(PHP 4 >= 4.0.2, PHP 5, PHP 7)
curl_exec — 执行 cURL 会话
成功时返回 TRUE， 或者在失败时返回 FALSE。 然而，如果 设置了 CURLOPT_RETURNTRANSFER 选项，函数执行成功时会返回执行的结果，失败时返回 FALSE 。
Example #1 获取网页
<?php
// 创建新的 cURL 资源
$ch = curl_init();
// 设置 URL 和相应的选项
curl_setopt($ch, CURLOPT_URL, "http://www.example.com/");
curl_setopt($ch, CURLOPT_HEADER, 0);
// 抓取 URL 并把它传递给浏览器
curl_exec($ch);
// 关闭 cURL 资源，并且释放系统资源
curl_close($ch);
?>
Don't disable SSL verification! You don't need to, and it's super easy to stay secure! If you found that turning off "CURLOPT_SSL_VERIFYHOST" and "CURLOPT_SSL_VERIFYPEER" solved your problem, odds are you're just on a Windows box. Takes 2 min to solve the problem. Walkthrough here:
https://snippets.webaware.com.au/howto/stop-turning-off-curlopt_ssl_verifypeer-and-fix-your-php-config/
If you are accessing HTTPS URLs and you do not receive any contents, try to disable verifying SSL.
<?php
curl_setopt($ch, CURLOPT_SSL_VERIFYHOST, 0);
curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, 0);
?>
Be always aware that CURLOPT_SSL_VERIFYPEER set to FALSE or 0 should never be used for production as it makes the link inmediately vulnerable to man-in-the-middle attack, still you can use it during development, but I would suggest that only if you KNOW what are you doing, otherwise spend some more time making requests to HTTPS sites work without resorting to set that option to FALSE or 0.
If you want to get contents of a page through HTTP GET or POST, you can write a small function like below.
<?php
/**
* Read entire contents of a URL in a string
* @param string $url URL to fetch data from
* @param mixed $params Parameters to the URL, typically an array. In case of POST, can be a string
* @param string $method HTTP method - can be 'GET' (default) or 'POST'
* @return mixed Contents of the webpage. Returns <i>false</i> in case of failure.
*/
function url_get_contents($url, $params = null, $method = 'GET') {
$contents = false;
if (!in_array($method, array('GET', 'POST'))) {
error_log(__FUNCTION__ . ": Unknown method '$method'");
return false;
}
if ($method == 'GET') {
if (is_array($params) && count($params) > 0) {
if ($params === array_values($params)) {
error_log(__FUNCTION__ . ": Numerical array recieved for argument '\$params' (assoc array expected)");
return false;
}
else {
$url .= '?' . http_build_query($params);
}
}
elseif (!is_null($params)) {
error_log(__FUNCTION__ . ": If you're making a GET request, argument \$params must be null or assoc array.");
return false;
}
}
$ch = curl_init($url);
if ($ch !== false) {
curl_setopt_array($ch, array(
CURLOPT_HEADER => false,
CURLOPT_RETURNTRANSFER => true,
));
if ($method == 'POST') {
curl_setopt($ch, CURLOPT_POST, true);
if (is_string($params) || is_array($params)) {
curl_setopt($ch, CURLOPT_POSTFIELDS, $params);
}
else {
error_log(__FUNCTION__ . ": Argument \$params should be an array of parameters or (if you want to send raw data) a string");
return false;
}
}
$contents = curl_exec($ch);
curl_close($ch);
}
return $contents;
}
There will be times when you need to get the response from curl_exec and capture the transfer - this is not very well documented but you can do so with the CURLOPT_RETURNTRANSFER Option
curl_setopt($ch, CURLOPT_RETURNTRANSFER, TRUE);
Bert
I was having a problem, for almost a week, of curl_exec() freezing/hanging when I made a request with it to a page that spends over an hour converting a large video file, and only afterwards sends control-data back to the calling script.
I'm using windows 7, WampServer-2.1d-64.exe, PHP 5.3.4, libcurl-7.21.3.
Here's the final solution;
On Fri, Feb 11, 2011 at 1:32 PM, Daniel Stenberg <daniel-of-haxx.se> wrote:
> On Fri, 11 Feb 2011, Tolas Anon wrote:
>
>>> It is also very easy for an application to enable the options as I've
>>> shown.
