The problem ended up being that deedscraper determines the appropriate language
by checking for the "lang" attribute in the html node from the deed pages. I
changed it to look for "xml:lang" as well, and it seems to work correctly now.
Bumping and reassigning this ticket to Lunpa. This ticket is several months old
now. CC Korea brought it up back in April, and just brought it back up today.
To reproduce the bug do this:
1) Visit: http://www.cckorea.org/xe/?mid=main
2) Click on the license badge in the footer
The attribution information from the scraper is not localized. Apparently this
is because the url fetching data from the scraper does not have a lang=
parameter passed to it. cc.engine should write out a lang= parameter for the
scraper URL that is equivalent to the current language of the deed.