This may be obvious, but:
Note that is MUCH faster to use use a single instance to make a series of curl requests rather than creating a new instance for each request.
(PHP 4 >= 4.0.2, PHP 5, PHP 7, PHP 8)
curl_init — 初始化 cURL 会话
初始化新会话，返回 cURL 句柄，供 curl_setopt()、curl_exec() 和 curl_close() 函数使用。
url
如果提供了该参数，CURLOPT_URL 选项将会被设置成这个值。也可以使用 curl_setopt() 函数手动地设置这个值。
注意:
如果设置了 open_basedir，
file协议会被 cURL 禁用。
成功时返回 cURL 句柄，错误时返回 false。
| 版本 | 说明 |
|---|---|
| 8.0.0 | 此函数成功时现在返回 CurlHandle 实例；之前返回 resource。 |
| 8.0.0 |
url 现在可为 null。
|
示例 #1 初始化新 cURL 会话并获取网页
<?php
// 创建新 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);
?>This may be obvious, but:
Note that is MUCH faster to use use a single instance to make a series of curl requests rather than creating a new instance for each request.NextgenThemes' note is applicable for very very limited situations. For completeness's sake, let's consider the following code snippet:
<?php
/*
Your localhost has a default Apache which simply returns "It works!"
*/
$repeatCount = 1000;
// begin section
// this section is slow
// call localhost, create new handle each time
$time = microtime(true);
foreach (range(1, $repeatCount) as $ignored) {
$ch = curl_init("http://localhost");
curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
$response = curl_exec($ch);
// do something with the response
unset($response);
curl_close($ch);
}
unset($ch);
$elapsed = microtime(true) - $time;
echo "Recreate curl handle, time taken: " . $elapsed . "\n";
// end section
// begin section
// this section is much faster
// call localhost, but reuse the handle
$time = microtime(true);
$ch = curl_init("http://localhost");
foreach (range(1, $repeatCount) as $ignored) {
curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
$response = curl_exec($ch);
// do something with the response
unset($response);
}
curl_close($ch);
$elapsed = microtime(true) - $time;
echo "Reuse curl handle, time taken: " . $elapsed . "\n";
// end section
/*
Example output:
Recreate curl handle, time taken: 11.289301872253
Reuse curl handle, time taken: 0.53790807723999
*/
?>
The above code supports the claim by NextgenThemes, however the "send curl requests in sequence" method in general is unnecessarily slow because:
- network transfer time (e.g. 100ms)
- remote processing time (e.g. 50ms)
- usually, no need to send requests in specific sequence
So, in practice, when you need to send multiple curl requests at the same time, just use the curl_multi_init method. Don't consider the "send curl requests in sequence" method unless you have very very specific/special needs.