'KE' => [ N_("Kenya"),          'en_ZW', '3' ], #
'KG' => [ N_("Kyrgyzstan"),     'ky_KG', '2' ], #
'KH' => [ N_("Cambodia"),       'en_US', '2' ], # km_KH not released yet
'KI' => [ N_("Kiribati"),       'en_US', '3' ], #
'KM' => [ N_("Comoros"),        'en_US', '2' ], #
'KN' => [ N_("Saint Kitts and Nevis"), 'en_US', '5' ], #
'KP' => [ N_("Korea (North)"),  'ko_KR', '2' ], #
'KR' => [ N_("Korea"),          'ko_KR', '2' ],
'KW' => [ N_("Kuwait"),         'ar_KW', '2' ],
'KY' => [ N_("Cayman Islands"), 'en_US', '5' ], #
'KZ' => [ N_("Kazakhstan"),     'ru_RU', '2' ], #
'LA' => [ N_("Laos"),           'lo_LA', '2' ],
'LB' => [ N_("Lebanon"),        'ar_LB', '2' ],
'LC' => [ N_("Saint Lucia"),    'en_US', '5' ], #
'LI' => [ N_("Liechtenstein"),  'de_CH', '1' ], #
'LK' => [ N_("Sri Lanka"),      'en_IN', '2' ], #
'LR' => [ N_("Liberia"),        'en_US', '3' ], #
'LS' => [ N_("Lesotho"),        'en_BW', '3' ], #
'LT' => [ N_("Lithuania"),      'lt_LT', '1' ],
'LU' => [ N_("Luxembourg"),     'de_LU', '1' ],
'LV' => [ N_("Latvia"),         'lv_LV', '1' ],
'LY' => [ N_("Libya"),          'ar_LY', '3' ],
'MA' => [ N_("Morocco"),        'ar_MA', '3' ],
'MC' => [ N_("Monaco"),         'fr_FR', '1' ], #
'MD' => [ N_("Moldova"),        'ro_RO', '1' ], #
'MG' => [ N_("Madagascar"),     'fr_FR', '3' ], #
'MH' => [ N_("Marshall Islands"), 'en_US', '4' ], #
'MK' => [ N_("Macedonia"),      'mk_MK', '1' ],
'ML' => [ N_("Mali"),           'en_US', '3' ], #
'MM' => [ N_("Myanmar"),        'en_US', '2' ], #
'MN' => [ N_("Mongolia"),       'mn_MN', '2' ],
'MP' => [ N_("Northern Mariana Islands"), 'en_US', '2' ], #
'MQ' => [ N_("Martinique"),     'fr_FR', '5' ], #
'MR' => [ N_("Mauritania"),     'en_US', '3' ], #
'MS' => [ N_("Montserrat"),     'en_US', '5' ], #
'MT' => [ N_("Malta"),          'mt_MT', '1' ],
'MU' => [ N_("Mauritius"),      'en_US', '3' ], #
'MV' => [ N_("Maldives"),       'en_US', '4' ], #
'MW' => [ N_("Malawi"),         'en_US', '3' ], #
'MX' => [ N_("Mexico"),         'es_MX', '5' ],
'MY' => [ N_("Malaysia"),       'ms_MY', '2' ],
'MZ' => [ N_("Mozambique"),     'pt_PT', '3' ], #
'NA' => [ N_("Namibia"),        'en_US', '3' ], #
'NC' => [ N_("New Caledonia"),  'fr_FR', '4' ], #
'NE' => [ N_("Niger"),          'en_US', '3' ], #
'NF' => [ N_("Norfolk Island"), 'en_GB', '4' ], #
'NG' => [ N_("Nigeria"),        'en_US', '3' ], #
'NI' => [ N_("Nicaragua"),      'es_NI', '5' ],
'NL' => [ N_("Netherlands"),    'nl_NL', '1' ],
'NO' => [ N_("Norway"),         'nb_NO', '1' ],
'NP' => [ N_("Nepal"),          'ne_NP', '2' ],
'NR' => [ N_("Nauru"),          'en_US', '4' ], #
'NU' => [ N_("Niue"),           'en_US', '4' ], #
'NZ' => [ N_("New Zealand"),    'en_NZ', '4' ],
'OM' => [ N_("Oman"),           'ar_OM', '2' ],
'PA' => [ N_("Panama"),         'es_PA', '5' ],
'PE' => [ N_("Peru"),           'es_PE', '5' ],
'PF' => [ N_("French Polynesia"), 'fr_FR', '4' ], #
'PG' => [ N_("Papua New Guinea"), 'en_NZ', '4' ], #
'PH' => [ N_("Philippines"),    'ph_PH', '2' ],
'PK' => [ N_("Pakistan"),       'ur_PK', '2' ],
'PL' => [ N_("Poland"),         'pl_PL', '1' ],
'PM' => [ N_("Saint Pierre and Miquelon"), 'fr_CA', '5' ], #
'PN' => [ N_("Pitcairn"),      'en_US', '4' ], #
'PR' => [ N_("Puerto Rico"),    'es_PR', '5' ],
'PS' => [ N_("Palestine"),      'ar_JO', '2' ], #
'PT' => [ N_("Portugal"),       'pt_PT', '1' ],
'PY' => [ N_("Paraguay"),       'es_PY', '5' ],
'PW' => [ N_("Palau"),          'en_US', '2' ], #
'QA' => [ N_("Qatar"),          'ar_QA', '2' ],
'RE' => [ N_("Reunion"),        'fr_FR', '2' ], #
'RO' => [ N_("Romania"),        'ro_RO', '1' ],
'RU' => [ N_("Russia"),         'ru_RU', '1' ],
'RW' => [ N_("Rwanda"),         'fr_FR', '3' ], #
'SA' => [ N_("Saudi Arabia"),   'ar_SA', '2' ],
'SB' => [ N_("Solomon Islands"), 'en_US', '4' ], #
'SC' => [ N_("Seychelles"),     'en_US', '4' ], #
'SD' => [ N_("Sudan"),          'ar_SD', '5' ],
'SE' => [ N_("Sweden"),         'sv_SE', '1' ],
'SG' => [ N_("Singapore"),      'en_SG', '2' ],
'SH' => [ N_("Saint Helena"),   'en_GB', '5' ], #
'SI' => [ N_("Slovenia"),       'sl_SI', '1' ],
'SJ' => [ N_("Svalbard and Jan Mayen Islands"), 'en_US', '1' ], #
'SK' => [ N_("Slovakia"),       'sk_SK', '1' ],
'SL' => [ N_("Sierra Leone"),   'en_US', '3' ], #
'SM' => [ N_("San Marino"),     'it_IT', '1' ], #
'SN' => [ N_("Senegal"),        'fr_FR', '3' ], #
'SO' => [ N_("Somalia"),        'en_US', '3' ], # so_SO
'SR' => [ N_("Suriname"),       'nl_NL', '5' ], #
'ST' => [ N_("Sao Tome and Principe"), 'en_US', '5' ], #
'SV' => [ N_("El Salvador"),    'es_SV', '5' ],
'SY' => [ N_("Syria"),          'ar_SY', '2' ],
'SZ' => [ N_("Swaziland"),      'en_BW', '3' ], #
'TC' => [ N_("Turks and Caicos Islands"), 'en_US', '5' ], #
'TD' => [ N_("Chad"),           'en_US', '3' ], #
'TF' => [ N_("French Southern Territories"), 'fr_FR', '4' ], #
'TG' => [ N_("Togo"),           'fr_FR', '3' ], #
'TH' => [ N_("Thailand"),       'th_TH', '2' ],
'TJ' => [ N_("Tajikistan"),     'tg_TJ', '2' ],
'TK' => [ N_("Tokelau"),        'en_US', '4' ], #
'TL' => [ N_("East Timor"),     'pt_PT', '4' ], #
'TM' => [ N_("Turkmenistan"),   'en_US', '2' ], #
'TN' => [ N_("Tunisia"),        'ar_TN', '5' ],
'TO' => [ N_("Tonga"),          'en_US', '3' ], #
'TR' => [ N_("Turkey"),         'tr_TR', '2' ],
'TT' => [ N_("Trinidad and Tobago"), 'en_US', '5' ], #
'TV' => [ N_("Tuvalu"),         'en_US', '4' ], #
'TW' => [ N_("Taiwan"),         'zh_TW', '2' ],
'TZ' => [ N_("Tanzania"),       'en_US', '3' ], #
'UA' => [ N_("Ukraine"),        'uk_UA', '1' ],
'UG' => [ N_("Uganda"),         'en_US', '3' ], # lug_UG
'UM' => [ N_("United States Minor Outlying Islands"), 'en_US', '5' ], #
'US' => [ N_("United States"),  'en_US', '5' ],
'UY' => [ N_("Uruguay"),        'es_UY', '5' ],
'UZ' => [ N_("Uzbekistan"),     'uz_UZ', '2' ],
'VA' => [ N_("Vatican"),        'it_IT', '1' ], #
'VC' => [ N_("Saint Vincent and the Grenadines"), 'en_US', '5' ], 
'VE' => [ N_("Venezuela"),      'es_VE', '5' ],
'VG' => [ N_("Virgin Islands (British)"), 'en_GB', '5' ], #
'VI' => [ N_("Virgin Islands (U.S.)"), 'en_US', '5' ], #
'VN' => [ N_("Vietnam"),        'vi_VN', '2' ],
'VU' => [ N_("Vanuatu"),        'en_US', '4' ], #
'WF' => [ N_("Wallis and Futuna"), 'fr_FR', '4' ], #
'WS' => [ N_("Samoa"),          'en_US', '4' ], #
'YE' => [ N_("Yemen"),          'ar_YE', '2' ],
'YT' => [ N_("Mayotte"),        'fr_FR', '3' ], #
'CS' => [ N_("Serbia & Montenegro"), 'sr_CS', '1' ],
'ZA' => [ N_("South Africa"),   'en_ZA', '5' ],
'ZM' => [ N_("Zambia"),         'en_US', '3' ], #
'ZW' => [ N_("Zimbabwe"),       'en_ZW', '5' ],
);
sub c2name   { exists $countries{$_[0]} && translate($countries{$_[0]}[0]) }
sub c2locale { exists $countries{$_[0]} && $countries{$_[0]}[1] }
sub list_countries {
    my (%options) = @_;
    my @l = keys %countries;
    $options{exclude_non_installed} ? grep { -e "/usr/share/locale/".c2locale($_)."/LC_CTYPE" } @l : @l;
}

