return nil, err
-	}
-
-	tree := NewTree(repo, id)
-	tree.gogitTree = gogitTree
-	return tree, nil
-}
-
-// GetTree find the tree object in the repository.
-func (repo *Repository) GetTree(idStr string) (*Tree, error) {
-	objectFormat, err := repo.GetObjectFormat()
-	if err != nil {
-		return nil, err
-	}
-
-	if len(idStr) != objectFormat.FullLength() {
-		res, _, err := NewCommand(repo.Ctx, "rev-parse", "--verify").AddDynamicArguments(idStr).RunStdString(&RunOpts{Dir: repo.Path})
-		if err != nil {
-			return nil, err
-		}
-		if len(res) > 0 {
-			idStr = res[:len(res)-1]
-		}
-	}
-	id, err := NewIDFromString(idStr)
-	if err != nil {
-		return nil, err
-	}
-	resolvedID := id
-	commitObject, err := repo.gogitRepo.CommitObject(plumbing.Hash(id.RawValue()))
-	if err == nil {
-		id = ParseGogitHash(commitObject.TreeHash)
-	}
-	treeObject, err := repo.getTree(id)
-	if err != nil {
-		return nil, err
-	}
-	treeObject.ResolvedID = resolvedID
-	return treeObject, nil
-}

modules/git/repo_tree_nogogit.go

@@ -1,95 +0,0 @@
-// Copyright 2020 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build !gogit
-
-package git
-
-import (
-	"io"
-)
-
-func (repo *Repository) getTree(id ObjectID) (*Tree, error) {
-	wr, rd, cancel := repo.CatFileBatch(repo.Ctx)
-	defer cancel()
-
-	_, _ = wr.Write([]byte(id.String() + "\n"))
-
-	// ignore the SHA
-	_, typ, size, err := ReadBatchLine(rd)
-	if err != nil {
-		return nil, err
-	}
-
-	switch typ {
-	case "tag":
-		resolvedID := id
-		data, err := io.ReadAll(io.LimitReader(rd, size))
-		if err != nil {
-			return nil, err
-		}
-		tag, err := parseTagData(id.Type(), data)
-		if err != nil {
-			return nil, err
-		}
-		commit, err := tag.Commit(repo)
-		if err != nil {
-			return nil, err
-		}
-		commit.Tree.ResolvedID = resolvedID
-		return &commit.Tree, nil
-	case "commit":
-		commit, err := CommitFromReader(repo, id, io.LimitReader(rd, size))
-		if err != nil {
-			return nil, err
-		}
-		if _, err := rd.Discard(1); err != nil {
-			return nil, err
-		}
-		commit.Tree.ResolvedID = commit.ID
-		return &commit.Tree, nil
-	case "tree":
-		tree := NewTree(repo, id)
-		tree.ResolvedID = id
-		objectFormat, err := repo.GetObjectFormat()
-		if err != nil {
-			return nil, err
-		}
-		tree.entries, err = catBatchParseTreeEntries(objectFormat, tree, rd, size)
-		if err != nil {
-			return nil, err
-		}
-		tree.entriesParsed = true
-		return tree, nil
-	default:
-		if err := DiscardFull(rd, size+1); err != nil {
-			return nil, err
-		}
-		return nil, ErrNotExist{
-			ID: id.String(),
-		}
-	}
-}
-
-// GetTree find the tree object in the repository.
-func (repo *Repository) GetTree(idStr string) (*Tree, error) {
-	objectFormat, err := repo.GetObjectFormat()
-	if err != nil {
-		return nil, err
-	}
-	if len(idStr) != objectFormat.FullLength() {
-		res, err := repo.GetRefCommitID(idStr)
-		if err != nil {
-			return nil, err
-		}
-		if len(res) > 0 {
-			idStr = res
-		}
-	}
-	id, err := NewIDFromString(idStr)
-	if err != nil {
-		return nil, err
-	}
-
-	return repo.getTree(id)
-}

modules/git/signature.go

@@ -5,13 +5,31 @@
 package git
 
 import (
+	"fmt"
 	"strconv"
 	"strings"
 	"time"
 
 	"code.gitea.io/gitea/modules/log"
+	"code.gitea.io/gitea/modules/util"
 )
 
