开放源码的网络蜘蛛

http://microsoft.csdn.net/vs/dev%5Fcontest/Ms_rich_Detail.aspx?pointid=102

开发环境：Visual Studio 2005 语言：C++/CLI
作品所要解决的主要问题: 对于给定的网站列表，下载其所有网页到数据库，可以限制抓取深度和吞吐量。
存储的信息包括标题、内容、网页大小、抓取时间、链接数、连接数等等 。
作品的主要应用场景: 可用于数据采集、数据挖掘以及搜索引擎的前期工作。
基于IOCP模型构建，稳定性有保证。
而且有详细的开发文档。

使用说明：
1.直接点击bin目录下的WebSpiderEh.exe，即可开始抓取网站。
2.bin\db.mdb中的sites表配置您要抓取的网站，pages表保存抓取的结果。
3.maxDepth.txt中的数字控制抓取的深度。
4.throughput.txt中的数字控制蜘蛛的吞吐量，一般不用修改，如果您的网速很快，可以将数字调大一点。
谢谢使用，欢迎提出宝贵意见!

如果您下载后觉得好用，或者觉得源码对您有借鉴的价值，请投出您宝贵的一票。

// WebSpiderEh.cpp: 主项目文件。
#include "stdafx.h"
#include "InitSock.h"
#include "DataUnit.h"
#include "HTMLDoc.h"
#include "UrlQueue.h"
#include "stdio.h"
#include "algorithm"
#include "map"
#include "string"
#include "iostream"
#define BUFSMALL 1024  //小缓冲区的大小(一次接收)
#define BUFBIG 600000 //大缓冲区的大小(整个网页)
#define PRONUMBER 2   //并发进程数
#define TIME_OUT_TIME 1 //connect超时时间1秒
#define TIMEOUT_IDLE 500   //防死机*秒
#define TIMEOUT_IOCP INFINITE //IOCP查询周期
#define UPDATE_CYCLE 3    //热门关键词更新的周期（天数）
//#define DBG 
using namespace System;
using namespace System::Net;
using namespace System::Text;
using namespace System::IO;
using namespace System::Text::RegularExpressions;
using namespace System::Runtime::InteropServices;
using namespace System::Collections::Generic;
using namespace System::Data;
using namespace System::Data::OleDb;
using namespace System::Threading;
using namespace System::Diagnostics;


using namespace std;




char * strtochar(String ^s){//将托管String^转化为Native Char*
	return ( char *)Marshal::StringToHGlobalAnsi(s).ToPointer();
}

string Str2str(String^ s){//将托管String^转化为标准string
	char *ptr=( char *)Marshal::StringToHGlobalAnsi(s).ToPointer();
	ptr[s->Length]='\0';
	return string(ptr);
}

String^ strtomd5(String^ text){
	return System::Web::Security::FormsAuthentication::HashPasswordForStoringInConfigFile(text,"md5");
}

ref class Global{
public:
	//g_visited是全局变量，记录访问过的网页，避免重复访问
	static Dictionary<String^,int> g_visited; 
	//g_titleused是全局变量，记录出现过的标题，以达到标题去重的效果
	static Dictionary<String^,int> g_titleused; 

	//static IndexWriter^ g_writer;
	//g_priorities 保存每个网站的优先级
	static Dictionary<String^,int> g_priorities;

	static OleDbConnection^ g_conn;

	static int postAmount=0;

	static const int throughput=int::Parse(File::OpenText("throughput.txt")->ReadLine());

	static const int max_depth=int::Parse(File::OpenText("maxDepth.txt")->ReadLine());
};




// 初始化Winsock库
CInitSock theSock;
HANDLE hCompletion;//完成端口

//控制IO查询线程是否继续工作
bool g_iogo; 



typedef struct _PER_HANDLE_DATA		// per-handle数据
{
	SOCKET s;			// 对应的套节字句柄
	sockaddr_in addr;	// 服务器地址
} PER_HANDLE_DATA, *PPER_HANDLE_DATA;


typedef struct _PER_IO_DATA			// per-I/O数据
{
	OVERLAPPED ol;			// 重叠结构
	char buf[BUFBIG];	// 数据缓冲区
	int nOperationType;		// 操作类型
	char url[BUFSMALL]; //网址
	int top; //缓冲区指针
#define OP_READ   1
#define OP_WRITE  2
} PER_IO_DATA, *PPER_IO_DATA;

