Merge branch 'main' into spanner-lib-bidi-grpc

Author: olavloite

.github/workflows/build-shared-spanner-lib.yml

@@ -25,10 +25,10 @@ jobs:
             target: windows
 
     steps:
-    - uses: actions/checkout@v4
+    - uses: actions/checkout@v6
 
     - name: Set up Go
-      uses: actions/setup-go@v5
+      uses: actions/setup-go@v6
       with:
         go-version: ${{ matrix.go-version }}
         cache: true

.github/workflows/python-spanner-lib-wrapper-lint.yml

@@ -22,10 +22,10 @@ jobs:
         working-directory: spannerlib/wrappers/spannerlib-python/spannerlib-python
 
     steps:
-    - uses: actions/checkout@v4
+    - uses: actions/checkout@v6
 
     - name: Set up Python
-      uses: actions/setup-python@v5
+      uses: actions/setup-python@v6
       with:
         python-version: '3.10'
         cache: 'pip'

.github/workflows/python-spanner-lib-wrapper-system-tests.yml

@@ -22,15 +22,15 @@ jobs:
         working-directory: spannerlib/wrappers/spannerlib-python/spannerlib-python
 
     steps:
-    - uses: actions/checkout@v4
+    - uses: actions/checkout@v6
 
     - name: Set up Go
-      uses: actions/setup-go@v5
+      uses: actions/setup-go@v6
       with:
         go-version: '1.25.x'
 
     - name: Set up Python
-      uses: actions/setup-python@v5
+      uses: actions/setup-python@v6
       with:
         python-version: ${{ matrix.python-version }}
         cache: 'pip'

.github/workflows/python-spanner-lib-wrapper-unit-tests.yml

@@ -18,15 +18,15 @@ jobs:
         working-directory: spannerlib/wrappers/spannerlib-python/spannerlib-python
 
     steps:
-    - uses: actions/checkout@v4
+    - uses: actions/checkout@v6
 
     - name: Set up Go
-      uses: actions/setup-go@v5
+      uses: actions/setup-go@v6
       with:
         go-version: '1.25.x'
 
     - name: Set up Python
-      uses: actions/setup-python@v5
+      uses: actions/setup-python@v6
       with:
         python-version: '3.10'
         cache: 'pip'

.github/workflows/release-python-spanner-lib-wrapper.yml

@@ -23,7 +23,7 @@ jobs:
       contents: read
 
     steps:
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@v6
 
       - name: Download Shared Library Artifacts
         uses: actions/download-artifact@v4 
@@ -42,7 +42,7 @@ jobs:
           ls -R spannerlib/wrappers/spannerlib-python/spannerlib-python/google/cloud/spannerlib/internal/lib
       
       - name: Set up Python
-        uses: actions/setup-python@v5
+        uses: actions/setup-python@v6
         with:
           python-version: '3.10'
           cache: 'pip'

.github/workflows/release-spannermockserver.yml

@@ -19,10 +19,10 @@ jobs:
 
     steps:
     - name: Checkout
-      uses: actions/checkout@v4
+      uses: actions/checkout@v6
 
     - name: Set up Python
-      uses: actions/setup-python@v5
+      uses: actions/setup-python@v6
       with:
         python-version: "3.10"
 

checksum_row_iterator.go

@@ -18,8 +18,10 @@ import (
 	"bytes"
 	"context"
 	"crypto/sha256"
-	"encoding/gob"
-	"reflect"
+	"encoding/binary"
+	"hash"
+	"math"
+	"sort"
 
 	"cloud.google.com/go/spanner"
 	sppb "cloud.google.com/go/spanner/apiv1/spannerpb"
@@ -32,16 +34,6 @@ import (
 
 var errNextAfterSTop = status.Errorf(codes.FailedPrecondition, "Next called after Stop")
 
-// init registers the protobuf types with gob so they can be encoded.
-func init() {
-	gob.Register(structpb.Value_BoolValue{})
-	gob.Register(structpb.Value_NumberValue{})
-	gob.Register(structpb.Value_StringValue{})
-	gob.Register(structpb.Value_NullValue{})
-	gob.Register(structpb.Value_ListValue{})
-	gob.Register(structpb.Value_StructValue{})
-}
-
 var _ rowIterator = &checksumRowIterator{}
 