+// Signature represents the Author, Committer or Tagger information.
+type Signature struct {
+	Name  string    // the committer name, it can be anything
+	Email string    // the committer email, it can be anything
+	When  time.Time // the timestamp of the signature
+}
+
+func (s *Signature) String() string {
+	return fmt.Sprintf("%s <%s>", s.Name, s.Email)
+}
+
+// Decode decodes a byte array representing a signature to signature
+func (s *Signature) Decode(b []byte) {
+	*s = *parseSignatureFromCommitLine(util.UnsafeBytesToString(b))
+}
+
 // Helper to get a signature from the commit line, which looks like:
 //
 //	full name <user@example.com> 1378823654 +0200

modules/git/signature_gogit.go

@@ -1,14 +0,0 @@
-// Copyright 2015 The Gogs Authors. All rights reserved.
-// Copyright 2019 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build gogit
-
-package git
-
-import (
-	"github.com/go-git/go-git/v5/plumbing/object"
-)
-
-// Signature represents the Author or Committer information.
-type Signature = object.Signature

modules/git/signature_nogogit.go

@@ -1,30 +0,0 @@
-// Copyright 2015 The Gogs Authors. All rights reserved.
-// Copyright 2019 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build !gogit
-
-package git
-
-import (
-	"fmt"
-	"time"
-
-	"code.gitea.io/gitea/modules/util"
-)
-
-// Signature represents the Author, Committer or Tagger information.
-type Signature struct {
-	Name  string    // the committer name, it can be anything
-	Email string    // the committer email, it can be anything
-	When  time.Time // the timestamp of the signature
-}
-
-func (s *Signature) String() string {
-	return fmt.Sprintf("%s <%s>", s.Name, s.Email)
-}
-
-// Decode decodes a byte array representing a signature to signature
-func (s *Signature) Decode(b []byte) {
-	*s = *parseSignatureFromCommitLine(util.UnsafeBytesToString(b))
-}

modules/git/tree.go

@@ -6,9 +6,26 @@ package git
 
 import (
 	"bytes"
+	"io"
 	"strings"
 )
 
+// Tree represents a flat directory listing.
+type Tree struct {
+	ID         ObjectID
+	ResolvedID ObjectID
+	repo       *Repository
+
+	// parent tree
+	ptree *Tree
+
+	entries       Entries
+	entriesParsed bool
+
+	entriesRecursive       Entries
+	entriesRecursiveParsed bool
+}
+
 // NewTree create a new tree according the repository and tree id
 func NewTree(repo *Repository, id ObjectID) *Tree {
 	return &Tree{
@@ -17,6 +34,100 @@ func NewTree(repo *Repository, id ObjectID) *Tree {
 	}
 }
 
+// ListEntries returns all entries of current tree.
+func (t *Tree) ListEntries() (Entries, error) {
+	if t.entriesParsed {
+		return t.entries, nil
+	}
+
+	if t.repo != nil {
+		wr, rd, cancel := t.repo.CatFileBatch(t.repo.Ctx)
+		defer cancel()
+
+		_, _ = wr.Write([]byte(t.ID.String() + "\n"))
+		_, typ, sz, err := ReadBatchLine(rd)
+		if err != nil {
+			return nil, err
+		}
+		if typ == "commit" {
+			treeID, err := ReadTreeID(rd, sz)
+			if err != nil && err != io.EOF {
+				return nil, err
+			}
+			_, _ = wr.Write([]byte(treeID + "\n"))
+			_, typ, sz, err = ReadBatchLine(rd)
+			if err != nil {
+				return nil, err
+			}
+		}
+		if typ == "tree" {
+			t.entries, err = catBatchParseTreeEntries(t.ID.Type(), t, rd, sz)
+			if err != nil {
+				return nil, err
+			}
+			t.entriesParsed = true
+			return t.entries, nil
+		}
+
+		// Not a tree just use ls-tree instead
+		if err := DiscardFull(rd, sz+1); err != nil {
+			return nil, err
+		}
+	}
+
+	stdout, _, runErr := NewCommand(t.repo.Ctx, "ls-tree", "-l").AddDynamicArguments(t.ID.String()).RunStdBytes(&RunOpts{Dir: t.repo.Path})
+	if runErr != nil {
+		if strings.Contains(runErr.Error(), "fatal: Not a valid object name") || strings.Contains(runErr.Error(), "fatal: not a tree object") {
+			return nil, ErrNotExist{
+				ID: t.ID.String(),
+			}
+		}
+		return nil, runErr
+	}
+
+	var err error
+	t.entries, err = parseTreeEntries(stdout, t)
+	if err == nil {
+		t.entriesParsed = true
+	}
+
+	return t.entries, err
+}
+
+// listEntriesRecursive returns all entries of current tree recursively including all subtrees
+// extraArgs could be "-l" to get the size, which is slower
+func (t *Tree) listEntriesRecursive(extraArgs TrustedCmdArgs) (Entries, error) {
+	if t.entriesRecursiveParsed {
+		return t.entriesRecursive, nil
+	}
+
+	stdout, _, runErr := NewCommand(t.repo.Ctx, "ls-tree", "-t", "-r").
+		AddArguments(extraArgs...).
+		AddDynamicArguments(t.ID.String()).
+		RunStdBytes(&RunOpts{Dir: t.repo.Path})
+	if runErr != nil {
+		return nil, runErr
+	}
+
+	var err error
+	t.entriesRecursive, err = parseTreeEntries(stdout, t)
+	if err == nil {
+		t.entriesRecursiveParsed = true
+	}
+
+	return t.entriesRecursive, err
+}
+
+// ListEntriesRecursiveFast returns all entries of current tree recursively including all subtrees, no size
+func (t *Tree) ListEntriesRecursiveFast() (Entries, error) {
+	return t.listEntriesRecursive(nil)
+}
+
+// ListEntriesRecursiveWithSize returns all entries of current tree recursively including all subtrees, with size
+func (t *Tree) ListEntriesRecursiveWithSize() (Entries, error) {
+	return t.listEntriesRecursive(TrustedCmdArgs{"--long"})
+}
+
 // SubTree get a sub tree by the sub dir path
 func (t *Tree) SubTree(rpath string) (*Tree, error) {
 	if len(rpath) == 0 {

