http: sniffing placeholder

R=dsymonds
CC=golang-dev
https://golang.org/cl/4746041

src/pkg/http/Makefile

@@ -18,6 +18,7 @@ GOFILES=\
 	response.go\
 	reverseproxy.go\
 	server.go\
+	sniff.go\
 	status.go\
 	transfer.go\
 	transport.go\

src/pkg/http/server.go

@@ -99,6 +99,7 @@ type conn struct {
 	buf        *bufio.ReadWriter    // buffered rwc
 	hijacked   bool                 // connection has been hijacked by handler
 	tlsState   *tls.ConnectionState // or nil when not using TLS
+	body       []byte
 }
 
 // A response represents the server side of an HTTP response.
@@ -112,6 +113,7 @@ type response struct {
 	written       int64    // number of bytes written in body
 	contentLength int64    // explicitly-declared Content-Length; or -1
 	status        int      // status code passed to WriteHeader
+	needSniff     bool     // need to sniff to find Content-Type
 
 	// close connection after this reply.  set on request and
 	// updated after response from handler if there's a
@@ -147,6 +149,7 @@ func newConn(rwc net.Conn, handler Handler) (c *conn, err os.Error) {
 	c.remoteAddr = rwc.RemoteAddr().String()
 	c.handler = handler
 	c.rwc = rwc
+	c.body = make([]byte, sniffLen)
 	br := bufio.NewReader(rwc)
 	bw := bufio.NewWriter(rwc)
 	c.buf = bufio.NewReadWriter(br, bw)
@@ -209,6 +212,7 @@ func (c *conn) readRequest() (w *response, err os.Error) {
 	w.req = req
 	w.header = make(Header)
 	w.contentLength = -1
+	c.body = c.body[:0]
 	return w, nil
 }
 
@@ -249,9 +253,9 @@ func (w *response) WriteHeader(code int) {
 			}
 		}
 	} else {
-		// Default output is HTML encoded in UTF-8.
+		// If no content type, apply sniffing algorithm to body.
 		if w.header.Get("Content-Type") == "" {
-			w.header.Set("Content-Type", "text/html; charset=utf-8")
+			w.needSniff = true
 		}
 	}
 
@@ -337,8 +341,35 @@ func (w *response) WriteHeader(code int) {
 	}
 	io.WriteString(w.conn.buf, proto+" "+codestring+" "+text+"\r\n")
 	w.header.Write(w.conn.buf)
+
+	// If we need to sniff the body, leave the header open.
+	// Otherwise, end it here.
+	if !w.needSniff {
 		io.WriteString(w.conn.buf, "\r\n")
 	}
+}
+
+// sniff uses the first block of written data,
+// stored in w.conn.body, to decide the Content-Type
+// for the HTTP body.
+func (w *response) sniff() {
+	if !w.needSniff {
+		return
+	}
+	w.needSniff = false
+
+	data := w.conn.body
+	ctype := detectContentType(data)
+	if ctype != "" {
+		fmt.Fprintf(w.conn.buf, "Content-Type: %s\r\n", ctype)
+	}
+	io.WriteString(w.conn.buf, "\r\n")
+
+	if w.chunking && len(data) > 0 {
+		fmt.Fprintf(w.conn.buf, "%x\r\n", len(data))
+		w.conn.buf.Write(data)
+	}
+}
 
 // bodyAllowed returns true if a Write is allowed for this response type.
 // It's illegal to call this before the header has been flushed.
@@ -369,6 +400,18 @@ func (w *response) Write(data []byte) (n int, err os.Error) {
 		return 0, ErrContentLength
 	}
 
+	var m int
+	if w.needSniff {
+		body := w.conn.body
+		m = copy(body[len(body):], data)
+		w.conn.body = body[:len(body)+m]
+		if m == len(data) {
+			return m, nil
+		}
+		w.sniff()
+		data = data[m:]
+	}
+
 	// TODO(rsc): if chunking happened after the buffering,
 	// then there would be fewer chunk headers.
 	// On the other hand, it would make hijacking more difficult.
@@ -385,7 +428,7 @@ func (w *response) Write(data []byte) (n int, err os.Error) {
 		}
 	}
 
-	return n, err
+	return m + n, err
 }
 
 // If this is an error reply (4xx or 5xx)
@@ -449,6 +492,9 @@ func (w *response) finishRequest() {
 	if !w.wroteHeader {
 		w.WriteHeader(StatusOK)
 	}
+	if w.needSniff {
+		w.sniff()
+	}
 	errorKludge(w)
 	if w.chunking {
 		io.WriteString(w.conn.buf, "0\r\n")
@@ -471,6 +517,7 @@ func (w *response) Flush() {
 	if !w.wroteHeader {
 		w.WriteHeader(StatusOK)
 	}
+	w.sniff()
 	w.conn.buf.Flush()
 }
 
@@ -517,6 +564,7 @@ func (c *conn) serve() {
 			if req.ContentLength == 0 {
 				w.Header().Set("Connection", "close")
 				w.WriteHeader(StatusBadRequest)
+				w.finishRequest()
 				break
 			}
 			req.Header.Del("Expect")
@@ -535,6 +583,7 @@ func (c *conn) serve() {
 			// respond with a 417 (Expectation Failed) status."
 			w.Header().Set("Connection", "close")
 			w.WriteHeader(StatusExpectationFailed)
+			w.finishRequest()
 			break
 		}
 

src/pkg/http/sniff.go

@@ -0,0 +1,18 @@
+// Copyright 2011 The Go Authors.  All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package http
+
+// Content-type sniffing algorithm.
+// http://tools.ietf.org/html/draft-ietf-websec-mime-sniff-03
+
+// The algorithm prefers to use sniffLen bytes to make its decision.
+const sniffLen = 1024
+
+// detectContentType returns the sniffed Content-Type string
+// for the given data.
+func detectContentType(data []byte) string {
+	// TODO(dsymonds,rsc): Implement algorithm from draft.
+	return "text/html; charset=utf-8"
+}