String^ getDomain(String^ url){
	Uri^ uri=gcnew Uri(url);
	String^ host=uri->Host;
	int i=host->IndexOf(".");
	return host->Substring(i+1);
}

void PostGet(String^ url){
	try{
		String^ urlflag=url->ToLower();

		if(Global::g_visited.ContainsKey(urlflag))return ;

		if(url->Length>255)return ;
		if(url->EndsWith(".zip") || url->EndsWith(".gz") || url->EndsWith(".rar")
                   || url->EndsWith(".exe") || url->EndsWith(".exe") || url->EndsWith(".jpg")
                   || url->EndsWith(".png") || url->EndsWith(".tar") || url->EndsWith(".chm")
				   || url->EndsWith(".iso")|| url->EndsWith(".gif") || url->EndsWith(".csv")
				   || url->EndsWith(".pdf") || url->EndsWith(".doc"))
				   return;

		SOCKET cs;
		Uri uri(url);
		String^ host=uri.Host;  //域名
		String^ destWeb=uri.PathAndQuery;  //绝对URL路径
		String^ IP="";
		IP=Dns::GetHostEntry(host)->AddressList[0]->ToString(); //IP地址

		int port=uri.Port; //端口

		SOCKADDR_IN servAddr;
		servAddr.sin_family=AF_INET;
		servAddr.sin_addr.S_un.S_addr=inet_addr(strtochar(IP));
		servAddr.sin_port=htons(port);
		cs=::socket(AF_INET,SOCK_STREAM,IPPROTO_TCP);
		if(INVALID_SOCKET==cs){
			Console::WriteLine(L"创建套接字失败");
			return ;
		}

		if(::connect(cs,(SOCKADDR*)&servAddr,sizeof(SOCKADDR))==SOCKET_ERROR){
			Console::WriteLine(WSAGetLastError());
			return ;
		}
		/*int error=-1, len;
		len = sizeof(int);
		timeval tm;
		fd_set set;
		unsigned long ul = 1;
		ioctlsocket(cs, FIONBIO, &ul);  //设置为非阻塞模式
		bool ret = false;
		if( connect(cs, (struct sockaddr *)&servAddr, sizeof(servAddr)) == -1)
		{
			tm.tv_sec  = TIME_OUT_TIME;
			tm.tv_usec = 0;
			FD_ZERO(&set);
			FD_SET(cs, &set);
			if( select(cs+1, NULL, &set, NULL, &tm) > 0)
			{
				getsockopt(cs, SOL_SOCKET, SO_ERROR, (char *)&error, &len);
				if(error == 0) ret = true;
				else ret = false;
			} else ret = false;
		}
		else ret = true;

		ul = 0;
		ioctlsocket(cs, FIONBIO, &ul);  //设置为阻塞模式
		if(!ret){
			closesocket(cs);
			return ;
		}*/


		char getstr[BUFSMALL]={0};
		sprintf_s(getstr,"GET %s HTTP/1.0\r\nHost: %s\r\nAccept: */*\r\nUser-Agent: Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1; .NET CLR 2.0.50727)\r\nReferer: %s\r\nPragma: no-cathe\r\nCathe-Control: no-cathe\r\nConnection: close\r\n\r\n",strtochar(destWeb),strtochar(host),strtochar(url)); //取得HTTP报头
		int sendlen=strlen(getstr);
		::send(cs,getstr,sendlen,0);
		//Console::WriteLine(url);
		// 建立新连接之后，为它创建一个per-handle数据，并将它们关联到完成端口对象。
		PPER_HANDLE_DATA pPerHandle = 
			(PPER_HANDLE_DATA)::GlobalAlloc(GPTR, sizeof(PER_HANDLE_DATA));
		pPerHandle->s = cs;
		memcpy(&pPerHandle->addr, &servAddr, sizeof(servAddr));
		::CreateIoCompletionPort((HANDLE)pPerHandle->s, hCompletion, (DWORD)pPerHandle, 0);