#- this list is built with the following command on the compile cluster:
#- rpm -qpl /cooker/RPMS/locales-* | grep LC_CTYPE | cut -d'/' -f5 | grep '_' | grep -v '\.' | sort | tr '\n' ' ' ; echo
our @locales = qw(ad_ET af_ZA am_ET an_ES ar_AE ar_BH ar_DZ ar_EG ar_IN ar_IQ ar_JO ar_KW ar_LB ar_LY ar_MA ar_OM ar_QA ar_SA ar_SD ar_SY ar_TN ar_YE as_IN az_AZ be_BY bg_BG bn_BD bn_IN br_FR bs_BA ca_ES cs_CZ cy_GB da_DK de_AT de_BE de_CH de_DE de_LU el_GR en_AU en_BE en_BW en_CA en_DK en_GB en_HK en_IE en_IN en_NZ en_PH en_SG en_US en_ZA en_ZW eo_XX es_AR es_BO es_CL es_CO es_CR es_DO es_EC es_ES es_GT es_HN es_MX es_NI es_PA es_PE es_PR es_PY es_SV es_US es_UY es_VE et_EE eu_ES fa_IR fi_FI fo_FO fr_BE fr_CA fr_CH fr_FR fr_LU ga_IE gd_GB gez_ER gez_ER@abegede gez_ET gez_ET@abegede gl_ES gu_IN gv_GB he_IL hi_IN hr_HR hu_HU hy_AM id_ID is_IS it_CH it_IT iu_CA ja_JP ka_GE kl_GL kn_IN ko_KR ku_TR kw_GB li_BE li_NL lo_LA lt_LT lv_LV mi_NZ mk_MK ml_IN mn_MN mr_IN ms_MY mt_MT nb_NO nds_DE nds_DE@traditional nds_NL ne_NP nl_BE nl_NL nn_NO oc_FR om_ET om_KE pa_IN ph_PH pl_PL pt_BR pt_PT qo_ET ro_RO ru_RU ru_UA se_NO sh_YU sid_ET sk_SK sl_SI sq_AL sr_CS sr_CS@Latn sr_YU sr_YU@Latn st_ZA sv_FI sv_SE sx_ET sz_ET ta_IN te_IN tg_TJ th_TH ti_ER ti_ET tig_ER tl_PH tr_TR tt_RU uk_UA ur_PK uz_UZ uz_UZ@Cyrl uz_UZ@Latn vi_VN wa_BE xh_ZA yi_US zh_CN zh_HK zh_SG zh_TW zu_ZA);
	
