refactor(tvix/nar-bridge): drop reader package

Make the import function usable on any reader.

Change-Id: I84d2004cb73cdd7a11fe8efb0f2efb6335d5e6b0
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9527
Reviewed-by: Connor Brewster <cbrewster@hey.com>
Tested-by: BuildkiteCI
Autosubmit: flokli <flokli@flokli.de>

tvix/nar-bridge/pkg/importer/importer.go

@@ -0,0 +1,254 @@
+package importer
+
+import (
+	"context"
+	"crypto/sha256"
+	"errors"
+	"fmt"
+	"io"
+	"path"
+	"strings"
+
+	castorev1pb "code.tvl.fyi/tvix/castore/protos"
+	"code.tvl.fyi/tvix/nar-bridge/pkg/hashers"
+	storev1pb "code.tvl.fyi/tvix/store/protos"
+	"github.com/nix-community/go-nix/pkg/nar"
+	"lukechampine.com/blake3"
+)
+
+// An item on the directories stack
+type stackItem struct {
+	path      string
+	directory *castorev1pb.Directory
+}
+
+// Import reads NAR from a reader, and returns a (sparsely populated) PathInfo
+// object.
+func Import(
+	// a context, to support cancellation
+	ctx context.Context,
+	// The reader the data is read from
+	r io.Reader,
+	// callback function called with each regular file content
+	blobCb func(fileReader io.Reader) error,
+	// callback function called with each finalized directory node
+	directoryCb func(directory *castorev1pb.Directory) error,
+) (*storev1pb.PathInfo, error) {
+	// wrap the passed reader in a reader that records the number of bytes read and
+	// their sha256 sum.
+	hr := hashers.NewHasher(r, sha256.New())
+
+	// construct a NAR reader from the underlying data.
+	narReader, err := nar.NewReader(hr)
+	if err != nil {
+		return nil, fmt.Errorf("failed to instantiate nar reader: %w", err)
+	}
+	defer narReader.Close()
+
+	// If we store a symlink or regular file at the root, these are not nil.
+	// If they are nil, we instead have a stackDirectory.
+	var rootSymlink *castorev1pb.SymlinkNode
+	var rootFile *castorev1pb.FileNode
+	var stackDirectory *castorev1pb.Directory
+
+	var stack = []stackItem{}
+
+	// popFromStack is used when we transition to a different directory or
+	// drain the stack when we reach the end of the NAR.
+	// It adds the popped element to the element underneath if any,
+	// and passes it to the directoryCb callback.
+	// This function may only be called if the stack is not already empty.
+	popFromStack := func() error {
+		// Keep the top item, and "resize" the stack slice.
+		// This will only make the last element unaccessible, but chances are high
+		// we're re-using that space anyways.
+		toPop := stack[len(stack)-1]
+		stack = stack[:len(stack)-1]
+
+		// if there's still a parent left on the stack, refer to it from there.
+		if len(stack) > 0 {
+			dgst, err := toPop.directory.Digest()
+			if err != nil {
+				return fmt.Errorf("unable to calculate directory digest: %w", err)
+			}
+
+			topOfStack := stack[len(stack)-1].directory
+			topOfStack.Directories = append(topOfStack.Directories, &castorev1pb.DirectoryNode{
+				Name:   []byte(path.Base(toPop.path)),
+				Digest: dgst,
+				Size:   toPop.directory.Size(),
+			})
+		}
+		// call the directoryCb
+		if err := directoryCb(toPop.directory); err != nil {
+			return fmt.Errorf("failed calling directoryCb: %w", err)
+		}
+		// Keep track that we have encounter at least one directory
+		stackDirectory = toPop.directory
+		return nil
+	}
+
+	getBasename := func(p string) string {
+		// extract the basename. In case of "/", replace with empty string.
+		basename := path.Base(p)
+		if basename == "/" {
+			basename = ""
+		}
+		return basename
+	}
+
+	for {
+		select {
+		case <-ctx.Done():
+			return nil, ctx.Err()
+		default:
+			// call narReader.Next() to get the next element
+			hdr, err := narReader.Next()
+
+			// If this returns an error, it's either EOF (when we're done reading from the NAR),
+			// or another error
+			if err != nil {
+				// if this returns no EOF, bail out
+				if !errors.Is(err, io.EOF) {
+					return nil, fmt.Errorf("failed getting next nar element: %w", err)
+				}
+
+				// The NAR has been read all the way to the end…
+				// Make sure we close the nar reader, which might read some final trailers.
+				if err := narReader.Close(); err != nil {
+					return nil, fmt.Errorf("unable to close nar reader: %w", err)
