php测试链接是否为404页面

帝国后台生成网站的sitemap，不知道其中是否有死链接，于是写了php脚本测试下链接的情况

<?php 

//从txt中获取图片地址和采集页网址
function getUrls($fileName){
	$lineArray=array();
	$fp = fopen($fileName, "r") or die("Unable to open file ".$fileName);
	
	while(! feof($fp)){
		$line=fgets($fp);
		$lineArr[]=trim($line);

	}

	fclose($fp);
	return $lineArr;

}


$urls=getUrls('urls.txt');
foreach ($urls as $url) {
	
	$headers = get_headers($url);  
	//print_r($headers); 
	if (strpos($headers[0], '404')){
		echo $url."-------->404".PHP_EOL;
	}else{  
	 	echo $url.'-------->正常访问'.PHP_EOL;  
	}  
}




 ?>

测试结果如下

测试页面是否能打开，并将统计结果输出

<?php 
set_time_limit(0);
//从txt中获取图片地址和采集页网址
function getUrls($fileName){
	$lineArray=array();
	$fp = fopen($fileName, "r") or die("Unable to open file ".$fileName);
	
	while(! feof($fp)){
		$line=fgets($fp);
		$lineArr[]=trim($line);

	}

	fclose($fp);
	return $lineArr;

}
function getStatus($url){


	$ch = curl_init();
	curl_setopt($ch, CURLOPT_URL, $url);
	curl_setopt($ch, CURLOPT_TIMEOUT, 30);
	curl_setopt($ch, CURLOPT_RETURNTRANSFER, 1);
	curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, FALSE);
	curl_setopt($ch, CURLOPT_SSL_VERIFYHOST, FALSE);
	curl_exec($ch);  // $resp = curl_exec($ch);
	$curl_code = curl_getinfo($ch, CURLINFO_HTTP_CODE);
	curl_close($ch);
	return $curl_code;

	
}

$urls=getUrls('urls.txt');
$fp=fopen("200urls.txt", "a+");
foreach ($urls as $url) {
	
	//$headers = get_headers($url); 
	$curl_code=getStatus($url); 
	if ($curl_code == 200 || $curl_code == 302|| $curl_code == 301) {
	    echo $url."-------->{$curl_code}".PHP_EOL."</br>";
		fwrite($fp, $url.PHP_EOL);
	} else {
	    echo $url.'-------->不能正常访问'.PHP_EOL."</br>";
	}
	
}

fclose($fp);


 ?>

测试结果如下

http://www.experienceproject.com——–>200
http://www.bcz.com——–>301
http://www.wordpress.com/——–>301
http://www.blogger.com——–>不能正常访问
https://www.facebook.com/——–>302
http://www.linkedin.com——–>301
http://www.tumblr.com/dashboard——–>不能正常访问
https://twitter.com/——–>200
https://myspace.com/home——–>302
http://www.arto.com——–>200
http://www.pysznosci.org——–>不能正常访问
http://www.skyrock.com——–>不能正常访问
http://www.livejournal.com——–>200
http://www.createblog.com——–>200
http://over-blog.com——–>不能正常访问
http://www.folkd.com——–>200
http://iblog.co.za——–>不能正常访问
http://friendsite.com——–>301
http://www.onsugar.com/——–>200