 // checksumRowIterator implements rowIterator and keeps track of a running
@@ -66,12 +58,12 @@ type checksumRowIterator struct {
 	// the retry has finished.
 	stopped bool
 
-	// checksum contains the current checksum for the results that have been
+	// hash contains the current hash for the results that have been
 	// seen. It is calculated as a SHA256 checksum over all rows that so far
 	// have been returned.
-	checksum *[32]byte
-	buffer   *bytes.Buffer
-	enc      *gob.Encoder
+	hash       hash.Hash
+	int32Buf   [4]byte
+	float64Buf [8]byte
 
 	// errIndex and err indicate any error and the index in the result set
 	// where the error occurred.
@@ -110,7 +102,7 @@ func (it *checksumRowIterator) Next() (row *spanner.Row, err error) {
 			// checksum of the columns that are included in this result. This is
 			// also used to detect the possible difference between two empty
 			// result sets with a different set of columns.
-			it.checksum, err = createMetadataChecksum(it.enc, it.buffer, it.metadata)
+			it.hash, err = it.createMetadataChecksum(it.metadata)
 			if err != nil {
 				return err
 			}
@@ -119,45 +111,79 @@ func (it *checksumRowIterator) Next() (row *spanner.Row, err error) {
 			return it.err
 		}
 		// Update the current checksum.
-		it.checksum, err = updateChecksum(it.enc, it.buffer, it.checksum, row)
-		return err
+		return it.updateChecksum(it.hash, row)
 	})
 	return row, err
 }
 
 // updateChecksum calculates the following checksum based on a current checksum
 // and a new row.
-func updateChecksum(enc *gob.Encoder, buffer *bytes.Buffer, currentChecksum *[32]byte, row *spanner.Row) (*[32]byte, error) {
-	buffer.Reset()
-	buffer.Write(currentChecksum[:])
+func (it *checksumRowIterator) updateChecksum(hash hash.Hash, row *spanner.Row) error {
 	for i := 0; i < row.Size(); i++ {
 		var v spanner.GenericColumnValue
 		err := row.Column(i, &v)
 		if err != nil {
-			return nil, err
+			return err
 		}
-		err = enc.Encode(v)
-		if err != nil {
-			return nil, err
+		it.hashValue(v.Value, hash)
+	}
+	return nil
+}
+
+func (it *checksumRowIterator) hashValue(value *structpb.Value, digest hash.Hash) {
+	switch value.GetKind().(type) {
+	case *structpb.Value_StringValue:
+		digest.Write(intToByte(it.int32Buf, len(value.GetStringValue())))
+		digest.Write([]byte(value.GetStringValue()))
+	case *structpb.Value_NullValue:
+		digest.Write([]byte{0})
+	case *structpb.Value_NumberValue:
+		digest.Write(float64ToByte(it.float64Buf, value.GetNumberValue()))
+	case *structpb.Value_BoolValue:
+		if value.GetBoolValue() {
+			digest.Write([]byte{1})
+		} else {
+			digest.Write([]byte{0})
+		}
+	case *structpb.Value_StructValue:
+		fields := make([]string, 0, len(value.GetStructValue().Fields))
+		for field := range value.GetStructValue().Fields {
+			fields = append(fields, field)
+		}
+		sort.Strings(fields)
+		for _, field := range fields {
+			digest.Write(intToByte(it.int32Buf, len(field)))
+			digest.Write([]byte(field))
+			it.hashValue(value.GetStructValue().Fields[field], digest)
+		}
+	case *structpb.Value_ListValue:
+		for _, v := range value.GetListValue().GetValues() {
+			it.hashValue(v, digest)
 		}
 	}
-	res := sha256.Sum256(buffer.Bytes())
-	return &res, nil
+}
+
+func intToByte(buf [4]byte, v int) []byte {
+	binary.BigEndian.PutUint32(buf[:], uint32(v))
+	return buf[:]
+}
+
+func float64ToByte(buf [8]byte, f float64) []byte {
+	binary.BigEndian.PutUint64(buf[:], math.Float64bits(f))
+	return buf[:]
 }
 