modules/git/tree_blob.go

@@ -5,7 +5,48 @@
 
 package git
 
-import "strings"
+import (
+	"path"
+	"strings"
+)
+
+// GetTreeEntryByPath get the tree entries according the sub dir
+func (t *Tree) GetTreeEntryByPath(relpath string) (*TreeEntry, error) {
+	if len(relpath) == 0 {
+		return &TreeEntry{
+			ptree:     t,
+			ID:        t.ID,
+			name:      "",
+			fullName:  "",
+			entryMode: EntryModeTree,
+		}, nil
+	}
+
+	// FIXME: This should probably use git cat-file --batch to be a bit more efficient
+	relpath = path.Clean(relpath)
+	parts := strings.Split(relpath, "/")
+	var err error
+	tree := t
+	for i, name := range parts {
+		if i == len(parts)-1 {
+			entries, err := tree.ListEntries()
+			if err != nil {
+				return nil, err
+			}
+			for _, v := range entries {
+				if v.Name() == name {
+					return v, nil
+				}
+			}
+		} else {
+			tree, err = tree.SubTree(name)
+			if err != nil {
+				return nil, err
+			}
+		}
+	}
+	return nil, ErrNotExist{"", relpath}
+}
 
 // GetBlobByPath get the blob object according the path
 func (t *Tree) GetBlobByPath(relpath string) (*Blob, error) {

modules/git/tree_blob_gogit.go

@@ -1,65 +0,0 @@
-// Copyright 2015 The Gogs Authors. All rights reserved.
-// Copyright 2019 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build gogit
-
-package git
-
-import (
-	"path"
-	"strings"
-
-	"github.com/go-git/go-git/v5/plumbing"
-	"github.com/go-git/go-git/v5/plumbing/filemode"
-	"github.com/go-git/go-git/v5/plumbing/object"
-)
-
-// GetTreeEntryByPath get the tree entries according the sub dir
-func (t *Tree) GetTreeEntryByPath(relpath string) (*TreeEntry, error) {
-	if len(relpath) == 0 {
-		return &TreeEntry{
-			ID: t.ID,
-			// Type: ObjectTree,
-			gogitTreeEntry: &object.TreeEntry{
-				Name: "",
-				Mode: filemode.Dir,
-				Hash: plumbing.Hash(t.ID.RawValue()),
-			},
-		}, nil
-	}
-
-	relpath = path.Clean(relpath)
-	parts := strings.Split(relpath, "/")
-	var err error
-	tree := t
-	for i, name := range parts {
-		if i == len(parts)-1 {
-			entries, err := tree.ListEntries()
-			if err != nil {
-				if err == plumbing.ErrObjectNotFound {
-					return nil, ErrNotExist{
-						RelPath: relpath,
-					}
-				}
-				return nil, err
-			}
-			for _, v := range entries {
-				if v.Name() == name {
-					return v, nil
-				}
-			}
-		} else {
-			tree, err = tree.SubTree(name)
-			if err != nil {
-				if err == plumbing.ErrObjectNotFound {
-					return nil, ErrNotExist{
-						RelPath: relpath,
-					}
-				}
-				return nil, err
-			}
-		}
-	}
-	return nil, ErrNotExist{"", relpath}
-}

modules/git/tree_blob_nogogit.go

@@ -1,49 +0,0 @@
-// Copyright 2020 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build !gogit
-
-package git
-
-import (
-	"path"
-	"strings"
-)
-
-// GetTreeEntryByPath get the tree entries according the sub dir
-func (t *Tree) GetTreeEntryByPath(relpath string) (*TreeEntry, error) {
-	if len(relpath) == 0 {
-		return &TreeEntry{
-			ptree:     t,
-			ID:        t.ID,
-			name:      "",
-			fullName:  "",
-			entryMode: EntryModeTree,
-		}, nil
-	}
-
-	// FIXME: This should probably use git cat-file --batch to be a bit more efficient
-	relpath = path.Clean(relpath)
-	parts := strings.Split(relpath, "/")
-	var err error
-	tree := t
-	for i, name := range parts {
-		if i == len(parts)-1 {
-			entries, err := tree.ListEntries()
-			if err != nil {
-				return nil, err
-			}
-			for _, v := range entries {
-				if v.Name() == name {
-					return v, nil
-				}
-			}
-		} else {
-			tree, err = tree.SubTree(name)
-			if err != nil {
-				return nil, err
-			}
-		}
-	}
-	return nil, ErrNotExist{"", relpath}
-}

