写了一个对xml文件的词法分析,程序的分析引擎用switch+state写的,主要是为加深对词法分析的理解.
while (sr.Peek()!=-1)
{
//int c = sr.Read(); //Char ch = (char)c; //Console.WriteLine("{0}\t {1}\t {2}",ch,c, Convert.ToString(c,2).PadLeft(16,'0') );
////switch
c = (char)sr.Read(); switch(state) { case 0://普通 jumpspace(); if(c=='<') state = 10; break; case 10://<[?] switch(c) { case '?': state =11;//<? break; case ' ': case '\t': case '\r': case '\n': throw new Exception("该位置上不允许有空白"); break; case '/': state = 50;//结束<[/]
.> break; default: state = 20;//元素名 sb.Remove(0,sb.Length);//加元素 sb.Append(c); while (sr.Peek() != -1) { char nextchar = (char)sr.Peek(); if (char.IsLetter(nextchar)) { sb.Append(nextchar);//element c = (char)sr.Read();
} else { //完成元素 //string temp = sb.ToString(); Console.WriteLine(sb.ToString()); StID.Push(sb.ToString()); sb.Remove(0, sb.Length); state = 15;//后接 / > ' ' jumpspace(); break; } } break; } break; case 11://<?[x] --系统 switch(c) { case 'x': state =12; break; } break; case 15: jumpspace(); if (c == '/' ) { Console.WriteLine( StID.Pop()); //完成 c =(char)sr.Read(); if (c != '>') throw new Exception("/>"); //下来可能是节点,也可能是文本 state = 0; } else if (c == '>') { state = 0;//原始状态 } else if (char.IsLetterOrDigit(c)) { //sb.Append(c);//属性 //属性开始 state = 30; goto case 30; } else { throw new Exception("错误的属性"); } break; case 20://元素名-第二个 只是元素 if (char.IsLetterOrDigit(c)) { sb.Append(c);//element } else if (char.IsWhiteSpace(c))//遍历空白 { //完成 //string temp = sb.ToString(); Console.WriteLine(sb.ToString()); StID.Push(sb.ToString()); sb.Remove(0, sb.Length); while (true) { c = (char)sr.Peek(); if (!char.IsWhiteSpace(c))//如果不空白,跳出 break; sr.Read(); } state = 30;//属性 } else if (c == '>') { //完成 //string temp = sb.ToString(); Console.WriteLine(sb.ToString()); StID.Push(sb.ToString()); sb.Remove(0, sb.Length); state = 15;//中间状态,下一个可能是元素也可能是text } else { throw new Exception("无效的字符"); } break; case 30://属性名first 后继 if (char.IsLetterOrDigit(c)) { sb.Append(c);//属性 } else { goto case 31; } break; case 31: jumpspace(); if(c=='=') { state =32; } else { throw new Exception("属性无效字符!"); } break; case 32://属性引号 jumpspace(); switch(c) { case '\'': state =36;//单引号; break; case '\"': state = 37;//双引号 break; default: throw new Exception("应该是引号"); } break; case 36: switch (c) { case '\'': state = 30;//复原,下一属性 jumpspace(); state = 15; break; case '>': throw new Exception("没有结束引号"); break; } break; case 37: switch (c) { case '\"': state = 30;//复原,下一属性 jumpspace(); state = 15; break; case '>': throw new Exception("没有结束引号"); break; } break; case 50: sb.Remove(0,sb.Length); if (char.IsLetterOrDigit(c)) sb.Append(c); else throw new Exception("错误的结束字符"); while (sr.Peek() != -1) { tempc = (char)sr.Peek(); if (char.IsLetterOrDigit(tempc)) { sb.Append(tempc); c = (char)sr.Read(); } else { //Console.WriteLine(StID.Peek tempstr = StID.Peek(); if(StID.Pop()!=sb.ToString()) throw new Exception (string.Format( "结束标记 '{0}' 与开始标记 '{1}' 不匹配" ,sb.ToString() ,tempstr)); state =51;//处理结尾 ' ' > break; } } break; case 51: jumpspace(); switch (c) { case '>': state = 0;//普通状态 break; default: throw new Exception("名称包含无效字符"); } break; }