Bug#99933: second attempt at more comprehensive unicode policy
At 11:55 PM 1/9/2003 -0500, Colin Walters wrote:
What do you expect GNOME programs to do? Since they fully support
UTF-8, you can input any Unicode character you want. Also, a program
like Evolution may receive a file in mail whose name uses Unicode
characters. And a lot of locale charsets (like ISO-8859-1) will not be
able to encode the string. The only sane solution is to just use UTF-8
for filenames.
You can input any Unicode character you want, but you probably have
to out of your way to input something outside your charset (i.e. probably
not on your keyboard or standard IM.) If I receive a file in the mail whose
name is not in ASCII (which has never happened to me), I would rename
it before saving it, so I could access it easily. How many people in a
Latin-1 locale who got an email with a Chinese file name would want it
saved with the original name? A simple hash - say, out of charset
characters to _ - would probably be fine.
But I am curious about your feelings on programs writing data in general
to the terminal; you feel they should not never to convert it to the
locale's charset, and we should just mandate that people using legacy
terminals use that filterm or whatever thing?
If you're dealing with a web browser, or a mail reader or anything else
that handles tagged data, it should convert it, of course. Anything else
should be in the locale charset, and manually recoded if necessary.
I'm not sure I really understand what you're asking here.
David Starner - starner@okstate.edu
(starner@okstate.edu may be disappearing soon - dvdeug@email.ro will work,
but is not suitable for high-volume traffic.)
Reply to: