refactor(tvix/nar-bridge): have Export return root node r/6780

… and nar size / sha256 digest.

Instead of producing sparse PathInfo messages when NARs are sent to
nar-bridge, the nar-bridge http server now keeps a lookup table
(narsha256) -> (rootNode, narSize)

This removes a whole bunch of noise, because we don't need to keep
sparse fields around.

A convenience function
`GenPathInfo(rootNode *castorev1pb.Node, narInfo *narinfo.NarInfo)` is
added, which is used to produce PathInfo messages, either when receiving
a NAR file over http and uploading it to a remote PathInfoService, or to
synthesize the PathInfoMessage to return to the client, if nar-bridge is
acting as a PathInfoService for a remove Nix HTTP Binary cache.

Change-Id: Ibba1ab6238a050816c4fab29cb21ae88877d8613
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9651
Tested-by: BuildkiteCI
Reviewed-by: Brian McGee <brian@bmcgee.ie>

4 files changed, 145 insertions, 132 deletions

diff --git a/tvix/nar-bridge/pkg/importer/gen_pathinfo.go b/tvix/nar-bridge/pkg/importer/gen_pathinfo.go
new file mode 100644
index 000000000000..bf8d08615e42
--- /dev/null
+++ b/tvix/nar-bridge/pkg/importer/gen_pathinfo.go
@@ -0,0 +1,62 @@
+package importer
+
+import (
+	castorev1pb "code.tvl.fyi/tvix/castore/protos"
+	storev1pb "code.tvl.fyi/tvix/store/protos"
+	"fmt"
+	"github.com/nix-community/go-nix/pkg/narinfo"
+	"github.com/nix-community/go-nix/pkg/storepath"
+)
+
+// GenPathInfo takes a rootNode and narInfo and assembles a PathInfo.
+// The rootNode is renamed to match the StorePath in the narInfo.
+func GenPathInfo(rootNode *castorev1pb.Node, narInfo *narinfo.NarInfo) (*storev1pb.PathInfo, error) {
+	// parse the storePath from the .narinfo
+	storePath, err := storepath.FromAbsolutePath(narInfo.StorePath)
+	if err != nil {
+		return nil, fmt.Errorf("unable to parse StorePath: %w", err)
+	}
+
+	// construct the references, by parsing ReferenceNames and extracting the digest
+	references := make([][]byte, len(narInfo.References))
+	for i, referenceStr := range narInfo.References {
+		// parse reference as store path
+		referenceStorePath, err := storepath.FromString(referenceStr)
+		if err != nil {
+			return nil, fmt.Errorf("unable to parse reference %s as storepath: %w", referenceStr, err)
+		}
+		references[i] = referenceStorePath.Digest
+	}
+
+	// construct the narInfo.Signatures[*] from pathInfo.Narinfo.Signatures[*]
+	narinfoSignatures := make([]*storev1pb.NARInfo_Signature, len(narInfo.Signatures))
+	for i, narinfoSig := range narInfo.Signatures {
+		narinfoSignatures[i] = &storev1pb.NARInfo_Signature{
+			Name: narinfoSig.Name,
+			Data: narinfoSig.Data,
+		}
+	}
+
+	// assemble the PathInfo.
+	pathInfo := &storev1pb.PathInfo{
+		// embed a new root node with the name set to the store path basename.
+		Node:       castorev1pb.RenamedNode(rootNode, storePath.String()),
+		References: references,
+		Narinfo: &storev1pb.NARInfo{
+			NarSize:        narInfo.NarSize,
+			NarSha256:      narInfo.FileHash.Digest(),
+			Signatures:     narinfoSignatures,
+			ReferenceNames: narInfo.References,
+		},
+	}
+
+	// run Validate on the PathInfo, more as an additional sanity check our code is sound,
+	// to make sure we populated everything properly, before returning it.
+	// Fail hard if we fail validation, this is a code error.
+	if _, err = pathInfo.Validate(); err != nil {
+		panic(fmt.Sprintf("PathInfo failed validation: %v", err))
+	}
+
+	return pathInfo, nil
+
+}
diff --git a/tvix/nar-bridge/pkg/importer/importer.go b/tvix/nar-bridge/pkg/importer/importer.go
index 9d6a7178a2c5..ce2798215a67 100644
--- a/tvix/nar-bridge/pkg/importer/importer.go
+++ b/tvix/nar-bridge/pkg/importer/importer.go
@@ -10,7 +10,6 @@ import (
 	"strings"
 
 	castorev1pb "code.tvl.fyi/tvix/castore/protos"
-	storev1pb "code.tvl.fyi/tvix/store/protos"
 	"github.com/nix-community/go-nix/pkg/nar"
 )
 
