path: root/i18npool/source/localedata/data/locale.dtd

<!--
 * This file is part of the LibreOffice project.
 *
 * This Source Code Form is subject to the terms of the Mozilla Public
 * License, v. 2.0. If a copy of the MPL was not distributed with this
 * file, You can obtain one at http://mozilla.org/MPL/2.0/.
 *
 * This file incorporates work covered by the following license notice:
 *
 *   Licensed to the Apache Software Foundation (ASF) under one or more
 *   contributor license agreements. See the NOTICE file distributed
 *   with this work for additional information regarding copyright
 *   ownership. The ASF licenses this file to you under the Apache
 *   License, Version 2.0 (the "License"); you may not use this file
 *   except in compliance with the License. You may obtain a copy of
 *   the License at http://www.apache.org/licenses/LICENSE-2.0 .
-->

<!-- ............................................................... -->
<!-- Locale data specification DTD ................................. -->
<!-- ............................................................... -->

<!--
    ===========================================================================
    ATTENTION! PLEASE! HEADS UP! IMPORTANT!
    ===========================================================================

    Please validate your locale data contribution.

    A quick validation check can be done if you have xmllint installed and the
    current locale.dtd file at hand, you can download the locale.dtd file from
    https://cgit.freedesktop.org/libreoffice/core/plain/i18npool/source/localedata/data/locale.dtd

    xmllint -dtdvalid locale.dtd -noout your_data.xml

    (Note that instead of one - hyphen minus two consecutive hyphen minus
    characters should be given to start an option, but a double hyphen in a XML
    comment technically ends a comment (though parsers look for a matching
    closing one with > as well) and let xmllint complain already about the .dtd
    itself. Some versions of xmllint also accept a single hyphen minus).


    Second, using a validating parser. A validating parser, for example, may be found at
    http://unicode.org/cldr/data/tools/java/org/unicode/cldr/util/XMLValidator.java
    Compile it into a class-jar and call it in the
    i18npool/source/localedata/data/ directory:
    java -cp <your_path>/XMLValidator.jar org.unicode.cldr.util.XMLValidator your_data.xml


    A third possibility is:

    - temporarily (!) change the DOCTYPE of your file to read (all on one line)
      <!DOCTYPE Locale SYSTEM "https://cgit.freedesktop.org/libreoffice/core/plain/i18npool/source/localedata/data/locale.dtd">

    - upload it to the form available at http://www.validome.org/xml/

    This will validate the file against the HEAD revision of locale.dtd


    Please test locale data files either in an enable-dbgutil build, which
    implements some checks and pops up assertion message boxes if the tests
    fail, or by setting the environment variable OOO_ENABLE_LOCALE_DATA_CHECKS
    to 'Y' or 'Yes' (or any other string starting with 'Y') or '1' before
    starting the application, which outputs the same messages to stderr and
    also works in a product build.

    Then follow this procedure:
    1. Create a new spreadsheet document.
    2. On a cell use context menu -> Format Cells -> Numbers.
    3. Select the locale in the Language list box => MUST be assertion free.
    3.a. Assertions are only shown at the very first time a number formatter
         is created, respectively the first time a specific locale data is
         used. To repeat steps 1.-3. you'd need to create another spreadsheet
         document.

    ===========================================================================
    NOTE the FormatElement comments further down.
    ===========================================================================

-->

<!-- ............................................................... -->
<!-- Entities for characters and symbols ........................... -->

<!ENTITY % UNOModule
        'unoid                  CDATA           #IMPLIED'>

<!ENTITY % MessageID
        'msgid                  CDATA           #REQUIRED'>

<!ENTITY % RefLocale
        'ref                    CDATA           #IMPLIED'>
<!-- Where given, an element can be inherited from another locale, e.g.
     ref="en_US" -->

<!ENTITY % LIBModule
        'module                 CDATA           #IMPLIED'>
<!-- The locale referred to for the sub categories, implementation detail. -->

