2016-11-22



Had a look at sort(1) today, because someone discovered  that  it
ignores spaces for the sorting order in a UTF-8 locale, [0]  thus
the  following  names  are  sorted wrong, according to the German
dictionary rules:

     $ LC_COLLATE=de_DE.UTF-8 sort <<!
     Bachmann Peter
     Bach Silke
     Bach Anton
     !
     Bach Anton
     Bachmann Peter
     Bach Silke

A workaround for this special case is sorting by specific fields:

     LC_COLLATE=de_DE.UTF-8 sort -t ' ' -k 1,1 -k 2,2

But *why* are spaces ignored in UTF-8 sorting?


[0]  https://debianforum.de/forum/viewtopic.php?f=34&t=163046


http://marmaro.de/lue/        markus schnalke <meillo@marmaro.de>