Merge pull request #237 from lijw30/master

[refactor] split ArrayEnd from ObjectEnd of rtmp amf0 to make code more readable

pkg/rtmp/amf0.go

@@ -48,7 +48,11 @@ const (
 	//Amf0TypeMarkerTypedObject = uint8(0x10)
 )
 
-var Amf0TypeMarkerObjectEndBytes = []byte{0, 0, Amf0TypeMarkerObjectEnd}
+var (
+	// object-end-type(0x00 0x00 0x09) 表示Object和EcmaArray类型的结束标识
+	Amf0TypeMarkerObjectEndBytes = []byte{0, 0, Amf0TypeMarkerObjectEnd}
+	Amf0TypeMarkerArrayEndBytes  = []byte{0, 0, Amf0TypeMarkerObjectEnd}
+)
 
 // ---------------------------------------------------------------------------------------------------------------------
 
@@ -396,10 +400,9 @@ func (amf0) ReadArray(b []byte) (ObjectPairArray, int, error) {
 		}
 	}
 
-	if len(b)-index >= 3 && bytes.Equal(b[index:index+3], Amf0TypeMarkerObjectEndBytes) {
+	if len(b)-index >= 3 && bytes.Equal(b[index:index+3], Amf0TypeMarkerArrayEndBytes) {
 		index += 3
 	} else {
-		// 测试时发现Array最后也是以00 00 09结束，不确定是否是标准规定的，加个日志在这
 		Log.Warn("amf ReadArray without suffix Amf0TypeMarkerObjectEndBytes.")
 	}
 	return ops, index, nil

pkg/rtmp/chunk_composer.go

@@ -270,7 +270,7 @@ func (c *ChunkComposer) RunLoop(reader io.Reader, cb OnCompleteMessage) error {
 
 		// TODO(chef): 这里应该永远执行不到，可以删除掉
 		if stream.msg.Len() > stream.header.MsgLen {
-			return base.NewErrRtmpShortBuffer(int(stream.header.MsgLen), int(stream.msg.Len()), "len of msg bigger tthan msg len of header")
+			return base.NewErrRtmpShortBuffer(int(stream.header.MsgLen), int(stream.msg.Len()), "len of msg bigger than msg len of header")
 		}
 	}
 }