modules/git/tree_entry.go

@@ -8,8 +8,98 @@ import (
 	"io"
 	"sort"
 	"strings"
+
+	"code.gitea.io/gitea/modules/log"
 )
 
+// TreeEntry the leaf in the git tree
+type TreeEntry struct {
+	ID ObjectID
+
+	ptree *Tree
+
+	entryMode EntryMode
+	name      string
+
+	size     int64
+	sized    bool
+	fullName string
+}
+
+// Name returns the name of the entry
+func (te *TreeEntry) Name() string {
+	if te.fullName != "" {
+		return te.fullName
+	}
+	return te.name
+}
+
+// Mode returns the mode of the entry
+func (te *TreeEntry) Mode() EntryMode {
+	return te.entryMode
+}
+
+// Size returns the size of the entry
+func (te *TreeEntry) Size() int64 {
+	if te.IsDir() {
+		return 0
+	} else if te.sized {
+		return te.size
+	}
+
+	wr, rd, cancel := te.ptree.repo.CatFileBatchCheck(te.ptree.repo.Ctx)
+	defer cancel()
+	_, 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
+	}
+	_, _, te.size, err = ReadBatchLine(rd)
+	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
+	}
+
+	te.sized = true
+	return te.size
+}
+
+// IsSubModule if the entry is a sub module
+func (te *TreeEntry) IsSubModule() bool {
+	return te.entryMode == EntryModeCommit
+}
+
+// IsDir if the entry is a sub dir
+func (te *TreeEntry) IsDir() bool {
+	return te.entryMode == EntryModeTree
+}
+
+// IsLink if the entry is a symlink
+func (te *TreeEntry) IsLink() bool {
+	return te.entryMode == EntryModeSymlink
+}
+
+// IsRegular if the entry is a regular file
+func (te *TreeEntry) IsRegular() bool {
+	return te.entryMode == EntryModeBlob
+}
+
+// IsExecutable if the entry is an executable file (not necessarily binary)
+func (te *TreeEntry) IsExecutable() bool {
+	return te.entryMode == EntryModeExec
+}
+
+// Blob returns the blob object the entry
+func (te *TreeEntry) Blob() *Blob {
+	return &Blob{
+		ID:      te.ID,
+		name:    te.Name(),
+		size:    te.size,
+		gotSize: te.sized,
+		repo:    te.ptree.repo,
+	}
+}
+
 // Type returns the type of the entry (commit, tree, blob)
 func (te *TreeEntry) Type() string {
 	switch te.Mode() {

modules/git/tree_entry_gogit.go

@@ -1,95 +0,0 @@
-// Copyright 2015 The Gogs Authors. All rights reserved.
-// Copyright 2019 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build gogit
-
-package git
-
-import (
-	"github.com/go-git/go-git/v5/plumbing"
-	"github.com/go-git/go-git/v5/plumbing/filemode"
-	"github.com/go-git/go-git/v5/plumbing/object"
-)
-
-// TreeEntry the leaf in the git tree
-type TreeEntry struct {
-	ID ObjectID
-
-	gogitTreeEntry *object.TreeEntry
-	ptree          *Tree
-
-	size     int64
-	sized    bool
-	fullName string
-}
-
-// Name returns the name of the entry
-func (te *TreeEntry) Name() string {
-	if te.fullName != "" {
-		return te.fullName
-	}
-	return te.gogitTreeEntry.Name
-}
-
-// Mode returns the mode of the entry
-func (te *TreeEntry) Mode() EntryMode {
-	return EntryMode(te.gogitTreeEntry.Mode)
-}
-
-// Size returns the size of the entry
-func (te *TreeEntry) Size() int64 {
-	if te.IsDir() {
-		return 0
-	} else if te.sized {
-		return te.size
-	}
-
-	file, err := te.ptree.gogitTree.TreeEntryFile(te.gogitTreeEntry)
-	if err != nil {
-		return 0
-	}
-
-	te.sized = true
-	te.size = file.Size
-	return te.size
-}
-
-// IsSubModule if the entry is a sub module
-func (te *TreeEntry) IsSubModule() bool {
-	return te.gogitTreeEntry.Mode == filemode.Submodule
-}
-
-// IsDir if the entry is a sub dir
-func (te *TreeEntry) IsDir() bool {
-	return te.gogitTreeEntry.Mode == filemode.Dir
-}
-
-// IsLink if the entry is a symlink
-func (te *TreeEntry) IsLink() bool {
-	return te.gogitTreeEntry.Mode == filemode.Symlink
-}
-
-// IsRegular if the entry is a regular file
-func (te *TreeEntry) IsRegular() bool {
-	return te.gogitTreeEntry.Mode == filemode.Regular
-}
-
-// IsExecutable if the entry is an executable file (not necessarily binary)
-func (te *TreeEntry) IsExecutable() bool {
-	return te.gogitTreeEntry.Mode == filemode.Executable
-}
-
-// Blob returns the blob object the entry
-func (te *TreeEntry) Blob() *Blob {
-	encodedObj, err := te.ptree.repo.gogitRepo.Storer.EncodedObject(plumbing.AnyObject, te.gogitTreeEntry.Hash)
-	if err != nil {
-		return nil
-	}
-
-	return &Blob{
-		ID:              ParseGogitHash(te.gogitTreeEntry.Hash),
-		gogitEncodedObj: encodedObj,
-		name:            te.Name(),
-	}
-}

