fix: improve url parsing

pkg/app/server/hertz_test.go

@@ -259,7 +259,7 @@ func TestNotAbsolutePath(t *testing.T) {
 	go engine.Run()
 	time.Sleep(200 * time.Microsecond)
 
-	s := "POST ?a=b HTTP/1.1\r\nContent-Length: 5\r\nContent-Type: foo/bar\r\n\r\nabcdef4343"
+	s := "POST ?a=b HTTP/1.1\r\nHost: a.b.c\r\nContent-Length: 5\r\nContent-Type: foo/bar\r\n\r\nabcdef4343"
 	zr := mock.NewZeroCopyReader(s)
 
 	ctx := app.NewContext(0)
@@ -270,7 +270,7 @@ func TestNotAbsolutePath(t *testing.T) {
 	assert.DeepEqual(t, consts.StatusOK, ctx.Response.StatusCode())
 	assert.DeepEqual(t, ctx.Request.Body(), ctx.Response.Body())
 
-	s = "POST a?a=b HTTP/1.1\r\nContent-Length: 5\r\nContent-Type: foo/bar\r\n\r\nabcdef4343"
+	s = "POST a?a=b HTTP/1.1\r\nHost: a.b.c\r\nContent-Length: 5\r\nContent-Type: foo/bar\r\n\r\nabcdef4343"
 	zr = mock.NewZeroCopyReader(s)
 
 	ctx = app.NewContext(0)
@@ -291,7 +291,7 @@ func TestNotAbsolutePathWithRawPath(t *testing.T) {
 	go engine.Run()
 	time.Sleep(200 * time.Microsecond)
 
-	s := "POST ?a=b HTTP/1.1\r\nContent-Length: 5\r\nContent-Type: foo/bar\r\n\r\nabcdef4343"
+	s := "POST ?a=b HTTP/1.1\r\nHost: a.b.c\r\nContent-Length: 5\r\nContent-Type: foo/bar\r\n\r\nabcdef4343"
 	zr := mock.NewZeroCopyReader(s)
 
 	ctx := app.NewContext(0)
@@ -302,7 +302,7 @@ func TestNotAbsolutePathWithRawPath(t *testing.T) {
 	assert.DeepEqual(t, consts.StatusBadRequest, ctx.Response.StatusCode())
 	assert.DeepEqual(t, default400Body, ctx.Response.Body())
 
-	s = "POST a?a=b HTTP/1.1\r\nContent-Length: 5\r\nContent-Type: foo/bar\r\n\r\nabcdef4343"
+	s = "POST a?a=b HTTP/1.1\r\nHost: a.b.c\r\nContent-Length: 5\r\nContent-Type: foo/bar\r\n\r\nabcdef4343"
 	zr = mock.NewZeroCopyReader(s)
 
 	ctx = app.NewContext(0)

pkg/protocol/uri.go

@@ -49,6 +49,7 @@
 	"github.com/cloudwego/hertz/internal/bytesconv"
 	"github.com/cloudwego/hertz/internal/bytestr"
 	"github.com/cloudwego/hertz/internal/nocopy"
+	"github.com/cloudwego/hertz/pkg/common/hlog"
 )
 
 // AcquireURI returns an empty URI instance from the pool.
@@ -373,6 +374,34 @@
 	u.parse(host, uri, false)
 }
 
+// Maybe rawURL is of the form scheme:path.
+// (Scheme must be [a-zA-Z][a-zA-Z0-9+-.]*)
+// If so, return scheme, path; else return nil, rawURL.
+func getScheme(rawURL []byte) (scheme, path []byte) {
+	for i := 0; i < len(rawURL); i++ {
+		c := rawURL[i]
+		switch {
+		case 'a' <= c && c <= 'z' || 'A' <= c && c <= 'Z':
+		// do nothing
+		case '0' <= c && c <= '9' || c == '+' || c == '-' || c == '.':
+			if i == 0 {
+				return nil, rawURL
+			}
+		case c == ':':
+			if i == 0 {
+				hlog.Errorf("error happened when try to parse the rawURL(%s): missing protocol scheme", rawURL)
+				return nil, nil
+			}
+			return rawURL[:i], rawURL[i+1:]
+		default:
+			// we have encountered an invalid character,
+			// so there is no valid scheme
+			return nil, rawURL
+		}
+	}
+	return nil, rawURL
+}
+
 func (u *URI) parse(host, uri []byte, isTLS bool) {
 	u.Reset()
 