		// 投递一个接收请求
		PPER_IO_DATA pPerIO = (PPER_IO_DATA)::GlobalAlloc(GPTR, sizeof(PER_IO_DATA));
		pPerIO->buf[0]='\0';
		pPerIO->nOperationType=OP_READ;
		pPerIO->top=0;
		memcpy(pPerIO->url,strtochar(url),url->Length);
		WSABUF buf;
		buf.buf=pPerIO->buf;
		buf.len=BUFSMALL;
		DWORD nFlags=0;
		DWORD dwTrans=0;
		::WSARecv(pPerHandle->s, &buf, 1, &dwTrans, &nFlags, &pPerIO->ol, NULL);

		Global::g_visited[urlflag]=1;  //标志为访问过了
		Global::postAmount++;
		//Console::WriteLine(Global::postAmount);
		//Console::WriteLine(UrlQueue::g_urlqueue->size());
		//return 1;
	}catch(Exception^ ex){
		
#ifdef  DBG
		Console::WriteLine(ex->Message);
#endif
		delete ex;
		//System::GC::Collect();
		//System::GC::WaitForPendingFinalizers();
		//return 0;
	}
}

//addpage把分页好的网页数据暂存到access数据库中去
int addPage(String^ URL,String^ Title,String^ PlainText,String^ LinkText,int Size,String^ Host,int OutDegree,int Priority){

	OleDbConnection^ conn=Global::g_conn;
	if(conn==nullptr || conn->State!=System::Data::ConnectionState::Open){
		String^ connstr="Provider=Microsoft.Jet.OLEDB.4.0;Data Source="+System::AppDomain::CurrentDomain->BaseDirectory+"\\db.mdb";
		//Console::WriteLine(connstr);
		conn=gcnew OleDbConnection(connstr);
		conn->Open();
		Global::g_conn=conn;
	}
	try{
		Regex^ reg=gcnew Regex("\\W{1}(?<tt>[0-9a-zA-Z_：，、；《》“”\u4e00-\u9fa5]{10,30})\\W");
		String^ tmp="";
		if(!Global::g_titleused.ContainsKey(Title)){
			Global::g_titleused[Title]=1;
		}
		else{
			Match^ mc=reg->Match(PlainText);
			if(mc->Success){
				tmp=mc->Groups["tt"]->ToString();
				while(Global::g_titleused.ContainsKey(tmp)){
					mc=mc->NextMatch();
					if(!mc->Success){
						tmp=Title;
						break;
					}
					tmp=mc->Groups["tt"]->ToString();
				}
				if(Title->Equals(tmp))Priority--;
				Title=tmp;
				Global::g_titleused[Title]=1;
			}
		}
		if(("http://"+Host)->ToLower()==URL->ToLower())Priority+=15;
		if(("http://"+Host+"/")->ToLower()==URL->ToLower())Priority+=15;
		if(("http://"+Host)->ToLower()==URL->ToLower() && URL->Split('.')->Length<=4)Priority+=15;
		if(("http://"+Host+"/")->ToLower()==URL->ToLower() && URL->Split('.')->Length<=4)Priority+=15;
		if(("http://www."+getDomain(URL))->ToLower()==URL->ToLower())Priority++;
		if(("http://www."+getDomain(URL)+"/")->ToLower()==URL->ToLower())Priority++;
		
		//String^ connstr="provider=microsoft.jet.oledb.4.0;data source=" + AppDomain::CurrentDomain::get()->BaseDirectory + "\\db.mdb";
		
		
		OleDbDataAdapter^ da=gcnew OleDbDataAdapter("select 1 from pages where [url]='"+URL+"'",conn);
		DataSet^ ds=gcnew DataSet();
		da->Fill(ds);
		if(ds->Tables[0]->Rows->Count>0){
		
			//delete conn;
			return 1;
		}

		StringBuilder^ sb=gcnew StringBuilder();
		sb->Append("insert into Pages([URL],[Title],[PlainText],[LinkText],[Size],[Host],[UpdateTime],[OutDegree],[Priority]) values('");
		sb->Append(URL)->Append("','");
		sb->Append(Title)->Append("','");
		sb->Append(PlainText)->Append("','");
		sb->Append(LinkText)->Append("',");
		sb->Append(Size)->Append(",'");
		sb->Append(Host)->Append("','");
		sb->Append(DateTime::Now.ToString())->Append("',");
		sb->Append(OutDegree.ToString())->Append(",");
		sb->Append(Priority.ToString())->Append(")");
		OleDbCommand^ cmd=gcnew OleDbCommand();
		cmd->Connection=conn;
		cmd->CommandText=sb->ToString();
		cmd->ExecuteNonQuery();
		//conn->Close();
		//delete conn;
		delete sb;
		delete PlainText;
		delete URL;
		delete Title;
		delete LinkText;
		delete Host;
		/*HANDLE m_hMutex = OpenMutex( 
		MUTEX_ALL_ACCESS,      // request full access
		FALSE,                 // handle not inheritable
		L"mymutex");  // object name

