SEO and Duplicate Content

This is potentially an issue, but it has more to do with how the module is implemented than the module itself.

I'm open to using this thread to explore options for mitigating this problem. Perhaps force the SiteMap to one site.

Really what you want are teaser pages, views, and the like that share elements across different affiliates. Some other answers below.

With your modules can you specify a "parent" or "source" site? Sort of like a default...

No. I was actually just thinking about this. With the new hook_url_alter() [patch required], we can force this behavior: Select the first match in the domain list and send all links to that domain. This might be made an optional setting for people worried about SEO. Take a look at domain_url_alter() and you can see how this would work.

Is your module aware if the current subdomain exists or not? If it is aware and the subdomain doesn't exist, can it redirect to a parent or default site (usually www. I'd imagine)...

In the versions released on Drupal.org, yes. You cannot navigate to an "unregistered" domain. Skirt still runs on alpha software -- not the beta release candidate.

Is there some kind of HTML element which you can use to make a page "claim" its source as another? This way when Google visits it can see that although houston.skirt.com is hosting node/20, its actually blah.skirt.com which is the REAL host and deserves all the credit.

I don't know. I'd love it if you could answer that question.

someone could maliciously release a page on their site pointing to hundreds of random sub-domains

This can't happen in the version released through drupal.org. Take a look at domain_init() and the 'valid' flag.

I think that the hook_url_alter() function will solve most of this issue.

For testing, change line 951 of domain.module:

  // This routine only needs to be run from certain urls.
  $check = domain_grant_all();
  if ($check) {
    // Check to see if this is a node or comment link and set $nid accordingly.
    // We static the $nid results to make this more efficient.

To:

  // This routine only needs to be run from certain urls.
  $check = domain_grant_all();
  if ($check || TRUE) {
    // Check to see if this is a node or comment link and set $nid accordingly.
    // We static the $nid results to make this more efficient.

In the release version, I'll change that TRUE to a setting option.

Working in DEV. Will commit to HEAD later today. Here's the behavior:

-- If SEO settings are turned on, all node links are rewritten as absolute URLs.
-- If the node is assigned to 'all affiliates' the link goes to the root domain.
-- If the node is assigned to a single affiliate, the link goes to that affiliate.
-- If the node is assigned to multiple affiliates, the link goes to the first matching domain.

By default, this behavior is turned off.

You probably could add this feature, but for most uses it is really overkill.

Perhaps the domain sorting function could be used here as well.

Committed to HEAD. If we need more advanced functionality, let's open a new issue.

Really. That doesn't happen on my system. Some questions:

-- Are you using clean urls?

-- Are you running Drupal in a subfolder?

That check happens in the function below, which may need some alteration.

Can you try editing this on your install?

/**
 * Determine an absolute path for a domain
 *
 * @param $domain
 *  The currently active $domain array, provided by domain_lookup().
 *
 * @ingroup domain 
 */
function domain_get_path($domain) {
  global $base_url;
  $_url = parse_url($base_url);
  // If in a subfolder, the web server needs the trailing slash.
  (!empty($_url['path'])) ? $_path = $_url['path'] .'/' : $_path = $_url['path'];
  $path = $domain['scheme'] .'://'. $domain['subdomain'] . $_path;
  return $path;
}

I'm testing with subfolders, so it may be that $_path _always_ needs the trailing slash, or that the !empty check is not sufficient.

Well, in a subfolder installation, $base_url returns: example.com/folder

So I was checking for that and adding a '/'

On my test machine, I run in a subfolder and do not set $base_url manually.

global $base_url;
print $base_url;

Returns:

http://example.com/drupal-5.3

So 'path' tests as TRUE in my usage.

print_r(parse_url($base_url));

Returns:

Array ( [scheme] => http [host] => example.com [path] => /drupal-5.3 )

Testing some alternative code now.

Perhaps this is the right answer:

function domain_get_path($domain) {
  global $base_url;
  $_url = parse_url($base_url);
  // If in a subfolder, the web server needs the trailing slash.
  if (empty($_url['path'])) {
    $_path = '/';
  }
  else {
    $_path = $_url['path'] .'/';
  }
  $path = $domain['scheme'] .'://'. $domain['subdomain'] . $_path;
  return $path;
}

Marking as fixed.