@@ -20,8 +19,8 @@ type stackItem struct {
 	directory *castorev1pb.Directory
 }
 
-// Import reads NAR from a reader, and returns a (sparsely populated) PathInfo
-// object.
+// Import reads a NAR from a reader, and returns a the root node,
+// NAR size and NAR sha256 digest.
 func Import(
 	// a context, to support cancellation
 	ctx context.Context,
@@ -31,7 +30,7 @@ func Import(
 	blobCb func(fileReader io.Reader) ([]byte, error),
 	// callback function called with each finalized directory node
 	directoryCb func(directory *castorev1pb.Directory) ([]byte, error),
-) (*storev1pb.PathInfo, error) {
+) (*castorev1pb.Node, uint64, []byte, error) {
 	// We need to wrap the underlying reader a bit.
 	// - we want to keep track of the number of bytes read in total
 	// - we calculate the sha256 digest over all data read
@@ -42,7 +41,7 @@ func Import(
 	multiW := io.MultiWriter(narCountW, sha256W)
 	narReader, err := nar.NewReader(io.TeeReader(r, multiW))
 	if err != nil {
-		return nil, fmt.Errorf("failed to instantiate nar reader: %w", err)
+		return nil, 0, nil, fmt.Errorf("failed to instantiate nar reader: %w", err)
 	}
 	defer narReader.Close()
 
@@ -98,7 +97,7 @@ func Import(
 	for {
 		select {
 		case <-ctx.Done():
-			return nil, ctx.Err()
+			return nil, 0, nil, ctx.Err()
 		default:
 			// call narReader.Next() to get the next element
 			hdr, err := narReader.Next()
@@ -108,60 +107,49 @@ func Import(
 			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)
+					return nil, 0, 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)
+					return nil, 0, 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)
+						return nil, 0, 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:        narCountW.BytesWritten(),
-						NarSha256:      sha256W.Sum(nil),
-						Signatures:     []*storev1pb.NARInfo_Signature{},
-						ReferenceNames: []string{},
-					},
-				}
+				// Stack is empty.
+				// Now either root{File,Symlink,Directory} is not nil,
+				// and we can return the root node.
+				narSize := narCountW.BytesWritten()
+				narSha256 := sha256W.Sum(nil)
 
 				if rootFile != nil {
-					pi.Node = &castorev1pb.Node{
+					return &castorev1pb.Node{
 						Node: &castorev1pb.Node_File{
 							File: rootFile,
 						},
-					}
-				}
-				if rootSymlink != nil {
-					pi.Node = &castorev1pb.Node{
+					}, narSize, narSha256, nil
+				} else if rootSymlink != nil {
+					return &castorev1pb.Node{
 						Node: &castorev1pb.Node_Symlink{
 							Symlink: rootSymlink,
 						},
-					}
-				}
-				if stackDirectory != nil {
+					}, narSize, narSha256, nil
+				} else 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)
+						return nil, 0, nil, fmt.Errorf("unable to calculate root directory digest: %w", err)
 					}
 
-					pi.Node = &castorev1pb.Node{
+					return &castorev1pb.Node{
 						Node: &castorev1pb.Node_Directory{
 							Directory: &castorev1pb.DirectoryNode{
 								Name:   []byte{},
@@ -169,9 +157,10 @@ func Import(
 								Size:   stackDirectory.Size(),
 							},
 						},
-					}
+					}, narSize, narSha256, nil
+				} else {
+					return nil, 0, nil, fmt.Errorf("no root set")
 				}
-				return pi, nil
 			}
 
 			// Check for valid path transitions, pop from stack if needed
@@ -185,7 +174,7 @@ func Import(
 			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)
+					return nil, 0, nil, fmt.Errorf("unable to pop from stack: %w", err)
 				}
 			}
 
@@ -209,7 +198,7 @@ func Import(
 
 				blobDigest, err := blobCb(blobReader)
 				if err != nil {
-					return nil, fmt.Errorf("failure from blobCb: %w", err)
+					return nil, 0, nil, fmt.Errorf("failure from blobCb: %w", err)
 				}
 
 				// ensure blobCb did read all the way to the end.
diff --git a/tvix/nar-bridge/pkg/importer/importer_test.go b/tvix/nar-bridge/pkg/importer/importer_test.go
index 6754fd005551..bf16c9926e3e 100644
--- a/tvix/nar-bridge/pkg/importer/importer_test.go
+++ b/tvix/nar-bridge/pkg/importer/importer_test.go
@@ -10,7 +10,6 @@ import (
 
 	castorev1pb "code.tvl.fyi/tvix/castore/protos"
 	"code.tvl.fyi/tvix/nar-bridge/pkg/importer"
-	storev1pb "code.tvl.fyi/tvix/store/protos"
 	"github.com/stretchr/testify/require"
 )
 
