Re: Huge auto-fill-mode bug intro'd in 20.0 still not fixed in 21.2.19
>>>> In
<14309.45522.386054.259893(a)turnbull.sk.tsukuba.ac.jp>
>>>> Stephen = "Stephen J. Turnbull"
<turnbull(a)sk.tsukuba.ac.jp> wrote:
Stephen> " GAI KOKU JIN " should become " GAI KOKUJIN ", the
Stephen> meaning taken will possibly change.
...
Stephen> (I use JIS X 0208 "full-width" characters for Japanese romaji,
If this means "ＧＡＩ ＫＯＫＵ ＪＩＮ" you can (or should) change
word-across-newline (WAN) not to include full-width alphabets,
e.g. "\\(\\cK\\|\\cH\\|\\cC\\|\\cS\\| \\|\\cc\\|\\ct\\)".
# I don't know how to express "\cj except alphabets and arabic
# numerals"...
Stephen> Personally, I make a habit of setting spaces around
Stephen> English embedded in Japanese;
Me, too.
Currently, newline to space replacing rules are below (from
fill.el):
(1) WAN +NL+WAN --> WAN + WAN
(2) not(WAN)+NL+WAN --> not(WAN) + WAN
(3) WAN +NL+not(WAN) --> WAN + not(WAN)
(4) SPC +NL+not(WAN) --> SPC + not(WAN)
(5) not(WAN)+NL+not(WAN) --> not(WAN) + SPC + not(WAN)
If (2) and (3) rules can be changed to insert SPC by option,
it will be acceptable for more people, IMHO.
--
大石一雄 (OISHI Kazuo)