sub standard_locale {
    my ($lang, $country, $prefer_lang) = @_;
    member("${lang}_${country}", @locales) and return "${lang}_${country}";
    $prefer_lang && member($lang, @locales) and return $lang;
    my $main_locale = locale_to_main_locale($lang);
    if ($main_locale ne $lang) {
	standard_locale($main_locale, $country, $prefer_lang);
    }
    '';
}

sub fix_variant {
    my ($locale) = @_;
    #- uz@Cyrl_UZ -> uz_UZ@Cyrl
    $locale =~ s/(.*)(\@\w+)(_.*)/$1$3$2/;
    $locale;
}

sub analyse_locale_name {
    my ($locale) = @_;
    $locale =~ /^(.*?) (?:_(.*?))? (?:\.(.*?))? (?:\@(.*?))? $/x &&
      { main => $1, country => $2, charset => $3, variant => $4 };
}

sub locale_to_main_locale {
    my ($locale) = @_;
    lc(analyse_locale_name($locale)->{main});
}

sub getlocale_for_lang {
    my ($lang, $country, $o_utf8) = @_;
    fix_variant((standard_locale($lang, $country, 'prefer_lang') || l2locale($lang)) . ($o_utf8 ? '.UTF-8' : ''));
}

sub getlocale_for_country {
    my ($lang, $country, $o_utf8) = @_;
    fix_variant((standard_locale($lang, $country, '') || c2locale($country)) . ($o_utf8 ? '.UTF-8' : ''));
}

sub getLANGUAGE {
    my ($lang, $o_country, $o_utf8) = @_;
    l2language($lang) || join(':', uniq(getlocale_for_lang($lang, $o_country, $o_utf8), 
					$lang, 
					locale_to_main_locale($lang)));
}


my %im_xim_program =
  (
   chinput => {
               'zh_CN' => '"chinput -gb"',
               'zh_CN.UTF-8' => '"chinput -gb"',
               'zh_HK' => '"chinput -big5"',
               'zh_HK.UTF-8' => '"chinput -big5"',
               'zh_SG' => '"chinput -gb"',
               'zh_SG.UTF-8' => '"chinput -gb"',
               'zh_TW' => '"chinput -big5"',
               'zh_TW.UTF-8' => '"chinput -big5"',
              },
   xcin => {
            'zh_TW' => '"@im=xcin-zh_TW"'
           },
  );

my %gtkqt_im =
  (
   ami => {
           XIM => 'Ami',
           #- NOTE: there are several possible versions of ami, for the different
           #- desktops (kde, gnome, etc). So XIM_PROGRAM isn't defined; it will
           #- be the xinitrc script, XIM section, that will choose the right one 
           #- XIM_PROGRAM => 'ami',
           XMODIFIERS => '"@im=Ami"',
           GTK_IM_MODULE => 'xim',
          },
   chinput => {
               GTK_IM_MODULE => 'xim',
               XIM => 'xcin',
               XMODIFIERS => '"@im=Chinput"',
               },
   fctix => {
             XIM => 'fcitx',
             XIM_PROGRAM => 'fcitx',
             XMODIFIERS => '"@im=fcitx"',
            },
   kinput2 => {   
               XIM => 'kinput2',
               XIM_PROGRAM => 'kinput2',
               XMODIFIERS => '"@im=kinput2"',
              },
   scim => {
            GTK_IM_MODULE => 'scim',
            XIM_PROGRAM => 'scim -d',
            XMODIFIERS => '"@im=SCIM"',
           },
   uim => {
           GTK_IM_MODULE => 'uim-anthy',
           XIM => 'uim-anthy',
           XIM_PROGRAM => 'uim-xim',
           XMODIFIERS => '"@im=uim-anthy"',
          },
   xcin => {
            XIM => 'xcin',
            XIM_PROGRAM => 'xcin',
            GTK_IM_MODULE => 'xim',
           },

);
           

my %xim = (
#- xcin only works with 'zh_TW', 'zh_TW.Big5', 'zh_CN', 'zh_CN.GB2312'
#- all other locale names, in particular 'zh_HK' or 'zh_TW.UTF-8'
#- are unknown to it. So chinput is used for all but 'zh_TW'
  'ja_JP' => {
	ENC => 'eucj',
  },
  'ja_JP.UTF-8' => {
	ENC => 'utf8',
  },
  'ko_KR' => {
	ENC => 'kr',
	CONSOLE_NOT_LOCALIZED => 'yes',
  },
  'ko_KR.UTF-8' => {
	ENC => 'utf8',
	CONSOLE_NOT_LOCALIZED => 'yes',
  },
  'zh_TW' => { 
 	ENC => 'big5',
 	CONSOLE_NOT_LOCALIZED => 'yes',
  },
  'zh_TW.UTF-8' => {
	ENC => 'utf8',
	CONSOLE_NOT_LOCALIZED => 'yes',
  },
  'zh_CN' => {
	ENC => 'gb',
	CONSOLE_NOT_LOCALIZED => 'yes',
  },
  'zh_CN.UTF-8' => {
	ENC => 'utf8',
	CONSOLE_NOT_LOCALIZED => 'yes',
  },
  'zh_HK' => {
	ENC => 'big5',
	CONSOLE_NOT_LOCALIZED => 'yes',
  },
  'zh_HK.UTF-8' => {
	ENC => 'utf8',
	CONSOLE_NOT_LOCALIZED => 'yes',
  },
  'zh_SG' => {
	ENC => 'gb',
	CONSOLE_NOT_LOCALIZED => 'yes',
  },
  'zh_SG.UTF-8' => {
	ENC => 'utf8',
	CONSOLE_NOT_LOCALIZED => 'yes',
  },
  #-�XFree86 has an internal XIM for Thai that enables syntax checking etc.
  #-�'Passthroug' is no check at all, 'BasicCheck' accepts bad sequences
  #-�and convert them to right ones, 'Strict' refuses bad sequences
  'th_TH' => {
	XIM_PROGRAM => '/bin/true', #- it's an internal module
	XMODIFIERS => '"@im=BasicCheck"',
  },
  #-�xvnkb is not an XIM input method; but an input method of another
  #-�kind, only XIM_PROGRAM needs to be defined
  #- ! xvnkb doesn't work in UTF-8 !
#-  'vi_VN.VISCII' => {
#-	XIM_PROGRAM => 'xvnkb',
#-  },
);

sub set_default_im {
    my ($im, @langs) = @_;
    add2hash($xim{$_}, $gtkqt_im{$im}) foreach @langs;
    add2hash($xim{$_}, { XIM_PROGRAM => $im_xim_program{$im}{$_} }) foreach @langs;
}

