Introduced sha256 support for git-sizer

Author: fcharlie

git/git.go

@@ -24,6 +24,8 @@ type Repository struct {
 	// gitBin is the path of the `git` executable that should be used
 	// when running commands in this repository.
 	gitBin string
+	// hashAgo is repository hash algo
+	hashAlgo HashAlgo
 }
 
 // smartJoin returns `relPath` if it is an absolute path. If not, it
@@ -49,9 +51,18 @@ func NewRepositoryFromGitDir(gitDir string) (*Repository, error) {
 		)
 	}
 
+	hashAlgo := HashSHA1
+	cmd := exec.Command(gitBin, "--git-dir", gitDir, "rev-parse", "--show-object-format")
+	if out, err := cmd.Output(); err == nil {
+		if string(bytes.TrimSpace(out)) == "sha256" {
+			hashAlgo = HashSHA256
+		}
+	}
+
 	repo := Repository{
-		gitDir: gitDir,
-		gitBin: gitBin,
+		gitDir:   gitDir,
+		gitBin:   gitBin,
+		hashAlgo: hashAlgo,
 	}
 
 	full, err := repo.IsFull()
@@ -170,3 +181,15 @@ func (repo *Repository) GitPath(relPath string) (string, error) {
 	// current directory, we can use it as-is:
 	return string(bytes.TrimSpace(out)), nil
 }
+
+func (repo *Repository) HashAlgo() HashAlgo {
+	return repo.hashAlgo
+}
+
+func (repo *Repository) HashSize() int {
+	return repo.hashAlgo.HashSize()
+}
+
+func (repo *Repository) NullOID() OID {
+	return repo.hashAlgo.NullOID()
+}

git/obj_iter.go

@@ -30,7 +30,7 @@ func (repo *Repository) NewObjectIter(ctx context.Context) (*ObjectIter, error)
 		errCh:    make(chan error),
 		headerCh: make(chan BatchHeader),
 	}
-
+	hashHexSize := repo.HashSize() * 2
 	iter.p.Add(
 		// Read OIDs from `iter.oidCh` and write them to `git
 		// rev-list`:
@@ -68,10 +68,10 @@ func (repo *Repository) NewObjectIter(ctx context.Context) (*ObjectIter, error)
 		pipe.LinewiseFunction(
 			"copy-oids",
 			func(_ context.Context, _ pipe.Env, line []byte, stdout *bufio.Writer) error {
-				if len(line) < 40 {
+				if len(line) < hashHexSize {
 					return fmt.Errorf("line too short: '%s'", line)
 				}
-				if _, err := stdout.Write(line[:40]); err != nil {
+				if _, err := stdout.Write(line[:hashHexSize]); err != nil {
 					return fmt.Errorf("writing OID to 'git cat-file': %w", err)
 				}
 				if err := stdout.WriteByte('\n'); err != nil {

git/obj_resolver.go

@@ -9,12 +9,12 @@ func (repo *Repository) ResolveObject(name string) (OID, error) {
 	cmd := repo.GitCommand("rev-parse", "--verify", "--end-of-options", name)
 	output, err := cmd.Output()
 	if err != nil {
-		return NullOID, fmt.Errorf("resolving object %q: %w", name, err)
+		return repo.NullOID(), fmt.Errorf("resolving object %q: %w", name, err)
 	}
 	oidString := string(bytes.TrimSpace(output))
 	oid, err := NewOID(oidString)
 	if err != nil {
-		return NullOID, fmt.Errorf("parsing output %q from 'rev-parse': %w", oidString, err)
+		return repo.NullOID(), fmt.Errorf("parsing output %q from 'rev-parse': %w", oidString, err)
 	}
 	return oid, nil
 }

git/oid.go

@@ -1,32 +1,75 @@
 package git
 
 import (
+	"bytes"
+	"crypto/sha1"
+	"crypto/sha256"
 	"encoding/hex"
 	"errors"
 )
 
+const (
+	HashSizeSHA256 = sha256.Size
+	HashSizeSHA1   = sha1.Size
+	HashSizeMax    = HashSizeSHA256
+)
+
+type HashAlgo int
+
+const (
+	HashUnknown HashAlgo = iota
+	HashSHA1
+	HashSHA256
+)
+
 // OID represents the SHA-1 object ID of a Git object, in binary
 // format.
 type OID struct {
-	v [20]byte
+	v        [HashSizeMax]byte
+	hashSize int
 }
 