modules/git/tree_entry_nogogit.go

@@ -1,96 +0,0 @@
-// Copyright 2020 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build !gogit
-
-package git
-
-import "code.gitea.io/gitea/modules/log"
-
-// TreeEntry the leaf in the git tree
-type TreeEntry struct {
-	ID ObjectID
-
-	ptree *Tree
-
-	entryMode EntryMode
-	name      string
-
-	size     int64
-	sized    bool
-	fullName string
-}
-
-// Name returns the name of the entry
-func (te *TreeEntry) Name() string {
-	if te.fullName != "" {
-		return te.fullName
-	}
-	return te.name
-}
-
-// Mode returns the mode of the entry
-func (te *TreeEntry) Mode() EntryMode {
-	return te.entryMode
-}
-
-// Size returns the size of the entry
-func (te *TreeEntry) Size() int64 {
-	if te.IsDir() {
-		return 0
-	} else if te.sized {
-		return te.size
-	}
-
-	wr, rd, cancel := te.ptree.repo.CatFileBatchCheck(te.ptree.repo.Ctx)
-	defer cancel()
-	_, 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
-	}
-	_, _, te.size, err = ReadBatchLine(rd)
-	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
-	}
-
-	te.sized = true
-	return te.size
-}
-
-// IsSubModule if the entry is a sub module
-func (te *TreeEntry) IsSubModule() bool {
-	return te.entryMode == EntryModeCommit
-}
-
-// IsDir if the entry is a sub dir
-func (te *TreeEntry) IsDir() bool {
-	return te.entryMode == EntryModeTree
-}
-
-// IsLink if the entry is a symlink
-func (te *TreeEntry) IsLink() bool {
-	return te.entryMode == EntryModeSymlink
-}
-
-// IsRegular if the entry is a regular file
-func (te *TreeEntry) IsRegular() bool {
-	return te.entryMode == EntryModeBlob
-}
-
-// IsExecutable if the entry is an executable file (not necessarily binary)
-func (te *TreeEntry) IsExecutable() bool {
-	return te.entryMode == EntryModeExec
-}
-
-// Blob returns the blob object the entry
-func (te *TreeEntry) Blob() *Blob {
-	return &Blob{
-		ID:      te.ID,
-		name:    te.Name(),
-		size:    te.size,
-		gotSize: te.sized,
-		repo:    te.ptree.repo,
-	}
-}

