On 13 Mar 2017, at 20:51, Julian Reschke <julian.reschke@xxxxxx> wrote: > > Well, if we allow UTF-16 and UTF-32 - even if we discourage their use - we should say how to detect those... Consistency is the hobgoblin of the little minds. (Apparently, not having that nonsense in there did not hurt for 7159? If people want to implement something useless and cannot figure this stuff out, should we help them?) Putting back any form of apparent support for UTF-16 and UTF-32 is a regression. Grüße, Carsten