Description of problem: LC_COLLATE for cs (Czech) locale is buggy wrt. uppercase letters, it incorrectly sorts uppercase letters before lowercase ones. Version-Release number of selected component (if applicable): glibc-2.4-8 Steps to Reproduce: 1. LC_ALL=cs_CZ.UTF-8 sort <<EOF jakub Jakub EOF Actual results: Jakub jakub Expected results: jakub Jakub Additional info: The Czech sorting algorithm should put lowercase letters first (when the sorted words are otherwise the same except possible non-alphanumeric characters). For more info about the Czech sorting, see Petr Olsak's article in the CSTUG bulletin (in Czech). http://math.feld.cvut.cz/ftp/olsak/bulletin/csr.ps Or see the examples from Jan Pazdziora's Czech sorting page: http://www.fi.muni.cz/~adelton/l10n/cssort/sort.tab Except for the above error, I think the Czech locale provides correct sorting. This behaviour is actually documented as known bug in /usr/share/i18n/locales/cs_CZ (see the comment labeled "3.7").
It's trivial to change but you have the get the last maintainer (Vladimir Michl) to agree. Only if he does or doesn't reply will I change the file.
Sorry for the delay - I have now tried to cotnact Vladimir Michl, and asked him to add a note to this bug or contact you by email.
Hello, as I see Yenya is right. And I think we should change it as he requests. Just after change, so lowercase letters are before uppercase, can we delete point 3.7 at the start as well. Thanks Vladimir Michl
Should be fixed in upstream cvs.
In glibc-2.6.90-17 in rawhide, please test.
Works for me, thanks!