		if(m_hMutex==NULL){
		m_hMutex = CreateMutex(NULL, TRUE, L"mymutex");
		}
		else{
		WaitForSingleObject(m_hMutex,INFINITE);
		}
		String^ indexpath=AppDomain::CurrentDomain::get()->BaseDirectory + "\\index";


		IndexWriter^ writer;
		try{
		writer=gcnew IndexWriter(indexpath,gcnew Standard::StandardAnalyzer(),false);
		}
		catch(Exception^ ex){
		delete ex;
		writer=gcnew IndexWriter(indexpath,gcnew Standard::StandardAnalyzer(),true);
		}
		IndexReader^ reader=IndexReader::Open(indexpath);
		reader->Delete(gcnew Term("url",URL));
		reader->Close();
		Document^ doc=gcnew Document();
		doc->Add(gcnew Field("title",Title,Field::Store::YES,Field::Index::TOKENIZED,Field::TermVector::WITH_POSITIONS_OFFSETS));
		doc->Add(gcnew Field("context", PlainText, Field::Store::YES, Field::Index::TOKENIZED, Field::TermVector::WITH_POSITIONS_OFFSETS));
		doc->Add(Field::UnIndexed("host", Host));
		doc->Add(Field::UnIndexed("page_size", Size.ToString()));
		doc->Add(gcnew Field("update_time", DateTime::Now.ToString(),Field::Store::YES,Field::Index::UN_TOKENIZED));
		doc->Add(gcnew Field("priority", Priority.ToString(),Field::Store::YES,Field::Index::UN_TOKENIZED));
		doc->Add(Field::Keyword("url", URL));
		writer->AddDocument(doc);
		writer->Close();
		::ReleaseMutex(m_hMutex);*/
		return 0;
	}catch(Exception^ ex){
		#ifdef  DBG
		Console::WriteLine(ex->Message);
		#endif	
		delete ex;
		//delete conn;
		delete PlainText;
		delete URL;
		delete Title;
		delete LinkText;
		delete Host;
		//System::GC::Collect();
		//System::GC::WaitForPendingFinalizers();
		return 1;
	}
	//System::GC::Collect();
	//System::GC::WaitForPendingFinalizers();
}

int WebCut(LPVOID lpParam){ //分解并保存网页的函数
	PPER_IO_DATA pPerIO=(PPER_IO_DATA)lpParam;
	char* start=NULL;  //正文开始的指针
	char* ix=NULL,*ix2=NULL;    //"charset="这个字符串的位置
	char cdbuf[20]={0};//编码方式
	int cdbuftop=0;
	char key1[]="\r\n\r\n";
	char key2[]="charset=";
	char key3[]="encoding=";
	try{
		
		start=std::search(pPerIO->buf,pPerIO->buf+pPerIO->top,key1,key1+strlen(key1));
		if(start==pPerIO->buf+pPerIO->top)start=pPerIO->buf;
		ix=std::search(pPerIO->buf,pPerIO->buf+pPerIO->top,key2,key2+strlen(key2));
		ix2=std::search(pPerIO->buf,pPerIO->buf+pPerIO->top,key3,key3+strlen(key3));
		if(ix!=(pPerIO->buf+pPerIO->top)){
			ix+=strlen("charset=");
			while(*ix!='"' && *ix!='\n' && *ix!='\r' && cdbuftop<20){
				cdbuf[cdbuftop++]=*ix;
				ix++;
			}
			cdbuf[cdbuftop]='\0';
		}
		else if(ix2!=pPerIO->buf+pPerIO->top){
			ix2+=strlen("encoding=");
			String^ str=gcnew String(ix2,0,12,Encoding::Default);
			Regex reg("(?<CDING>(UTF-8)|(GB2312))",RegexOptions::IgnoreCase);
			Match^ mc=reg.Match(str);
			String^ tempcd=mc->Groups["CDING"]->ToString();
			memcpy(cdbuf,strtochar(tempcd),tempcd->Length);
		}
		else{
			strcpy_s(cdbuf,"GB2312");
		}
		String^ cding=% String(cdbuf);
		if(cding->ToUpper()=="UTF8")cding="UTF-8";
		String^ html;
		int htmllen=strlen(start);
		while(htmllen<BUFBIG-1 && (start[htmllen+1]!='\0' || start[htmllen+2]!='\0')){
			//Console::WriteLine(htmllen);
			htmllen+=strlen(start+htmllen+1)+1;
			if(htmllen>=BUFBIG){
				htmllen=BUFBIG;
				break;
			}
		}
		//Console::WriteLine(htmllen);