>>
>> Just not for the platforms i use, apparently.. :(
>
> Why not? A quick search for "tcp keepalive windows" shows this:
> http://msdn.microsoft.com/en-us/library/ms819735.aspx
This fixed my problem in the simple test AND in my application!
No need to send keep-alive bytes on the text/html level either.
I just added the "KeepAliveTime" setting with windows 7 regedit.exe in
HKEY_LOCAL_MACHINE/SYSTEM/CurrentControlSet/services/Tcpip, as a
REG_DWORD value, set it to decimal 25000 (so 25 seconds), rebooted,
and it all works as i want it now..
Note that windows registry entries key names are case-sensitive, wrong
casing and they'll be deleted on restart.
Lots of thanks, Daniel. I would've never solved this on my own..
------------------------------------------------------
(now useless) details at:
http://curl.haxx.se/mail/lib-2011-02/0101.html
http://readlist.com/lists/lists.php.net/php-general/16/81195.html
Just in case anyone is looking for a a couple of simple functions [to help automate cURL processes for POST and GET queries] I thought I'd post these.
<?php
/**
* Send a POST requst using cURL
* @param string $url to request
* @param array $post values to send
* @param array $options for cURL
* @return string
*/
function curl_post($url, array $post = NULL, array $options = array())
{
$defaults = array(
CURLOPT_POST => 1,
CURLOPT_HEADER => 0,
CURLOPT_URL => $url,
CURLOPT_FRESH_CONNECT => 1,
CURLOPT_RETURNTRANSFER => 1,
CURLOPT_FORBID_REUSE => 1,
CURLOPT_TIMEOUT => 4,
CURLOPT_POSTFIELDS => http_build_query($post)
);
$ch = curl_init();
curl_setopt_array($ch, ($options + $defaults));
if( ! $result = curl_exec($ch))
{
trigger_error(curl_error($ch));
}
curl_close($ch);
return $result;
}
/**
* Send a GET requst using cURL
* @param string $url to request
* @param array $get values to send
* @param array $options for cURL
* @return string
*/
function curl_get($url, array $get = NULL, array $options = array())
{
$defaults = array(
CURLOPT_URL => $url. (strpos($url, '?') === FALSE ? '?' : ''). http_build_query($get),
CURLOPT_HEADER => 0,
CURLOPT_RETURNTRANSFER => TRUE,
CURLOPT_TIMEOUT => 4
);
$ch = curl_init();
curl_setopt_array($ch, ($options + $defaults));
if( ! $result = curl_exec($ch))
{
trigger_error(curl_error($ch));
}
curl_close($ch);
return $result;
}
?>
If having problems with special chars or entities (like á, ?, à, etc.), using the ISO encode, just decode the values given with the function utf8_decode().
For example:
The returned string is the following:
<xml><name>Iv??n</name></xml>
Using utf8_decode, the result in ISO is
<xml><name>Iván</name></xml>
so far i have not come across any code or library file that will
extract cookie information from a http header, so i've written
one.
