Last modified: 2014-02-12 23:38:32 UTC
During editing there might be warnings reported back from the API, and those should be fatal for the present request but be non-fatal if requested anew with some kind of token. This is basically the same as whats done during image uploads. Examples for this kind of warnings are similar/equal label-description -pairs, overly long labels and descriptions, exceptions from the AbuseFilter, and similar.