1、htmlCheck类
using System;
using System.Collections.Generic;
using System.Text;
using System.Collections;
using System.Diagnostics;
namespace htmlCheck
{
/// <summary>
/// Asp.Net检查HTML是否闭合以及自动修复 http://webapi.cnblogs.com/
/// </summary>
class TagsList
{
private ArrayList data; public int Size
{
get
{
return data.Count;
}
} public TagsList()
{
data = new ArrayList();
} public void add(String str)
{
data.Add(str);
} public string get(int index)
{
if (index < data.Count)
return (string)data[index];
else
return null;
} public bool remove(string str)
{
if (data.IndexOf(str) == -) return false;
data.Remove(str);
return true;
} public void remove(int index)
{
data.RemoveAt(index);
}
} public class TagsChecker
{
/// <summary>
/// 检查html标签是闭合
/// </summary>
/// <param name="str"></param>
/// <returns></returns>
public static bool check(string str)
{
TagsList[] unclosedTags = getUnclosedTags(str); if (unclosedTags[].Size != )
{
return false;
}
for (int i = ; i < unclosedTags[].Size; i++)
{
if (unclosedTags[].get(i) != null)
return false;
} return true;
}
/// <summary>
/// 处理未闭合的html代码
/// </summary>
/// <param name="str"></param>
/// <returns></returns>
public static string fix(String str)
{
StringBuilder fixeds = new StringBuilder(); // 存放修复后的字符串
TagsList[] unclosedTags = getUnclosedTags(str); // 生成新字符串
for (int i = unclosedTags[].Size - ; i > -; i--)
{
fixeds.Append("<" + unclosedTags[].get(i) + ">");
} fixeds.Append(str); for (int i = unclosedTags[].Size - ; i > -; i--)
{
String s = null;
if ((s = unclosedTags[].get(i)) != null)
{
fixeds.Append("</" + s + ">");
}
} return fixeds.ToString();
} private static TagsList[] getUnclosedTags(String str)
{
StringBuilder temp = new StringBuilder(); // 存放标签
TagsList[] unclosedTags = new TagsList[];
unclosedTags[] = new TagsList(); // 前不闭合，如有</div>而前面没有<div>
unclosedTags[] = new TagsList(); // 后不闭合，如有<div>而后面没有</div>
bool flag = false; // 记录双引号"或单引号'
char currentJump = ' '; // 记录需要跳过''还是"" char current = ' ', last = ' '; // 当前 & 上一个 // 开始判断
for (int i = ; i < str.Length; )
{
current = str[i++]; // 读取一个字符
if (current == '"' || current == '\'')
{
flag = flag ? false : true; // 若为引号，flag翻转
currentJump = current;
if (flag)
{
while (i < str.Length && str[i++] != currentJump)
; // 跳过引号之间的部分
flag = false;
}
}
else if (current == '<')
{ // 开始提取标签
current = str[i++];
if (current == '/')
{ // 标签的闭合部分，如</div>
current = str[i++]; // 读取标签
while (i < str.Length && current != '>')
{
temp.Append(current);
current = str[i++];
} // 从tags_bottom移除一个闭合的标签
if (!unclosedTags[].remove(temp.ToString()))
{ // 若移除失败，说明前面没有需要闭合的标签
unclosedTags[].add(temp.ToString()); // 此标签需要前闭合
}
temp.Remove(, temp.Length); // 清空temp
}
else
{ // 标签的前部分，如<div>
last = current;
while (i < str.Length && current != ' '
&& current != ' ' && current != '>')
{
temp.Append(current);
last = current;
current = str[i++];
} // 已经读取到标签，跳过其他内容，如<div id=test>跳过id=test
while (i < str.Length && current != '>')
{
last = current;
current = str[i++];
if (current == '"' || current == '\'')
{ // 判断双引号
flag = flag ? false : true;
currentJump = current;
if (flag)
{ // 若引号不闭合，跳过到下一个引号之间的内容
while (i < str.Length && str[i++] != currentJump)
;
current = str[i++];
flag = false;
}
}
}
if (last != '/' && current == '>') // 判断这种类型：<TagName />
unclosedTags[].add(temp.ToString());
temp.Remove(, temp.Length);
}
}
}
return unclosedTags;
}
}
}
2、使用方法
using System;
using System.Collections.Generic;
using System.Web;
using System.Web.UI;
using System.Web.UI.WebControls;
using htmlCheck;//注意，添加这个引用
using System.Text.RegularExpressions; public partial class ceshi : System.Web.UI.Page
{
protected void Page_Load(object sender, EventArgs e)
{
string str1 = "<p>tt<table><tr><td>webapi</td></tr><tr><td>323</td></p>";
string str2 = "tt<u>ss</u><div id=test name=\"<test>\"><a>fds</a></div>";
Response.Write("\r\n检查文本 " + str1);
Response.Write("\r\n结果：" + TagsChecker.check(str1));
Response.Write("\r\n检查文本 " + str2);
Response.Write("\r\n结果：" + TagsChecker.check(str2));
Response.Write("\r\n修复文本 " + str1);
Response.Write("\r\n结果：" + TagsChecker.fix(str1)); }
}
说明：转载的，已测试，确实闭合修复了，但是闭合的位置不正确