cURL allows transfer of data across a wide variety of protocols, and is a very powerful system. It’s widely used as a way to send data across websites, including things like API interaction and oAuth. cURL is unrestricted in what it can do, from the basic HTTP request, to the more complex FTP upload or interaction with an authentication enclosed HTTPS site. We’ll be looking at the simple difference between sending a
POST request and dealing with the returned response, as well as highlighting some useful parameters.
Before we can do anything with a cURL request, we need to first instantiate an instance of cURL – we can do this by calling the function
curl_init();, which returns a cURL resource. This function takes one parameter which is the URL that you want to send the request to, however, in our case, we’ll hold off doing that for now and set it an alternatively way later.
Once we’ve got a cURL resource, we can begin to assign some settings, below is a list of some of the core ones that I set
CURLOPT_RETURNTRANSFER– Return the response as a string instead of outputting it to the screen
CURLOPT_CONNECTTIMEOUT– Number of seconds to spend attempting to connect
CURLOPT_TIMEOUT– Number of seconds to allow cURL to execute
CURLOPT_USERAGENT– Useragent string to use for request
CURLOPT_URL– URL to send request to
CURLOPT_POST– Send request as
CURLOPT_POSTFIELDS– Array of data to POST in request
We can set a setting by using the
curl_setopt() method, which takes three parameters, the cURL resource, the setting and the value. So, to set the URL that we’re sending the request to as http://testcURL.com:
As mentioned, we can set the URL by sending a parameter through when getting the cURL resource:
It is possible to set multiple settings at one time by passing through an array of settings and values to the function
When all of the options are sent, and the request is ready to send, we can call the
curl_exec() method which will execute the cURL request. This function can return three different things:
false– if there is an error executing the request
true– if the request executed without error and
CURLOPT_RETURNTRANSFERis set to
- The result – if the request executed without error and
CURLOPT_RETURNTRANSFERis set to
Using the previous example, where we are wanting to get the result back, we would use the following:
$result now containing the response from the page – which might be JSON, a string or a full blown site’s HTML.
When you’ve sent a request and got the result back, you should look to close the cURL request so that you can free up some system resources, this is as simple as calling the
curl_close() method which as with all other functions takes the resource as its parameter.
GET request is the default request method that is used, and is very straight forward to use, infact all of the examples so far have been
GET requests. If you want to send parameters along in the request you simply append them to the URL as a query string such as
So for example to send a
GET request to the above URL and return the result we would use:
The sole difference between the
GET request syntax is the addition of one setting, two if you want to send some data. We’ll be setting
true and sending an array of data through with the setting
So for example switching the above
GET request to be a
POST request, we would use the following code:
There you have a
POST request that will work the same as our
GET request above and return the response back to the script so that you can use it as you want.
As much as we all hate errors, you really need to take care to account for any eventuality with cURL as ultimately you will not have control over the site(s) that you are sending your request to, you cannot guarantee that the response will be in the format that you want, or that the site will even be available.
There are a few functions that you can use to handle errors and these are:
curl_error()– returns a string error message, will be blank
''if the request doesn’t fail.
curl_errno()– which will return the cURL error number which you can then look up on this page listing error codes.
An example usage would be:
You might want to look at using the setting
true if you want any HTTP response code greater than 400 to cause an error, instead of returning the page HTML.
cURL is a behemoth, and has many many possibilities. Some sites might only serve pages to some user agents, and when working with APIs, some might request you send a specfici user agent, this is something to be aware of. If you want to have a play with some cURL requests, why not have a go at playing with oAuth with Instagram. Or you can just reach us using this link.