<!ENTITY % replaceFrom  'replaceFrom    CDATA #IMPLIED'>
<!ENTITY % replaceTo    'replaceTo      CDATA #IMPLIED'>
<!-- See below for the LC_FORMAT element. -->

<!ELEMENT DefaultName  (#PCDATA)>

<!-- Locale is made of different sub categories -->
<!ELEMENT Locale (LC_INFO, LC_CTYPE, LC_FORMAT, LC_FORMAT_1?, LC_COLLATION, LC_SEARCH, LC_INDEX, LC_CALENDAR, LC_CURRENCY,  LC_TRANSLITERATION, LC_MISC, LC_NumberingLevel, LC_OutLineNumberingLevel)>
<!ATTLIST Locale versionDTD CDATA #FIXED "2.0.3">
<!-- Version identifier to prevent mismatching data files being submitted
     because of older files being copied as templates. The version should be
     less than or equal to the LibreOffice release, or the LibreOffice release
     number plus some extension, like "2.0.enhanced", to be able to easily
     determine the corresponding LibreOffice version. New versions of the DTD
     with new required elements SHOULD REALLY result in a new versionDTD here
     and LOCALE_VERSION_DTD in ../LocaleNode.cxx, and *.xml files MUST be
     adapted then, otherwise building the data or checking it with a validating
     parser will throw an error.
-->
<!ATTLIST Locale allowUpdateFromCLDR (yes|no) #REQUIRED>
<!-- Whether some data elements may be (automatically) updated from the Common
     Locale Data Repository, see http://cldr.unicode.org/
     Note: This mechanism currently (2010-02-21) is outdated.
-->
<!ATTLIST Locale  version CDATA #REQUIRED>
<!-- Valid number, may designate versioned data -->


<!ELEMENT LC_INFO (Language, Country, Platform?, Variant?)>

<!ELEMENT Language (LangID, DefaultName) >
<!ELEMENT LangID (#PCDATA) >
<!-- LangID must be a valid two or three letter language identifier defined by
     ISO 639. Use ISO 639-1 two letter code where available, else ISO 639-2 or
     639-3 three letter code.

     If the Variant element designates a BCP 47 language tag (because the
     locale is not expressible as a Language,Country pair, for example contains
     a script tag) then the LangID value MUST be the code 'qlt' (reserved by
     ISO 639-3 for private use) to tell the language tag processor that the
     actual language tag is in the Variant element.
-->

<!ELEMENT Country (CountryID, DefaultName) >
<!ELEMENT CountryID (#PCDATA) >
<!-- CountryID must be a valid two letter country identifier defined by ISO 3166. -->

<!ELEMENT Platform (PlatformID) >
<!ELEMENT PlatformID (#PCDATA) >
<!-- Unused, deprecated, can be generic|unix|win32|macos, best Platform element
     be absent.
-->

<!ELEMENT Variant (#PCDATA) >
<!-- If the LangID element contains the value 'qlt' then the Variant element
     MUST contain the valid BCP 47 language tag of the locale. If LangID is
     another ISO 639 code than 'qlt' then the Variant element must be empty or
     not present.
-->


<!-- The LC_FORMAT element contains number format codes and may actually appear
     twice (with the second occurrence named LC_FORMAT_1). One section is
     mandatory and MUST contain at least all required format codes
     formatindex="0" to formatindex="47" (except 10 and 11 MUST NOT be
     defined), MUST NOT contain formatindex 48 and 49, MUST contain formatindex
     50, and MAY contain other format codes. See below ATTLIST FormatElement
     formatindex.

     A second LC_FORMAT_1 section may follow containing other format codes. The
     difference between the two sections is that they are inherited
     independently if a locale uses the RefLocale mechanism (ref="..."). This
     may be used to not offer locale dependent format codes to other locales
     that otherwise inherit the format codes.

