Last modified: 2014-04-29 03:00:29 UTC
I have a database that contains several titles with unicode non-breaking spaces. The following query shows 13 rows: mysql> select page_id, page_title from page where page_title like CONCAT('%', convert(unhex('c2a0') using utf8), '%'); CKEditor used to be used on this wiki and I suspect it is to blame. These pages aren't accessible in the wiki, but they do turn up search results. Converting the rows so that the character is replaced with "_" makes the page accessible. A conversion should probably be performed by upgrade.php.
Where is upgrade.php in the codebase?
I should have written "update.php". And I discovered maintenance/cleanupTitles.php which fixes the problem. Adding a notice to update.php that the user should run cleanupTitles.php would help.