# CJK default input methods:
set_default_im('ami',   qw(ko_KR ko_KR.UTF-8));
set_default_im('scim',  qw(ja_JP ja_JP.UTF-8 zh_CN zh_CN.UTF-8 zh_HK zh_HK.UTF-8 zh_SG zh_SG.UTF-8 zh_TW zh_TW.UTF-8));


#- [0]: console font name
#- [1]: sfm map for console font (if needed)
#- [2]: acm file for console font (none if utf8)
#- [3]: iocharset param for mount (utf8 if utf8)
#- [4]: codepage parameter for mount (none if utf8)
my %charsets = (
#- chinese needs special console driver for text mode
"Big5"        => [ undef,         undef,   undef,           "big5",       "950" ],
"gb2312"      => [ undef,         undef,   undef,           "gb2312",     "936" ],
"C"           => [ "lat0-16",     undef,   "iso15",         "iso8859-1",  "850" ],
"iso-8859-1"  => [ "lat1-16",     undef,   "iso01",         "iso8859-1",  "850" ],
"iso-8859-2"  => [ "lat2-sun16",  undef,   "iso02",         "iso8859-2",  "852" ],
"iso-8859-5"  => [ "UniCyr_8x16", undef,   "iso05",         "iso8859-5",  "866" ],
"iso-8859-7"  => [ "iso07.f16",   undef,   "iso07",         "iso8859-7",  "869" ],
"iso-8859-9"  => [ "lat5u-16",    undef,   "iso09",         "iso8859-9",  "857" ],
"iso-8859-13" => [ "tlat7",       undef,   "iso13",         "iso8859-13", "775" ],
"iso-8859-15" => [ "lat0-16",     undef,   "iso15",         "iso8859-15", "850" ],
#- japanese needs special console driver for text mode [kon2]
"jisx0208"    => [ undef,         undef,   "trivial.trans", "euc-jp",     "932" ],
"koi8-r"      => [ "UniCyr_8x16", undef,   "koi8-r",        "koi8-r",     "866" ],
"koi8-u"      => [ "UniCyr_8x16", undef,   "koi8-u",        "koi8-u",     "866" ],
"cp1251"      => [ "UniCyr_8x16", undef,   "cp1251",        "cp1251",     "866" ],
#- korean needs special console driver for text mode
"ksc5601"     => [ undef,         undef,   undef,           "euc-kr",     "949" ],
#- I have no console font for Thai...
"tis620"      => [ undef,         undef,   "trivial.trans", "tis-620",    "874" ],
# UTF-8 encodings here; they differ in the console font mainly.
"utf_am"      => [ "Agafari-16",     undef,   undef,      "utf8",    undef ],
"utf_ar"      => [ "iso06.f16",      undef,   undef,      "utf8",    undef ],
"utf_az"      => [ "tiso09e",        undef,   undef,      "utf8",    undef ],
"utf_bn"      => [ undef,            undef,   undef,      "utf8",    undef ],
"utf_cyr1"    => [ "UniCyr_8x16",    undef,   undef,      "utf8",    undef ],
"utf_cyr2"    => [ "koi8-k",         undef,   undef,      "utf8",    undef ],
"utf_he"      => [ "iso08.f16",      undef,   undef,      "utf8",    undef ],
"utf_hi"      => [ undef,            undef,   undef,      "utf8",    undef ],
"utf_hy"      => [ "arm8",           undef,   undef,      "utf8",    undef ],
"utf_ka"      => [ "t_geors",        undef,   undef,      "utf8",    undef ],
"utf_kn"      => [ undef,            undef,   undef,      "utf8",    undef ],
"utf_ml"      => [ undef,            undef,   undef,      "utf8",    undef ],
"utf_lo"      => [ undef,            undef,   undef,      "utf8",    undef ],
"utf_ta"      => [ "tamil",          undef,   undef,      "utf8",    undef ],
"utf_vi"      => [ "tcvn8x16",       undef,   undef,      "utf8",    undef ],
"utf_lat8"    => [ "iso14.f16",      undef,   undef,      "utf8",    undef ],
# default for utf-8 encodings
"unicode"     => [ "LatArCyrHeb-16", undef,   undef,      "utf8",    undef ],
);

#- for special cases not handled magically
my %charset2kde_charset = (
    gb2312 => 'gb2312.1980-0',
    jisx0208 => 'jisx0208.1983-0',
    ksc5601 => 'ksc5601.1987-0',
    Big5 => 'big5-0',
    cp1251 => 'microsoft-cp1251',
    utf8 => 'iso10646-1',
    tis620 => 'tis620-0',
    #- Tamil KDE translations still use TSCII, and KDE know it as iso-8859-1
    utf_ta => 'iso8859-1',
);

my @during_install__lang_having_their_LC_CTYPE = qw(ja ko ta);


#- -------------------

sub list { 
    my (%options) = @_;
    my @l = list_langs();
    if ($options{exclude_non_installed_langs}) {
	@l = grep { -e "/usr/share/locale/$_/LC_CTYPE" } @l;
    }
    @l;
}

sub l2console_font {
    my ($locale, $during_install) = @_;
    my $c = $charsets{l2charset($locale->{lang}) || return} or return;
    my ($name, $sfm, $acm) = @$c;
    undef $acm if $locale->{utf8} && !$during_install;
    ($name, $sfm, $acm);
}

sub get_kde_lang {
    my ($locale, $o_default) = @_;

    #- get it using 
    #- echo C $(rpm -qp --qf "%{name}\n" /RPMS/kde-i18n-*  | sed 's/kde-i18n-//')
    my @valid_kde_langs = qw(C af ar az be bg br bs ca cs cy da de el en_GB eo es et eu fa fi fo fr ga gl he hr hu id is it ja ko ku lo lt lv mi mk mn mt nb nl nn nso oc pl pt pt_BR ro ru se sk sl sr ss sv ta th tr uk uz ven vi wa wen xh zh_CN zh_TW zu);
    my %valid_kde_langs; @valid_kde_langs{@valid_kde_langs} = ();

    my $valid_lang = sub {
	my ($lang) = @_;
	#- fast & dirty solution to ensure bad entries do not happen
        my %fixlangs = (en => 'C', en_US => 'C',
                        'sr@Latn' => 'sr',
                        st => 'nso', ve => 'ven',
                        zh_CN => 'zh_CN', zh_SG => 'zh_CN', zh_TW => 'zh_TW', zh_HK => 'zh_TW');
        exists $fixlangs{$lang} ? $fixlangs{$lang} :
	  exists $valid_kde_langs{$lang} ? $lang :
	  exists $valid_kde_langs{locale_to_main_locale($lang)} ? locale_to_main_locale($lang) : '';
    };

    my $r;
    $r ||= $valid_lang->($locale->{lang});
    $r ||= find { $valid_lang->($_) } split(':', getlocale_for_lang($locale->{lang}, $locale->{country}));
    $r || $o_default || 'C';
}

sub charset2kde_charset {
    my ($charset, $o_default) = @_;
    my $iocharset = ($charsets{$charset} || [])->[3];

    my @valid_kde_charsets = qw(big5-0 gb2312.1980-0 iso10646-1 iso8859-1 iso8859-4 iso8859-6 iso8859-8 iso8859-13 iso8859-14 iso8859-15 iso8859-2 iso8859-3 iso8859-5 iso8859-7 iso8859-9 koi8-r koi8-u ksc5601.1987-0 jisx0208.1983-0 microsoft-cp1251 tis620-0);
    my %valid_kde_charsets; @valid_kde_charsets{@valid_kde_charsets} = ();

    my $valid_charset = sub {
	my ($charset) = @_;
	#- fast & dirty solution to ensure bad entries do not happen
	exists $valid_kde_charsets{$charset} && $charset;
    };

    my $r;
    $r ||= $valid_charset->($charset2kde_charset{$charset});
    $r ||= $valid_charset->($charset2kde_charset{$iocharset});
    $r ||= $valid_charset->($iocharset);
    $r || $o_default || 'iso10646-1';
}

#- font+size for different charsets; the field [0] is the default,
#- others are overrridens for fixed(1), toolbar(2), menu(3) and taskbar(4)
my %charset2kde_font = (
  'C' => [ "Sans,10", "Monospace,10" ],
  'iso-8859-1'  => [ "Sans,10", "Monospace,10" ],
  'iso-8859-2'  => [ "Sans,10", "Monospace,10" ],
  'iso-8859-7'  => [ "Helvetica,12", "courier,10", "Helvetica,11" ],
  'iso-8859-9'  => [ "Sans,10", "Monospace,10" ],
  'iso-8859-15' => [ "Sans,10", "Monospace,10" ],
  'iso-8859-13' => [ "Sans,10", "Monospace,10" ],
  'jisx0208' => [ "Sazanami Mincho,13", "Sazanami Gothic,13" ],
  'ksc5601' => [ "Baekmuk Gulim,16" ],
  'gb2312' => [ "AR PL SungtiL GB,13" ],
  'Big5' => [ "AR PL Mingti2L Big5,13" ],
  'tis620' => [ "Norasi,16", "Norasi,15" ],
  'utf_ar' => [ "Kacs_qr,14", "Courier New,13", "Kacs_qr,13" ], 
  'utf_am' => [ "GF Zemen Unicode,15" ],
  'utf_az' => [ "Nimbus Sans,12", "Nimbus Mono,10", "Nimbus Sans,11" ],
  'utf_bn' => [ "Mukti,14", ],
  'utf_he' => [ "Nachlieli,13", , "Miriam Mono,10", "Nachlieli,11" ],
  'utf_hi' => [ "Raghindi,14", ],
  'utf_hy' => [ "Artsounk,12", "Monospace,10", "Artsounk,11" ],
  'utf_kn' => [ "Sampige,14", ],
  'utf_ml' => [ "malayalam,14", ],
  'utf_ta' => [ "TSCu_Paranar,14", "Tsc_avarangalfxd,10", "TSCu_Paranar,12", ],
  #- the following should be changed to better defaults when better fonts
  #- get available
  'utf_vi' => [ "misc-fixed,13", "misc-fixed,13", "misc-fixed,10", ],
  'utf_ka' => [ "clearlyu,15" ],
  'utf_lo' => [ "clearlyu,15" ],
  'default' => [ "Nimbus Sans,12", "Nimbus Mono,10", "Nimbus Sans,11" ],
);

sub charset2kde_font {
    my ($charset, $type) = @_;

    my $font = $charset2kde_font{$charset} || $charset2kde_font{default};
    my $r = $font->[$type] || $font->[0];

    #- the format is "font-name,size,-1,5,0,0,0,0,0,0" I have no idea of the
    #- meaning of that "5"...
    "$r,-1,5,0,0,0,0,0,0";
}

# this define pango name fonts (like "NimbusSans L") depending
# on the "charset" defined by language array. This allows to selecting
# an appropriate font for each language.
my %charset2pango_font = (
  'tis620' =>      "Norasi 17",
  'utf_ar' =>      "Kacst-Qr 14",
  'utf_cyr2' =>    "Nimbus Sans L 12",
  'utf_he' =>      "Sans 12",
  'utf_hy' =>      "Artsounk 14",
  'utf_ka' =>      "Sans 14",
  'utf_lo' =>      "Sans 14",
  'utf_ta' =>      "TSCu_Paranar 14",
  'utf_vi' =>      "Sans 14",
  'iso-8859-7' =>  "Kerkis 14",
  'jisx0208' =>    "Sans 18",
  #- Nimbus Sans L is missing some chars used by some cyrillic languages,
  #- but tose haven't yet DrakX translations; it also misses vietnamese
  #- latin chars; all other latin and cyrillic are covered.
  'default' =>     "Nimbus Sans L 12"
);

sub charset2pango_font {
    my ($charset) = @_;
    