-->
<!ELEMENT LC_FORMAT (DateAcceptancePattern*, FormatElement*) >
<!-- All FormatElement elements must be given if the RefLocale mechanism is not used! -->
<!ATTLIST LC_FORMAT %RefLocale;>
<!ATTLIST LC_FORMAT %replaceFrom;>
<!-- Define placeholder for currency code, usually "[CURRENCY]" -->
<!ATTLIST LC_FORMAT %replaceTo;>
<!-- Currency code to be used to replace the placeholder, e.g. "[$R-1C09]".
     Note: The brackets and the leading $ character are mandatory, the
     hyphen-minus separates the currency symbol from the hexagesimal MS-LCID,
     letters contained in the LCID have to be in upper case, leading zeros are
     to be omitted. LCIDs are defined in include/i18nlangtag/lang.h (
     https://opengrok.libreoffice.org/xref/core/include/i18nlangtag/lang.h )
-->

<!ELEMENT DateAcceptancePattern (#PCDATA)>
<!-- Pattern that defines an input sequence match to be accepted as
     (abbreviated) date. For example, in en_US locale "M/D" accepts an input of
     11/23 as CurrentYear-November-23 whereas 11/23/ would not be a date. In
     de_DE locale "D.M." accepts an input of 23.11. as  CurrentYear-November-23
     whereas 23.11 would not be a date.

     For each locale one pattern that matches a full date is automatically
     generated from FormatElement formatIndex="21" and does not need to be
     defined, for example "M/D/Y" or "D.M.Y".

     At least one pattern for abbreviated date input must be defined, i.e.
     contain D and M but not Y.

     NOTE: use only single letter D,M,Y. Multiple patterns can be defined.
-->

<!ELEMENT LC_FORMAT_1 (FormatElement*) >
<!ATTLIST LC_FORMAT_1 %RefLocale;>
<!ATTLIST LC_FORMAT_1 %replaceFrom;>
<!ATTLIST LC_FORMAT_1 %replaceTo;>

<!ELEMENT FormatElement   ( FormatCode, DefaultName?)>
<!ATTLIST FormatElement    %MessageID;>
<!ATTLIST FormatElement    default  (true|false)        #REQUIRED >
<!ATTLIST FormatElement    type     (short|medium|long) #REQUIRED >
<!--
    There may be up to three groups (type="short", type="medium", type="long")
    for each usage category defined. Each group, if defined, needs exactly one
    default. The type roughly determines the display string length, for example
    short, medium, and long date formats.

    The number formatter determines an ultimate default format of a specific
    usage category by looking at the medium, long, and short default formats,
    in that very particular order.
-->
<!ATTLIST FormatElement    usage    (FIXED_NUMBER|FRACTION_NUMBER|PERCENT_NUMBER|SCIENTIFIC_NUMBER|CURRENCY|DATE|TIME|DATE_TIME) #REQUIRED >
<!ATTLIST FormatElement formatindex  CDATA #REQUIRED>
<!--
    The following FormatElements must follow specific rules:

    All:
        The format indices 0..65 are reserved and, for backwards compatibility,
        indices 0..49 MUST be used as stated in
        offapi/com/sun/star/i18n/NumberFormatIndex.idl (
        https://opengrok.libreoffice.org/xref/core/offapi/com/sun/star/i18n/NumberFormatIndex.idl )
        Note that indices 10 ("# ?/?"), 11 ("# ??/??"), 48 (BOOLEAN) and 49 (@
        Text) are generated internally, as they aren't locale dependent, and
        must not be used in locale data XML files. All other formats have to be
        present.

        Note also that "must be used as stated" does not mean that the format
        codes must be identical, of course the meaning of a format code should
        match, for example en_US MM/DD/YY matches de_DE DD.MM.YY
        Just imagine the same index being used with another locale shouldn't
        change the meaning of representation significantly.

