Stopping a simplexml_load_file() function after a few seconds

PHP
1

Hello everyone. I would really appreciate your help PHP masters as I was not able to find the solution by my self. The problem is as follows:

I’m using the simple_xml_load_file function to fetch some xml from Youtube in order to embed a video with the oembed technique. The code is:


$xml = simplexml_load_file("http://www.youtube.com/oembed?url=".$post['embed']."&format=xml")

Nothing at all wrong with that. When a connection is made to Youtube everything is great. However when it cannot make a connection it will spend a minute trying to do so without loading the rest of the script. That wouldn’t bother me a lot but for the fact that I’m in China and Youtube is banned here. So every time I’m testing a page with a Youtube embed I rather not because the script will hang up for a minute until it decides that it cannot get the requested XML . So my question is: How can I abort the code above after a given amount of seconds, is it a php.ini setting or is there a way I can do this with php code?

Your help would be really appreciated,

thanks :slight_smile:

2

There are a fair few ways to do this, here’s one. :slight_smile:


function simplexml_load_file_from_url($url, $timeout = 5){
  
  $context = array('http' => array('timeout' => (int)$timeout));
  
  $data = file_get_contents($url, false, $context);
  
  if(!$data){
    trigger_error('Cannot load data from url: ' . $url, E_USER_NOTICE);
    return false;
  }
  
  return simplexml_load_string($data);
}
3

CURL should be used because file_get_contents() only fails once the max execution time of php has been reached. Using CURL time before abort can be controlled. So you would first load the contents into a string using cURL than into a smplexml object when the request was successful otherwise handle the exception gracefully. Its never a good idea to use file_get_contents() to fetch external content from another site due to the execution time only be limited to the that of what is set for the php max execution time.

4

Does this apply even when a context has been applied stating otherwise Oddz?

5

Thank you for both answers , I will look into them :slight_smile: But of course if I can avoid using a library I will

6

No. I’ve no idea what oddz is talking about. There is a configuration option relating to this topic which is default_socket_timeout (default: 60 seconds). This setting affects file_get_contents() unless a stream context is used with a timeout.

Now, some more thoughts. First is that Anthony’s function should be using the [url=http://php.net/stream_context_create]stream_context_create() function to actually create a context.

Next, as mentioned above, you could temporarily set the default_socket_timeout INI option to something less than the 60 seconds currently being used. Or you could use a stream context with file_get_contents() as Anthony (nearly) showed. Or you could use a stream context with simplexml_load_file() via the libxml_set_streams_context() function.

7

Gah, sorry Salathe; a mishap on my part. :slight_smile:

8

Thanks Salathe and Anthony, very useful info!
As for performance are there any differences from using file_get_contents() or simplexml_load_file()? I noticed a slight delay on page load sometimes when using simplexml_load_file() although it might not have anything to do with how the function works.

Thanks again wise people!