在检查报头是否有404错误时，需要考虑的一件事是站点不会立即生成404错误。

很多网站会检查PHP/ASP(等等)源代码中是否存在某个页面，然后将您转到404页面。在这些情况下，头基本上是由生成的404头扩展的。在这种情况下，404错误不会出现在头文件的第一行，而是第10行。

$array = get_headers($url);
$string = $array[0];
print_r($string) // would generate:

Array ( 
[0] => HTTP/1.0 301 Moved Permanently 
[1] => Date: Fri, 09 Nov 2018 16:12:29 GMT 
[2] => Server: Apache/2.4.34 (FreeBSD) LibreSSL/2.7.4 PHP/7.0.31 
[3] => X-Powered-By: PHP/7.0.31 
[4] => Set-Cookie: landing=%2Freed-diffuser-fig-pudding-50; path=/; HttpOnly 
[5] => Location: /reed-diffuser-fig-pudding-50/ 
[6] => Content-Length: 0 
[7] => Connection: close 
[8] => Content-Type: text/html; charset=utf-8 
[9] => HTTP/1.0 404 Not Found 
[10] => Date: Fri, 09 Nov 2018 16:12:29 GMT 
[11] => Server: Apache/2.4.34 (FreeBSD) LibreSSL/2.7.4 PHP/7.0.31 
[12] => X-Powered-By: PHP/7.0.31 
[13] => Set-Cookie: landing=%2Freed-diffuser-fig-pudding-50%2F; path=/; HttpOnly 
[14] => Connection: close 
[15] => Content-Type: text/html; charset=utf-8 
) 

很快:

function http_response($url){
    $resURL = curl_init(); 
    curl_setopt($resURL, CURLOPT_URL, $url); 
    curl_setopt($resURL, CURLOPT_BINARYTRANSFER, 1); 
    curl_setopt($resURL, CURLOPT_HEADERFUNCTION, 'curlHeaderCallback'); 
    curl_setopt($resURL, CURLOPT_FAILONERROR, 1); 
    curl_exec ($resURL); 
    $intReturnCode = curl_getinfo($resURL, CURLINFO_HTTP_CODE); 
    curl_close ($resURL); 
    if ($intReturnCode != 200 && $intReturnCode != 302 && $intReturnCode != 304) { return 0; } else return 1;
}

echo 'google:';
echo http_response('http://www.google.com');
echo '/ ogogle:';
echo http_response('http://www.ogogle.com');

当从php中判断url是否存在时，有几件事需要注意:

Is the url itself valid (a string, not empty, good syntax), this is quick to check server side. Waiting for a response might take time and block code execution. Not all headers returned by get_headers() are well formed. Use curl (if you can). Prevent fetching the entire body/content, but only request the headers. Consider redirecting urls: Do you want the first code returned? Or follow all redirects and return the last code? You might end up with a 200, but it could redirect using meta tags or javascript. Figuring out what happens after is tough.

请记住，无论你使用什么方法，等待回复都需要时间。 所有代码都可能(很可能)停止，直到您知道结果或请求超时。

例如:如果url无效或不可达，下面的代码可能需要很长时间才能显示页面:

<?php
$urls = getUrls(); // some function getting say 10 or more external links

foreach($urls as $k=>$url){
  // this could potentially take 0-30 seconds each
  // (more or less depending on connection, target site, timeout settings...)
  if( ! isValidUrl($url) ){
    unset($urls[$k]);
  }
}

echo "yay all done! now show my site";
foreach($urls as $url){
  echo "<a href=\"{$url}\">{$url}</a><br/>";
}

下面的函数可能会有帮助，你可能想修改它们以适应你的需要:

    function isValidUrl($url){
        // first do some quick sanity checks:
        if(!$url || !is_string($url)){
            return false;
        }
        // quick check url is roughly a valid http request: ( http://blah/... ) 
        if( ! preg_match('/^http(s)?:\/\/[a-z0-9-]+(\.[a-z0-9-]+)*(:[0-9]+)?(\/.*)?$/i', $url) ){
            return false;
        }
        // the next bit could be slow:
        if(getHttpResponseCode_using_curl($url) != 200){
//      if(getHttpResponseCode_using_getheaders($url) != 200){  // use this one if you cant use curl
            return false;
        }
        // all good!
        return true;
    }
    