there are two files:
file1.php - acts as server and sets the cookies
file2.php - acts as browser and retrieves the cookies set by file1
/*****file1.php****/
<?php
/*set 3 cookies at the end of the execution of this script*/
setcookie("cookie1","cookie 1 data", time() - 123123);
setcookie("cookie2","cookie 2 data", time() + 54326);
setcookie("cookie3","cookie 3 data", time());
?>
/****end file1.php**/
/*****file2.php****/
<?php
$url = "http://[host][uri]/file1.php"; /*insert desired host and
uri*/
$browser_id = "some crazy browser";
$curl_handle = curl_init();
$options = array
(
CURLOPT_URL=>$url,
CURLOPT_HEADER=>true,
CURLOPT_RETURNTRANSFER=>true,
CURLOPT_FOLLOWLOCATION=>true,
CURLOPT_USERAGENT=>$browser_id
);
curl_setopt_array($curl_handle,$options);
$server_output = curl_exec($curl_handle);
curl_close($curl_handle);
/*construct the http search pattern for cookies*/
$pattern = "/Set-Cookie:";
$pattern .= "(?P<name>.*?)=(?P<value>.*?); ";
$pattern .= "expires=(?P<expiry_dayname>\w+), ";
$pattern .= "(?P<expiry_day>\d+)-
(?P<expiry_month>\w+)-(?P<expiry_year>\d+) ";
$pattern .= "(?P<expiry_hour>\d+):
(?P<expiry_minute>\d+):(?P<expiry_second>\d+) ";
$pattern .= "(?P<expiry_zone>\w+)/";
preg_match_all($pattern,$server_output,$matches);
$table_string = "
<h1>cookie information table</h1>
<table border='1'>
<tr>
<td>cookie name</td>
<td>value</td>
<td>expiry day</td>
<td>expiry date</td>
<td>expiry time</td>
<td>expiry timezone</td>
</tr>
";
$i=0;
foreach($matches[name] as $cookie_name)
{
$table_string .= "
<tr>
<td>$cookie_name</td>
<td>{$matches[value][$i]}</td>
<td>{$matches[expiry_dayname][$i]}</td>
<td>{$matches[expiry_day][$i]}-
{$matches[expiry_month][$i]}-
{$matches[expiry_year][$i]}</td>
<td>{$matches[expiry_hour][$i]}:
{$matches[expiry_minute][$i]}:
{$matches[expiry_second][$i]}</td>
<td>{$matches[expiry_zone][$i]}</td>
</tr>
";
$i++;
}
$table_string .= "</table>";
echo $table_string;
?>
/****end file2.php**/
i based this code on the following http header:
(obtained by going: echo $server_output; in file2.php)
HTTP/1.1 200 OK Date: Thu, 30 Jul 2009 07:10:07 GMT
Server: Apache/2.2.11 (Win32) DAV/2 mod_ssl/2.2.11
OpenSSL/0.9.8i PHP/5.2.9 X-Powered-By: PHP/5.2.9
Set-Cookie: cookie1=cookie+1+data; expires=Wed,
29-Jul-2009 03:23:27 GMT Set-Cookie
cookie2=cookie+2+data; expires=Wed, 29-Jul-2009 03:23:27
GMT Set-Cookie: cookie3=cookie+3+data; expires=Wed,
29-Jul-2009 03:23:27 GMT Content-Length: 196
Content-Type: text/html
if your header differs from this one then $pattern in file2.php
will need to be modified accordingly. hopefully this code will save
you a lot of time though!
example:
echo CurlTool::fetchContent('www.onet.pl');
CurlTool::downloadFile('http://download.gadu-gadu.pl/gg77.exe', 'c:/');
<?php
error_reporting(E_STRICT | E_ALL);
class CurlTool {
public static $userAgents = array(
'FireFox3' => 'Mozilla/5.0 (Windows; U; Windows NT 5.1; pl; rv:1.9) Gecko/2008052906 Firefox/3.0',
'GoogleBot' => 'Mozilla/5.0 (compatible; Googlebot/2.1; +http://www.google.com/bot.html)',
'IE7' => 'Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 6.0)',
'Netscape' => 'Mozilla/4.8 [en] (Windows NT 6.0; U)',
'Opera' => 'Opera/9.25 (Windows NT 6.0; U; en)'
);
public static $options = array(
CURLOPT_USERAGENT => 'Mozilla/5.0 (Windows; U; Windows NT 5.1; pl; rv:1.9) Gecko/2008052906 Firefox/3.0',
