|Problems with Like Button & Canonical URL|
| 8:58 pm on Mar 18, 2011 (gmt 0)|
Ok, we're having all kinds of problems with a FB Like button, and maybe someone else has had similar.
We have a site on a subdomain, call it splunge.example.com. There are no Facebook buttons, boxes, or code on this subdomain.
We created another subdomain, call it lumberjack.example.com. This subdomain was created by copying the contents of the splunge folder to the lumberjack folder, editing them, and then setting up the new subdomain.
We want a Facebook Like button on lumberjack.example.com. I got the code for the button, and the Open Graph code, and handed it off to someone to add, which he did. However, he mistaken left a canonical url meta tag with splunge.example.com on the home page.
The end result of this was that when you click on the Like button, the content that gets 'liked' is splunge.example.com - because the button will apparently use the canonical url if specified even over the one you specify in the Open Graph header code.
So okay, I think this is kinda stupid, but I went with it - we removed the canonical tag, regenerated all the FB code, but we're STILL having the same behavior - if you go to the lumberjack.example.com site and click on the Like button, the content that gets 'liked' is splunge.example.com. Even though there is NO mention of splunge anywhere on the lumberjack site.
So it is poisoned? I have tried this on numerous connections, browsers, several different FB accounts.
Really weird. And aggravating.
| 9:06 am on Mar 19, 2011 (gmt 0)|
perhaps FB will update the canonical url if you enter it in the URL Linter:
| 3:19 pm on Mar 19, 2011 (gmt 0)|
I can't remember if that requires every page link or not though, it's been awhile.
Cache may play a role here too. There are a few things you can do, one of them is using curl to auto-process the links. More information here [developers.facebook.com] under the When does Facebook scrape my page? section. And here [developers.facebook.com] under Editing Meta Tags.