-// NullOID is the null object ID; i.e., all zeros.
-var NullOID OID
+func (h HashAlgo) NullOID() OID {
+	switch h {
+	case HashSHA1:
+		return OID{hashSize: HashSizeSHA1}
+	case HashSHA256:
+		return OID{hashSize: HashSizeSHA256}
+	}
+	return OID{}
+}
+
+func (h HashAlgo) HashSize() int {
+	switch h {
+	case HashSHA1:
+		return HashSizeSHA1
+	case HashSHA256:
+		return HashSizeSHA256
+	}
+	return 0
+}
+
+// defaultNullOID is the null object ID; i.e., all zeros.
+var defaultNullOID OID
+
+func IsNullOID(o OID) bool {
+	return bytes.Equal(o.v[:], defaultNullOID.v[:])
+}
 
 // OIDFromBytes converts a byte slice containing an object ID in
 // binary format into an `OID`.
 func OIDFromBytes(oidBytes []byte) (OID, error) {
 	var oid OID
-	if len(oidBytes) != len(oid.v) {
+	oidSize := len(oidBytes)
+	if oidSize != HashSizeSHA1 && oidSize != HashSizeSHA256 {
 		return OID{}, errors.New("bytes oid has the wrong length")
 	}
-	copy(oid.v[0:20], oidBytes)
+	oid.hashSize = oidSize
+	copy(oid.v[0:oidSize], oidBytes)
 	return oid, nil
 }
 
-// NewOID converts an object ID in hex format (i.e., `[0-9a-f]{40}`)
-// into an `OID`.
+// NewOID converts an object ID in hex format (i.e., `[0-9a-f]{40,64}`) into an `OID`.
 func NewOID(s string) (OID, error) {
 	oidBytes, err := hex.DecodeString(s)
 	if err != nil {
@@ -37,18 +80,18 @@ func NewOID(s string) (OID, error) {
 
 // String formats `oid` as a string in hex format.
 func (oid OID) String() string {
-	return hex.EncodeToString(oid.v[:])
+	return hex.EncodeToString(oid.v[:oid.hashSize])
 }
 
 // Bytes returns a byte slice view of `oid`, in binary format.
 func (oid OID) Bytes() []byte {
-	return oid.v[:]
+	return oid.v[:oid.hashSize]
 }
 
 // MarshalJSON expresses `oid` as a JSON string with its enclosing
 // quotation marks.
 func (oid OID) MarshalJSON() ([]byte, error) {
-	src := oid.v[:]
+	src := oid.v[:oid.hashSize]
 	dst := make([]byte, hex.EncodedLen(len(src))+2)
 	dst[0] = '"'
 	dst[len(dst)-1] = '"'

git/tree.go

@@ -10,13 +10,14 @@ import (
 
 // Tree represents a Git tree object.
 type Tree struct {
-	data string
+	data     string
+	hashSize int
 }
 
 // ParseTree parses the tree object whose contents are contained in
 // `data`. `oid` is currently unused.
 func ParseTree(oid OID, data []byte) (*Tree, error) {
-	return &Tree{string(data)}, nil
+	return &Tree{string(data), oid.hashSize}, nil
 }
 
 // Size returns the size of the tree object.
@@ -36,13 +37,15 @@ type TreeEntry struct {
 // TreeIter is an iterator over the entries in a Git tree object.
 type TreeIter struct {
 	// The as-yet-unread part of the tree's data.
-	data string
+	data     string
+	hashSize int
 }
 
 // Iter returns an iterator over the entries in `tree`.
 func (tree *Tree) Iter() *TreeIter {
 	return &TreeIter{
-		data: tree.data,
+		data:     tree.data,
+		hashSize: tree.hashSize,
 	}
 }
 