+				}
+
+				// Check the stack. While it's not empty, we need to pop things off the stack.
+				for len(stack) > 0 {
+					err := popFromStack()
+					if err != nil {
+						return nil, fmt.Errorf("unable to pop from stack: %w", err)
+					}
+
+				}
+
+				// Stack is empty. We now either have a regular or symlink root node,
+				// or we encountered at least one directory assemble pathInfo with these and
+				// return.
+				pi := &storev1pb.PathInfo{
+					Node:       nil,
+					References: [][]byte{},
+					Narinfo: &storev1pb.NARInfo{
+						NarSize:        uint64(hr.BytesWritten()),
+						NarSha256:      hr.Sum(nil),
+						Signatures:     []*storev1pb.NARInfo_Signature{},
+						ReferenceNames: []string{},
+					},
+				}
+
+				if rootFile != nil {
+					pi.Node = &castorev1pb.Node{
+						Node: &castorev1pb.Node_File{
+							File: rootFile,
+						},
+					}
+				}
+				if rootSymlink != nil {
+					pi.Node = &castorev1pb.Node{
+						Node: &castorev1pb.Node_Symlink{
+							Symlink: rootSymlink,
+						},
+					}
+				}
+				if stackDirectory != nil {
+					// calculate directory digest (i.e. after we received all its contents)
+					dgst, err := stackDirectory.Digest()
+					if err != nil {
+						return nil, fmt.Errorf("unable to calculate root directory digest: %w", err)
+					}
+
+					pi.Node = &castorev1pb.Node{
+						Node: &castorev1pb.Node_Directory{
+							Directory: &castorev1pb.DirectoryNode{
+								Name:   []byte{},
+								Digest: dgst,
+								Size:   stackDirectory.Size(),
+							},
+						},
+					}
+				}
+				return pi, nil
+			}
+
+			// Check for valid path transitions, pop from stack if needed
+			// The nar reader already gives us some guarantees about ordering and illegal transitions,
+			// So we really only need to check if the top-of-stack path is a prefix of the path,
+			// and if it's not, pop from the stack. We do this repeatedly until the top of the stack is
+			// the subdirectory the new entry is in, or we hit the root directory.
+
+			// We don't need to worry about the root node case, because we can only finish the root "/"
+			// If we're at the end of the NAR reader (covered by the EOF check)
+			for len(stack) > 1 && !strings.HasPrefix(hdr.Path, stack[len(stack)-1].path+"/") {
+				err := popFromStack()
+				if err != nil {
+					return nil, fmt.Errorf("unable to pop from stack: %w", err)
+				}
+			}
+
+			if hdr.Type == nar.TypeSymlink {
+				symlinkNode := &castorev1pb.SymlinkNode{
+					Name:   []byte(getBasename(hdr.Path)),
+					Target: []byte(hdr.LinkTarget),
+				}
+				if len(stack) > 0 {
+					topOfStack := stack[len(stack)-1].directory
+					topOfStack.Symlinks = append(topOfStack.Symlinks, symlinkNode)
+				} else {
+					rootSymlink = symlinkNode
+				}
+
+			}
+			if hdr.Type == nar.TypeRegular {
+				// wrap reader with a reader calculating the blake3 hash
+				fileReader := hashers.NewHasher(narReader, blake3.New(32, nil))
+
+				err := blobCb(fileReader)
+				if err != nil {
+					return nil, fmt.Errorf("failure from blobCb: %w", err)
+				}
+
+				// drive the file reader to the end, in case the CB function doesn't read
+				// all the way to the end on its own
+				if fileReader.BytesWritten() != uint32(hdr.Size) {
+					_, err := io.ReadAll(fileReader)
+					if err != nil {
+						return nil, fmt.Errorf("unable to read until the end of the file content: %w", err)
+					}
+				}
+
+				// read the blake3 hash
+				dgst := fileReader.Sum(nil)
+
+				fileNode := &castorev1pb.FileNode{
+					Name:       []byte(getBasename(hdr.Path)),
+					Digest:     dgst,
+					Size:       uint32(hdr.Size),
+					Executable: hdr.Executable,
+				}
+				if len(stack) > 0 {
+					topOfStack := stack[len(stack)-1].directory
+					topOfStack.Files = append(topOfStack.Files, fileNode)
+				} else {
+					rootFile = fileNode
+				}
+			}
+			if hdr.Type == nar.TypeDirectory {
+				directory := &castorev1pb.Directory{
+					Directories: []*castorev1pb.DirectoryNode{},
+					Files:       []*castorev1pb.FileNode{},
+					Symlinks:    []*castorev1pb.SymlinkNode{},
+				}
+				stack = append(stack, stackItem{
+					directory: directory,
+					path:      hdr.Path,
+				})
+			}
+		}
+	}
+}