modules/git/tree_entry_test.go

@@ -1,103 +0,0 @@
-// Copyright 2017 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build gogit
-
-package git
-
-import (
-	"testing"
-
-	"github.com/go-git/go-git/v5/plumbing/filemode"
-	"github.com/go-git/go-git/v5/plumbing/object"
-	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
-)
-
-func getTestEntries() Entries {
-	return Entries{
-		&TreeEntry{gogitTreeEntry: &object.TreeEntry{Name: "v1.0", Mode: filemode.Dir}},
-		&TreeEntry{gogitTreeEntry: &object.TreeEntry{Name: "v2.0", Mode: filemode.Dir}},
-		&TreeEntry{gogitTreeEntry: &object.TreeEntry{Name: "v2.1", Mode: filemode.Dir}},
-		&TreeEntry{gogitTreeEntry: &object.TreeEntry{Name: "v2.12", Mode: filemode.Dir}},
-		&TreeEntry{gogitTreeEntry: &object.TreeEntry{Name: "v2.2", Mode: filemode.Dir}},
-		&TreeEntry{gogitTreeEntry: &object.TreeEntry{Name: "v12.0", Mode: filemode.Dir}},
-		&TreeEntry{gogitTreeEntry: &object.TreeEntry{Name: "abc", Mode: filemode.Regular}},
-		&TreeEntry{gogitTreeEntry: &object.TreeEntry{Name: "bcd", Mode: filemode.Regular}},
-	}
-}
-
-func TestEntriesSort(t *testing.T) {
-	entries := getTestEntries()
-	entries.Sort()
-	assert.Equal(t, "v1.0", entries[0].Name())
-	assert.Equal(t, "v12.0", entries[1].Name())
-	assert.Equal(t, "v2.0", entries[2].Name())
-	assert.Equal(t, "v2.1", entries[3].Name())
-	assert.Equal(t, "v2.12", entries[4].Name())
-	assert.Equal(t, "v2.2", entries[5].Name())
-	assert.Equal(t, "abc", entries[6].Name())
-	assert.Equal(t, "bcd", entries[7].Name())
-}
-
-func TestEntriesCustomSort(t *testing.T) {
-	entries := getTestEntries()
-	entries.CustomSort(func(s1, s2 string) bool {
-		return s1 > s2
-	})
-	assert.Equal(t, "v2.2", entries[0].Name())
-	assert.Equal(t, "v2.12", entries[1].Name())
-	assert.Equal(t, "v2.1", entries[2].Name())
-	assert.Equal(t, "v2.0", entries[3].Name())
-	assert.Equal(t, "v12.0", entries[4].Name())
-	assert.Equal(t, "v1.0", entries[5].Name())
-	assert.Equal(t, "bcd", entries[6].Name())
-	assert.Equal(t, "abc", entries[7].Name())
-}
-
-func TestFollowLink(t *testing.T) {
-	r, err := openRepositoryWithDefaultContext("tests/repos/repo1_bare")
-	require.NoError(t, err)
-	defer r.Close()
-
-	commit, err := r.GetCommit("37991dec2c8e592043f47155ce4808d4580f9123")
-	require.NoError(t, err)
-
-	// get the symlink
-	lnk, err := commit.Tree.GetTreeEntryByPath("foo/bar/link_to_hello")
-	require.NoError(t, err)
-	assert.True(t, lnk.IsLink())
-
-	// should be able to dereference to target
-	target, err := lnk.FollowLink()
-	require.NoError(t, err)
-	assert.Equal(t, "hello", target.Name())
-	assert.False(t, target.IsLink())
-	assert.Equal(t, "b14df6442ea5a1b382985a6549b85d435376c351", target.ID.String())
-
-	// should error when called on normal file
-	target, err = commit.Tree.GetTreeEntryByPath("file1.txt")
-	require.NoError(t, err)
-	_, err = target.FollowLink()
-	assert.EqualError(t, err, "file1.txt: not a symlink")
-
-	// should error for broken links
-	target, err = commit.Tree.GetTreeEntryByPath("foo/broken_link")
-	require.NoError(t, err)
-	assert.True(t, target.IsLink())
-	_, err = target.FollowLink()
-	assert.EqualError(t, err, "broken_link: broken link")
-
-	// should error for external links
-	target, err = commit.Tree.GetTreeEntryByPath("foo/outside_repo")
-	require.NoError(t, err)
-	assert.True(t, target.IsLink())
-	_, err = target.FollowLink()
-	assert.EqualError(t, err, "outside_repo: points outside of repo")
-
-	// testing fix for short link bug
-	target, err = commit.Tree.GetTreeEntryByPath("foo/link_short")
-	require.NoError(t, err)
-	_, err = target.FollowLink()
-	assert.EqualError(t, err, "link_short: broken link")
-}

