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>

5 files changed, 60 insertions, 117 deletions

diff --git a/tvix/nar-bridge/pkg/http/nar_get.go b/tvix/nar-bridge/pkg/http/nar_get.go
index eecc59900784..a226c9da01eb 100644
--- a/tvix/nar-bridge/pkg/http/nar_get.go
+++ b/tvix/nar-bridge/pkg/http/nar_get.go
@@ -29,16 +29,18 @@ func renderNar(
 	log *log.Entry,
 	directoryServiceClient castorev1pb.DirectoryServiceClient,
 	blobServiceClient castorev1pb.BlobServiceClient,
-	narHashToPathInfoMu *sync.Mutex,
-	narHashToPathInfo map[string]*storev1pb.PathInfo,
+	narHashDbMu *sync.Mutex,
+	narHashDb map[string]*narData,
 	w io.Writer,
 	narHash *nixhash.Hash,
 	headOnly bool,
 ) error {
 	// look in the lookup table
-	narHashToPathInfoMu.Lock()
-	pathInfo, found := narHashToPathInfo[narHash.SRIString()]
-	narHashToPathInfoMu.Unlock()
+	narHashDbMu.Lock()
+	narData, found := narHashDb[narHash.SRIString()]
+	narHashDbMu.Unlock()
+
+	rootNode := narData.rootNode
 
 	// if we didn't find anything, return 404.
 	if !found {
@@ -53,7 +55,7 @@ func renderNar(
 	directories := make(map[string]*castorev1pb.Directory)
 
 	// If the root node is a directory, ask the directory service for all directories
-	if pathInfoDirectory := pathInfo.GetNode().GetDirectory(); pathInfoDirectory != nil {
+	if pathInfoDirectory := rootNode.GetDirectory(); pathInfoDirectory != nil {
 		rootDirectoryDigest := pathInfoDirectory.GetDigest()
 		log = log.WithField("root_directory", base64.StdEncoding.EncodeToString(rootDirectoryDigest))
 
@@ -95,7 +97,7 @@ func renderNar(
 	// render the NAR file
 	err := storev1pb.Export(
 		w,
-		pathInfo.Node,
+		rootNode,
 		func(directoryDigest []byte) (*castorev1pb.Directory, error) {
 			log.WithField("directory", base64.StdEncoding.EncodeToString(directoryDigest)).Debug("Get directory")
 			directoryRefStr := hex.EncodeToString(directoryDigest)
@@ -177,7 +179,7 @@ func registerNarGet(s *Server) {
 			log := log.WithField("narhash_url", narHash.SRIString())
 
 			// TODO: inline more of that function here?
-			err = renderNar(ctx, log, s.directoryServiceClient, s.blobServiceClient, &s.narHashToPathInfoMu, s.narHashToPathInfo, w, narHash, isHead)
+			err = renderNar(ctx, log, s.directoryServiceClient, s.blobServiceClient, &s.narDbMu, s.narDb, w, narHash, isHead)
 			if err != nil {
 				if errors.Is(err, fs.ErrNotExist) {
 					w.WriteHeader(http.StatusNotFound)
diff --git a/tvix/nar-bridge/pkg/http/nar_put.go b/tvix/nar-bridge/pkg/http/nar_put.go
index 52ab425dfcf9..16e257537898 100644
--- a/tvix/nar-bridge/pkg/http/nar_put.go
+++ b/tvix/nar-bridge/pkg/http/nar_put.go
@@ -39,7 +39,7 @@ func registerNarPut(s *Server) {
 		directoriesUploader := importer.NewDirectoriesUploader(ctx, s.directoryServiceClient)
 		defer directoriesUploader.Done() //nolint:errcheck
 
-		pathInfo, err := importer.Import(
+		rootNode, narSize, narSha256, err := importer.Import(
 			ctx,
 			// buffer the body by 10MiB
 			bufio.NewReaderSize(r.Body, 10*1024*1024),
@@ -80,7 +80,7 @@ func registerNarPut(s *Server) {
 		// This check ensures the server-side came up with the same root hash.
 
 		if directoriesPutResponse != nil {
-			rootDigestPathInfo := pathInfo.GetNode().GetDirectory().GetDigest()
+			rootDigestPathInfo := rootNode.GetDirectory().GetDigest()
 			rootDigestDirectoriesPutResponse := directoriesPutResponse.GetRootDigest()
 
 			log := log.WithFields(logrus.Fields{
@@ -102,17 +102,18 @@ func registerNarPut(s *Server) {
 
