Allow to override Curl options passed to Guzzle request

[antonbabenko]

Currently curl options are hardcoded in Client.php, but sometimes more options have to be set (increase timeout or max. redirects).

Code from Client.php:

 $guzzleRequest->getCurlOptions()
        ->set(CURLOPT_FOLLOWLOCATION, false)
        ->set(CURLOPT_MAXREDIRS, 0)
        ->set(CURLOPT_TIMEOUT, 30);

[mtdowling]

You can add custom curl options to clients when injecting a Guzzle client into a Goutte client. If you prefix a client configuration option name with curl. it will be treated as a Curl option and added to each request created by the client.

See http://guzzlephp.org/tour/http.html#low-level-curl-access

[antonbabenko]

@mtdowling - Thanks. It works :)

[seejohnrun]

Note: This is no longer the correct answer for redirects (and is seemingly unreliable now), there is now redirect.disable, which you can set to true:
http://guzzlephp.org/tour/http.html#low-level-curl-access

[bkuhl]

Unfortunately both of these links are no longer valid.

[antonbabenko]

@bkuhl It is here - http://docs.guzzlephp.org/en/latest/http-client/request.html#custom-curl-options

[hgraca]

It's not there any more, can you find the correct link? tkx.

[antonbabenko]

I think this is it - http://guzzle.readthedocs.org/en/latest/clients.html#config

[jcroll]

If you are using a recent release of "fabpot/goutte" which is using "guzzlehttp/guzzle": "4.*" I believe you can alter the curl options like below:

$client = new \Goutte\Client();
$client->getClient()->setDefaultOption('config', ['curl' => ['CURLOPT_TIMEOUT' => 60]]);
$crawler = $client->request('GET', $url);

[hgraca]

I was trying to use codeception to test my api call which can take a long time and I wanted to put the timeout to 0.
I ended up doing this:

private function setCurlOption($curlOption, $value)
{
    if (!isset($this->client)) {
        if (isset($this->config['url']) && !strpos($this->config['url'], '://')) {
            // not valid url
            foreach ($this->getModules() as $module) {
                if ($module instanceof \Codeception\Util\Framework) {
                    $this->client = $module->client;
                    break;
                }
            }
        } else {
            if (!$this->hasModule('PhpBrowser')) {
                throw new \Exception(__CLASS__, "For REST testing via HTTP please enable PhpBrowser module");
            }
            $this->client = $this->getModule('PhpBrowser')->session->getDriver()->getClient();
        }
        if (!isset($this->client)) {
            throw new \Exception(__CLASS__, "Client for REST requests not initialized.\nProvide either PhpBrowser module, or a framework module which shares FrameworkInterface");
        }
    }

    if (method_exists($this->client, 'getClient')) {
        /** @var \Guzzle\Common\Collection $clientConfig */
        $clientConfig             = $this->client->getClient()->getConfig();
        $curlOptions              = $clientConfig->get('curl.options');
        $curlOptions[$curlOption] = $value;
        $clientConfig->set('curl.options', $curlOptions);
    }
}

I got most of this from somewhere else though.

[ssanders]

For Goutte 1 / Guzzle 3 / PHP 5.3:

$client->getClient()->getConfig()->set('curl.options', array(CURLOPT_TIMEOUT => 30));

[fabpot]

@ssanders Could you add a note in the README file?

[mtdowling]

That's not how it works in Guzzle 4. See: http://docs.guzzlephp.org/en/latest/faq.html#how-can-i-add-custom-curl-options and http://docs.guzzlephp.org/en/latest/clients.html#creating-a-client.

This could be set after a client is created using:

$client->getClient()->setDefaultOption('config/curl/' . CURLOPT_XYZ, 'Foo');

There is no need to set things like timeout as that can be handled by Guzzle for various adapters using request options: http://docs.guzzlephp.org/en/latest/clients.html#timeout

[ssanders]

I don’t think that actually works in Goutte 2.

I set CURLOPT_TIMEOUT and _CONNECTTIMEOUT to 0 yet get cURL error 28 on a URL.

Guzzle itself, which I think has no default timeout, works OK, no options changed.

[YuvrajHinger]

Modify your scrapping url.
I did in a way like
http://www.yuvrajhinger.in
for every domain add http://www.

It works for me.

Hope it helps you.