#290 Japanese support enhancement

[Junya Ishihara]

The enhancement which users can see:

1) Support Japanese emails in other charset(utf8, euc-
jp, shift_jis) than iso-2022-jp.

2) Parse Japanese subject with Kakasi.

The enhancement for source code maintenance:

3) Decode and parse Japanese emails line by line. Not
the whole content at once. This will enable support of
base64 and quoted-printable Japanese emails later.

4) Explicitly declare the locale in start sub function in
Bayes.pm. This removed all "no locale" branch in
Japanese mode, and made the source code more simple
and clear. Also, it can avoid bugs which may be going to
be inserted in the future.

[Junya Ishihara]

diff file from Bayes.pm in 0.21.1

[Junya Ishihara]

diff file from MailParse.pm in 0.21.1

[Junya Ishihara]

Modified Bayes.pm based on 0.21.1

[Junya Ishihara]

Modified MailParse.pm based on 0.21.1

[Junya Ishihara]

Modified Bayes.pm based on 0.21.1 (2)

[Junya Ishihara]

Logged In: YES
user_id=723403

setlocale( LC_COLLATE, 'C' );

instead of

setlocale( LC_COLLATE, '20932' );

This is to avoid Perl crash when non-Japanese emails(eg.
Korean) arrive.

[Junya Ishihara]

diff file from Bayes.pm in 0.21.1 (2)

[Junya Ishihara]

Modified MailParse.pm based on 0.21.1 (2)

[Junya Ishihara]

Logged In: YES
user_id=723403

Modified MailParse.pm based on 0.21.1 (2)

This is not to show warning message when $argument is null.

$argument = parse_line_with_kakasi( $self, $argument ) if (
$self->{lang__} eq 'Nihongo' && $argument ne '' );

[Junya Ishihara]

diff file from MailParse.pm in 0.21.1 (2)

[John Graham-Cumming]

Logged In: YES
user_id=578491

This is now merged.