 // createMetadataChecksum calculates the checksum of the metadata of a result.
 // Only the column names and types are included in the checksum. Any transaction
 // metadata is not included.
-func createMetadataChecksum(enc *gob.Encoder, buffer *bytes.Buffer, metadata *sppb.ResultSetMetadata) (*[32]byte, error) {
-	buffer.Reset()
+func (it *checksumRowIterator) createMetadataChecksum(metadata *sppb.ResultSetMetadata) (hash.Hash, error) {
+	digest := sha256.New()
 	for _, field := range metadata.RowType.Fields {
-		err := enc.Encode(field)
-		if err != nil {
-			return nil, err
-		}
+		digest.Write(intToByte(it.int32Buf, len(field.Name)))
+		digest.Write([]byte(field.Name))
+		digest.Write(intToByte(it.int32Buf, int(field.Type.Code.Number())))
 	}
-	res := sha256.Sum256(buffer.Bytes())
-	return &res, nil
+	return digest, nil
 }
 
 // retry implements retriableStatement.retry for queries. It will execute the
@@ -167,8 +193,6 @@ func createMetadataChecksum(enc *gob.Encoder, buffer *bytes.Buffer, metadata *sp
 // initial iterator was also returned by the new iterator, and that the errors
 // were returned by the same row index.
 func (it *checksumRowIterator) retry(ctx context.Context, tx *spanner.ReadWriteStmtBasedTransaction) error {
-	buffer := &bytes.Buffer{}
-	enc := gob.NewEncoder(buffer)
 	retryIt := tx.QueryWithOptions(ctx, it.stmt, it.options)
 	// If the original iterator had been stopped, we should also always stop the
 	// new iterator.
@@ -193,12 +217,13 @@ func (it *checksumRowIterator) retry(ctx context.Context, tx *spanner.ReadWriteS
 	// Iterate over the new result set as many times as we iterated over the initial
 	// result set. The checksums of the two should be equal. Also, the new result set
 	// should return any error on the same index as the original.
-	var newChecksum *[32]byte
+	// var newChecksum *[32]byte
 	var checksumErr error
+	newHash := sha256.New()
 	for n := int64(0); n < it.nc; n++ {
 		row, err := retryIt.Next()
 		if n == 0 && (err == nil || err == iterator.Done) {
-			newChecksum, checksumErr = createMetadataChecksum(enc, buffer, retryIt.Metadata)
+			newHash, checksumErr = it.createMetadataChecksum(retryIt.Metadata)
 			if checksumErr != nil {
 				return failRetry(checksumErr)
 			}
@@ -211,14 +236,14 @@ func (it *checksumRowIterator) retry(ctx context.Context, tx *spanner.ReadWriteS
 			}
 			if errorsEqualForRetry(err, it.err) && n == it.errIndex {
 				// Check that the checksums are also equal.
-				if !checksumsEqual(newChecksum, it.checksum) {
+				if !checksumsEqual(newHash, it.hash) {
 					return failRetry(ErrAbortedDueToConcurrentModification)
 				}
 				return replaceIt(nil)
 			}
 			return failRetry(ErrAbortedDueToConcurrentModification)
 		}
-		newChecksum, err = updateChecksum(enc, buffer, newChecksum, row)
+		err = it.updateChecksum(newHash, row)
 		if err != nil {
 			return failRetry(err)
 		}
@@ -230,16 +255,21 @@ func (it *checksumRowIterator) retry(ctx context.Context, tx *spanner.ReadWriteS
 	if it.err != nil {
 		return failRetry(ErrAbortedDueToConcurrentModification)
 	}
-	if !checksumsEqual(newChecksum, it.checksum) {
+	if !checksumsEqual(newHash, it.hash) {
 		return failRetry(ErrAbortedDueToConcurrentModification)
 	}
 	// Everything seems to be equal, replace the underlying iterator and return
 	// a nil error.
 	return replaceIt(nil)
 }
 
-func checksumsEqual(c1, c2 *[32]byte) bool {
-	return (reflect.ValueOf(c1).IsNil() && reflect.ValueOf(c2).IsNil()) || *c1 == *c2
+func checksumsEqual(h1, h2 hash.Hash) bool {
+	if h1 == nil || h2 == nil {
+		return h1 == h2
+	}
+	c1 := h1.Sum(nil)
+	c2 := h2.Sum(nil)
+	return bytes.Equal(c1, c2)
 }
 
 func (it *checksumRowIterator) Stop() {