yet another Curl question

[marcnyc]

Hello folks,
there is a website where I have an account. On this website I log in with my email as username and a password.
I am trying to retrieve some data from this website with curl and show it on my website.
I am using a regular curl job (I followed the php manual and its notes plus some tutorials I found on the web) but for some reason I am unable to log in with curl.
I have been trying setting CURL to POST the variables but had not luck.
I noticed that I can log into the website by entering an URL in the browser of this type: domain.tld/user=[e-mail]&pw=mypass&go=1 so I also tried with GET in the CURL.
Anything that I tried had the same outcome, I simply get "1". Anything I do returns a blank page with the number 1 on it.
Reading the manual about Curl I found that option to set the connection type as GET or as CONNECT and I also tried those, but I get the same result all the times.
I was simply wondering if anybody knows of any other way of connecting to a site besides GET/POST/CONNECT. Curl seems so easy and straightforward to me that I can't really imagine why it wouldn't work.
Thanks

[tsinka]

Do you login through a HTML form or do you get a httpauth popup dialog where you login ?

Check this thread (hope it points in the right direction):

http://www.phpbuilder.com/board/showthread.php?s=&threadid=10288364

Thomas

[AstroTeg]

Can we see just your curl code (the good stuff that references curl_curl function). I think you might be missing a curl_setopt() setting which tells curl to return the page results. Make sure you have this:

curl_setopt($curl, CURLOPT_RETURNTRANSFER, 1);				// allows us to capture the HTML

Without it, I think you'll end up getting either a 1 or a 0 if curl was successful in accessing the page or not.

[marcnyc]

Ok here is the code:

$anti_spam_measure = 'myusername';
$url_calendar_login = 'http://industrial.org/login.php';
$urlstring_calendar_login = 'u_email='.$anti_spam_measure.'%40mydomain.com&u_password=mypassword&go=1';
	$ch_calendar_login = curl_init();
	curl_setopt($ch_calendar_login, CURLOPT_URL,$url_calendar_login);
	curl_setopt($ch_calendar_login, CURLOPT_HEADER, 0);
	curl_setopt($ch_calendar_login, CURLOPT_POST, 1);
	curl_setopt($ch_calendar_login, CURLOPT_POSTFIELDS, $urlstring_calendar_login);
	curl_setopt($ch_calendar_login, CURLOPT_FOLLOWLOCATION, 0);
	curl_setopt($ch_calendar_login, CURLOPT_REFERER, 'http://industrial.org/login.php');
	curl_setopt($ch_calendar_login, CURLOPT_RETURNTRANSFER, 1);
	curl_setopt($ch_calendar_login, CURLOPT_COOKIEFILE, "");
	$data_calendar_login = curl_exec($ch_calendar_login);

if ( substr_count(strtolower($data_calendar_login), strtolower('logged in')) >= 1 ) echo "YEAH!!!";

echo 'data is '.$data_calendar_login;

curl_close($ch_calendar_login);

exit;

[tsinka]

Two problems:

1. you need FOLLOWLOCATION because there's a redirect header
2. use www.industrial.org. curl sees .industrial.org as cookie domain the leading dot confuses curl. www.industrial.org works.

<?PHP
$anti_spam_measure = 'myuser';
$url_calendar_login = 'http://www.industrial.org/login.php';
$urlstring_calendar_login = 'u_email='.$anti_spam_measure.'%40mydomain.com&u_password=mypassword&go=1';
    $ch_calendar_login = curl_init();
    curl_setopt($ch_calendar_login, CURLOPT_URL,$url_calendar_login);
    curl_setopt($ch_calendar_login, CURLOPT_HEADER, 0);
    curl_setopt($ch_calendar_login, CURLOPT_POST, 1);
    curl_setopt($ch_calendar_login, CURLOPT_POSTFIELDS, $urlstring_calendar_login);
    curl_setopt($ch_calendar_login, CURLOPT_FOLLOWLOCATION, 1);
    curl_setopt($ch_calendar_login, CURLOPT_REFERER, 'http://www.industrial.org/login.php');
    curl_setopt($ch_calendar_login, CURLOPT_RETURNTRANSFER, 1);
    curl_setopt($ch_calendar_login, CURLOPT_COOKIEFILE, "");
    $data_calendar_login = curl_exec($ch_calendar_login);

if ( substr_count(strtolower($data_calendar_login), strtolower('logged in')) >= 1 ) echo "YEAH!!!";

echo 'data is '.$data_calendar_login;

curl_close($ch_calendar_login);

exit;
?>

Thomas

[marcnyc]

Wow, thanks a lot, those two things really did fix it.
I understand the thing about www but can you explain the thing about FOLLOWLOCATION (just so I understand and learn from this).

After reading a few tutorials about Curl, I thought that setting it to 0 is needed to avoid that on execution the Curl script follows header location commands and leaves my site (where the script is) to go to the site the script is accessing... I now obviously know I am wrong because I have seen that it doesn't go there anyway, but if you cared to explain how I misunderstood this option it'd be great...
Also how did you know there is a redirect? Couldn't it have been a refresh meta tag?

[tsinka]

I enabled

CURLOPT_HEADER

That prints the headers and that showed me that there was a redirect.

Without FOLLOWLOCATION curl doesn't follow that redirect. With FOLLOWLOCATION set to 1 curl follows the redirect.

In PHP that redirect would look like:

header("Location: /somescript.php");

The login script of that site checks if the submitted data is valid and then redirects to (I think) user.php.

The use of CURLOPT_FOLLOWLOCATION shouldn't affect the behaviour of your script. The curl options affect the curl session only ... not your script.

One example of when to set this option to 0:

You login with curl and then don't want to get the contents of the page the site redirect you to. You just want to get the session cookie and then to get the contents of another page that needs a valid session cookie.

Thomas

[marcnyc]

Thank you for the explanation. I understand clearly now. Thanks for your time. I appreciate.