Always specify UTF-8 targt charset for html_entity_decode(); default is 8-bit ISO-8859-1 which causes things to break when we later pass them through things that expect to work with UTF-8. For instance, running through preg_replace() with the /u option results in NULL, leading to problems with OStatus and SubMirror generating their plaintext versions and doing length-cropping.

This commit is contained in:
Brion Vibber
2010-09-30 11:29:31 -07:00
parent 91f25ca817
commit 1acc7d66c6
6 changed files with 7 additions and 7 deletions

View File

@@ -89,7 +89,7 @@ function importActivityStream($user, $doc)
$html = htmLawed($html, $config);
$content = html_entity_decode(strip_tags($html));
$content = html_entity_decode(strip_tags($html), ENT_QUOTES, 'UTF-8');
$notice = Notice::saveNew($user->id,
$content,