 		// Compare the nar hash specified in the URL with the one that has been
 		// calculated while processing the NAR file
-		piNarHash, err := nixhash.ParseNixBase32(
-			"sha256:" + nixbase32.EncodeToString(pathInfo.GetNarinfo().NarSha256),
+		// TODO: bump go-nix and remove the parsing
+		narHash, err := nixhash.ParseNixBase32(
+			"sha256:" + nixbase32.EncodeToString(narSha256),
 		)
 		if err != nil {
 			panic("must parse nixbase32")
 		}
 
-		if !bytes.Equal(narHashFromUrl.Digest(), piNarHash.Digest()) {
+		if !bytes.Equal(narHashFromUrl.Digest(), narHash.Digest()) {
 			log := log.WithFields(logrus.Fields{
-				"narhash_received_sha256": piNarHash.SRIString(),
-				"narsize":                 pathInfo.GetNarinfo().GetNarSize(),
+				"narhash_received_sha256": narHash.SRIString(),
+				"narsize":                 narSize,
 			})
 			log.Error("received bytes don't match narhash from URL")
 
@@ -123,7 +124,6 @@ func registerNarPut(s *Server) {
 			}
 
 			return
-
 		}
 
 		// Insert the partial pathinfo structs into our lookup map,
@@ -131,9 +131,12 @@ func registerNarPut(s *Server) {
 		// The same  might exist already, but it'll have the same contents (so
 		// replacing will be a no-op), except maybe the root node Name field value, which
 		// is safe to ignore (as not part of the NAR).
-		s.narHashToPathInfoMu.Lock()
-		s.narHashToPathInfo[piNarHash.SRIString()] = pathInfo
-		s.narHashToPathInfoMu.Unlock()
+		s.narDbMu.Lock()
+		s.narDb[narHash.SRIString()] = &narData{
+			rootNode: rootNode,
+			narSize:  narSize,
+		}
+		s.narDbMu.Unlock()
 
 		// Done!
 	})
diff --git a/tvix/nar-bridge/pkg/http/narinfo_get.go b/tvix/nar-bridge/pkg/http/narinfo_get.go
index 8338d43676f5..6a537237a88b 100644
--- a/tvix/nar-bridge/pkg/http/narinfo_get.go
+++ b/tvix/nar-bridge/pkg/http/narinfo_get.go
@@ -29,7 +29,7 @@ func renderNarinfo(
 	log *log.Entry,
 	pathInfoServiceClient storev1pb.PathInfoServiceClient,
 	narHashToPathInfoMu *sync.Mutex,
-	narHashToPathInfo map[string]*storev1pb.PathInfo,
+	narHashToPathInfo map[string]*narData,
 	outputHash []byte,
 	w io.Writer,
 	headOnly bool,
@@ -51,6 +51,7 @@ func renderNarinfo(
 		return fmt.Errorf("unable to get pathinfo: %w", err)
 	}
 
+	// TODO: don't parse
 	narHash, err := nixhash.ParseNixBase32("sha256:" + nixbase32.EncodeToString(pathInfo.GetNarinfo().GetNarSha256()))
 	if err != nil {
 		// TODO: return proper error
@@ -59,7 +60,10 @@ func renderNarinfo(
 
 	// add things to the lookup table, in case the same process didn't handle the NAR hash yet.
 	narHashToPathInfoMu.Lock()
-	narHashToPathInfo[narHash.SRIString()] = pathInfo
+	narHashToPathInfo[narHash.SRIString()] = &narData{
+		rootNode: pathInfo.GetNode(),
+		narSize:  pathInfo.GetNarinfo().GetNarSize(),
+	}
 	narHashToPathInfoMu.Unlock()
 
 	if headOnly {
@@ -102,7 +106,7 @@ func registerNarinfoGet(s *Server) {
 			return
 		}
 