CURLOPT_AUTOREFERER => true,
CURLOPT_COOKIEFILE => '',
CURLOPT_FOLLOWLOCATION => true
);
private static $proxyServers = array();
private static $proxyCount = 0;
private static $currentProxyIndex = 0;
public static function addProxyServer($url) {
self::$proxyServers[] = $url;
++self::$proxyCount;
}
public static function fetchContent($url, $verbose = false) {
if (($curl = curl_init($url)) == false) {
throw new Exception("curl_init error for url $url.");
}
if (self::$proxyCount > 0) {
$proxy = self::$proxyServers[self::$currentProxyIndex++ % self::$proxyCount];
curl_setopt($curl, CURLOPT_PROXY, $proxy);
if ($verbose === true) {
echo "Reading $url [Proxy: $proxy] ... ";
}
} else if ($verbose === true) {
echo "Reading $url ... ";
}
curl_setopt($curl, CURLOPT_RETURNTRANSFER, true);
curl_setopt_array($curl, self::$options);
$content = curl_exec($curl);
if ($content === false) {
throw new Exception("curl_exec error for url $url.");
}
curl_close($curl);
if ($verbose === true) {
echo "Done.\n";
}
$content = preg_replace('#\n+#', ' ', $content);
$content = preg_replace('#\s+#', ' ', $content);
return $content;
}
public static function downloadFile($url, $fileName, $verbose = false) {
if (($curl = curl_init($url)) == false) {
throw new Exception("curl_init error for url $url.");
}
if (self::$proxyCount > 0) {
$proxy = self::$proxyServers[self::$currentProxyIndex++ % self::$proxyCount];
curl_setopt($curl, CURLOPT_PROXY, $proxy);
if ($verbose === true) {
echo "Downloading $url [Proxy: $proxy] ... ";
}
} else if ($verbose === true) {
echo "Downloading $url ... ";
}
curl_setopt_array($curl, self::$options);
if (substr($fileName, -1) == '/') {
$targetDir = $fileName;
$fileName = tempnam(sys_get_temp_dir(), 'c_');
}
if (($fp = fopen($fileName, "wb")) === false) {
throw new Exception("fopen error for filename $fileName");
}
curl_setopt($curl, CURLOPT_FILE, $fp);
curl_setopt($curl, CURLOPT_BINARYTRANSFER, true);
if (curl_exec($curl) === false) {
fclose($fp);
unlink($fileName);
throw new Exception("curl_exec error for url $url.");
} elseif (isset($targetDir)) {
$eurl = curl_getinfo($curl, CURLINFO_EFFECTIVE_URL);
preg_match('#^.*/(.+)$#', $eurl, $match);
fclose($fp);
rename($fileName, "$targetDir{$match[1]}");
$fileName = "$targetDir{$match[1]}";
} else {
fclose($fp);
}
curl_close($curl);
if ($verbose === true) {
echo "Done.\n";
}
return $fileName;
}
}
?>
Great class Roman - just one fix:
Replace the following line:
<?php
if (isset($params['host']) && $params['host']) $header[]="Host: ".$host;
?>
with this:
<?php
if (isset($params['host']) && $params['host']) $header[]="Host: " . $params['host'];
?>
CURL automatically creates the host parameter (since it is required for HTTP/1.1 requests), so you don't need to set it. But if you created a custom host parameter, the above bug would cause a '400 Bad Request' response due to invalid host specified.
Also when copying and pasting the class code, make sure that no line breaks occur (for example in the $header and $user_agent definitions etc.). It will still be valid PHP, but the HTTP request will not be valid, and you may get a '400 Bad Request' response from the server.
It took me a little playing around with an HTTP Sniffer before I finally got an HTTP POST request fully working!