    function getHttpResponseCode_using_curl($url, $followredirects = true){
        // returns int responsecode, or false (if url does not exist or connection timeout occurs)
        // NOTE: could potentially take up to 0-30 seconds , blocking further code execution (more or less depending on connection, target site, and local timeout settings))
        // if $followredirects == false: return the FIRST known httpcode (ignore redirects)
        // if $followredirects == true : return the LAST  known httpcode (when redirected)
        if(! $url || ! is_string($url)){
            return false;
        }
        $ch = @curl_init($url);
        if($ch === false){
            return false;
        }
        @curl_setopt($ch, CURLOPT_HEADER         ,true);    // we want headers
        @curl_setopt($ch, CURLOPT_NOBODY         ,true);    // dont need body
        @curl_setopt($ch, CURLOPT_RETURNTRANSFER ,true);    // catch output (do NOT print!)
        if($followredirects){
            @curl_setopt($ch, CURLOPT_FOLLOWLOCATION ,true);
            @curl_setopt($ch, CURLOPT_MAXREDIRS      ,10);  // fairly random number, but could prevent unwanted endless redirects with followlocation=true
        }else{
            @curl_setopt($ch, CURLOPT_FOLLOWLOCATION ,false);
        }
//      @curl_setopt($ch, CURLOPT_CONNECTTIMEOUT ,5);   // fairly random number (seconds)... but could prevent waiting forever to get a result
//      @curl_setopt($ch, CURLOPT_TIMEOUT        ,6);   // fairly random number (seconds)... but could prevent waiting forever to get a result
//      @curl_setopt($ch, CURLOPT_USERAGENT      ,"Mozilla/5.0 (Windows NT 6.0) AppleWebKit/537.1 (KHTML, like Gecko) Chrome/21.0.1180.89 Safari/537.1");   // pretend we're a regular browser
        @curl_exec($ch);
        if(@curl_errno($ch)){   // should be 0
            @curl_close($ch);
            return false;
        }
        $code = @curl_getinfo($ch, CURLINFO_HTTP_CODE); // note: php.net documentation shows this returns a string, but really it returns an int
        @curl_close($ch);
        return $code;
    }
    
    function getHttpResponseCode_using_getheaders($url, $followredirects = true){
        // returns string responsecode, or false if no responsecode found in headers (or url does not exist)
        // NOTE: could potentially take up to 0-30 seconds , blocking further code execution (more or less depending on connection, target site, and local timeout settings))
        // if $followredirects == false: return the FIRST known httpcode (ignore redirects)
        // if $followredirects == true : return the LAST  known httpcode (when redirected)
        if(! $url || ! is_string($url)){
            return false;
        }
        $headers = @get_headers($url);
        if($headers && is_array($headers)){
            if($followredirects){
                // we want the last errorcode, reverse array so we start at the end:
                $headers = array_reverse($headers);
            }
            foreach($headers as $hline){
                // search for things like "HTTP/1.1 200 OK" , "HTTP/1.0 200 OK" , "HTTP/1.1 301 PERMANENTLY MOVED" , "HTTP/1.1 400 Not Found" , etc.
                // note that the exact syntax/version/output differs, so there is some string magic involved here
                if(preg_match('/^HTTP\/\S+\s+([1-9][0-9][0-9])\s+.*/', $hline, $matches) ){// "HTTP/*** ### ***"
                    $code = $matches[1];
                    return $code;
                }
            }
            // no HTTP/xxx found in headers:
            return false;
        }
        // no headers :
        return false;
    }

function url_exists($url) {
    $headers = @get_headers($url);
    return (strpos($headers[0],'200')===false)? false:true;
}

cURL可以返回HTTP代码，我不认为所有额外的代码是必要的?

function urlExists($url=NULL)
    {
        if($url == NULL) return false;
        $ch = curl_init($url);
        curl_setopt($ch, CURLOPT_TIMEOUT, 5);
        curl_setopt($ch, CURLOPT_CONNECTTIMEOUT, 5);
        curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
        $data = curl_exec($ch);
        $httpcode = curl_getinfo($ch, CURLINFO_HTTP_CODE);
        curl_close($ch); 
        if($httpcode>=200 && $httpcode<300){
            return true;
        } else {
            return false;
        }
    }

我运行一些测试，看看我的网站上的链接是否有效-提醒我当第三方改变他们的链接。我有一个网站的问题，有一个配置不良的证书，这意味着php的get_headers不能工作。

所以，我读到卷曲更快，并决定给一个尝试。然后我在领英上遇到了一个问题，给了我一个999错误，后来证明是用户代理的问题。

我不关心证书是否对该测试无效，也不关心响应是否为重定向。

然后我认为使用get_headers无论如何，如果卷曲失败....

试试看....

/**
 * returns true/false if the $url is valid.
 *
 * @param string $url assumes this is a valid url.
 *
 * @return bool
 */
private function urlExists(string $url): bool
{
    $ch = curl_init($url);
    curl_setopt($ch, CURLOPT_URL, $url);
    curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);     // do not output response in stdout
    curl_setopt($ch, CURLOPT_NOBODY, true);             // this does a head request to make it faster.
    curl_setopt($ch, CURLOPT_HEADER, true);             // just the headers
    curl_setopt($ch, CURLOPT_SSL_VERIFYSTATUS, false);  // turn off that pesky ssl stuff - some sys admins can't get it right.
    curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, false);
    // set a real user agent to stop linkedin getting upset.
    curl_setopt($ch, CURLOPT_USERAGENT, 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_14_1) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/70.0.3538.77 Safari/537.36');
    curl_exec($ch);
    $http_code = curl_getinfo($ch, CURLINFO_HTTP_CODE);
    if (($http_code >= 200 && $http_code < 400) || $http_code === 999) {
        curl_close($ch);
        return true;
    }
    //$error = curl_error($ch); // used for debugging.
    curl_close($ch);

    // just try the get_headers - it might work!
    stream_context_set_default(
        ['http' => ['method' => 'HEAD']]
    );
    $file_headers = @get_headers($url);

    if ($file_headers !== false) {
        $response_code = substr($file_headers[0], 9, 3);
        return $response_code >= 200 && $response_code < 400;
    }

    return false;
}