[EMAIL PROTECTED] wrote:

> So, maybe you can save time and instead of installing something
> new, describe exactly what you do, what you would like to
> happen but what happens instead.

OK.. here�s the detailed and boring description...

My locale is en_GB.UTF-8. My keyboard is us_intl (in X) and
us-intl (on the console).

A) I leave X and go to console by means of CTRL-ALT-F2.

I enter the following:

   cat >test.1 [enter]
   � [enter]
   CTRL-D

The � character is made by first typing ', then e.

I check the result by means of hexdump -C test.1, which gives:

00000000  e9 0a                                             |..|
00000002

B) The same on the console, after running unicode_start, gives the
same result:

00000000  e9 0a                                             |..|
00000002

I.e., the accented character is entered into the file as *one
byte* (0xE9; the 0x0A is of course the result of typing ENTER).
The only difference is that the � character is not *visible*
immediately when I type ', e.

C) Going back to X, after the same in an xterm, hexdump -C test.1
produces:

00000000  c3 a9 0a                                          |...|
00000003

In other words, in xterm (with UTF-8 locale) a proper UTF-8
representation of the � character is generated. Also the character
is visible when it is typed (like A, but unlike B).

Now I assume that the idea of 'putting the console in UTF-8 mode'
is to make the unicode console (case B) behave like the unicode
xterm (case C). But it doesn�t.

For completeness, there is:

D) In a xterm, I type LANG=C xterm, creating a new xterm with "C"
locale. This behaves exactly like case A.

The console *screen* gets properly switched by unicode_start and
unicode_stop, as can be seen by cat-ing the test file in the two
modes. The two-byte version of � is only displayed when unicode is
on, the one-byte version only when it is off.

Regards, Jan



--
Linux-UTF8:   i18n of Linux on all levels
Archive:      http://mail.nl.linux.org/linux-utf8/

Reply via email to