Add NextReader to BlockReader

v2/block_reader.go

@@ -4,6 +4,7 @@ import (
 	"errors"
 	"fmt"
 	"io"
+	"math"
 
 	blocks "github.com/ipfs/go-block-format"
 	"github.com/ipfs/go-cid"
@@ -141,6 +142,23 @@ func (br *BlockReader) Next() (blocks.Block, error) {
 	return blocks.NewBlockWithCid(data, c)
 }
 
+// NextReader returns a CID, io.Reader and length of what should be the next block
+// the CID itself is not verified, and the reader is limited to the size of the block
+// The user of this function HAS TO consume all of the bytes in the returned reader before using any other function
+// on the BlockReader.
+// The returned length might be larger than MaxAllowedSectionSize, it is up to the user to check before loading the data into memory.
+func (br *BlockReader) NextReader() (cid.Cid, io.Reader, uint64, error) {
+	c, length, err := util.ReadNodeHeader(br.r, br.opts.ZeroLengthSectionAsEOF, math.MaxUint64)
+	if err != nil {
+		return cid.Undef, nil, 0, err
+	}
+	limitReader := io.LimitReader(br.r, int64(length))
+
+	ss := uint64(c.ByteLen()) + length
+	br.offset += uint64(varint.UvarintSize(ss)) + ss
+	return c, limitReader, length, nil
+}
+
 // BlockMetadata contains metadata about a block's section in a CAR file/stream.
 //
 // There are two offsets for the block section which will be the same if the
@@ -171,26 +189,14 @@ type BlockMetadata struct {
 // If the underlying reader used by the BlockReader is actually a ReadSeeker, this method will attempt to
 // seek over the underlying data rather than reading it into memory.
 func (br *BlockReader) SkipNext() (*BlockMetadata, error) {
-	sectionSize, err := util.LdReadSize(br.r, br.opts.ZeroLengthSectionAsEOF, br.opts.MaxAllowedSectionSize)
+	c, blockSize, err := util.ReadNodeHeader(br.r, br.opts.ZeroLengthSectionAsEOF, br.opts.MaxAllowedSectionSize)
 	if err != nil {
 		return nil, err
 	}
-	if sectionSize == 0 {
-		_, _, err := cid.CidFromBytes([]byte{}) // generate zero-byte CID error
-		if err == nil {
-			panic("expected zero-byte CID error")
-		}
-		return nil, err
-	}
-
+	cidSize := uint64(c.ByteLen())
+	sectionSize := blockSize + cidSize
 	lenSize := uint64(varint.UvarintSize(sectionSize))
 
-	cidSize, c, err := cid.CidFromReader(io.LimitReader(br.r, int64(sectionSize)))
-	if err != nil {
-		return nil, err
-	}
-
-	blockSize := sectionSize - uint64(cidSize)
 	blockOffset := br.offset
 
 	// move our reader forward; either by seeking or slurping

v2/block_reader_test.go

@@ -352,6 +352,23 @@ func TestBlockReader(t *testing.T) {
 			}
 			_, err = car.Next()
 			req.ErrorIs(err, io.EOF)
+
+			car, err = carv2.NewBlockReader(testCase.reader())
+			req.NoError(err)
+			req.ElementsMatch(roots, car.Roots)
+
+			for i := 0; i < 100; i++ {
+				cid, r, length, err := car.NextReader()
+				req.NoError(err)
+				req.Equal(blks[i].block.Cid(), cid)
+				req.Equal(uint64(len(blks[i].block.RawData())), length)
+				data := make([]byte, length)
+				_, err = io.ReadFull(r, data)
+				req.NoError(err)
+				req.Equal(blks[i].block.RawData(), data)
+			}
+			_, err = car.Next()
+			req.ErrorIs(err, io.EOF)
 		})
 	}
 }

v2/internal/carv1/util/util.go

@@ -3,6 +3,7 @@ package util
 import (
 	"errors"
 	"io"
+	"math"
 
 	internalio "github.com/ipld/go-car/v2/internal/io"
 
@@ -33,6 +34,33 @@ func ReadNode(r io.Reader, zeroLenAsEOF bool, maxReadBytes uint64) (cid.Cid, []b
 	return c, data[n:], nil
 }
 
+// ReadNodeHeader returns the specified CID of the node and the length of data to be read.
+func ReadNodeHeader(r io.Reader, zeroLenAsEOF bool, maxReadBytes uint64) (cid.Cid, uint64, error) {
+	size, err := LdReadSize(r, zeroLenAsEOF, maxReadBytes)
+	if err != nil {
+		return cid.Cid{}, 0, err
+	}
+
+	if size == 0 {
+		_, _, err := cid.CidFromBytes([]byte{}) // generate zero-byte CID error
+		if err == nil {
+			panic("expected zero-byte CID error")
+		}
+		return cid.Undef, 0, err
+	}
+
+	if size > math.MaxInt64 {
+		return cid.Cid{}, 0, ErrHeaderTooLarge
+	}
+	limitReader := io.LimitReader(r, int64(size))
+	n, c, err := cid.CidFromReader(limitReader)
+	if err != nil {
+		return cid.Cid{}, 0, err
+	}
+
+	return c, size - uint64(n), nil
+}
+
 func LdWrite(w io.Writer, d ...[]byte) error {
 	var sum uint64
 	for _, s := range d {