#3622 Japanese locale of [clock format] is wrong

[ohtsuka yoshio]

Tcl8.5a5/Windows2000

Hi! I'm Japanese.
I'm glad that clock have japanese locale.
But the locale have some mistakes.

1. In Modern Japan, \u5eff and \u5345 are not used.
2. Japanese month names from ICU2.8 have arabic numbers,
but other parts(years, days, times, etc) have chinese
numbers. So they should be unified by arabic or chinese.
In my opinion, arabic number is good. Because other
date command(GNU,Solaris,etc) use arabic number for %Ec.
3. Japanese Era days are wrong. need +0900 timezone.
Meiji 1873-01-01 00:00:00 +0900
Taisyo 1912-07-30 00:00:00 +0900
Shouwa 1926-12-25 00:00:00 +0900
Heisei 1989-01-08 00:00:00 +0900 (not 1989-01-07)
4. This year is Heisei 19, but output was Heisei 20.
5. some formats(%Ec,%Ex,etc) aren't formats used well in Japan.
I patched this. See attached files, please.

Attached files are patched loadICU.tcl and outputted ja.msg.

[ohtsuka yoshio]

patched loadICU.tcl and ja.msg

[Kevin B KENNY]

Logged In: YES
user_id=99768
Originator: NO

Domo arigato gozaimashita, 'nyademo'-san.

I'm applying your patch essentially unchanged. I'm a little confused by (2),
though, because %Om, %Od, %Oh, %Om, %Os use Han numerals for the other
parts of the time. If you think that it's correct that dates should all
be Romanji even if the user has requested locale-specific formatting,
then I'll go with it, but I thought that perhaps it was wise to have
Kanji available.

I'm also adding a test case for the transition dates.

I suppose I can excuse the mistake with the +0900 time zone by
observing that the last day of Showa is also the first day of Heisei?

[Kevin B KENNY]

Logged In: YES
user_id=99768
Originator: NO

Domo arigato gozaimashita, 'nyademo'-san.

I'm applying your patch essentially unchanged.
I'm a little confused by (2), though, because %Om, %Od,
%Oh, %Om, %Os use Han numerals for the other
parts of the time. If you think that it's correct
that dates should all be Romanji even if the user has
requested locale-specific formatting, then I'll go with
it, but I thought that perhaps it was wise to have
Kanji available.

I'm also adding a test case for the transition dates.

I suppose I can excuse the mistake with the +0900 time zone by
observing that the last day of Showa is also the first day of Heisei?

[ohtsuka yoshio]

Logged In: YES
user_id=932722
Originator: YES

Thank you for your reply!!

>If you think that it's correct
>that dates should all be Romanji even if the user has
>requested locale-specific formatting, then I'll go with
>it, but I thought that perhaps it was wise to have
>Kanji available.

Ah, I forgot to think deeply about %Om, %Od, %Oh, %Om, %Os.
Fuumm, It seems to me that it's preferred that these formats
output Kanji Numbers, as you said. sorry.

>I suppose I can excuse the mistake with the +0900 time zone by
>observing that the last day of Showa is also the first day of Heisei?

It seems to me that if both of Showa and Heisei era days
are wrong, the test goes success...