    $charset2pango_font{exists $charset2pango_font{$charset} ? $charset : 'default'};
}

sub l2pango_font {
    my ($lang) = @_;

    my $charset = l2charset($lang) or log::l("no charset found for lang $lang!"), return;
    my $font = charset2pango_font($charset);
    log::l("lang:$lang charset:$charset font:$font sfm:$charsets{$charset}[0]");

    if (common::usingRamdisk()) {
	if ($charsets{$charset}[0] !~ /lat|koi|UniCyr/) {
	    install_any::remove_bigseldom_used();
	    unlink glob_('/usr/share/langs/*');  #- remove langs images
	    my @generic_fontfiles = qw(/usr/X11R6/lib/X11/fonts/12x13mdk.pcf.gz /usr/X11R6/lib/X11/fonts/18x18mdk.pcf.gz);
	    #- need to unlink first because the files actually exist (and are void); they must exist
	    #- because if not, when gtk starts, pango will recompute its cache file and exclude them
	    unlink($_), install_any::getAndSaveFile($_) foreach @generic_fontfiles;
	}

	my %pango_modules = (arabic => 'ar|fa|ur', hangul => 'ko', hebrew => 'he|yi', indic => 'hi|bn|ta|te|mr', thai => 'th');
	foreach my $module (keys %pango_modules) {
	    next if $lang !~ /$pango_modules{$module}/;
	    install_any::remove_bigseldom_used();
	    my ($pango_modules_dir) = glob('/usr/lib/pango/*/modules');
	    install_any::getAndSaveFile("$pango_modules_dir/pango-$module-xft.so");
	}
    }
    
    return $font;
}

sub set {
    my ($locale, $b_translate_for_console) = @_;
    
    if ($::move) {
	move::handleI18NClp($locale->{lang});
	put_in_hash(\%ENV, i18n_env($locale));
	return;
    }

    my $lang = $locale->{lang};
    exists $langs{$lang} or log::l("lang::set: trying to set to $lang but I don't know it!"), return;

    my $dir = "$ENV{SHARE_PATH}/locale";
    if (!-e "$dir/$lang" && common::usingRamdisk()) {
	@ENV{qw(LANG LC_ALL LANGUAGE LINGUAS)} = ();

	my @LCs = qw(LC_ADDRESS LC_COLLATE LC_IDENTIFICATION LC_MEASUREMENT LC_MONETARY LC_NAME LC_NUMERIC LC_PAPER LC_TELEPHONE LC_TIME);
	
	my $charset = during_install__l2charset($lang) || $lang;
	
	#- there are 3 main charsets containing everything for all locales, except LC_CTYPE
	#- by default, there is UTF-8.
	#- when asked for GB2312 or BIG5, removing the other main charsets
	my $main_charset = member($charset, 'GB2312', 'BIG5') ? $charset : 'UTF-8';
	
	#- removing everything
	#- except in main charset: only removing LC_CTYPE if it is there
	eval { rm_rf($_ eq $main_charset ? "$dir/$_/LC_CTYPE" : "$dir/$_") } foreach all($dir);
	
	if (!-e "$dir/$main_charset") {
	    #- getting the main charset
	    mkdir "$dir/$main_charset";
	    mkdir "$dir/$main_charset/LC_MESSAGES";
	    install_any::getAndSaveFile("$dir/$main_charset/$_") foreach @LCs, 'LC_MESSAGES/SYS_LC_MESSAGES';
	}
	mkdir "$dir/$lang";
	
	#- linking to the main charset
	symlink "../$main_charset/$_", "$dir/$lang/$_" foreach @LCs, 'LC_MESSAGES';	    
	
	#- getting LC_CTYPE (putting it directly in $lang)
	install_any::getAndSaveFile("Mandrake/mdkinst$dir/$charset/LC_CTYPE", "$dir/$lang/LC_CTYPE");
    }
    
    #- set all LC_* variables to a unique locale ("C"), and only redefine
    #- LC_CTYPE (for X11 choosing the fontset) and LANGUAGE (for the po files)
    $ENV{$_} = 'C' foreach qw(LC_NUMERIC LC_TIME LC_COLLATE LC_MONETARY LC_PAPER LC_NAME LC_ADDRESS LC_TELEPHONE LC_MEASUREMENT LC_IDENTIFICATION);
    
    $ENV{LC_CTYPE}    = $lang;
    $ENV{LC_MESSAGES} = $lang;
    $ENV{LANG}        = $lang;
    
    if ($b_translate_for_console && $lang =~ /^(ko|ja|zh|th)/) {
	log::l("not translating in console");
	$ENV{LANGUAGE}  = 'C';
    } else {
	$ENV{LANGUAGE}  = getLANGUAGE($lang);
    }
    load_mo();
    $lang;
}

sub langs {
    my ($l) = @_;
    $l->{all} ? list_langs() : grep { $l->{$_} } keys %$l;
}

sub langsLANGUAGE {
    my ($l, $o_c) = @_;
    uniq(map { split ':', getLANGUAGE($_, $o_c) } langs($l));
}

sub utf8_should_be_needed {
    my ($locale) = @_; 
    my @l = uniq(grep { $_ ne 'C' } map { l2charset($_) } langs($locale->{langs}));
    @l > 1 || any { /utf|unicode/ } @l;
}

sub pack_langs { 
    my ($l) = @_; 
    my $s = $l->{all} ? 'all' : join ':', uniq(map { getLANGUAGE($_) } langs($l));
    $ENV{RPM_INSTALL_LANG} = $s;
    $s;
}

sub system_locales_to_ourlocale {
    my ($locale_lang, $locale_country) = @_;
    my $locale = {};
    my $h = analyse_locale_name($locale_lang);
    my $locale_lang_no_encoding = join('_', $h->{main}, if_($h->{country}, $h->{country}));
    $locale->{lang} = member($locale_lang_no_encoding, list_langs()) ?
	$locale_lang_no_encoding : #- special lang's such as en_US pt_BR
	$h->{main};
    $locale->{lang} .= '@' . $h->{variant} if $h->{variant};
    $locale->{country} = analyse_locale_name($locale_country)->{country};
    $locale->{utf8} = $h->{encoding} && $h->{encoding} eq 'UTF-8';
    #- safe fallbacks
    $locale->{lang} ||= 'en_US';
    $locale->{country} ||= 'US';
    $locale;
}