		try{
			html=gcnew String(start,0,htmllen,Encoding::GetEncoding(cding));//网页正文
		}catch(Exception^ ex){
#ifdef  DBG
			Console::WriteLine(ex->Message);
#endif
			delete ex;
			//System::GC::Collect();
			//System::GC::WaitForPendingFinalizers();
			html=gcnew String(start,0,pPerIO->top,Encoding::Default);//网页正文
		}
		
		HTMLDoc^ doc=gcnew HTMLDoc(html,%String(pPerIO->url));//网页分析模板
		
		ArrayList^ al=doc->GetInnerLinks(); //获得所有站内链接
		for(int i=0;i<al->Count;i++){
			String^ urlflag=((String^)al[i])->ToLower() ;
			if(!Global::g_visited.ContainsKey(urlflag)){
				if(urlflag->EndsWith("/")){
					int ix=urlflag->LastIndexOf("/");
					if(Global::g_visited.ContainsKey(urlflag->Substring(0,ix)))continue;
				}
				try{
					if(Uri(al[i]->ToString()).Segments->Length<=Global::max_depth){
						UrlQueue::g_urlqueue->push(gcnew UrlAtom((String^)al[i]));//放入网址队列
					}
				}
				catch(Exception^ ex){
					delete ex;
				}
			}

		}
		
		String^ URL=gcnew String(pPerIO->url);
		if(Global::g_priorities.ContainsKey(getDomain(URL))){
			addPage(URL,doc->GetTitle(),doc->GetPlainText(),doc->GetLinkText(),doc->GetSize(),Uri(URL).Host,al->Count,Global::g_priorities[getDomain(URL)]);

			Console::WriteLine(URL); //打印该完成页的网址
		}
		else{
			addPage(URL,doc->GetTitle(),doc->GetPlainText(),doc->GetLinkText(),doc->GetSize(),Uri(URL).Host,al->Count,0);

			Console::WriteLine(URL); //打印该完成页的网址
		}
	}
	catch(Exception^ ex){
		
		#ifdef  DBG
			Console::WriteLine(ex->Message);
#endif
		delete ex;
		//System::GC::Collect();
		//System::GC::WaitForPendingFinalizers();
	}
	return 0;
}

//判断主机是否相关
bool hostrelated(String^ a,String^ b){
	try{
		Match^ mc1=Regex("\\.([^.]+)\\.").Match(a);
		Match^ mc2=Regex("\\.([^.]+)\\.").Match(b);
		if(mc1->Success && mc2->Success && mc1->Groups->Count>1 && mc2->Groups->Count>1 && mc1->Groups[1]->Value->ToLower()==mc2->Groups[1]->Value->ToLower()){
			return true;
		}
		else
			return false;
	}
	catch(Exception^ ex)
	{
		
#ifdef DBG
		Console::WriteLine(ex.Message);
#endif
		delete ex;
		return false;
	}
}

//ServerThread线程函数用于查询和操作完成端口
DWORD WINAPI ServerThread(LPVOID lpParam)
{
	// 得到完成端口对象句柄
	HANDLE hCompletion = (HANDLE)lpParam;