@@ -455,20 +484,14 @@
 }
 
 func splitHostURI(host, uri []byte) ([]byte, []byte, []byte) {
-	n := bytes.Index(uri, bytestr.StrSlashSlash)
-	if n < 0 {
-		return bytestr.StrHTTP, host, uri
-	}
-	scheme := uri[:n]
-	if bytes.IndexByte(scheme, '/') >= 0 {
+	scheme, path := getScheme(uri)
+
+	if scheme == nil {
 		return bytestr.StrHTTP, host, uri
 	}
-	if len(scheme) > 0 && scheme[len(scheme)-1] == ':' {
-		scheme = scheme[:len(scheme)-1]
-	}
-	n += len(bytestr.StrSlashSlash)
-	uri = uri[n:]
-	n = bytes.IndexByte(uri, '/')
+
+	uri = path[len(bytestr.StrSlashSlash):]
+	n := bytes.IndexByte(uri, '/')
 	if n < 0 {
 		// A hack for bogus urls like foobar.com?a=b without
 		// slash after host.
@@ -587,6 +610,9 @@
 		if len(u.scheme) > 0 {
 			schemeOriginal = append([]byte(nil), u.scheme...)
 		}
+		if n == 0 {
+			newURI = bytes.Join([][]byte{u.scheme, bytestr.StrColon, newURI}, nil)
+		}
 		u.Parse(nil, newURI)
 		if len(schemeOriginal) > 0 && len(u.scheme) == 0 {
 			u.scheme = append(u.scheme[:0], schemeOriginal...)

pkg/protocol/uri_test.go

@@ -42,6 +42,7 @@
 package protocol
 
 import (
+	"bytes"
 	"path/filepath"
 	"reflect"
 	"runtime"
@@ -468,3 +469,35 @@ func TestParseURI(t *testing.T) {
 	uri := string(ParseURI(expectURI).FullURI())
 	assert.DeepEqual(t, expectURI, uri)
 }
+
+func TestSplitHostURI(t *testing.T) {
+	cases := []struct {
+		host, uri                      []byte
+		wantScheme, wantHost, wantPath []byte
+	}{
+		{
+			[]byte("example.com"), []byte("/foobar"),
+			[]byte("http"), []byte("example.com"), []byte("/foobar"),
+		},
+		{
+			[]byte("example2.com"), []byte("http://example2.com"),
+			[]byte("http"), []byte("example2.com"), []byte("/"),
+		},
+		{
+			[]byte("example2.com"), []byte("http://example3.com"),
+			[]byte("http"), []byte("example3.com"), []byte("/"),
+		},
+		{
+			[]byte("example3.com"), []byte("https://foobar.com?a=b"),
+			[]byte("https"), []byte("foobar.com"), []byte("?a=b"),
+		},
+	}
+
+	for _, c := range cases {
+		gotScheme, gotHost, gotPath := splitHostURI(c.host, c.uri)
+		if !bytes.Equal(gotScheme, c.wantScheme) || !bytes.Equal(gotHost, c.wantHost) || !bytes.Equal(gotPath, c.wantPath) {
+			t.Errorf("splitHostURI(%q, %q) == (%q, %q, %q), want (%q, %q, %q)",
+				c.host, c.uri, gotScheme, gotHost, gotPath, c.wantScheme, c.wantHost, c.wantPath)
+		}
+	}
+}

pkg/route/engine.go

@@ -89,6 +89,8 @@
 	default404Body = []byte("404 page not found")
 	default405Body = []byte("405 method not allowed")
 	default400Body = []byte("400 bad request")
+
+	requiredHostBody = []byte("missing required Host header")
 )
 
 type hijackConn struct {
@@ -721,6 +723,13 @@
 	}
 
 	rPath := string(ctx.Request.URI().Path())
+
+	// align with https://datatracker.ietf.org/doc/html/rfc2616#section-5.2
+	if len(ctx.Request.Host()) == 0 && ctx.Request.Header.IsHTTP11() && bytesconv.B2s(ctx.Request.Method()) != consts.MethodConnect {
+		serveError(c, ctx, consts.StatusBadRequest, requiredHostBody)
+		return
+	}
+
 	httpMethod := bytesconv.B2s(ctx.Request.Header.Method())
 	unescape := false
 	if engine.options.UseRawPath {