CODEWITHSUNDEEP
phpmiddlewareguzzleguzzle6
halfer
halfer

Reputation: 20487

Can I add middleware to the default Guzzle 6 HandlerStack, rather than creating a new stack?

I am using the Spatie\Crawler crawler software in a fairly standard way, like so:

$client = new Client([
    RequestOptions::COOKIES => true,
    RequestOptions::CONNECT_TIMEOUT => 10,
    RequestOptions::TIMEOUT => 10,
    RequestOptions::ALLOW_REDIRECTS => true,
]);
$crawler = new Crawler($client, 1);
$crawler->
    setCrawlProfile(new MyCrawlProfile($startUrl, $pathRegex))->
    setCrawlObserver(new MyCrawlObserver())->
    startCrawling($url);

I've omitted the definition of the classes MyCrawlProfile of MyCrawlObserver for brevity, but anyway, this works as it stands.

I want to add some middleware in order to change some requests before they are made, so I added this demo code:

$stack = new HandlerStack();
$stack->setHandler(new CurlHandler());
$stack->push(
    Middleware::mapRequest(function (RequestInterface $request) {
        echo "Middleware running\n";

        return $request;
    })
);
$client = new Client([
    RequestOptions::COOKIES => true,
    RequestOptions::CONNECT_TIMEOUT => 10,
    RequestOptions::TIMEOUT => 10,
    RequestOptions::ALLOW_REDIRECTS => true,
    'handler' => $stack,
]);
// ... rest of crawler code here ...

However, it falls on the first hurdle - it scrapes the root of the site (/) which is actually a Location redirect, and then stops. It turns out that I am now missing the RedirectMiddleware despite not having removed it deliberately.

So, my problem is fixed by also adding this:

$stack->push(Middleware::redirect());

I wonder now what other things are set up by default in Guzzle that I have accidentally removed by creating a fresh HandlerStack. Cookies? Retry mechanisms? Other stuff? I don't need those things right now, but I'd be a bit more confident about my system's long-term reliability if my code merely modified the existing stack.

Is there a way to do that? As far as I can tell, I'm doing things as per the manual.

Upvotes: 1

Views: 1890

Answers (1)

Alexey Shokov
Alexey Shokov

Reputation: 5010

$stack = HandlerStack::create();

instead of

$stack = new HandlerStack();
$stack->setHandler(new CurlHandler());

It's important, because create() adds additional middlewares, especially for redirects.

Upvotes: 3

Related Questions