        You'll notice differences of non-matching format codes only if
        documents use the Default language in number formats and either are
        stored in old SO5 binary file format and loaded on another system where
        languages/locales aren't the same, or if the default locale is switched
        under menu Tools.Options.LanguageSettings.Languages.LocaleSetting
        dialog, which exchanges formats on the fly in the spreadsheet
        application. Please check it out! So far only very few locale data file
        we received got that right, especially not in date formats!

        For easier comparison between locales in future please sort the
        FormatElements by their formatindex="..." value within a usage group.
        This isn't necessary to be technically correct and isn't done in many
        locales yet, but will certainly help.

        If you want to define yet more formats than the preset and reserved
        0..49 range that's fine, only make sure those formatindex="..." values
        are each >=66 and all values are unique within one locale.

    Of usage="FIXED_NUMBER":
        formatindex="0" MUST be the format containing the 'General' keyword.
        The keyword itself may be localized, it is good practice though to
        stick with a wording known from another spreadsheet application for
        better user experience. Like other format codes it may be prepended
        with a [NatNum1] modifier if values are to be displayed using native
        numbering if no specific format was applied. The format must have the
        default="true" and type="medium" attributes.

    Of usage="DATE":
        formatindex="21" is used to edit already existing date data. In order
        to always edit the full century the long year YYYY code must be used.
        Furthermore, the format has to be of an editable type, of course, which
        means parseable. Therefore it should only contain DD, MM, YYYY and date
        separators, and the YMD default order is determined from the order
        encountered in this format.
        TODO: Future versions should make use of an edit="true" attribute
        instead of relying on this special index requirement.

        formatindex="33" must be ISO 8601 YYYY-MM-DD format code.

    Of usage="DATE_TIME":
        formatindex="46" should contain a short year YY code and only HH and MM
        without SS seconds.

        formatindex="47" is used to edit already existing combined date/time
        data. The requirements are the same as for formatindex="21" above, plus
        HH and MM and SS codes.

        formatindex="50" must contain a long year YYYY code and only HH and MM
        without SS seconds.

    Of usage="TIME":
        * The FormatElement with default="true" type="medium" is used to edit
          times and thus must contain all HH and MM and SS codes, e.g. HH:MM:SS
        * Formatindices 43, 44, 45 are special in the sense that they are
          programmatically used to automatically display values that meet
          certain criteria:
          * Formatindex="43" contains the [HH] format code that displays hour
            values greater than or equal to 24 (as opposed to a simple HH that
            displays modulo 24).
          * Formatindex="44" uses no hour code but 100th seconds and the
            Time100SecSeparator.
          * Formatindex="45" uses both [HH] hour code and 100th seconds to be
            able to edit such time values without losing information.

    Of usage="CURRENCY":
        formatindices 12, 13, 14, 15, 17 with [$xxx-yyy] notation must use the
        xxx currency symbol that has the attribute
        usedInCompatibleFormatCodes="true".  The hexadecimal yyy LANGID must be
        properly set. It can be found in the file include/i18nlangtag/lang.h (
        https://opengrok.libreoffice.org/xref/core/include/i18nlangtag/lang.h )
        You may verify the proper use of the xxx currency symbol with the AWK
        script i18npool/source/localedata/data/currency-check.awk, it mustn't
        display any output. If it does, then there's something wrong.

-->
<!ELEMENT FormatCode      (#PCDATA)>


<!-- The LC_CALENDAR element defines calendars used with a locale. -->
<!ELEMENT LC_CALENDAR (Calendar* ) >
<!-- At least one Calendar element must be given if the RefLocale mechanism is not used! -->
<!ATTLIST LC_CALENDAR %RefLocale;>

