It appears as though there is a conflict with globalredirect. If the following stack is enabled:

* pathauto
* path_redirect
* globalredirect
* PURL

... then any path created via pathauto will invoke an infinite redirect loop. Disabling globalredirect fixes the problem.

However, disabling globalredirect means that individuals can access (and link to) 'node/##' paths without it redirecting to the aliased path, causing search engines to see duplicate content.

FYI, a redirected path via path_redirect does not invoke an infinite redirect loop, only globalredirect.

Comments

yhahn’s picture

Assigned: Unassigned » yhahn

Will look at this... in the meantime, implementing canonical URLs is a good way to ensure that search engines get one version of the page.

yhahn’s picture

Status: Active » Closed (duplicate)

Some research indicates that this is an issue in globalredirect that can be fixed with patch in #13 on this queue. Maintainer indicates that it will be committed and included in the next release.

http://drupal.org/node/346911

I have confirmed that using this patch in conjunction with PURL works.