StatusNet chooses the first content element in an Atom feed, while
it should really choose the 'html' representation for its 'rendered'
and 'text' representation for the (text-only) 'content'.
GNU social will implement a better algorithm for retrieving Atom
feeds, but that is yet to be done. So to avoid having link-less posts
on remote nodes, we'll just do the old switch-a-roo.
Other Atom readers, such as Mozilla Firefox, has the reverse priority
(choosing the last of the content elements).
// XXX: assuming HTML content here
$xo->element(
ActivityUtils::CONTENT,
// XXX: assuming HTML content here
$xo->element(
ActivityUtils::CONTENT,
- array('type' => 'text'),
- html_entity_decode(strip_tags($this->content), ENT_QUOTES, 'UTF-8')
+ array('type' => 'html'),
+ common_xml_safe_str($this->content)
);
$xo->element(
ActivityUtils::CONTENT,
);
$xo->element(
ActivityUtils::CONTENT,
- array('type' => 'html'),
- common_xml_safe_str($this->content)
+ array('type' => 'text'),
+ html_entity_decode(strip_tags($this->content), ENT_QUOTES, 'UTF-8')