http://qa.mandrakesoft.com/show_bug.cgi?id=5581





------- Additional Comments From [EMAIL PROTECTED]  2003-13-09 17:14 -------
In doing some more testing, it seems that if you are already using a UTF-8
encoded locale, it will keep UTF-8 encoding.  So when all my UTF-8 encoding was
taken out of my current locale, localedrake chose zh_TW, zh_CN and en_US (with
ENC=big-5 for the TW locale).  If I was currently using a UTF-8 locale, then
localedrake instead chose zh_TW.UTF-8, zh_CN.UTF-8 and en_US.UTF-8 (with ENC=utf-8).

I suppose this kind of behavior seems reasonable, especially on the surface. 
Then the entire system can keep a UTF-8 encoding no matter the language.  This
would be especially good for things like filenames.  It is definitely important,
however, that we have an option in localedrake to choose or reject UTF-8 when
choosing the language. 

But I still don't think UTF-8 encoding is reasonable for users of Chinese quite
yet (although I hope we will be there soon with Linux).  Let me give three
examples why we should use GB right now and not UTF-8:

1) Currently pretty much everyone in the mainland and Singapore uses GB
encoding, and most everyone else (HK and TW) uses Big-5.  When sharing text
files, or reading files from their Windows partitions, they will have troubles
with UTF-8.  People who know how to convert a file will probaby know how to
change their locale to UTF-8.  Most people have no clue.

2) When doing a Chinese install, Mandrake appropriately uses a zh_CN locale, and
GB encoding.  If another person on the same network installs in English with
Chinese support (which will be fairly common), then uses localedrake to change
their language, it will be zh_CN.UTF-8.  If these two computers share files (via
nfs), then they will have troubles reading each other's files and filenames.

3) I had my friend who is somewhat computer savvy, but not an English speaker,
come over and play with Mandrake 9.2 rc2 to see if it gave him any problems (and
it did very well, by the way, once I fixed a few problems).  I did notice that
when he would open a terminal, he would look at the list, which included
Konsole, and four rxvt terminals: Simplified Chinese, Traditional Chinese,
Japanese, and Korean (the languages I installed), and always choose the
Simplified Chinese one, which is gbrxvt.  So if using a utf-8 locale, gbrxvt,
assuming it is all gb, will display garbage.  I am sure the other rxvt consoles
would be the same.  This is actually a fairly serious issue.

I am interested in these things working out of the box for users of Chinese.  If
it is too complicated, or too broken, they will go elsewhere (as they are
already doing since Chinese was broken in 9.1).

Please make the system so that when using an English install, but also
installing support for Chinese, that the locales are NOT zh_??.UTF-8, but
instead are then zh_CN and zh_TW.  Basically Chinese should not use UTF-8 unless
explicitly requested.


-- 
Configure bugmail: http://qa.mandrakesoft.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.


------- Reminder: -------
assigned_to: [EMAIL PROTECTED]
status: ASSIGNED
creation_date: 
description: 
When installing in English with added support for Simplified Chinese, then
changing the user's language with localedrake, the locale chosen is zh_CN.UTF-8
However, if the installation is purely Chinese, the locale is zh_CN which then
defaults to a GB variant codepage.

Although things seem to work with UTF-8 encoding (except when using a GB rxvt
terminal), it is not currently an appropriate encoding for Chinese as most
people using Chinese use either GB or Big-5 encoding.  And especially since the
default Simplified Chinese installation of Mandrake uses GB encoding,
localedrake should only use zh_CN (not zh_CN.UTF-8) when changing from
en_US.UTF-8 to Simplified Chinese.  I assume the same problem would exist for
Big-5 encoding.

Reply via email to