<!ELEMENT Calendar (DaysOfWeek, MonthsOfYear, GenitiveMonths*, PartitiveMonths*, Eras, StartDayOfWeek, MinimalDaysInFirstWeek) >
<!ATTLIST Calendar %UNOModule;>
<!-- The unoid of a gregorian calendar MUST be lower case "gregorian",
     calendars MUST match the names defined in the OASIS OpenDocument Format
     (ODF) 1.2 or later specification. The implementation name registered with
     the LibreOffice service registry MUST match, e.g.
     com.sun.star.i18n.Calendar_gregorian
-->
<!ATTLIST Calendar default (true|false) #REQUIRED >
<!-- Exactly one Calendar element has to be the default calendar. -->

<!ELEMENT DaysOfWeek (Day*)>
<!-- All Day elements of a Calendar must be given if the RefLocale mechanism is not used! -->
<!ATTLIST DaysOfWeek %RefLocale;>
<!-- Sequence of days is important, MUST start with Sunday. -->
<!ELEMENT Day (DayID, DefaultAbbrvName, DefaultFullName, DefaultNarrowName*)>
<!ELEMENT DayID (#PCDATA)>
<!-- Preferably the lower case abbreviated English name like sun for Sunday. -->
<!ELEMENT DefaultAbbrvName (#PCDATA)>
<!-- The abbreviated day name, e.g. Sun for Sunday. -->
<!ELEMENT DefaultFullName (#PCDATA)>
<!-- The full day name, e.g. Sunday for Sunday. -->
<!ELEMENT DefaultNarrowName (#PCDATA)>
<!-- The narrow day name, e.g. S for Sunday.
     If not specified, the first letter of the corresponding DefaultFullName is taken.
 -->

<!ELEMENT MonthsOfYear (Month*)>
<!-- Nominative month names.
     All Month elements of a Calendar must be given if the RefLocale mechanism is not used!
 -->
<!ATTLIST MonthsOfYear %RefLocale;>
<!-- Sequence of months is important, MUST start with the first month of a
     year, e.g. January in a Gregorian calendar.
 -->

<!ELEMENT GenitiveMonths (Month*)>
<!-- Possessive genitive case month names, for example in Slavic locales. The
     element is optional, but if present all Month elements of a Calendar must
     be given if the RefLocale mechanism is not used! If not specified,
     the MonthsOfYear names will be used in the context of the number
     formatter's genitive case. -->
<!ATTLIST GenitiveMonths %RefLocale;>
<!-- Sequence of months is important, MUST start with the first month of a
     year, e.g. January in a Gregorian calendar.
 -->

<!ELEMENT PartitiveMonths (Month*)>
<!-- Partitive case month names, for example in Finnish locales. The
     element is optional, but if present all Month elements of a Calendar must
     be given if the RefLocale mechanism is not used! If not specified,
     GenitiveMonths names will be used, or if those are not specified then
     MonthsOfYear, in the context of the number formatter's partitive case. -->
<!ATTLIST PartitiveMonths %RefLocale;>
<!-- Sequence of months is important, MUST start with the first month of a
     year, e.g. January in a Gregorian calendar.
 -->

<!-- Rules for use of nominative / genitive / partitive case month names in
     number formatter when encountering MMM or MMMM:

     * MMM or MMMM immediately preceded or followed by a literal character
       other than space => nominative month name (noun), for Excel and
       backwards compatibility such as Finnish MMMM"ta"
     * no day of month (D or DD) present in format code => nominative name
     * day of month (D or DD) after MMM or MMMM => genitive name
       * no genitive names defined => nominative name
     * day of month (D or DD) before MMM or MMMM => partitive name
       * no partitive names defined => genitive name
         * no genitive names defined => nominative name

     NOTE:

     If only <MonthsOfYear> and <PartitiveMonths> are specified but not
     <GenitiveMonths>, then for MMM(M) D(D) formats the <MonthsOfYear>
     nominative name is displayed. Only for D(D) MMM(M) formats the
     <PartitiveMonths> name is displayed.

     If only for MMM(M) D(D) formats the <GenitiveMonths> are to be displayed
     but nominative names for D(D) MMM(M), then specify <PartitiveMonths>
     identical to <MonthsOfYear>, do not omit it as otherwise it would inherit
     from <GenitiveMonths> again.