-		err = renderNarinfo(ctx, log, s.pathInfoServiceClient, &s.narHashToPathInfoMu, s.narHashToPathInfo, outputHash, w, false)
+		err = renderNarinfo(ctx, log, s.pathInfoServiceClient, &s.narDbMu, s.narDb, outputHash, w, false)
 		if err != nil {
 			if errors.Is(err, fs.ErrNotExist) {
 				w.WriteHeader(http.StatusNotFound)
diff --git a/tvix/nar-bridge/pkg/http/narinfo_put.go b/tvix/nar-bridge/pkg/http/narinfo_put.go
index 6494bca6f20d..fd588bec8644 100644
--- a/tvix/nar-bridge/pkg/http/narinfo_put.go
+++ b/tvix/nar-bridge/pkg/http/narinfo_put.go
@@ -2,14 +2,11 @@ package http
 
 import (
 	"net/http"
-	"path"
 
-	castorev1pb "code.tvl.fyi/tvix/castore/protos"
-	storev1pb "code.tvl.fyi/tvix/store/protos"
+	"code.tvl.fyi/tvix/nar-bridge/pkg/importer"
 	"github.com/go-chi/chi/v5"
 	"github.com/nix-community/go-nix/pkg/narinfo"
 	"github.com/nix-community/go-nix/pkg/nixbase32"
-	"github.com/nix-community/go-nix/pkg/storepath"
 	"github.com/sirupsen/logrus"
 	log "github.com/sirupsen/logrus"
 )
@@ -43,12 +40,10 @@ func registerNarinfoPut(s *Server) {
 			"output_path": narInfo.StorePath,
 		})
 
-		var pathInfo *storev1pb.PathInfo
-
 		// look up the narHash in our temporary map
-		s.narHashToPathInfoMu.Lock()
-		pathInfo, found := s.narHashToPathInfo[narInfo.NarHash.SRIString()]
-		s.narHashToPathInfoMu.Unlock()
+		s.narDbMu.Lock()
+		narData, found := s.narDb[narInfo.NarHash.SRIString()]
+		s.narDbMu.Unlock()
 		if !found {
 			log.Error("unable to find referred NAR")
 			w.WriteHeader(http.StatusBadRequest)
@@ -60,10 +55,12 @@ func registerNarinfoPut(s *Server) {
 			return
 		}
 
+		rootNode := narData.rootNode
+
 		// compare fields with what we computed while receiving the NAR file
 
 		// NarSize needs to match
-		if pathInfo.Narinfo.NarSize != narInfo.NarSize {
+		if narData.narSize != narInfo.NarSize {
 			log.Error("narsize mismatch")
 			w.WriteHeader(http.StatusBadRequest)
 			_, err := w.Write([]byte("unable to parse narinfo"))
@@ -73,90 +70,23 @@ func registerNarinfoPut(s *Server) {
 
 			return
 		}
-		// We know the narhash in the .narinfo matches one of the two narhashes in the partial pathInfo,
-		// because that's how we found it.
-
-		// FUTUREWORK: We can't compare References yet, but it'd be a good idea to
-		// do reference checking on .nar files server-side during upload.
-		// We however still need to be parse them, because we store
-		// the bytes in pathInfo.References, and the full strings in pathInfo.Narinfo.ReferenceNames.
-		referencesBytes := make([][]byte, 0)
-		for _, reference := range narInfo.References {
-			storePath, err := storepath.FromString(reference)
-			if err != nil {
-				log.WithField("reference", reference).WithError(err).Error("unable to parse reference")
-				w.WriteHeader(http.StatusBadRequest)
-				_, err := w.Write([]byte("unable to parse reference"))
-				if err != nil {
-					log.WithError(err).Errorf("unable to write error message to client")
-				}
-
-				return
-			}
-			referencesBytes = append(referencesBytes, storePath.Digest)
-		}
 
-		// assemble the []*storev1pb.NARInfo_Signature{} from narinfo.Signatures.
-		pbNarinfoSignatures := make([]*storev1pb.NARInfo_Signature, 0)
-		for _, narinfoSig := range narInfo.Signatures {
+		pathInfo, err := importer.GenPathInfo(rootNode, narInfo)
+		if err != nil {
+			log.WithError(err).Error("unable to generate PathInfo")
 
-			pbNarinfoSignatures = append(pbNarinfoSignatures, &storev1pb.NARInfo_Signature{
-				Name: narinfoSig.Name,
-				Data: narinfoSig.Data,
-			})
-		}
+			w.WriteHeader(http.StatusInternalServerError)
+			_, err := w.Write([]byte("unable to generate PathInfo"))
+			if err != nil {
+				log.WithError(err).Errorf("unable to write error message to client")
+			}
 
