Serve compressed content directly if client indicates support.

src/http.go

@@ -0,0 +1,89 @@
+package git_pages
+
+import (
+	"cmp"
+	"regexp"
+	"slices"
+	"strconv"
+	"strings"
+)
+
+var httpAcceptEncodingRegexp = regexp.MustCompile(`` +
+	// token optionally prefixed by whitespace
+	`^[ \t]*([a-zA-Z0-9$!#$%&'*+.^_\x60|~-]+)` +
+	// quality value prefixed by a semicolon optionally surrounded by whitespace
+	`(?:[ \t]*;[ \t]*q=(0(?:\.[0-9]{1,3})?|1(?:\.0{1,3})?))?` +
+	// optional whitespace followed by comma or end of line
+	`[ \t]*(?:,|$)`,
+)
+
+type httpEncoding struct {
+	code string
+	qval float64
+}
+
+type httpEncodings struct {
+	encodings []httpEncoding
+}
+
+func parseHTTPEncodings(headerValue string) (result httpEncodings) {
+	for headerValue != "" {
+		matches := httpAcceptEncodingRegexp.FindStringSubmatch(headerValue)
+		if matches == nil {
+			return httpEncodings{}
+		}
+		enc := httpEncoding{strings.ToLower(matches[1]), 1.0}
+		if matches[2] != "" {
+			enc.qval, _ = strconv.ParseFloat(matches[2], 64)
+		}
+		result.encodings = append(result.encodings, enc)
+		headerValue = headerValue[len(matches[0]):]
+	}
+	if len(result.encodings) == 0 {
+		// RFC 9110 says (https://httpwg.org/specs/rfc9110.html#field.accept-encoding):
+		// "If no Accept-Encoding header field is in the request, any content
+		// coding is considered acceptable by the user agent."
+		// In practice, no client expects to receive a compressed response
+		// without having sent Accept-Encoding in the request.
+	}
+	return
+}
+
+// Negotiate returns the most preferred encoding that is acceptable by the
+// client, or an empty string if no encodings are acceptable.
+func (e *httpEncodings) Negotiate(codes ...string) string {
+	prefs := make(map[string]float64, len(codes))
+	for _, code := range codes {
+		prefs[code] = 0
+	}
+	implicitIdentity := true
+	for _, enc := range e.encodings {
+		if enc.code == "*" {
+			for code := range prefs {
+				prefs[code] = enc.qval
+			}
+			implicitIdentity = false
+		} else if _, ok := prefs[enc.code]; ok {
+			prefs[enc.code] = enc.qval
+		}
+		if enc.code == "*" || enc.code == "identity" {
+			implicitIdentity = false
+		}
+	}
+	if _, ok := prefs["identity"]; ok && implicitIdentity {
+		prefs["identity"] = -1 // sort last
+	}
+	encs := make([]httpEncoding, len(codes))
+	for idx, code := range codes {
+		encs[idx] = httpEncoding{code, prefs[code]}
+	}
+	slices.SortStableFunc(encs, func(a, b httpEncoding) int {
+		return -cmp.Compare(a.qval, b.qval)
+	})
+	for _, enc := range encs {
+		if enc.qval != 0 {
+			return enc.code
+		}
+	}
+	return ""
+}

src/pages.go

@@ -13,6 +13,7 @@ import (
 	"net/url"
 	"os"
 	"path"
+	"strconv"
 	"strings"
 	"time"
 
@@ -229,21 +230,44 @@ func getPage(w http.ResponseWriter, r *http.Request) error {
 		defer closer.Close()
 	}
 
+	acceptedEncodings := parseHTTPEncodings(r.Header.Get("Accept-Encoding"))
+	negotiatedEncoding := true
+
 	switch entry.GetTransform() {
 	case Transform_None:
-		// nothing to do
-	case Transform_Zstandard:
-		// Ideally, we would serve zstd-compressed data to a client that indicates support with
-		// an `Accept-Encoding: zstd` header. Unfortunately we can't because we rely on MIME
-		// type detection done in `http.ServeContent`.
-		compressedData, _ := io.ReadAll(reader)
-		decompressedData, err := zstdDecoder.DecodeAll(compressedData, []byte{})
-		if err != nil {
-			w.WriteHeader(http.StatusInternalServerError)
-			fmt.Fprintf(w, "internal server error: %s\n", err)
-			return err
+		if acceptedEncodings.Negotiate("identity") != "identity" {
+			negotiatedEncoding = false
 		}
-		reader = bytes.NewReader(decompressedData)
+	case Transform_Zstandard:
+		supported := []string{"zstd", "identity"}
+		if entry.ContentType == nil {
+			// If Content-Type is unset, `http.ServeContent` will try to sniff
+			// the file contents. That won't work if it's compressed.
+			supported = []string{"identity"}
+		}
+		switch acceptedEncodings.Negotiate(supported...) {
+		case "zstd":
+			// Set Content-Length ourselves since `http.ServeContent` only sets
+			// it if Content-Encoding is unset or if it's a range request.
+			w.Header().Set("Content-Length", strconv.FormatInt(*entry.Size, 10))
+			w.Header().Set("Content-Encoding", "zstd")
+		case "identity":
+			compressedData, _ := io.ReadAll(reader)
+			decompressedData, err := zstdDecoder.DecodeAll(compressedData, []byte{})
+			if err != nil {
+				w.WriteHeader(http.StatusInternalServerError)
+				fmt.Fprintf(w, "internal server error: %s\n", err)
+				return err
+			}
+			reader = bytes.NewReader(decompressedData)
+		default:
+			negotiatedEncoding = false
+		}
+	}
+	if !negotiatedEncoding {
+		w.WriteHeader(http.StatusNotAcceptable)
+		return fmt.Errorf("no supported content encodings (accept-encoding: %q)",
+			r.Header.Get("Accept-Encoding"))
 	}
 
 	// decide on the HTTP status
@@ -253,6 +277,11 @@ func getPage(w http.ResponseWriter, r *http.Request) error {
 			io.Copy(w, reader)
 		}
 	} else {
+		if entry.ContentType != nil {
+			// don't let http.ServeContent mime-sniff compressed data
+			w.Header().Set("Content-Type", *entry.ContentType)
+		}
+
 		// allow the use of multi-threading in WebAssembly
 		w.Header().Set("Cross-Origin-Embedder-Policy", "credentialless")
 		w.Header().Set("Cross-Origin-Opener-Policy", "same-origin")
@@ -265,7 +294,7 @@ func getPage(w http.ResponseWriter, r *http.Request) error {
 		w.Header().Set("Cache-Control", "max-age=60, stale-while-revalidate=3600")
 		// see https://web.dev/articles/stale-while-revalidate for details
 
-		// http.ServeContent handles content type and caching
+		// http.ServeContent handles conditional requests and range requests
 		http.ServeContent(w, r, entryPath, mtime, reader)
 	}
 	return nil