modules/git/tree_gogit.go

@@ -1,98 +0,0 @@
-// Copyright 2015 The Gogs Authors. All rights reserved.
-// Copyright 2019 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build gogit
-
-package git
-
-import (
-	"io"
-
-	"github.com/go-git/go-git/v5/plumbing"
-	"github.com/go-git/go-git/v5/plumbing/object"
-)
-
-// Tree represents a flat directory listing.
-type Tree struct {
-	ID         ObjectID
-	ResolvedID ObjectID
-	repo       *Repository
-
-	gogitTree *object.Tree
-
-	// parent tree
-	ptree *Tree
-}
-
-func (t *Tree) loadTreeObject() error {
-	gogitTree, err := t.repo.gogitRepo.TreeObject(plumbing.Hash(t.ID.RawValue()))
-	if err != nil {
-		return err
-	}
-
-	t.gogitTree = gogitTree
-	return nil
-}
-
-// ListEntries returns all entries of current tree.
-func (t *Tree) ListEntries() (Entries, error) {
-	if t.gogitTree == nil {
-		err := t.loadTreeObject()
-		if err != nil {
-			return nil, err
-		}
-	}
-
-	entries := make([]*TreeEntry, len(t.gogitTree.Entries))
-	for i, entry := range t.gogitTree.Entries {
-		entries[i] = &TreeEntry{
-			ID:             ParseGogitHash(entry.Hash),
-			gogitTreeEntry: &t.gogitTree.Entries[i],
-			ptree:          t,
-		}
-	}
-
-	return entries, nil
-}
-
-// ListEntriesRecursiveWithSize returns all entries of current tree recursively including all subtrees
-func (t *Tree) ListEntriesRecursiveWithSize() (Entries, error) {
-	if t.gogitTree == nil {
-		err := t.loadTreeObject()
-		if err != nil {
-			return nil, err
-		}
-	}
-
-	var entries []*TreeEntry
-	seen := map[plumbing.Hash]bool{}
-	walker := object.NewTreeWalker(t.gogitTree, true, seen)
-	for {
-		fullName, entry, err := walker.Next()
-		if err == io.EOF {
-			break
-		}
-		if err != nil {
-			return nil, err
-		}
-		if seen[entry.Hash] {
-			continue
-		}
-
-		convertedEntry := &TreeEntry{
-			ID:             ParseGogitHash(entry.Hash),
-			gogitTreeEntry: &entry,
-			ptree:          t,
-			fullName:       fullName,
-		}
-		entries = append(entries, convertedEntry)
-	}
-
-	return entries, nil
-}
-
-// ListEntriesRecursiveFast is the alias of ListEntriesRecursiveWithSize for the gogit version
-func (t *Tree) ListEntriesRecursiveFast() (Entries, error) {
-	return t.ListEntriesRecursiveWithSize()
-}

modules/git/tree_nogogit.go

@@ -1,121 +0,0 @@
-// Copyright 2020 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build !gogit
-
-package git
-
-import (
-	"io"
-	"strings"
-)
-
-// Tree represents a flat directory listing.
-type Tree struct {
-	ID         ObjectID
-	ResolvedID ObjectID
-	repo       *Repository
-
-	// parent tree
-	ptree *Tree
-
-	entries       Entries
-	entriesParsed bool
-
-	entriesRecursive       Entries
-	entriesRecursiveParsed bool
-}
-
-// ListEntries returns all entries of current tree.
-func (t *Tree) ListEntries() (Entries, error) {
-	if t.entriesParsed {
-		return t.entries, nil
-	}
-
-	if t.repo != nil {
-		wr, rd, cancel := t.repo.CatFileBatch(t.repo.Ctx)
-		defer cancel()
-
-		_, _ = wr.Write([]byte(t.ID.String() + "\n"))
-		_, typ, sz, err := ReadBatchLine(rd)
-		if err != nil {
-			return nil, err
-		}
-		if typ == "commit" {
-			treeID, err := ReadTreeID(rd, sz)
-			if err != nil && err != io.EOF {
-				return nil, err
-			}
-			_, _ = wr.Write([]byte(treeID + "\n"))
-			_, typ, sz, err = ReadBatchLine(rd)
-			if err != nil {
-				return nil, err
-			}
-		}
-		if typ == "tree" {
-			t.entries, err = catBatchParseTreeEntries(t.ID.Type(), t, rd, sz)
-			if err != nil {
-				return nil, err
-			}
-			t.entriesParsed = true
-			return t.entries, nil
-		}
-
-		// Not a tree just use ls-tree instead
-		if err := DiscardFull(rd, sz+1); err != nil {
-			return nil, err
-		}
-	}
-
-	stdout, _, runErr := NewCommand(t.repo.Ctx, "ls-tree", "-l").AddDynamicArguments(t.ID.String()).RunStdBytes(&RunOpts{Dir: t.repo.Path})
-	if runErr != nil {
-		if strings.Contains(runErr.Error(), "fatal: Not a valid object name") || strings.Contains(runErr.Error(), "fatal: not a tree object") {
-			return nil, ErrNotExist{
-				ID: t.ID.String(),
-			}
-		}
-		return nil, runErr
-	}
-
-	var err error
-	t.entries, err = parseTreeEntries(stdout, t)
-	if err == nil {
-		t.entriesParsed = true
-	}
-
-	return t.entries, err
-}
-
-// listEntriesRecursive returns all entries of current tree recursively including all subtrees
-// extraArgs could be "-l" to get the size, which is slower
-func (t *Tree) listEntriesRecursive(extraArgs TrustedCmdArgs) (Entries, error) {
-	if t.entriesRecursiveParsed {
-		return t.entriesRecursive, nil
-	}
-
-	stdout, _, runErr := NewCommand(t.repo.Ctx, "ls-tree", "-t", "-r").
-		AddArguments(extraArgs...).
-		AddDynamicArguments(t.ID.String()).
-		RunStdBytes(&RunOpts{Dir: t.repo.Path})
-	if runErr != nil {
-		return nil, runErr
-	}
-
-	var err error
-	t.entriesRecursive, err = parseTreeEntries(stdout, t)
-	if err == nil {
-		t.entriesRecursiveParsed = true
-	}
-
-	return t.entriesRecursive, err
-}
-
-// ListEntriesRecursiveFast returns all entries of current tree recursively including all subtrees, no size
-func (t *Tree) ListEntriesRecursiveFast() (Entries, error) {
-	return t.listEntriesRecursive(nil)
-}
-
-// ListEntriesRecursiveWithSize returns all entries of current tree recursively including all subtrees, with size
-func (t *Tree) ListEntriesRecursiveWithSize() (Entries, error) {
-	return t.listEntriesRecursive(TrustedCmdArgs{"--long"})
-}

modules/git/utils_test.go

@@ -13,7 +13,7 @@ import (
 // but not in production code.
 