Thanks,
Alan
<?php
class CurlRequest
{
private $ch;
/**
* Init curl session
*
* $params = array('url' => '',
* 'host' => '',
* 'header' => '',
* 'method' => '',
* 'referer' => '',
* 'cookie' => '',
* 'post_fields' => '',
* ['login' => '',]
* ['password' => '',]
* 'timeout' => 0
* );
*/
public function init($params)
{
$this->ch = curl_init();
$user_agent = 'Mozilla/5.0 (Windows; U;
Windows NT 5.1; ru; rv:1.8.0.9) Gecko/20061206 Firefox/1.5.0.9';
$header = array(
"Accept: text/xml,application/xml,application/xhtml+xml,
text/html;q=0.9,text/plain;q=0.8,image/png,*/*;q=0.5",
"Accept-Language: ru-ru,ru;q=0.7,en-us;q=0.5,en;q=0.3",
"Accept-Charset: windows-1251,utf-8;q=0.7,*;q=0.7",
"Keep-Alive: 300");
if (isset($params['host']) && $params['host']) $header[]="Host: ".$host;
if (isset($params['header']) && $params['header']) $header[]=$params['header'];
@curl_setopt ( $this -> ch , CURLOPT_RETURNTRANSFER , 1 );
@curl_setopt ( $this -> ch , CURLOPT_VERBOSE , 1 );
@curl_setopt ( $this -> ch , CURLOPT_HEADER , 1 );
if ($params['method'] == "HEAD") @curl_setopt($this -> ch,CURLOPT_NOBODY,1);
@curl_setopt ( $this -> ch, CURLOPT_FOLLOWLOCATION, 1);
@curl_setopt ( $this -> ch , CURLOPT_HTTPHEADER, $header );
if ($params['referer']) @curl_setopt ($this -> ch , CURLOPT_REFERER, $params['referer'] );
@curl_setopt ( $this -> ch , CURLOPT_USERAGENT, $user_agent);
if ($params['cookie']) @curl_setopt ($this -> ch , CURLOPT_COOKIE, $params['cookie']);
if ( $params['method'] == "POST" )
{
curl_setopt( $this -> ch, CURLOPT_POST, true );
curl_setopt( $this -> ch, CURLOPT_POSTFIELDS, $params['post_fields'] );
}
@curl_setopt( $this -> ch, CURLOPT_URL, $params['url']);
@curl_setopt ( $this -> ch , CURLOPT_SSL_VERIFYPEER, 0 );
@curl_setopt ( $this -> ch , CURLOPT_SSL_VERIFYHOST, 0 );
if (isset($params['login']) & isset($params['password']))
@curl_setopt($this -> ch , CURLOPT_USERPWD,$params['login'].':'.$params['password']);
@curl_setopt ( $this -> ch , CURLOPT_TIMEOUT, $params['timeout']);
}
/**
* Make curl request
*
* @return array 'header','body','curl_error','http_code','last_url'
*/
public function exec()
{
$response = curl_exec($this->ch);
$error = curl_error($this->ch);
$result = array( 'header' => '',
'body' => '',
'curl_error' => '',
'http_code' => '',
'last_url' => '');
if ( $error != "" )
{
$result['curl_error'] = $error;
return $result;
}
$header_size = curl_getinfo($this->ch,CURLINFO_HEADER_SIZE);
$result['header'] = substr($response, 0, $header_size);
$result['body'] = substr( $response, $header_size );
$result['http_code'] = curl_getinfo($this -> ch,CURLINFO_HTTP_CODE);
$result['last_url'] = curl_getinfo($this -> ch,CURLINFO_EFFECTIVE_URL);
return $result;
}
}
?>
Example of use:
<?php
..........
try
{
$params = array('url' => 'http://www.google.com',
'host' => '',
'header' => '',
'method' => 'GET', // 'POST','HEAD'
'referer' => '',
'cookie' => '',
'post_fields' => '', // 'var1=value&var2=value
'timeout' => 20
);
$this->curl->init($params);
$result = $this->curl->exec();
if ($result['curl_error']) throw new Exception($result['curl_error']);
if ($result['http_code']!='200') throw new Exception("HTTP Code = ".$result['http_code']);
if (!$result['body']) throw new Exception("Body of file is empty");
...............
}
catch (Exception $e)
{
echo $e->getMessage();
}
?>
Note that when you use CURL to POST things....e.g:
curl_setopt($ch, CURLOPT_POSTFIELDS, "string=This is a string");
The data part (e.g. "This is a string") inside the 3rd parameter should be applied with urlencode()
Otherwise, if you intend to send a string like "%2F", you will end up with a "/" on the receiving end, which can cause troubles. (e.g. serialize() data cannot be unserialize() becase of the change in string length).
If you see a "0" at the end of the output, you might want to switch to HTTP/1.0:
curl_setopt($ch, CURLOPT_HTTP_VERSION, 1.0);
Thank you for sharing this. I was wondering why my result was 1.
To get around this in a safe way, this is how I check if the result is valid.
$ch = curl_init(); /// initialize a cURL session
curl_setopt ($ch, CURLOPT_RETURNTRANSFER, true);
$xmlResponse = curl_exec ($ch);
curl_close ($ch);
if (!is_string($xmlResponse) || !strlen($xmlResponse)) {
return $this->_set_error( "Failure Contacting Server" );
} else {
return $xmlResponse;
}
Be careful when using curl_exec() and the CURLOPT_RETURNTRANSFER option. According to the manual and assorted documentation:
Set CURLOPT_RETURNTRANSFER to TRUE to return the transfer as a string of the return value of curl_exec() instead of outputting it out directly.