 -->

<!ELEMENT Month (MonthID, DefaultAbbrvName, DefaultFullName, DefaultNarrowName*)>
<!ELEMENT MonthID (#PCDATA)>
<!-- Preferably the lower case abbreviated English name like jan for January. -->

<!ELEMENT Eras (Era*)>
<!-- All Era elements of a Calendar must be given if the RefLocale mechanism is not used! -->
<!ATTLIST Eras %RefLocale;>
<!-- The eras MUST be in chronological order, e.g. first BC then AC. -->
<!ELEMENT Era (EraID, DefaultAbbrvName, DefaultFullName)>
<!ELEMENT EraID (#PCDATA)>
<!-- If a calendar has special eras (like zh_TW ROC or ja_JP Gengou calendar)
     and a date before those eras is undefined, a leading (first) dummy era
     with EraID="Dummy" has to be defined to enable the number formatter to
     fall back to a Gregorian calendar for those date values if the XCalendar
     implementation returns an era value of 0.
-->

<!ELEMENT StartDayOfWeek (DayID)>
<!-- MUST exactly match (case significant!) one of the DayID of DaysOfWeek -->

<!ELEMENT MinimalDaysInFirstWeek (#PCDATA)>
<!-- The number of days of a week that must reside in the beginning of a year
     to make a week the first week of the year. For example, a value of 4 means
     that at least 4 days of a week must be in the new year. So if the week
     starts on Monday, the first week of a year will be the week where Thursday
     is in the new year.
-->


<!-- The LC_CURRENCY element defines currencies used with a locale. -->
<!ELEMENT LC_CURRENCY (Currency* ) >
<!-- At least one Currency element must be given if the RefLocale mechanism is not used! -->
<!ATTLIST LC_CURRENCY %RefLocale;>
<!ELEMENT Currency (CurrencyID, CurrencySymbol, BankSymbol, CurrencyName, DecimalPlaces)>
<!ATTLIST Currency  default (true|false) #REQUIRED >
<!-- Exactly one Currency element has to be the default currency. -->
<!ATTLIST Currency  usedInCompatibleFormatCodes (true|false) #REQUIRED >
<!-- If this currency is the one used in compatible number format codes with
     <member>FormatElement::formatIndex</member> values in the range 12..17.
     Those format codes are used to generate some old style currency format
     codes for compatibility with StarOffice5 and StarOffice4.
     Every locale data file MUST contain exactly one currency having this set to "true",
     and that currency MUST be used in format codes 12..17.
     For European countries using EUR it MUST be the old currency, for example, DM.
-->
<!ATTLIST Currency  legacyOnly (true|false) #IMPLIED >
<!-- If this Currency element exists only to be able to correctly load legacy
     documents and is not selectable in the UI otherwise. Defaults to "false"
     if not specified. If this attribute is "true", 'default' and
     'usedInCompatibleFormatCodes' must both be "false".

     Currency elements must not be changed to contain only a different
     CurrencySymbol element without changing the CurrencyID and BankSymbol
     elements, instead the entire Currency element must be duplicated, the old
     element needs this 'legacyOnly' attribute be added and 'default' and
     'usedInCompatibleFormatCodes' attributes must be set to "false", and only
     in the new duplicated Currency element the CurrencySymbol element be
     changed.
-->
<!ELEMENT CurrencyID (#PCDATA)>
<!-- The ISO 4217 three letter currency code, e.g. USD or EUR. -->
<!ELEMENT CurrencySymbol (#PCDATA)>
<!-- The currency symbol, e.g. $ or €. -->
<!ELEMENT BankSymbol (#PCDATA)>
<!-- The ISO 4217 three letter currency code, e.g. USD or EUR. -->
<!ELEMENT CurrencyName (#PCDATA)>
<!-- The native currency name, e.g. Dollar or Euro. -->
<!ELEMENT DecimalPlaces (#PCDATA)>
<!-- Number of decimal places used with the currency, usually 2 or 0, e.g. 2
     for cents.
-->


