Bruno Jargot wrote at 01/24/07 10:26:
Which of the LC* variables is responsible to set the encoding which an application or shell uses? I never understood why there are different LC* variables.
Hello Bruno, The LC_ALL has the highest priority. And then LC_* (such as LC_CTYPE and LC_COLLATE) and then lastly LANG. This priority definition is described at the setlocale(3C) man page. In C/C++ and such programming, people just call setlocale(3C) to set the locale as you know in their application. In shell, *if needed*, checking on LC_ALL, LC_CTYPE, and then LANG would be a good idea for figuring out the locale and its codeset. Of course, you could also check on LC_ALL, LC_MESSAGES, and then LANG if you mainly concern about messages assuming that people won't put incompatible locale at LC_MESSAGES. The original idea of having multiple categories was that people could mix up different locales as long as they are compatible in terms of codeset such as LANG=en_US.UTF-8 and LC_COLLATE=fr_FR.UTF-8 so that I would be able to use all American English UTF-8 but with French collation or sorting order. Or, LANG=fr_FR.UTF-8 and LC_MESSAGES=en_US.UTF-8 or C and so on so that multiple locales could be used to comprise an environment of my choice. Ienup _______________________________________________ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
