Sniff Content-Type during site update.

This isn't yet used in the code responding to GET requests because we
do not yet have a migration path for legacy code.

src/manifest.go

@@ -9,7 +9,10 @@ import (
 	"errors"
 	"fmt"
 	"log"
+	"mime"
+	"net/http"
 	"path"
+	"path/filepath"
 	"strings"
 	"sync"
 
@@ -122,6 +125,24 @@ again:
 	}
 }
 
+// Sniff content type using the same algorithm as `http.ServeContent`.
+func DetectContentType(manifest *Manifest) {
+	for path, entry := range manifest.Contents {
+		if entry.GetType() == Type_Directory || entry.GetType() == Type_Symlink {
+			// no Content-Type
+		} else if entry.GetType() == Type_InlineFile && entry.GetTransform() == Transform_None {
+			contentType := mime.TypeByExtension(filepath.Ext(path))
+			if contentType == "" {
+				contentType = http.DetectContentType(entry.Data[:512])
+			}
+			entry.ContentType = proto.String(contentType)
+		} else {
+			panic(fmt.Errorf("DetectContentType encountered invalid entry: %v, %v",
+				entry.GetType(), entry.GetTransform()))
+		}
+	}
+}
+
 // The `clauspost/compress/zstd` package recommends reusing a compressor to avoid repeated
 // allocations of internal buffers.
 var zstdEncoder, _ = zstd.NewWriter(nil, zstd.WithEncoderLevel(zstd.SpeedBetterCompression))
@@ -133,13 +154,13 @@ func CompressFiles(ctx context.Context, manifest *Manifest) {
 
 	var originalSize, transformedSize int64
 	for _, entry := range manifest.Contents {
-		if entry.GetType() == Type_InlineFile && entry.GetXfrm() == Transform_None {
+		if entry.GetType() == Type_InlineFile && entry.GetTransform() == Transform_None {
 			originalSize += entry.GetSize()
 			compressedData := zstdEncoder.EncodeAll(entry.GetData(), make([]byte, 0, entry.GetSize()))
 			if len(compressedData) < int(*entry.Size) {
 				entry.Data = compressedData
 				entry.Size = proto.Int64(int64(len(entry.Data)))
-				entry.Xfrm = Transform_Zstandard.Enum()
+				entry.Transform = Transform_Zstandard.Enum()
 			}
 			transformedSize += entry.GetSize()
 		}
@@ -163,6 +184,8 @@ func PrepareManifest(ctx context.Context, manifest *Manifest) error {
 		log.Printf("redirects ok: %d rules\n", len(manifest.Redirects))
 	}
 
+	DetectContentType(manifest)
+
 	if config.Feature("compress") {
 		CompressFiles(ctx, manifest)
 	}
@@ -196,10 +219,11 @@ func StoreManifest(ctx context.Context, name string, manifest *Manifest) (*Manif
 		if cannotBeInlined {
 			dataHash := sha256.Sum256(entry.Data)
 			extManifest.Contents[name] = &Entry{
-				Type: Type_ExternalFile.Enum(),
-				Size: entry.Size,
-				Data: fmt.Appendf(nil, "sha256-%x", dataHash),
-				Xfrm: entry.Xfrm,
+				Type:        Type_ExternalFile.Enum(),
+				Size:        entry.Size,
+				Data:        fmt.Appendf(nil, "sha256-%x", dataHash),
+				Transform:   entry.Transform,
+				ContentType: entry.ContentType,
 			}
 			extObjectMap[string(dataHash[:])] = *entry.Size
 		} else {

src/pages.go

@@ -223,7 +223,7 @@ func getPage(w http.ResponseWriter, r *http.Request) error {
 		defer closer.Close()
 	}
 
-	switch entry.GetXfrm() {
+	switch entry.GetTransform() {
 	case Transform_None:
 		// nothing to do
 	case Transform_Zstandard:

src/schema.pb.go

@@ -144,7 +144,10 @@ type Entry struct {
 	Data []byte `protobuf:"bytes,3,opt,name=data" json:"data,omitempty"`
 	// Only present for `type == InlineFile` and `type == ExternalFile` that
 	// have been transformed.
-	Xfrm          *Transform `protobuf:"varint,4,opt,name=xfrm,enum=Transform" json:"xfrm,omitempty"`
+	Transform *Transform `protobuf:"varint,4,opt,name=transform,enum=Transform" json:"transform,omitempty"`
+	// Only present for `type == InlineFile` and `type == ExternalFile`.
+	// Currently, optional (not present on certain legacy manifests).
+	ContentType   *string `protobuf:"bytes,5,opt,name=content_type,json=contentType" json:"content_type,omitempty"`
 	unknownFields protoimpl.UnknownFields
 	sizeCache     protoimpl.SizeCache
 }
@@ -200,13 +203,20 @@ func (x *Entry) GetData() []byte {
 	return nil
 }
 
