Last modified: 2014-10-22 20:39:05 UTC
When displaying 'non valid' usernames after a cohort validation we are displaying a stream of bytes rather than the user name. See below: { "mediawiki_username": "\u062e\u0644\u0648\u062f \u0637\u0627\u0631\u0642 \u0639\u0628\u062f \u0627\u0644\u062d\u0643\u064a\u0645 \u0645\u062d\u0645\u062f", "reason_invalid": "invalid user_name" },
Since your example looks like perfectly fine JSON for an arabic username to me (ignoring the spaces, which I attribute to copy/paste), I assume you are referring to that you'd prefer to see the unicode characters unescaped?
Right, cause this is a report that we give to our users so names have to be "legible".
This issue has been solved together with 60628. Adding an HTML page (instead of JSON) for invalid user lists renders non-ascii characters as unescaped unicode.
