bug report^3 on GREP

[上西 俊雄]

Dear Dr. Stepan Kasal

Thank you again for your prompt answer to my second bug report on GREP.

Thank you also for enlightening me for ``export'' and the method of
accessing outside directories in Cygwin Window.

I appreciate your explanation that if it is a bug, how long solution
will take is anyone's guess.

I am an amateur programmer myself in C and C++, and I think I might be
able to  fix it myself if the code is stripped down to sjis locale only,
which would be only an ad-hoc solution at best. I am told that we can
use NTS (null-terminated string) in sjis locale but not otherwise.

Oh, I am not a doctor. I am an ex-editor or ex-employee of a publishing
house. I understand ``locale'' is very important in this I18N world, but
don't you think if we have plural locales for a single Japananese
language, it is a case of oxymoron?

The PDF file of my proposal of Expanded Hepurn System is made by dvipdfm
from dvi file, and the multibyte characters (sjis) are not embedded.

                                           Sincerely yours,

                                                        KAMINISHI Toshio

======================== <address@hidden> ========================