<!ELEMENT LC_CTYPE (Separators?, Markers?, TimeAM?, TimePM?,  MeasurementSystem?)>
<!-- All elements must be given if the RefLocale mechanism is not used! -->
<!ATTLIST LC_CTYPE  %RefLocale;>
<!ATTLIST LC_CTYPE  %UNOModule;>

<!ELEMENT Separators (DateSeparator, ThousandSeparator, DecimalSeparator, DecimalSeparatorAlternative?, TimeSeparator, Time100SecSeparator, ListSeparator, LongDateDayOfWeekSeparator, LongDateDaySeparator, LongDateMonthSeparator, LongDateYearSeparator)>
<!ELEMENT DateSeparator (#PCDATA)>
<!ELEMENT ThousandSeparator (#PCDATA)>
<!ELEMENT DecimalSeparator (#PCDATA)>
<!ELEMENT DecimalSeparatorAlternative (#PCDATA)>
<!ELEMENT TimeSeparator (#PCDATA)>
<!ELEMENT Time100SecSeparator (#PCDATA)>
<!ELEMENT ListSeparator (#PCDATA)>
<!ELEMENT LongDateDayOfWeekSeparator (#PCDATA)>
<!ELEMENT LongDateDaySeparator (#PCDATA)>
<!ELEMENT LongDateMonthSeparator (#PCDATA)>
<!ELEMENT LongDateYearSeparator (#PCDATA)>

<!ELEMENT Markers (QuotationStart, QuotationEnd, DoubleQuotationStart, DoubleQuotationEnd)>
<!ELEMENT QuotationStart (#PCDATA)>
<!ELEMENT QuotationEnd  (#PCDATA)>
<!ELEMENT DoubleQuotationStart (#PCDATA)>
<!ELEMENT DoubleQuotationEnd (#PCDATA)>

<!ELEMENT TimeAM (#PCDATA)>
<!ELEMENT TimePM (#PCDATA)>
<!ELEMENT MeasurementSystem  (#PCDATA)>


<!ELEMENT LC_COLLATION (Collator*, CollationOptions?)>
<!-- All elements must be given if the RefLocale mechanism is not used! -->
<!ATTLIST LC_COLLATION %RefLocale;>
<!--
	Optional ICU tailoring.

	See Collation Customization in ICU User Guide for syntax,
	https://unicode-org.github.io/icu/userguide/collation/customization/

	There are two ways to add language specific tailoring in LibreOffice.
	For small tailoring, you can directly add it in locale data here. For
	large tailoring, it is suggested to put the data under collator/data, so
	it will be compiled to a binary format in build time and improve performance
	in run time.

-->
<!ELEMENT Collator (#PCDATA)>
<!ATTLIST Collator  %UNOModule;>
<!ATTLIST Collator  default (true|false) #REQUIRED >
<!ELEMENT CollationOptions (TransliterationModules+)>
<!ELEMENT TransliterationModules (#PCDATA)>


<!ELEMENT LC_SEARCH (SearchOptions?)>
<!-- All elements must be given if the RefLocale mechanism is not used! -->
<!ATTLIST LC_SEARCH %RefLocale;>
<!ELEMENT SearchOptions (TransliterationModules+)>


<!ELEMENT LC_INDEX (IndexKey*, UnicodeScript*, FollowPageWord*)>
<!ATTLIST LC_INDEX %RefLocale;>

<!--
    The IndexKey element is optional, but should be given if the locale
    requires a specific sort order in Writer's index tables or entries are to
    be combined under keys.

    Index key for the algorithm and language, like >A-Z< for English => A, B,
    C, ..., Y, Z. The letters specify under which key an entry goes and the
    order the keys are sorted. Keys may be reordered or letters inserted to
    form a specific order, for example (ve_ZA) >A-D Ḓ E-L Ḽ M N Ṋ Ṅ O-T Ṱ U-Z<
    Entries that don't match a defined key are appended to the index list. Used
    in Writer textprocessor.