When retrieving a document with no content (ie. 0 byte file), curl_exec() will return bool(true), not an empty string. I've not seen any mention of this in the manual.
Example code to reproduce this:
<?php
// fictional URL to an existing file with no data in it (ie. 0 byte file)
$url = 'http://www.example.com/empty_file.txt';
$curl = curl_init();
curl_setopt($curl, CURLOPT_URL, $url);
curl_setopt($curl, CURLOPT_RETURNTRANSFER, true);
curl_setopt($curl, CURLOPT_HEADER, false);
// execute and return string (this should be an empty string '')
$str = curl_exec($curl);
curl_close($curl);
// the value of $str is actually bool(true), not empty string ''
var_dump($str);
?>
If you've got problems with curl_exec not working, you should rather check curl_errno and curl_error than using commandline curl, like so:
(since this is easier, and also allows you to check for errors runtime, which is a vital part of any well-design piece of code. ;)
<?php
$creq = curl_init();
curl_setopt($creq, CURLOPT_URL, "http://www.foo.internal");
curl_exec($creq);
/* To quote curl_errno documentation:
Returns the error number for the last cURL operation on the resource ch, or 0 (zero) if no error occurred. */
if (curl_errno($creq)) {
print curl_error($creq);
} else {
curl_close($creq);
}
?>
fyi:
It returns false if there's an error while executing the curl session, no matter how CURLOPT_RETURNTRANSFER is set.
Checking the source, curl_exec seems to return FALSE on failure, TRUE on success (unless CURLOPT_RETURNTRANSFER is set 1, and then it returns the returned data).
fyi - if you are having problems getting a
webpage to display in your webpage with
curl_setopt(CURLOPT_RETURNTRANSFER, 1);
due to version bugginess perhaps,
you may can use output control functions
like this to show a web page
inside your webpage:
<html><head><title>whatever</title></head>
<body>
<script language="php">
$ch = curl_init("http://www.cocoavillage.com/");
// use output buffering instead of returntransfer -itmaybebuggy
ob_start();
curl_exec($ch);
curl_close($ch);
$retrievedhtml = ob_get_contents();
ob_end_clean();
// if you intend to print this page with meta tags, better clear out any expiration tag
// $result = preg_replace('/(?s)<meta http-equiv="Expires"[^>]*>/i', '', $retrievedhtml);
// for now I just want what is between the body tags so need
// somehow cut the header footer
$bodyandend = stristr($retrievedhtml,"<body");
// not needed- $positionstartbodystring = strlen($retrievedhtml)-strlen($bodyandend);
$positionendstartbodytag = strpos($bodyandend,">") + 1;
// got to change all to lowercase temporarily
// because end body may be upperlowercasemix
// to bad strirstr does not exist
$temptofindposition=strtolower($bodyandend);
$positionendendbodytag=strpos($temptofindposition,"</body");
//now to get the endbetween body tags
$grabbedbody=substr($bodyandend,
$positionendstartbodytag,
$positionendendbodytag);
//be sure to fix syntax broke by display on phpwebsite... like above line
print("$grabbedbody");
</script>
</body></html>
tada
If you retrieve a web page and print it (so you can see it in your browser), and the page has an expiration, this expiration now applies to MyProgram.php and next time your program/page is called, even if it's grabbing a different web page, it will show what it just displayed. In Netscape you can get rid of this by going into Edit, Options, Advanced, Cache, and clear out the Disk Cache. But this is really annoying after short order. The following prevents the above scenario:
<?php
function GetCurlPage ($pageSpec) {
$ch = curl_init($pageSpec);
curl_setopt($ch, CURLOPT_RETURNTRANSFER, 1);
$tmp = curl_exec ($ch);
curl_close ($ch);
// if you intend to print this page, better clear out expiration tag
$tmp = preg_replace('/(?s)<meta http-equiv="Expires"[^>]*>/i', '', $tmp);
return $tmp;
}
?>