[SalesForce] handle PDF content in HTTP callout

Callouts cannot be made from triggers as that would hold up the database transaction until the callout completed, which can be up to 120 seconds from a limits perspective. This could obviously cause significant contention with other transactions and impact performance.

The only way that you can execute a callout from a trigger is to schedule it to run asynchronously, for example by executing a method with the @future method as you have found. The key word here is asynchronously - you can't wait for a response from the callout as that would require the transaction to stall until it completes, which is the reason that synchronous callouts aren't allowed.

The way that I've handled this kind of thing in the past is to have a field on the object that captures the response. I pass the id(s) of the record(s) that I am processing to the @future method. Then when the callout completes, the @future method can retrieve the record(s) from the database and update the field(s).

If you are trying to return a response to the user (as part of a visualforce page maybe) you can still utilise this mechanism, you just have to poll the controller to check the object to see if a response has been received - I've used an actionpoller for this and provided the user with a 'checking' spinner to keep them (hopefully) interested.

It has been a month since the original post, but hopefully this helps someone else out in the same situation. I handle a similar situation when logging into Site Minder protected pages. I use the code below to deal with sites that throw out an unknown amount of redirects.

Http http = new Http();
HttpRequest req = new HttpRequest();
HttpResponse res = new HttpResponse();
// do your initial http call here
req.setMethod('GET');
req.setEndpoint('http://www.cloudywithachanceofcode.com/redirect1.php');
res = http.send(req);
// redirection checking
boolean redirect = false;
if(res.getStatusCode() >=300 && res.getStatusCode() <= 307 && res.getStatusCode() != 306) {
    do {
        redirect = false; // reset the value each time
        String loc = res.getHeader('Location'); // get location of the redirect
        if(loc == null) {
            redirect = false;
            continue;
        }
        req = new HttpRequest();
        req.setEndpoint(loc);
        req.setMethod('GET');
        res = http.send(req);
        if(res.getStatusCode() != 500) { // 500 = fail
            if(res.getStatusCode() >=300 && res.getStatusCode() <= 307 && res.getStatusCode() != 306) {
                redirect= true;
            }
            // I do special handling here with cookies
            // if you need to bring a session cookie over to the
            // redirected page, this is the place to grab that info
        }
    } while (redirect && Limits.getCallouts() != Limits.getLimitCallouts());
}
//congratulations you're outside of the redirects now
//read what you need from the res object
system.debug(res.getBody());

It looks for a redirect HTTP response code and then grabs the location from the headers, and repeats the process until it gets a normal response or falls through without a proper response.

You could add some more checks for other status codes as well. Check here for the status codes if you're not familiar with them.

https://developer.mozilla.org/en-US/docs/HTTP/HTTP_response_codes