Returns canonical link with success, but is the wrong URL?

[jobjorn]

Hopefully I've just missed something, but URLResolver appears to be unable to process some links correctly. For example this news article: http://www.bbc.co.uk/news/uk-politics-33686708#sa-ns_mchannel=rss&ns_source=PublicRSS20-sa

If I run URLResolver and compare it with fetching the canonical link another way as in this gist,
https://gist.github.com/jobjorn/a1c67b4eb374bc99fdea , I get different results:

URLResolver: http://www.bbc.co.uk/news/uk-politics-33686708#sa-ns_mchannel=rss&ns_source=PublicRSS20-sa
cURL and DOM: http://www.bbc.co.uk/news/uk-politics-33686708

As you can verify by looking at the article yourself, the latter one is the correct one.

Any idea what that could be about?

Regards and with thanks for your work,

Jobjörn

[mattwright]

Hi Jobjörn,

The library was coded this way because typically the stuff after a '#' in a URL is a 'bookmark' (fragment identified) within the page, or with some web sites, can even alter what is displayed with JavaScript. You can read about that here:

https://stackoverflow.com/questions/5283395/url-hash-is-persisting-between-redirects

I'm hesitant to just strip these out, but I could consider adding a couple routines that would let you pull the Canonical URL without the fragment identifier... The place in the code where we add it back in is within the fullyQualifyURI() method if you want to change the behavior yourself. Thoughts?

Matt

[jobjorn]

That sounds reasonable to me!