Am Montag, 27. März 2017, 09:36:39 CEST schrieb Philipp Thomas:
- Carlos E. R. (email@example.com) [20170325 15:44]:
the rest is posix
And that is sensible. Note that setting locale for root to posix was done on purpose mainly because sorting behaves according to locale so a regex like [a-z] only works as expected when sorting is done according to posix. Still today many scripts silently assume a posix locale instead of either setting the affecting locale fascet or using something like [::lower::] which Posix introduced for exactly thios case.
....then the admin should make sure the script works before he runs it.
For a normal user the broken UTF8-characters are more of a problem than a sort that may behave different (reminds me that Python has a Natsort-Lib for this - I assume that deals exactly with this problem).
So, if we want to focus on useability, we should challenge the 'by purpose'.