Re: Locale problem and sent index
On Saturday, May 19, 2007 at 7:41:21 +0900, Henry Nelson wrote:
> Content-Type: text/plain; charset=unknown-8bit
Ahum? You turned this knob with the red [don't touch] warning? ;-)
>> the /a ring/ in EUC-JP: The 3 bytes sequence 8F AB A9.
> But it's not EUC-JP, right? Didn't we decide it was EUC-JP-MS?
Straight EUC-JP. The -MS variant, being a perfect superset, also
contains the same character coded by the same sequence.
> I 'cat' "ring-above-a.euc-jp", I get that double-width centered dot.
With Gothic, me too. With Lucida, a with ring above.
> In 'less', I see "H<8F><AB><A9>kedal"
Probably either a broken LESSCHARSET/CHARDEF setup, or a not good
enough EUC-JP locale on the NetBSD libc.
>>> With "Lucida Console", are those double-width glyphs?
> what do you actually get displayed on _your_ monitor.
I see double-width glyphs when the selected charset is EUC-JP. But
single-width glyphs when the charset is Latin-1, UTF-8, and such.
Those characters are said to have an "East Asian Ambiguous" width:
- They naturally take 1 column in their original language. Exactly like
their straight not accented versions, the ASCII letters a and o (and K,
and c), also take 1 column. Of course the libc wcwidth(0xE5) call
accordingly returns 1 (U+00E5 is the /a ring/).
- But CJK terminals display them on 2 columns (for no good reason, but
self-compatibility). And CJK locales declare them wide: The libc
wcwidth(0xE5) call then returns 2.
Bye! Alain.
--
Give your computer's unused idle processor cycles to a scientific goal:
The Folding@home project at <URL:http://folding.stanford.edu/>.