sub read {
    my ($prefix, $user_only) = @_;
    $prefix ||= "";
    my ($f1, $f2) = ("$prefix$ENV{HOME}/.i18n", "$prefix/etc/sysconfig/i18n");
    my %h = getVarsFromSh($user_only && -e $f1 ? $f1 : $f2);
    system_locales_to_ourlocale($h{LC_MESSAGES} || 'en_US', $h{LC_MONETARY} || 'en_US');
}

sub write_langs {
    my ($langs) = @_;
    my $s = pack_langs($langs);
    symlink "$::prefix/etc/rpm", "/etc/rpm" if $::prefix;
    substInFile { s/%_install_langs.*//; $_ .= "%_install_langs $s\n" if eof && $s } "$::prefix/etc/rpm/macros";
}

sub i18n_env {
    my ($locale) = @_;

    my $locale_lang = getlocale_for_lang($locale->{lang}, $locale->{country}, $locale->{utf8});
    my $locale_country = getlocale_for_country($locale->{lang}, $locale->{country}, $locale->{utf8});

    my $h = {
	XKB_IN_USE => '',
	(map { $_ => $locale_lang } qw(LANG LC_COLLATE LC_CTYPE LC_MESSAGES LC_TIME)),
	LANGUAGE => getLANGUAGE($locale->{lang}, $locale->{country}, $locale->{utf8}),
	(map { $_ => $locale_country } qw(LC_NUMERIC LC_MONETARY LC_ADDRESS LC_MEASUREMENT LC_NAME LC_PAPER LC_IDENTIFICATION LC_TELEPHONE))
    };

    log::l("lang::write: lang:$locale->{lang} country:$locale->{country} locale|lang:$locale_lang locale|country:$locale_country language:$h->{LANGUAGE}");

    $h;
}

sub write { 
    my ($prefix, $locale, $b_user_only, $b_dont_touch_kde_files) = @_;

    $locale && $locale->{lang} or return;

    my $h = i18n_env($locale);

    my ($name, $sfm, $acm) = l2console_font($locale, 0);
    if ($name && !$b_user_only) {
	my $p = "$prefix/usr/lib/kbd";
	if ($name) {
	    eval {
		my $font = "$p/consolefonts/$name.psf";
		$font .= ".gz" if ! -e $font;
		cp_af($font, "$prefix/etc/sysconfig/console/consolefonts");
		add2hash $h, { SYSFONT => $name };
	    };
	    $@ and log::l("missing console font $name");
	}
	if ($sfm) {
	    eval {
		cp_af(glob_("$p/consoletrans/$sfm*"), "$prefix/etc/sysconfig/console/consoletrans");
		add2hash $h, { UNIMAP => $sfm };
	    };
	    $@ and log::l("missing console unimap file $sfm");
	}
	if ($acm) {
	    eval {
		cp_af(glob_("$p/consoletrans/$acm*"), "$prefix/etc/sysconfig/console/consoletrans");
		add2hash $h, { SYSFONTACM => $acm };
	    };
	    $@ and log::l("missing console acm file $acm");
	}
	
    }
    add2hash $h, $xim{$h->{LANG}};

    #- deactivate translations on console for RTL languages
    if ($h->{LANG} =~ /ar|fa|he|ur|yi/) {
	add2hash $h, { CONSOLE_NOT_LOCALIZED => 'yes' }
    }

    setVarsInSh($prefix . ($b_user_only ? "$ENV{HOME}/.i18n" : '/etc/sysconfig/i18n'), $h);
    substInFile {
        s!^function lang\b.*!function lang()="$h->{LANG}"!g;
    } "$::prefix/etc/menu-methods/lang.h";

    eval {
	my $confdir = $prefix . ($b_user_only ? "$ENV{HOME}/.kde" : '/usr') . '/share/config';

	-d $confdir or die 'not configuring kde config files since it is not installed/used';

	configure_kdeglobals($locale, $confdir);

	my %qt_xim = (zh => 'Over The Spot', ko => 'On The Spot', ja => 'On The Spot');
	if ($b_user_only && (my $qt_xim = $qt_xim{locale_to_main_locale($locale->{lang})})) {
	    update_gnomekderc("$ENV{HOME}/.qt/qtrc", General => (XIMInputStyle => $qt_xim));
	}

	if (!$b_user_only) {
	    my $kde_charset = charset2kde_charset(l2charset($locale->{lang}));
	    my $welcome = c::to_utf8(N("Welcome to %s", '%n'));
	    substInFile { 
		s/^(GreetString)=.*/$1=$welcome/;
		s/^(Language)=.*/$1=$locale->{lang}/;
		if (!member($kde_charset, 'iso8859-1', 'iso8859-15')) { 
		    #- don't keep the default for those
    		    my $font_list = $charset2kde_font{l2charset($locale->{lang})} || $charset2kde_font{default};
		    my $font_small = $font_list->[0];
		    my $font_huge = $font_small;
		    $font_huge =~ s/(.*?),\d+/$1,24/;
		    s/^(StdFont)=.*/$1=$font_small,5,$kde_charset,50,0/;
		    s/^(FailFont)=.*/$1=$font_small,5,$kde_charset,75,0/;
		    s/^(GreetFont)=.*/$1=$font_huge,5,$kde_charset,50,0/;
		}
	    } "$::prefix/usr/share/config/kdm/kdmrc";
	}

    } if !$b_dont_touch_kde_files;
}

sub configure_kdeglobals {
    my ($locale, $confdir) = @_;
    my $kdeglobals = "$confdir/kdeglobals";

    my $charset = l2charset($locale->{lang});
    my $kde_charset = charset2kde_charset($charset);
    my ($prev_kde_charset) = cat_($kdeglobals) =~ /^Charset=(.*)/mi;

    mkdir_p($confdir);

    update_gnomekderc($kdeglobals, Locale => (
    	      Charset => $kde_charset,
    	      Country => lc($locale->{country}),
    	      Language => get_kde_lang($locale),
    	  ));

    if ($prev_kde_charset ne $kde_charset) {
        update_gnomekderc($kdeglobals, WM => (
       		      activeFont => charset2kde_font($charset,0),
       		  ));
        update_gnomekderc($kdeglobals, General => (
       		      fixed => charset2kde_font($charset, 1),
       		      font => charset2kde_font($charset, 0),
       		      menuFont => charset2kde_font($charset, 3),
       		      taskbarFont => charset2kde_font($charset, 4),
       		      toolBarFont => charset2kde_font($charset, 2),
       	          ));
        update_gnomekderc("$confdir/konquerorrc", FMSettings => (
       		      StandardFont => charset2kde_font($charset, 0),
       		  ));
        update_gnomekderc("$confdir/kdesktoprc", FMSettings => (
       		      StandardFont => charset2kde_font($charset, 0),
       		  ));
    }
}

sub bindtextdomain() {
    #- if $::prefix is set, search for libDrakX.mo in locale_special
    #- NB: not using $::isInstall to make it work more easily at install and standalone
    my $localedir = "$ENV{SHARE_PATH}/locale" . ($::prefix ? "_special" : '');

    c::setlocale();
    c::bind_textdomain_codeset('libDrakX', 'UTF-8');
    $::need_utf8_i18n = 1;
    c::bindtextdomain('libDrakX', $localedir);

