[SOLVED] CURLOPT_SETTIMEOUT trouble

[Fyorl]

I've got a function that gets some XML data from a server. Unfortunately the server isn't very reliable so I've used curl_setopt() with a CURLOPT_SETTIMEOUT value of 5.

 

Here's some reference code to better understand:

$ch = curl_init();
curl_setopt($ch, CURLOPT_URL, $this -> url);
curl_setopt($ch, CURLOPT_RETURNTRANSFER, 1);
curl_setopt($ch, CURLOPT_USERAGENT, $userAgent);
curl_setopt($ch, CURLOPT_CONNECTTIMEOUT, 5);
$xml = curl_exec($ch);
curl_close($ch);

 

I'm unsure as what it supposed to happen. If it times-out, will $xml be false? At the moment I'm assuming that behaviour and loading the data from cache if $xml is false. The only problem is that whenever the server is unavailable, the script just takes forever to execute and doesn't seem to be timing-out at all.

 

In actual fact there's some really weird things happening which I have no idea about but I just wanted to check first that this part of the code isn't the problem because it seems like the most obvious place where things are going wrong.

[tinker]

In your example your using CURLOPT_CONNECTTIMEOUT, in my limited experience with cURL I use:

curl_setopt($ch, CURLOPT_TIMEOUT, 16);

[Fyorl]

What's the significance of 16? I'll try it but I don't think it will make any difference as the script seems to be ignoring the timeout and waiting 20 minutes for the absent server to respond.

 

Have you had any success with using 16? What I mean is, has there been a time when the script has actually had to stop trying to connect due to a timeout?

[tinker]

16 is just a sweet arbitrary value, i've got mine high at the moment because actual connections seem to be taking a long time for me, but the original tutorial was to 3, meaning 4 seconds, go figure? However CURLOPT_TIMEOUT works for me!

[Fyorl]

I just had a quick look through the documentation. CURLOPT_SETTIMEOUT isn't even on the curl_setopt page so I have no idea where I got it from now.

 

From the description CURLOPT_CONNECTTIMEOUT looks like the one I want but it obviously doesn't work for some reason as I was already using it. I might just use your suggestion and go with CURLOPT_TIMEOUT to make sure the load times are snappy though.

 

Thanks though!