@@ -74,12 +77,12 @@ func (iter *TreeIter) NextEntry() (TreeEntry, bool, error) {
 	entry.Name = iter.data[:nulAt]
 
 	iter.data = iter.data[nulAt+1:]
-	if len(iter.data) < 20 {
+	if len(iter.data) < iter.hashSize {
 		return TreeEntry{}, false, errors.New("tree entry ends unexpectedly")
 	}
-
-	copy(entry.OID.v[0:20], iter.data[0:20])
-	iter.data = iter.data[20:]
+	entry.OID.hashSize = iter.hashSize
+	copy(entry.OID.v[0:iter.hashSize], iter.data[0:iter.hashSize])
+	iter.data = iter.data[iter.hashSize:]
 
 	return entry, true, nil
 }

git_sizer_test.go

@@ -849,3 +849,40 @@ func TestSubmodule(t *testing.T) {
 	assert.Equal(t, counts.Count32(2), h.UniqueBlobCount, "unique blob count")
 	assert.Equal(t, counts.Count32(3), h.MaxExpandedBlobCount, "max expanded blob count")
 }
+
+func TestSHA256(t *testing.T) {
+	t.Parallel()
+
+	ctx := context.Background()
+
+	t.Helper()
+
+	path, err := os.MkdirTemp("", "sha256")
+	require.NoError(t, err)
+
+	testRepo := testutils.TestRepo{Path: path}
+	defer testRepo.Remove(t)
+
+	// Don't use `GitCommand()` because the directory might not
+	// exist yet:
+	cmd := exec.Command("git", "init", "--object-format", "sha256", testRepo.Path)
+	cmd.Env = testutils.CleanGitEnv()
+	err = cmd.Run()
+	require.NoError(t, err)
+
+	timestamp := time.Unix(1112911993, 0)
+
+	testRepo.AddFile(t, "hello.txt", "Hello, world!\n")
+	cmd = testRepo.GitCommand(t, "commit", "-m", "initial")
+	testutils.AddAuthorInfo(cmd, &timestamp)
+	require.NoError(t, cmd.Run(), "creating initial commit")
+
+	cmd = testRepo.GitCommand(t, "commit", "-m", "initial", "--allow-empty")
+	testutils.AddAuthorInfo(cmd, &timestamp)
+	require.NoError(t, cmd.Run(), "creating commit")
+
+	repo := testRepo.Repository(t)
+
+	_, err = sizes.CollectReferences(ctx, repo, refGrouper{})
+	require.NoError(t, err)
+}

internal/testutils/repoutils.go

@@ -165,7 +165,7 @@ func (repo *TestRepo) UpdateRef(t *testing.T, refname string, oid git.OID) {
 
 	var cmd *exec.Cmd
 
-	if oid == git.NullOID {
+	if git.IsNullOID(oid) {
 		cmd = repo.GitCommand(t, "update-ref", "-d", refname)
 	} else {
 		cmd = repo.GitCommand(t, "update-ref", refname, oid.String())

sizes/graph.go

@@ -134,7 +134,7 @@ func ScanRepositoryUsingGraph(
 		case "tree":
 			trees = append(trees, ObjectHeader{obj.OID, obj.ObjectSize})
 		case "commit":
-			commits = append(commits, CommitHeader{ObjectHeader{obj.OID, obj.ObjectSize}, git.NullOID})
+			commits = append(commits, CommitHeader{ObjectHeader{obj.OID, obj.ObjectSize}, repo.NullOID()})
 		case "tag":
 			tags = append(tags, ObjectHeader{obj.OID, obj.ObjectSize})
 		default:

sizes/output.go

@@ -155,7 +155,7 @@ func (i *item) Emit(t *table) {
 }
 
 func (i *item) Footnote(nameStyle NameStyle) string {
-	if i.path == nil || i.path.OID == git.NullOID {
+	if i.path == nil || git.IsNullOID(i.path.OID) {
 		return ""
 	}
 	switch nameStyle {
@@ -214,7 +214,7 @@ func (i *item) MarshalJSON() ([]byte, error) {
 		LevelOfConcern: float64(value) / i.scale,
 	}
 
-	if i.path != nil && i.path.OID != git.NullOID {
+	if i.path != nil && !git.IsNullOID(i.path.OID) {
 		stat.ObjectName = i.path.OID.String()
 		stat.ObjectDescription = i.path.Path()
 	}