Re: i18n + widechars
Thomas Dickey <[email protected]>
| Newsgroups | gmane.network.tin.devel |
|---|---|
| Message-ID | <[email protected]> |
On Wed, 6 Apr 2005, Michael Bienia wrote: > On 2005-04-06 13:40:23 +0200, Urs Janßen wrote: >> In <[email protected]>, Urs Janßen wrote: >>>> I can't reproduce it in my normal utf-8 environment. Do you have some >>>> more information about how to reproduce it? >>> >>> not really. my normal env. is a non utf-8 env. I started a new shell in >>> an xterm via >>> xterm -u8 -fn "-Misc-Fixed-Medium-R-Normal--18-120-100-100-C-90-ISO10646-1" >>> and set the floowing locales: >>> LC_CTYPE=de_DE.utf-8 >>> LC_MESSAGES=de_DE.utf-8 >>> tin was linked against libncursesw and I got the 'grabbled' output. >> >> /s/libncursesw/libncurses - the problem only shows up if linked >> against libncurses, not if linked against libncursesw. > > I assume this is a problem within libncurses. If I let tin print the > complete string (options_menu.c::311) also to stderr I get there the > expected result: all ':' are in one column. It's not a "problem", but a limitation. UTF-8 locales are designed for storing more than 8-bits worth of information in a cell. The line-drawing characters for instance do that. Depending on how old the libncurses is, it will/won't have some code to make it do the expected thing with line-drawing characters. (ncurses 5.4 has some fixes for this, and I made other fixes after 5.4 which some distributions have packaged). -- Thomas E. Dickey http://invisible-island.net ftp://invisible-island.net