Quiz URLs/Paths using Pathauto for Logged-in and Non-logged in users

The reason for the long token on the second is that the token is basically used to prevent caching. Pathauto will not recognize that URL as matching it's automatically generated path, and so you will get the ugly URLs.

If you are up for trying an experiment, you can try this:

1. In a text editor or IDE, open quiz/quiz.module
2. Go to line 988. Find the code that looks like this:

// If anonymous user and no unique hash, refresh with a unique string to prevent caching.
  if (!$quiz->name && arg(4) == NULL) {
    drupal_goto('node/'. $quiz->nid .'/quiz/start/'. md5(mt_rand() . time()));
  }

3. Change it to this:

  if (!$quiz->name && empty($_GET['quizkey'])) {
    drupal_goto('node/' . $node->nid, array('quizkey' => md5(mt_rand() . time()));
  }

4. Save the file and test.

What this should do (if it works at all) is add a query string instead of adding a path. I believe this will generate URLs that look like this:

domain.com/quiz/title-of-quiz?quizkey=717530d18d6e8ec1a1f28af8d6c8d94c

Looking at the code, we might be able to generate something more compact than a full MD5 checksum. But I'll delve into that if the above works.

If the above works (well), I may try and work that into the official Quiz module. I'm not big on embedding params into the URL for no good reason. The strategy I'm using above is closer to what Drupal Core does.

Matt

Waiting on response.

@mbutcher

Your suggestion seems to be working, actually you are suppose to change line no 1094 which looks as below

if (!$user->uid) {  
  //drupal_goto('node/'. $quiz->nid .'/quiz/start/'. md5(mt_rand() . time()));  #was
  drupal_goto('node/' . $quiz->nid, array('quizkey' => md5(mt_rand() . time())));  #now
}

your point 3 in #1 says $node. Its not $node it should be $quiz.

If this works, we should check it in. It would be nice to have pathauto working correctly all the time.

I believe sivaji is going to add his patch into Quiz when his exams are done.

If I understsand your second request, you want some way of figuring out whether the current page is the results page? Is that right?

cheuyungphat: Yeah, file that as a separate issue and we'll see what we can do.

thanks code has been committed, changing its status to fixed.