> On necessity of isHalfwidth/isFullwidth (or XPG/POSIX wcwidth/wcswidth),
> even though I understand and agree to your point, I believe we
> need those kind of interfaces to support at least text terminal and
> their emulators that need to have fixed width font(s) in practice
> probably one for each different locales or at least territories group.
int wcwidth(wint_t wc);
"This function determines the number of column positions required for the
wide-character code wc. ..." -- X/Open Internationalisation Guide
All those who think it is silly to try to define a table across
all of Unicode to specify the return value for this function,
please raise your hands. If not, please tell me what the correct
answers are for Devanagari or Arabic.
Such POSIX interfaces have grown up around an impoverished concept
> I also think, in that sense, it might be a good idea that if UTC will
> provide a few sample width tables with a disclaimer of restricted usage
> of such tables. (I believe any vendor that implements POSIX and
> Unicode/UTF-8 locale already has it for their wcwidth/wcswidth
I would suggest rather that anyone who feels (for POSIX compliance)
that they must support such things, should convert their wcwidth
tables for their ja_JP locale, etc. to UTF-8 and just use that
as their answer.
Or perhaps a vendor that implements POSIX and Unicode/UTF-8 locale
would care to volunteer instead what their suggested answer is
and post a public table that other such vendors could agree on.
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:37 EDT