Re: 16bit ctype table

[yamt%mwd.biglobe.ne.jp@localhost (YAMAMOTO Takashi)]

hi,

> Takehiko NOZAKI <takehiko.nozaki <at> gmail.com> writes:
> 
>> > ftp://ftp.netbsd.org/pub/NetBSD/misc/tnozaki/patch-insufficient_ctype_bits
> 
> It’s not insufficient. 8 bits are enough.

while it might be possible (i haven't checked), i'm not sure if it's
a good idea as a conversion from rune would be more costly than proposed
16 bit one.

YAMAMOTO Takashi

> 
> /* from Bruno Haible's libutf8 */
> 
> /* How the 12 character attributes are encoded in 8 bits: Every attribute is
>    represented by an "include bitmask" and an "exclude bitmask".
>     Attribute   bit/formula             comment
>      upper      upper && !lower         implies towlower(x) != x == 
> towupper(x)
>      lower      !upper && lower         implies towlower(x) == x != 
> towupper(x)
>      title      upper && lower          implies towlower(x) != x != 
> towupper(x)
>      alpha      bit                     superset of upper || lower
>      digit      xdigit && !alpha        '0'..'9' and more
>      xdigit     bit                     '0'..'9','a'..'f','A'..'F' and more
>      space      bit                     ' ', '\f', '\n', '\r', '\t', '\v'
>      print      bit
>      graph      print && !space
>      blank      bit                     ' ', '\t'
>      cntrl      bit                     0x00..0x1F,0x7F
>      punct      print && !(alpha || xdigit || space)
>      alnum      alpha || xdigit
> */
> 
> See https://www.mirbsd.org/cvs.cgi/src/lib/libc/include/mir18n.h?rev=HEAD
> 
> bye,
> //mirabilos