@@ -18,7 +17,7 @@ func TestSymlink(t *testing.T) {
 	f, err := os.Open("../../testdata/symlink.nar")
 	require.NoError(t, err)
 
-	actualPathInfo, err := importer.Import(
+	rootNode, narSize, narSha256, err := importer.Import(
 		context.Background(),
 		f,
 		func(blobReader io.Reader) ([]byte, error) {
@@ -28,35 +27,25 @@ func TestSymlink(t *testing.T) {
 		},
 	)
 	require.NoError(t, err)
-
-	expectedPathInfo := &storev1pb.PathInfo{
-		Node: &castorev1pb.Node{
-			Node: &castorev1pb.Node_Symlink{
-				Symlink: &castorev1pb.SymlinkNode{
-					Name:   []byte(""),
-					Target: []byte("/nix/store/somewhereelse"),
-				},
+	require.Equal(t, &castorev1pb.Node{
+		Node: &castorev1pb.Node_Symlink{
+			Symlink: &castorev1pb.SymlinkNode{
+				Name:   []byte(""),
+				Target: []byte("/nix/store/somewhereelse"),
 			},
 		},
-		References: [][]byte{},
-		Narinfo: &storev1pb.NARInfo{
-			NarSize: 136,
-			NarSha256: []byte{
-				0x09, 0x7d, 0x39, 0x7e, 0x9b, 0x58, 0x26, 0x38, 0x4e, 0xaa, 0x16, 0xc4, 0x57, 0x71, 0x5d, 0x1c, 0x1a, 0x51, 0x67, 0x03, 0x13, 0xea, 0xd0, 0xf5, 0x85, 0x66, 0xe0, 0xb2, 0x32, 0x53, 0x9c, 0xf1,
-			},
-			Signatures:     []*storev1pb.NARInfo_Signature{},
-			ReferenceNames: []string{},
-		},
-	}
-
-	requireProtoEq(t, expectedPathInfo, actualPathInfo)
+	}, rootNode)
+	require.Equal(t, []byte{
+		0x09, 0x7d, 0x39, 0x7e, 0x9b, 0x58, 0x26, 0x38, 0x4e, 0xaa, 0x16, 0xc4, 0x57, 0x71, 0x5d, 0x1c, 0x1a, 0x51, 0x67, 0x03, 0x13, 0xea, 0xd0, 0xf5, 0x85, 0x66, 0xe0, 0xb2, 0x32, 0x53, 0x9c, 0xf1,
+	}, narSha256)
+	require.Equal(t, uint64(136), narSize)
 }
 
 func TestRegular(t *testing.T) {
 	f, err := os.Open("../../testdata/onebyteregular.nar")
 	require.NoError(t, err)
 
-	actualPathInfo, err := importer.Import(
+	rootNode, narSize, narSha256, err := importer.Import(
 		context.Background(),
 		f,
 		func(blobReader io.Reader) ([]byte, error) {
@@ -68,7 +57,6 @@ func TestRegular(t *testing.T) {
 			panic("no directories expected!")
 		},
 	)
-	require.NoError(t, err)
 
 	// The blake3 digest of the 0x01 byte.
 	BLAKE3_DIGEST_0X01 := []byte{
@@ -77,29 +65,21 @@ func TestRegular(t *testing.T) {
 		0x65, 0x2b,
 	}
 
