头文字D,Spider

贴一段我用D写的Spider的代码。
越来越爱D了。

[code]
import std.stdio;
import std.string;
import std.conv;
import std.socket;
import std.socketstream;
import std.stream;
import std.regexp;
import std.thread;
import std.c.time;

char[][] g_queue;
int g_task_amount=0;
const char[] homepage="http://mobile.younet.com/";
const ushort max_thread=20;

alias std.string.find strfind;

char[] getHTML(char[] url){

char[] domain,html;
ubyte[1024] buf;
ushort port=80;
if(!isURL(url))return null;
int i=strfind(url,"://")+3;
url=url[i..$];
int j=strfind(url,":");
int e=strfind(url,"/");
if(e<0){
e=url.length;
}
if(j>0){
port=toUshort(url[j+1..e]);
domain=url[0..j];
}
else{
domain=url[0..e];
}

if(e==url.length){
url="/";
}
else{
url=url[e..$];
}

debug(younet){
writefln(toString(port) ~" "~ domain ~" "~ url);
}

Socket sock=new TcpSocket(new InternetAddress(domain,port));
Stream ss=new SocketStream(sock);
ss.writeString("GET " ~ url ~ " HTTP/1.0\r\n"
"Host: " ~ domain ~ "\r\n"
"Connection: close\r\n"
"Referer: http://" ~ domain ~ url ~ "\r\n"
"\r\n\r\n\r\n\r\n");
int recv_amount=ss.read(buf);
while(recv_amount>0){
html ~= cast(char[])buf[0..recv_amount];
recv_amount=ss.read(buf);
}

ss.close();
sock.close();

char[][] mc=RegExp("(URL=|Location: )(.*?)[\"\r]").match(html);
if(mc.length==3){
char[] new_location=mc[2];
html=getHTML(new_location);
return html;
}
int start_pos=strfind(html,"\r\n\r\n") ;

html=html[start_pos+4 .. $];
return html[0..$];
}

int crawl(void * p){
while(true){
char[] url,html;
synchronized{
if(g_queue.length==0)
sleep(1);
if(g_queue.length==0)
break;
url=g_queue[0];

writefln("begin:" ~ url);
if(g_queue.length>0)
g_queue=g_queue[1..$];
}
try{
html=getHTML(url);
}
catch(Exception ex){
synchronized{
if(g_task_amount>0)
g_task_amount-=1;
}
writefln(ex);
writefln("failed:" ~ url);
writefln("remains" ~ toString(g_task_amount));
continue;
}
debug(younet){
printf(toStringz("!!!" ~ html[0..200]));
}
if(strfind(url,"files/list")<0){
synchronized{
foreach(m;RegExp("files/list_\\d+\\.html").search(html)){
g_queue ~= homepage ~ m.match(0);
g_task_amount+=1;
}
g_task_amount-=1;
}
writefln("done:" ~ url);
debug(younet){
writefln(g_queue);
}
}
else{
writefln("done:" ~ url);
synchronized{
g_task_amount-=1;
writefln("remains" ~ toString(g_task_amount));
}
}

}
return 1;
}
int main(char[][] args){
//writefln("Hello");
g_queue ~= homepage;
g_task_amount+=1;
Thread[] tds;
for(int i=0;i<max_thread;i++){
Thread t=new Thread(&crawl,null);
t.start();
tds ~= t;
}
sleep(5);
while(true){
sleep(1);
if(g_task_amount<=0)break;
}
return 0;
}
[/code]
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值