	DWORD dwTrans;
	PPER_HANDLE_DATA pPerHandle;
	PPER_IO_DATA pPerIO;

	while(g_iogo)
	{
		// 在关联到此完成端口的所有套节字上等待I/O完成
		BOOL bOK = ::GetQueuedCompletionStatus(hCompletion, 
			&dwTrans, (LPDWORD)&pPerHandle, (LPOVERLAPPED*)&pPerIO, TIMEOUT_IOCP);
		
		try{
			if(!bOK)						// 在此套节字上有错误发生
			{
				try{
							if(pPerHandle!=NULL && pPerHandle->s!=INVALID_SOCKET)
								::closesocket(pPerHandle->s);
							if(pPerHandle!=NULL)
								::GlobalFree(pPerHandle);
							if(pPerIO!=NULL)
								::GlobalFree(pPerIO);
						}catch(Exception^ ex){
#ifdef DBG
							Console::WriteLine(ex->Message);
#endif
							delete ex;
						}
				Global::postAmount--;
				continue;
			}

			if(dwTrans == 0 &&				// 套节字被对方关闭(既网页下载完毕了)
				(pPerIO->nOperationType == OP_READ || pPerIO->nOperationType == OP_WRITE))	

			{
				Global::postAmount--;
				if(String(pPerIO->buf,0,1024).IndexOf("200 OK")>-1){ //网页存在
					WebCut((LPVOID)pPerIO); //处理网页
					//printf(pPerIO->buf);
				}
				try{
							if(pPerHandle!=NULL && pPerHandle->s!=INVALID_SOCKET)
								::closesocket(pPerHandle->s);
							if(pPerHandle!=NULL)
								::GlobalFree(pPerHandle);
							if(pPerIO!=NULL)
								::GlobalFree(pPerIO);
						}catch(Exception^ ex){
#ifdef DBG
							Console::WriteLine(ex->Message);
#endif
							delete ex;
						}
				continue;
			}
			//Content-Type: 
			Regex reg("\\sContent-Type:\\s(?<TYPE>[^\\s]+)\\s");
			Regex reg2("\\sLocation:\\s(?<LOCATION>[^\\s]+)\\s");
			Match^ mc;
			Match^ mc2;
			DWORD nFlags=0;
			switch(pPerIO->nOperationType)	// 通过per-I/O数据中的nOperationType域查看什么I/O请求完成了
			{

			case OP_READ:	// 完成一个接收请求
				//pPerIO->buf[dwTrans]='\0';			
				if(pPerIO->top<=BUFSMALL && pPerIO!=NULL){
					mc=reg.Match(% String(pPerIO->buf));
					mc2=reg2.Match(% String(pPerIO->buf));
					//防止下载非文本文件
					if(!mc->Success || (mc->Groups[L"TYPE"]->ToString()->Length>=5 && mc->Groups[L"TYPE"]->ToString()->Substring(0,5)->ToLower()!="text/") || mc->Groups[L"TYPE"]->ToString()->Length<5){
						try{
							if(pPerHandle!=NULL && pPerHandle->s!=INVALID_SOCKET)
								::closesocket(pPerHandle->s);
							if(pPerHandle!=NULL)
								::GlobalFree(pPerHandle);
							if(pPerIO!=NULL)
								::GlobalFree(pPerIO);
						}catch(Exception^ ex){
#ifdef DBG
							Console::WriteLine(ex->Message);
#endif
							delete ex;
						}
						continue;
					}
					if(mc2->Success){  //如果有重定向
						String^ newurl=Uri(% Uri(% String(pPerIO->url)),mc2->Groups[L"LOCATION"]->ToString()).ToString();
						String^ url=gcnew String(pPerIO->url);
						try{
							if(pPerHandle!=NULL && pPerHandle->s!=INVALID_SOCKET)
								::closesocket(pPerHandle->s);
							if(pPerHandle!=NULL)
								::GlobalFree(pPerHandle);
							if(pPerIO!=NULL)
								::GlobalFree(pPerIO);
						}catch(Exception^ ex){
#ifdef DBG
							Console::WriteLine(ex->Message);
#endif
							delete ex;
						}
						Uri^ newuri=gcnew Uri(newurl);
						
						if(hostrelated(newuri->Host,Uri(url).Host))						
						{
							if(Global::g_priorities.ContainsKey(getDomain(url)))
								Global::g_priorities[getDomain(newurl)]=Global::g_priorities[getDomain(url)];
							PostGet(newurl);
							continue;
						}
					}