 func skipIfSHA256NotSupported(t *testing.T) {
-	if isGogit || CheckGitVersionAtLeast("2.42") != nil {
+	if CheckGitVersionAtLeast("2.42") != nil {
 		t.Skip("skipping because installed Git version doesn't support SHA256")
 	}
 }

modules/gitrepo/walk.go

@@ -1,8 +1,6 @@
 // Copyright 2024 The Gitea Authors. All rights reserved.
 // SPDX-License-Identifier: MIT
 
-//go:build !gogit
-
 package gitrepo
 
 import (

modules/gitrepo/walk_gogit.go

@@ -1,40 +0,0 @@
-// Copyright 2024 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build gogit
-
-package gitrepo
-
-import (
-	"context"
-
-	"github.com/go-git/go-git/v5/plumbing"
-)
-
-// WalkReferences walks all the references from the repository
-// refname is empty, ObjectTag or ObjectBranch. All other values should be treated as equivalent to empty.
-func WalkReferences(ctx context.Context, repo Repository, walkfn func(sha1, refname string) error) (int, error) {
-	gitRepo := repositoryFromContext(ctx, repo)
-	if gitRepo == nil {
-		var err error
-		gitRepo, err = OpenRepository(ctx, repo)
-		if err != nil {
-			return 0, err
-		}
-		defer gitRepo.Close()
-	}
-
-	i := 0
-	iter, err := gitRepo.GoGitRepo().References()
-	if err != nil {
-		return i, err
-	}
-	defer iter.Close()
-
-	err = iter.ForEach(func(ref *plumbing.Reference) error {
-		err := walkfn(ref.Hash().String(), string(ref.Name()))
-		i++
-		return err
-	})
-	return i, err
-}

modules/lfs/pointer_scanner.go

@@ -1,8 +1,6 @@
 // Copyright 2021 The Gitea Authors. All rights reserved.
 // SPDX-License-Identifier: MIT
 
-//go:build !gogit
-
 package lfs
 
 import (

modules/lfs/pointer_scanner_gogit.go

@@ -1,62 +0,0 @@
-// Copyright 2021 The Gitea Authors. All rights reserved.
-// SPDX-License-Identifier: MIT
-
-//go:build gogit
-
-package lfs
-
-import (
-	"context"
-	"fmt"
-
-	"code.gitea.io/gitea/modules/git"
-
-	"github.com/go-git/go-git/v5/plumbing/object"
-)
-
-// SearchPointerBlobs scans the whole repository for LFS pointer files
-func SearchPointerBlobs(ctx context.Context, repo *git.Repository, pointerChan chan<- PointerBlob, errChan chan<- error) {
-	gitRepo := repo.GoGitRepo()
-
-	err := func() error {
-		blobs, err := gitRepo.BlobObjects()
-		if err != nil {
-			return fmt.Errorf("lfs.SearchPointerBlobs BlobObjects: %w", err)
-		}
-
-		return blobs.ForEach(func(blob *object.Blob) error {
-			select {
-			case <-ctx.Done():
-				return ctx.Err()
-			default:
-			}
-
-			if blob.Size > blobSizeCutoff {
-				return nil
-			}
-
-			reader, err := blob.Reader()
-			if err != nil {
-				return fmt.Errorf("lfs.SearchPointerBlobs blob.Reader: %w", err)
-			}
-			defer reader.Close()
-
-			pointer, _ := ReadPointer(reader)
-			if pointer.IsValid() {
-				pointerChan <- PointerBlob{Hash: blob.Hash.String(), Pointer: pointer}
-			}
-
-			return nil
-		})
-	}()
-	if err != nil {
-		select {
-		case <-ctx.Done():
-		default:
-			errChan <- err
-		}
-	}
-
-	close(pointerChan)
-	close(errChan)
-}

release-notes/4941.md

@@ -0,0 +1 @@
+Drop support to build Forgejo with the optional go-git Git backend. It only affects users who built Forgejo manually using `TAGS=gogits`, which no longer has any effect. Moving forward, we only support the default backend using the git binary. Please get in touch if you used the go-git backend and require any assistance moving away from it.