-func (x *Entry) GetXfrm() Transform {
-	if x != nil && x.Xfrm != nil {
-		return *x.Xfrm
+func (x *Entry) GetTransform() Transform {
+	if x != nil && x.Transform != nil {
+		return *x.Transform
 	}
 	return Transform_None
 }
 
+func (x *Entry) GetContentType() string {
+	if x != nil && x.ContentType != nil {
+		return *x.ContentType
+	}
+	return ""
+}
+
 // See https://docs.netlify.com/manage/routing/redirects/overview/ for details.
 // Only a subset of the Netlify specification is representable here.
 type Redirect struct {
@@ -337,8 +347,8 @@ type Manifest struct {
 	Commit  *string `protobuf:"bytes,3,opt,name=commit" json:"commit,omitempty"`
 	// Contents
 	Contents   map[string]*Entry `protobuf:"bytes,4,rep,name=contents" json:"contents,omitempty" protobuf_key:"bytes,1,opt,name=key" protobuf_val:"bytes,2,opt,name=value"`
-	TotalSize  *int64            `protobuf:"varint,5,opt,name=total_size,json=totalSize" json:"total_size,omitempty"`
-	StoredSize *int64            `protobuf:"varint,8,opt,name=stored_size,json=storedSize" json:"stored_size,omitempty"` // after deduplication
+	TotalSize  *int64            `protobuf:"varint,5,opt,name=total_size,json=totalSize" json:"total_size,omitempty"`    // simple sum of each `entry.size`
+	StoredSize *int64            `protobuf:"varint,8,opt,name=stored_size,json=storedSize" json:"stored_size,omitempty"` // external objects, after deduplication
 	// Netlify-style `_redirects`
 	Redirects []*Redirect `protobuf:"bytes,6,rep,name=redirects" json:"redirects,omitempty"`
 	// Diagnostics for non-fatal errors
@@ -437,13 +447,14 @@ var File_schema_proto protoreflect.FileDescriptor
 
 const file_schema_proto_rawDesc = "" +
 	"\n" +
-	"\fschema.proto\"j\n" +
+	"\fschema.proto\"\x97\x01\n" +
 	"\x05Entry\x12\x19\n" +
 	"\x04type\x18\x01 \x01(\x0e2\x05.TypeR\x04type\x12\x12\n" +
 	"\x04size\x18\x02 \x01(\x03R\x04size\x12\x12\n" +
-	"\x04data\x18\x03 \x01(\fR\x04data\x12\x1e\n" +
-	"\x04xfrm\x18\x04 \x01(\x0e2\n" +
-	".TransformR\x04xfrm\"\\\n" +
+	"\x04data\x18\x03 \x01(\fR\x04data\x12(\n" +
+	"\ttransform\x18\x04 \x01(\x0e2\n" +
+	".TransformR\ttransform\x12!\n" +
+	"\fcontent_type\x18\x05 \x01(\tR\vcontentType\"\\\n" +
 	"\bRedirect\x12\x12\n" +
 	"\x04from\x18\x01 \x01(\tR\x04from\x12\x0e\n" +
 	"\x02to\x18\x02 \x01(\tR\x02to\x12\x16\n" +
@@ -502,7 +513,7 @@ var file_schema_proto_goTypes = []any{
 }
 var file_schema_proto_depIdxs = []int32{
 	0, // 0: Entry.type:type_name -> Type
-	1, // 1: Entry.xfrm:type_name -> Transform
+	1, // 1: Entry.transform:type_name -> Transform
 	6, // 2: Manifest.contents:type_name -> Manifest.ContentsEntry
 	3, // 3: Manifest.redirects:type_name -> Redirect
 	4, // 4: Manifest.problems:type_name -> Problem

src/schema.proto

@@ -36,7 +36,10 @@ message Entry {
 	bytes data = 3;
 	// Only present for `type == InlineFile` and `type == ExternalFile` that
 	// have been transformed.
-	Transform xfrm = 4;
+	Transform transform = 4;
+	// Only present for `type == InlineFile` and `type == ExternalFile`.
+	// Currently, optional (not present on certain legacy manifests).
+	string content_type = 5;
 }
 
 // See https://docs.netlify.com/manage/routing/redirects/overview/ for details.