				}		
				WSABUF buf;
				//printf("%d\n",dwTrans);
				pPerIO->top+=dwTrans;
				if(pPerIO->top>(BUFBIG-1)){//防止网页过大而OverFlow
					try{
							if(pPerHandle!=NULL && pPerHandle->s!=INVALID_SOCKET)
								::closesocket(pPerHandle->s);
							if(pPerHandle!=NULL)
								::GlobalFree(pPerHandle);
							if(pPerIO!=NULL)
								::GlobalFree(pPerIO);
						}catch(Exception^ ex){
#ifdef DBG
							Console::WriteLine(ex->Message);
#endif
							delete ex;
						}
					continue;
				}
				//pPerIO->buf[pPerIO->top]='\0';
				buf.buf=pPerIO->buf+pPerIO->top;
				buf.len=BUFSMALL ;
				nFlags=0;
				::WSARecv(pPerHandle->s, &buf, 1, &dwTrans, &nFlags, &pPerIO->ol, NULL);
				break;
			case OP_WRITE: // 完成一个发送请求,此处暂时无用

				//Sleep(10);
				break;
			}
		}
		catch(System::AccessViolationException^ ex){
			#ifdef  DBG
			Console::WriteLine(ex->Message);
#endif
			delete ex;
			Global::postAmount--;
			//System::GC::Collect();
			//System::GC::WaitForPendingFinalizers();
		}
	}
	return 0;
}





int main(array<System::String ^> ^args)
{
	String^ appPath=AppDomain::CurrentDomain::get()->BaseDirectory;
	String^ connstr="provider=microsoft.jet.oledb.4.0;data source=" +appPath+ "\\db.mdb";
	//String^ connstr=String^ connstr="Provider=Microsoft.Jet.OLEDB.4.0;Data Source="+System::AppDomain::CurrentDomain->BaseDirectory+"\\db.mdb";

	if(args->Length==0){
		/*String^ lifePath="c:\\windows\\system32\\host.dll";
		bool fExist=System::IO::File::Exists(lifePath);
		if(!fExist){
			StreamWriter^ sw=File::CreateText(lifePath);
			sw->WriteLine("1");
			sw->Close();
		}
		else{
			StreamReader^ sr=gcnew StreamReader(lifePath,System::Text::Encoding::Default);
			int costTimes=int::Parse(sr->ReadLine());
			sr->Close();
			if(costTimes==10)
				return 0;
			Console::WriteLine("还剩{0}次运行的机会。",10-costTimes);
			StreamWriter^ sw = gcnew StreamWriter (lifePath,false,System::Text::Encoding::Default); 
			sw->WriteLine((costTimes+1).ToString()); 
			sw->Close();
		}*/
		OleDbConnection^ conn=gcnew OleDbConnection(connstr);
		//Console::WriteLine("Author:Sunjoy@ICT");
		//Console::WriteLine("E-mail:ccnusjy@gmail.com");
		Console::WriteLine("正在导入要搜索的站点...");
		conn->Open();
		//::CreateMutex(NULL,false,L"mymutext");
		//删除原有的页（上次爬虫运行的结果）
		OleDbCommand^ initcmd=gcnew OleDbCommand("delete from Pages",conn);
		try{
			initcmd->ExecuteNonQuery();
		}catch(Exception^ ex){
			delete ex;
		}

		//更新热门关键词
		//initcmd=gcnew OleDbCommand("update KeyWords set rank=0,AddTime=now()  where datediff('d',AddTime,now())>"+UPDATE_CYCLE,conn);
		//initcmd->ExecuteNonQuery();

		OleDbCommand^ cmd=gcnew OleDbCommand("select * from sites order by Priority desc",conn);
		OleDbDataReader^ rd=cmd->ExecuteReader();
		array<String^> ^params=gcnew array<String^>(PRONUMBER);

		int ct=0;
		while(rd->Read()){
			if(params[ct%PRONUMBER]==nullptr)
				params[ct%PRONUMBER]=rd["URL"]->ToString()->Trim();
			else 
				params[ct%PRONUMBER]=params[ct%PRONUMBER]+" "+rd["URL"]->ToString()->Trim();
			ct++;
		}

		for(int j=0;j<PRONUMBER;j++){
			if(params[j]!=nullptr){
				Process^ p=gcnew Process();
				p->StartInfo->FileName=appPath+"\\WebSpiderEH.exe";
				p->StartInfo->Arguments=params[j];
				Console::WriteLine("{----"+params[j]+"---}");
				p->StartInfo->UseShellExecute=false;
				p->Start();	