-	expectedPathInfo := &storev1pb.PathInfo{
-		Node: &castorev1pb.Node{
-			Node: &castorev1pb.Node_File{
-				File: &castorev1pb.FileNode{
-					Name:       []byte(""),
-					Digest:     BLAKE3_DIGEST_0X01,
-					Size:       1,
-					Executable: false,
-				},
+	require.NoError(t, err)
+	require.Equal(t, &castorev1pb.Node{
+		Node: &castorev1pb.Node_File{
+			File: &castorev1pb.FileNode{
+				Name:       []byte(""),
+				Digest:     BLAKE3_DIGEST_0X01,
+				Size:       1,
+				Executable: false,
 			},
 		},
-		References: [][]byte{},
-		Narinfo: &storev1pb.NARInfo{
-			NarSize: 120,
-			NarSha256: []byte{
-				0x73, 0x08, 0x50, 0xa8, 0x11, 0x25, 0x9d, 0xbf, 0x3a, 0x68, 0xdc, 0x2e, 0xe8, 0x7a, 0x79, 0xaa, 0x6c, 0xae, 0x9f, 0x71, 0x37, 0x5e, 0xdf, 0x39, 0x6f, 0x9d, 0x7a, 0x91, 0xfb, 0xe9, 0x13, 0x4d,
-			},
-			Signatures:     []*storev1pb.NARInfo_Signature{},
-			ReferenceNames: []string{},
-		},
-	}
-
-	requireProtoEq(t, expectedPathInfo, actualPathInfo)
+	}, rootNode)
+	require.Equal(t, []byte{
+		0x73, 0x08, 0x50, 0xa8, 0x11, 0x25, 0x9d, 0xbf, 0x3a, 0x68, 0xdc, 0x2e, 0xe8, 0x7a, 0x79, 0xaa, 0x6c, 0xae, 0x9f, 0x71, 0x37, 0x5e, 0xdf, 0x39, 0x6f, 0x9d, 0x7a, 0x91, 0xfb, 0xe9, 0x13, 0x4d,
+	}, narSha256)
+	require.Equal(t, uint64(120), narSize)
 }
 
 func TestEmptyDirectory(t *testing.T) {
@@ -111,7 +91,7 @@ func TestEmptyDirectory(t *testing.T) {
 		Files:       []*castorev1pb.FileNode{},
 		Symlinks:    []*castorev1pb.SymlinkNode{},
 	}
-	actualPathInfo, err := importer.Import(
+	rootNode, narSize, narSha256, err := importer.Import(
 		context.Background(),
 		f,
 		func(blobReader io.Reader) ([]byte, error) {
@@ -122,28 +102,19 @@ func TestEmptyDirectory(t *testing.T) {
 		},
 	)
 	require.NoError(t, err)
-
-	expectedPathInfo := &storev1pb.PathInfo{
-		Node: &castorev1pb.Node{
-			Node: &castorev1pb.Node_Directory{
-				Directory: &castorev1pb.DirectoryNode{
-					Name:   []byte(""),
-					Digest: mustDirectoryDigest(expectedDirectory),
-					Size:   expectedDirectory.Size(),
-				},
+	require.Equal(t, &castorev1pb.Node{
+		Node: &castorev1pb.Node_Directory{
+			Directory: &castorev1pb.DirectoryNode{
+				Name:   []byte(""),
+				Digest: mustDirectoryDigest(expectedDirectory),
+				Size:   expectedDirectory.Size(),
 			},
 		},
-		References: [][]byte{},
-		Narinfo: &storev1pb.NARInfo{
-			NarSize: 96,
-			NarSha256: []byte{
-				0xa5, 0x0a, 0x5a, 0xb6, 0xd9, 0x92, 0xf5, 0x59, 0x8e, 0xdd, 0x92, 0x10, 0x50, 0x59, 0xfa, 0xe9, 0xac, 0xfc, 0x19, 0x29, 0x81, 0xe0, 0x8b, 0xd8, 0x85, 0x34, 0xc2, 0x16, 0x7e, 0x92, 0x52, 0x6a,
-			},
-			Signatures:     []*storev1pb.NARInfo_Signature{},
-			ReferenceNames: []string{},
-		},
-	}
-	requireProtoEq(t, expectedPathInfo, actualPathInfo)
+	}, rootNode)
+	require.Equal(t, []byte{
+		0xa5, 0x0a, 0x5a, 0xb6, 0xd9, 0x92, 0xf5, 0x59, 0x8e, 0xdd, 0x92, 0x10, 0x50, 0x59, 0xfa, 0xe9, 0xac, 0xfc, 0x19, 0x29, 0x81, 0xe0, 0x8b, 0xd8, 0x85, 0x34, 0xc2, 0x16, 0x7e, 0x92, 0x52, 0x6a,
+	}, narSha256)
+	require.Equal(t, uint64(96), narSize)
 }
 