    $localedir;
}

sub load_mo {
    my ($lang) = @_;

    my $localedir = bindtextdomain();
    my $suffix = 'LC_MESSAGES/libDrakX.mo';

    $lang ||= $ENV{LANGUAGE} || $ENV{LC_ALL} || $ENV{LC_MESSAGES} || $ENV{LANG};

    my @possible_langs = map { { name => $_, mofile => "$localedir/$_/$suffix" } } split ':', $lang;

    -s $_->{mofile} and return $_->{name} foreach @possible_langs;

    if ($::isInstall && common::usingRamdisk()) {
        foreach (@possible_langs) {
            #- cleanup
            eval { rm_rf($localedir) };
            eval { mkdir_p(dirname($_->{mofile})) };

	    install_any::getAndSaveFile($_->{mofile});
	    -s $_->{mofile} and return $_->{name};
	}
    }
    '';
}


#- used in Makefile during "make get_needed_files"
sub console_font_files() {
    map { -e $_ ? $_ : "$_.gz" }
      (map { "/usr/lib/kbd/consolefonts/$_.psf" } uniq grep { $_ } map { $_->[0] } values %charsets),
      (map { -e $_ ? $_ : "$_.sfm" } map { "/usr/lib/kbd/consoletrans/$_" } uniq grep { $_ } map { $_->[1] } values %charsets),
      (map { -e $_ ? $_ : "$_.acm" } map { "/usr/lib/kbd/consoletrans/$_" } uniq grep { $_ } map { $_->[2] } values %charsets),
}

sub load_console_font {
    my ($locale) = @_;
    my ($name, $sfm, $acm) = l2console_font($locale, 1);

    require run_program;
    run_program::run(if_($ENV{LD_LOADER}, $ENV{LD_LOADER}), 
		     'consolechars', '-v', '-f', $name || 'lat0-16',
		     if_($sfm, '-u', $sfm), if_($acm, '-m', $acm));

    #- in console mode install, ensure we'll get translations in the right codeset
    #- (charset of locales reported by the glibc are UTF-8 during install)
    if ($acm) {
	c::bind_textdomain_codeset('libDrakX', l2charset($locale->{lang}));
	$::need_utf8_i18n = 0;
    }
}

sub fs_options {
    my ($locale) = @_;
    if ($locale->{utf8}) {
	(iocharset => 'utf8', codepage => undef);
    } else {
	my $c = $charsets{l2charset($locale->{lang}) || return} or return;
	my ($iocharset, $codepage) = @$c[3..4];
	(iocharset => $iocharset, codepage => $codepage);
    }
}

sub during_install__l2charset {
    my ($lang) = @_;
    return if member($lang, @during_install__lang_having_their_LC_CTYPE);

    my ($c) = l2charset($lang) or die "bad lang $lang\n";
    $c = 'UTF-8' if member($c, 'tis620', 'C', 'unicode');
    $c = 'UTF-8' if $c =~ /koi8-/;
    $c = 'UTF-8' if $c =~ /iso-8859/;
    $c = 'UTF-8' if $c =~ /cp125/;
    $c = 'UTF-8' if $c =~ /utf_/;
    uc($c);
}

sub check() {
    $^W = 0;
    my $ok = 1;
    my $warn = sub {
	print STDERR "$_[0]\n";
    };
    my $err = sub {
	&$warn;
	$ok = 0;
    };
    
    my @wanted_charsets = uniq map { l2charset($_) } list_langs();
    print "\tWarnings:\n";
    $warn->("unused charset $_ (given in \%charsets, but not used in \%langs)") foreach difference2([ keys %charsets ], \@wanted_charsets);

    $warn->("unused entry $_ in \%xim") foreach grep { !/UTF-8/ } difference2([ keys %xim ], [ map { l2locale($_) } list_langs() ]);

    #- consolefonts are checked during build via console_font_files()

    if (my @l = difference2([ 'default', keys %charsets ], [ keys %charset2kde_font ])) {
	$warn->("no kde font for charset " . join(" ", @l));
    }

    if (my @l = grep { get_kde_lang({ lang => $_, country => 'US' }, 'err') eq 'err' } list_langs()) {
	$warn->("no KDE lang for langs " . join(" ", @l));
    }
    if (my @l = grep { charset2kde_charset($_, 'err') eq 'err' } keys %charsets) {
	$warn->("no KDE charset for charsets " . join(" ", @l));
    }

    $warn->("no country corresponding to default locale $_->[1] of lang $_->[0]")
      foreach grep { $_->[1] =~ /.._(..)/ && !exists $countries{$1} } map { [ $_, l2locale($_) ] } list_langs();

    print "\tErrors:\n";

    $err->("invalid charset $_ ($_ does not exist in \%charsets)") foreach difference2(\@wanted_charsets, [ keys %charsets ]);
    $err->("invalid charset $_ in \%charset2kde_font ($_ does not exist in \%charsets)") foreach difference2([ keys %charset2kde_font ], [ 'default', keys %charsets ]);

    $err->("default locale $_->[1] of lang $_->[0] isn't listed in \@locales")
      foreach grep { !member($_->[1], @locales) } map { [ $_, l2locale($_) ] } list_langs();

    $err->("lang image for lang $_->[0] is missing (file $_->[1])")
      foreach grep { !(-e $_->[1]) } map { [ $_, "pixmaps/langs/lang-$_.png" ] } list_langs();

    $err->("default locale $_->[1] of country $_->[0] isn't listed in \@locales")
      foreach grep { !member($_->[1], @locales) } map { [ $_, c2locale($_) ] } list_countries();


    exit($ok ? 0 : 1);
}

1;