-		// If everything matches, We will add References, NAR signatures and the
-		// output path name, and then upload to the pathinfo service.
-		// We want a copy here, because we don't want to mutate the contents in the lookup table
-		// until we get things back from the remote store.
-		pathInfoToUpload := &storev1pb.PathInfo{
-			Node:       nil, // set below
-			References: referencesBytes,
-			Narinfo: &storev1pb.NARInfo{
-				NarSize:        pathInfo.Narinfo.NarSize,
-				NarSha256:      pathInfo.Narinfo.NarSha256,
-				Signatures:     pbNarinfoSignatures,
-				ReferenceNames: narInfo.References,
-			},
+			return
 		}
 
-		// We need to add the basename of the storepath from the .narinfo
-		// to the pathInfo to be sent.
-		switch v := (pathInfo.GetNode().GetNode()).(type) {
-		case *castorev1pb.Node_File:
-			pathInfoToUpload.Node = &castorev1pb.Node{
-				Node: &castorev1pb.Node_File{
-					File: &castorev1pb.FileNode{
-						Name:       []byte(path.Base(narInfo.StorePath)),
-						Digest:     v.File.Digest,
-						Size:       v.File.Size,
-						Executable: v.File.Executable,
-					},
-				},
-			}
-		case *castorev1pb.Node_Symlink:
-			pathInfoToUpload.Node = &castorev1pb.Node{
-				Node: &castorev1pb.Node_Symlink{
-					Symlink: &castorev1pb.SymlinkNode{
-						Name:   []byte(path.Base(narInfo.StorePath)),
-						Target: v.Symlink.Target,
-					},
-				},
-			}
-		case *castorev1pb.Node_Directory:
-			pathInfoToUpload.Node = &castorev1pb.Node{
-				Node: &castorev1pb.Node_Directory{
-					Directory: &castorev1pb.DirectoryNode{
-						Name:   []byte(path.Base(narInfo.StorePath)),
-						Digest: v.Directory.Digest,
-						Size:   v.Directory.Size,
-					},
-				},
-			}
-		}
+		log.WithField("pathInfo", pathInfo).Debug("inserted new pathInfo")
 
-		receivedPathInfo, err := s.pathInfoServiceClient.Put(ctx, pathInfoToUpload)
+		receivedPathInfo, err := s.pathInfoServiceClient.Put(ctx, pathInfo)
 		if err != nil {
 			log.WithError(err).Error("unable to upload pathinfo to service")
 			w.WriteHeader(http.StatusInternalServerError)
@@ -168,8 +98,6 @@ func registerNarinfoPut(s *Server) {
 			return
 		}
 
-		log.Debugf("received new pathInfo: %v+", receivedPathInfo)
-
-		// TODO: update the local temporary pathinfo with this?
+		log.WithField("pathInfo", receivedPathInfo).Debug("got back PathInfo")
 	})
 }
diff --git a/tvix/nar-bridge/pkg/http/server.go b/tvix/nar-bridge/pkg/http/server.go
index e8eb229ab41a..aa8037e2ba38 100644
--- a/tvix/nar-bridge/pkg/http/server.go
+++ b/tvix/nar-bridge/pkg/http/server.go
@@ -25,11 +25,17 @@ type Server struct {
 	// When uploading NAR files to a HTTP binary cache, the .nar
 	// files are uploaded before the .narinfo files.
 	// We need *both* to be able to fully construct a PathInfo object.
-	// Keep a in-memory map of narhash(es) (in SRI) to sparse PathInfo.
+	// Keep a in-memory map of narhash(es) (in SRI) to (unnamed) root node and nar
+	// size.
 	// This is necessary until we can ask a PathInfoService for a node with a given
 	// narSha256.
-	narHashToPathInfoMu sync.Mutex
-	narHashToPathInfo   map[string]*storev1pb.PathInfo
+	narDbMu sync.Mutex
+	narDb   map[string]*narData
+}
+
+type narData struct {
+	rootNode *castorev1pb.Node
+	narSize  uint64
 }
 
 func New(
@@ -64,7 +70,7 @@ func New(
 		directoryServiceClient: directoryServiceClient,
 		blobServiceClient:      blobServiceClient,
 		pathInfoServiceClient:  pathInfoServiceClient,
-		narHashToPathInfo:      make(map[string]*storev1pb.PathInfo),
+		narDb:                  make(map[string]*narData),
 	}
 
 	registerNarPut(s)