On Fri, Mar 11, 2011 at 13:00, Jakub Jelinek <jakub@xxxxxxxxxx> wrote: > On Fri, Mar 11, 2011 at 09:07:35AM +0200, Dotan Cohen wrote: >> I cannot reproduce that: > > That's because you forgot to call setlocale (LC_ALL, ""); > first. Without that the comparison is done in C locale always. > I thought that I could set the locale in the bash environment. Other apps, such as yum, do respond to that, though they may be coded to check. Thanks. > As for why in tr_TR this is different, in tr_TR.UTF-8 > wcscasecmp (L"hijkl", L"HİJKL") == 0 > and > wcscasecmp (L"hıjkl", L"HIJKL") == 0 > (i.e. uppercase for i is LATIN CAPITAL LETTER I WITH DOT ABOVE > and lowercase for I is LATIN SMALL LETTER DOTLESS I). > This is especially problematic because i and I are single byte > in UTF-8 while İ and ı are two byte. > Thank you for that explanation. This is very important to know, and I concede that it weakens my argument in favor of a case-insensitive option for yum. In fact, I am no longer of that opinion, I am now convinced. Thanks. This is why I enjoy discussing / arguing with intelligent people. I learn new things. -- Dotan Cohen http://gibberish.co.il http://what-is-what.com -- users mailing list users@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