 func TestFull(t *testing.T) {
@@ -458,7 +429,7 @@ func TestFull(t *testing.T) {
 
 	numDirectoriesReceived := 0
 
-	actualPathInfo, err := importer.Import(
+	rootNode, narSize, narSha256, err := importer.Import(
 		context.Background(),
 		f,
 		func(blobReader io.Reader) ([]byte, error) {
@@ -480,28 +451,19 @@ func TestFull(t *testing.T) {
 		},
 	)
 	require.NoError(t, err)
-
-	expectedPathInfo := &storev1pb.PathInfo{
-		Node: &castorev1pb.Node{
-			Node: &castorev1pb.Node_Directory{
-				Directory: &castorev1pb.DirectoryNode{
-					Name:   []byte(""),
-					Digest: mustDirectoryDigest(expectedDirectories["/"]),
-					Size:   expectedDirectories["/"].Size(),
-				},
+	require.Equal(t, &castorev1pb.Node{
+		Node: &castorev1pb.Node_Directory{
+			Directory: &castorev1pb.DirectoryNode{
+				Name:   []byte(""),
+				Digest: mustDirectoryDigest(expectedDirectories["/"]),
+				Size:   expectedDirectories["/"].Size(),
 			},
 		},
-		References: [][]byte{},
-		Narinfo: &storev1pb.NARInfo{
-			NarSize: 464152,
-			NarSha256: []byte{
-				0xc6, 0xe1, 0x55, 0xb3, 0x45, 0x6e, 0x30, 0xb7, 0x61, 0x22, 0x63, 0xec, 0x09, 0x50, 0x70, 0x81, 0x1c, 0xaf, 0x8a, 0xbf, 0xd5, 0x9f, 0xaa, 0x72, 0xab, 0x82, 0xa5, 0x92, 0xef, 0xde, 0xb2, 0x53,
-			},
-			Signatures:     []*storev1pb.NARInfo_Signature{},
-			ReferenceNames: []string{},
-		},
-	}
-	requireProtoEq(t, expectedPathInfo, actualPathInfo)
+	}, rootNode)
+	require.Equal(t, []byte{
+		0xc6, 0xe1, 0x55, 0xb3, 0x45, 0x6e, 0x30, 0xb7, 0x61, 0x22, 0x63, 0xec, 0x09, 0x50, 0x70, 0x81, 0x1c, 0xaf, 0x8a, 0xbf, 0xd5, 0x9f, 0xaa, 0x72, 0xab, 0x82, 0xa5, 0x92, 0xef, 0xde, 0xb2, 0x53,
+	}, narSha256)
+	require.Equal(t, uint64(464152), narSize)
 }
 
 // TestCallbackErrors ensures that errors returned from the callback function
@@ -514,7 +476,7 @@ func TestCallbackErrors(t *testing.T) {
 
 		targetErr := errors.New("expected error")
 
-		_, err = importer.Import(
+		_, _, _, err = importer.Import(
 			context.Background(),
 			f,
 			func(blobReader io.Reader) ([]byte, error) {
@@ -532,7 +494,7 @@ func TestCallbackErrors(t *testing.T) {
 
 		targetErr := errors.New("expected error")
 
-		_, err = importer.Import(
+		_, _, _, err = importer.Import(
 			context.Background(),
 			f,
 			func(blobReader io.Reader) ([]byte, error) {
@@ -562,7 +524,7 @@ func TestPopDirectories(t *testing.T) {
 	require.NoError(t, err)
 	defer f.Close()
 
-	_, err = importer.Import(
+	_, _, _, err = importer.Import(
 		context.Background(),
 		f,
 		func(blobReader io.Reader) ([]byte, error) { return mustBlobDigest(blobReader), nil },
diff --git a/tvix/nar-bridge/pkg/importer/roundtrip_test.go b/tvix/nar-bridge/pkg/importer/roundtrip_test.go
index 1daba449bd33..3f297d32dd1b 100644
--- a/tvix/nar-bridge/pkg/importer/roundtrip_test.go
+++ b/tvix/nar-bridge/pkg/importer/roundtrip_test.go
@@ -30,7 +30,7 @@ func TestRoundtrip(t *testing.T) {
 	blobsMap := make(map[string][]byte, 0)
 	directoriesMap := make(map[string]*castorev1pb.Directory)
 
-	pathInfo, err := importer.Import(
+	rootNode, _, _, err := importer.Import(
 		context.Background(),
 		bytes.NewBuffer(narContents),
 		func(blobReader io.Reader) ([]byte, error) {
@@ -56,10 +56,10 @@ func TestRoundtrip(t *testing.T) {
 	require.NoError(t, err)
 
 	// done populating everything, now actually test the export :-)
-	var buf bytes.Buffer
+	var narBuf bytes.Buffer
 	err = storev1pb.Export(
-		&buf,
-		pathInfo.Node,
+		&narBuf,
+		rootNode,
 		func(directoryDgst []byte) (*castorev1pb.Directory, error) {
 			d, found := directoriesMap[base64.StdEncoding.EncodeToString(directoryDgst)]
 			if !found {
@@ -77,5 +77,5 @@ func TestRoundtrip(t *testing.T) {
 	)
 
 	require.NoError(t, err, "exporter shouldn't fail")
-	require.Equal(t, narContents, buf.Bytes())
+	require.Equal(t, narContents, narBuf.Bytes())
 }