在c#中使用正則表達式進行匹配,有時候我們會遇到這種情況,cpu使用率100%,但是正則表達式并沒有異常拋出,正則一直處于匹配過程中,這將導致系統資源被耗盡,應用程序被卡住,這是由于正則不完全匹配,而且Regex中沒有Timeout屬性,使正則處理器陷入了死循環。
這種情況尤其可能發生在對非可靠的被匹配對象的匹配過程中,例如在我的個人網站www.eahan.com項目中,對多個網站頁面的自動采集匹配,就經常發生該問題。為了避免資源耗盡的情況發生,我寫了一個AsynchronousRegex類,顧名思義,異步的Regex。給該類一個設置一個Timeout屬性,將Regex匹配的動作置于單獨的線程中,AsynchronousRegex監控Regex匹配超過Timeout限定時銷毀線程。
using System;
using System.Text.RegularExPRessions;
using System.Threading;
namespace LZT.Eahan.Common
{
public class AsynchronousRegex
{
private MatchCollection mc;
private int _timeout; // 最長休眠時間(超時),毫秒
private int sleepCounter;
private int sleepInterval; // 休眠間隔,毫秒
private bool _isTimeout;
public bool IsTimeout
{
get {return this._isTimeout;}
}
public AsynchronousRegex(int timeout)
{
this._timeout = timeout;
this.sleepCounter = 0;
this.sleepInterval = 100;
this._isTimeout = false;
this.mc = null;
}
public MatchCollection Matchs(Regex regex, string input)
{
Reg r = new Reg(regex, input);
r.OnMatchComplete += new Reg.MatchCompleteHandler(this.MatchCompleteHandler);
Thread t = new Thread(new ThreadStart(r.Matchs));
t.Start();
this.Sleep(t);
t = null;
return mc;
}
private void Sleep(Thread t)
{
if (t != null && t.IsAlive)
{
Thread.Sleep(TimeSpan.FromMilliseconds(this.sleepInterval));
this.sleepCounter ++;
if (this.sleepCounter * this.sleepInterval >= this._timeout)
{
t.Abort();
this._isTimeout = true;
}
else
{
this.Sleep(t);
}
}
}
private void MatchCompleteHandler(MatchCollection mc)
{
this.mc = mc;
}
class Reg
{
internal delegate void MatchCompleteHandler(MatchCollection mc);
internal event MatchCompleteHandler OnMatchComplete;
public Reg(Regex regex, string input)
{
this._regex = regex;
this._input = input;
}
private string _input;
public string Input
{
get {return this._input;}
set {this._input = value;}
}
private Regex _regex;
public Regex Regex
{
get {return this._regex;}
set {this._regex = value;}
}
internal void Matchs()
{
MatchCollection mc = this._regex.Matches(this._input);
if (mc != null && mc.Count > 0) // 這里有可能造成cpu資源耗盡
{
this.OnMatchComplete(mc);
}
}
}
}
}
新聞熱點
疑難解答