On Mon, Jun 11, 2001 at 21:20:16 +0200, Joerg Wunsch wrote:
> > locale in each program which tries to deal with it directly, since
> > suppose that program have the same default assumption (from where?)
> > for missing information.
> Sorry, i can't follow you. Which program is trying to guess local
> names, and how could i get a problem by providing a user-friendly
> alias name (i. e., a symlink)? And, why do all the other Unices
> don't have a problem with it?
This is not locale data problem so can't be resolved via symlink. This is
locale name problem. With partial locale name given program can't sense
some info from missing parts of it, i.e. can't determine codeset and
territory in worst case.
All Unices have problem with it since it is standard relaxation provided
historycally. Recent standartization efforts attempts to compensate such
bad practice, f.e. nl_langinfo(CODESET) allows you to get codeset now, but
nothing yet allows you to get territory of shortest locale name. Old
programs are not aware about nl_langinfo(CODESET) and attempts to get all
info from env. variable value. In case there is short form, programs make
various assumptions which are different and often not right.
Andrey A. Chernov
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message