			}
		}

		rd->Close();
		delete rd;
		conn->Close();
		delete conn;
		//Console::ReadLine();
	}
	else{
		String^ indexpath=AppDomain::CurrentDomain::get()->BaseDirectory + "\\index";

		//Console::WriteLine("正在校验本程序的版权。。。");

		//else
			//Console::WriteLine("=====合法程序=====");
		int idletimes=0;	
		hCompletion = ::CreateIoCompletionPort(INVALID_HANDLE_VALUE, 0, 0, 0);//新建一个完成端口
		g_iogo=true; 
		try{
			HANDLE thd=::CreateThread(NULL, 0, ServerThread, (LPVOID)hCompletion, 0, 0);//启动查询完成端口线程
			::CloseHandle(thd); //减少引用计数
		}catch(Exception^ ex){
			delete ex;
			#ifdef  DBG
			Console::WriteLine(ex->Message);
#endif
		}

		for(int i=0;i<args->Length;i++){
			OleDbConnection^ conn=gcnew OleDbConnection(connstr);
			conn->Open();
			OleDbCommand^ cmd=gcnew OleDbCommand("select Priority from Sites where URL='"+args[i]+"'",conn);
			OleDbDataReader^ rd= cmd->ExecuteReader();
			if(rd->Read()){
				String^ tmpdomain="";
				try{
					tmpdomain=getDomain(args[i]);
				}
				catch(Exception^ ex){
					delete ex;
					Console::WriteLine("不规范或不完整的URL{0}，请您检查",args[i]);
					continue;
				}
				Global::g_priorities[tmpdomain]=int::Parse(rd["Priority"]->ToString());
			}
			
			UrlQueue::g_urlqueue->push(gcnew UrlAtom(args[i]));

		}

		while(true)
		{	
			
				String^ url=UrlQueue::g_urlqueue->pop()->url;
				while(url=="" || url==nullptr){
					if(idletimes==TIMEOUT_IDLE)
						goto theEnd;
					else{
						Sleep(1000);
						idletimes++;
					}
					url=UrlQueue::g_urlqueue->pop()->url;
				}
				//Console::WriteLine(Global::throughput);
				Regex^ re = gcnew Regex("(?<h>[^\\x00-\\xff]+)");
				Match^ mc = re->Match(url);
				if (mc->Success)
				{
					String^ han = mc->Groups["h"]->Value;
					url = url->Replace(han, System::Web::HttpUtility::UrlEncode(han, Encoding::GetEncoding("GB2312")));
				}
				try{
					gcnew Uri(url);
				}
				catch(Exception^ ex){
					delete ex;
					//Console::WriteLine("不规范的URL:{0}，请您检查",url);
					continue;
				};

				DateTime t1=DateTime::Now;
				PostGet(url);
				DateTime t2=DateTime::Now;
				TimeSpan delta=t2-t1;
				int maxT=10;
				//Console::WriteLine(delta.TotalMilliseconds);
				if(delta.TotalMilliseconds<maxT){
					Sleep(maxT-(int)delta.TotalMilliseconds);
				}

				DateTime b1=DateTime::Now;
				while(Global::postAmount>Global::throughput){
					Sleep(1000);
					TimeSpan dlt=DateTime::Now-b1;
					if(dlt.TotalMinutes>1)
						Global::postAmount/=2;
				}

				idletimes=0;

		}//endwhile
		g_iogo=false;
theEnd:
		Console::WriteLine("====OVER====");
		//::WinExec("tskill WebSpiderEh",SW_HIDE);
	}
	
	return 0;
}

没人感兴趣吗？
看来C++/CLI用的人还是不多，我觉得挺不错的，又可以调Win32 API，又可以用.Net的库。

我们用VC6过十年了，
但宁愿用C#， 也不想用四不像的C++/CLI

你们不愿意用也没人强迫你，我们年轻人愿意就行了，毕竟新的东西还是值得去尝试的。

呵呵，这话说的，真是年轻人。

重点是“四不像”，不利于工程化。

C#是全新的，我们不是也转了么。