    The initial data was setup according to the ICU collation chart at
    http://oss.software.ibm.com/icu/charts/collation/
    Note: ICU site was moved to https://icu.unicode.org/ and as ICU per
    default uses CLDR, collation charts are available at
    https://www.unicode.org/cldr/charts/latest/

    Possible notations in the syntax of the IndexKey element are:

    '-' (dash): Ellipsis, all letters elided by the ellipsis are included as
    index keys in alphabetic order. For example, 'A-Z' includes all ASCII
    letters A to Z.

    '[]' (square brackets): all letters included in square brackets are
    skipping letters. It is used for CTL languages, for example in Thai
    (th_TH), to skip prefix vowels or signs. For example, if ["] double quote
    is defined as skipping letter, index item '"Index"' will be under 'I', not
    '"'.

    '{}' (curly brackets): define multiple letters index key, for example
    '{Cs}' is one of the index keys for Hungarian (hu_HU).

    '()' (parentheses): define optional description for index key. If defined,
    the description will be shown as index key title, instead of the index key
    itself. For example, 'E(E, É)' in Hungarian adds entries with 'E' to the
    description 'E, É'.

-->
<!ELEMENT IndexKey (#PCDATA)>
<!ATTLIST IndexKey  %UNOModule;>
<!ATTLIST IndexKey  %LIBModule;>
<!ATTLIST IndexKey  default (true|false) #REQUIRED >
<!ATTLIST IndexKey  phonetic (true|false) #REQUIRED >

<!--
    The Unicode script types are those of
    offapi/com/sun/star/i18n/UnicodeScript.idl, they define the code range for
    the language.
-->
<!ELEMENT UnicodeScript (#PCDATA)>

<!--
    The FollowPageWord entries were originally hard-coded in
    ../../indexentry/indexentrysupplier.cxx, most locales used the English
    ``p.'' and ``pp.'', valid data should be provided by native speakers.
    These words or abbreviations are used in the Writer's index table. The
    first FollowPageWord element is the abbreviation for "page" (p.), the
    second FollowPageWord element the abbreviation for "page and following
    pages" (pp.).
-->
<!ELEMENT FollowPageWord (#PCDATA)>


<!ELEMENT LC_TRANSLITERATION (Transliteration*)>
<!ATTLIST LC_TRANSLITERATION %RefLocale;>
<!ELEMENT Transliteration EMPTY>
<!ATTLIST Transliteration  %UNOModule;>

<!ELEMENT LC_MISC (ForbiddenCharacters?, BreakIteratorRules?, ReservedWords?)>
<!ATTLIST LC_MISC %RefLocale;>
<!ELEMENT ForbiddenCharacters (ForbiddenLineBeginCharacters, ForbiddenLineEndCharacters, LineBreakHangingCharacters)>
<!ELEMENT ForbiddenLineBeginCharacters (#PCDATA)>
<!ELEMENT ForbiddenLineEndCharacters (#PCDATA)>

<!-- optional break iterator rules for the languages.
     if defined, 5 rules should be all listed, each of them can be blank and default will be used.
     order of the rules is significant!
-->
<!ELEMENT BreakIteratorRules (EditMode, DictionaryMode, WordCountMode, CharacterMode, LineMode)>
<!ELEMENT EditMode (#PCDATA)>
<!ELEMENT DictionaryMode (#PCDATA)>
<!ELEMENT WordCountMode (#PCDATA)>
<!ELEMENT CharacterMode (#PCDATA)>
<!ELEMENT LineMode (#PCDATA)>

<!ELEMENT ReservedWords  (trueWord, falseWord, quarter1Word, quarter2Word, quarter3Word, quarter4Word, aboveWord, belowWord, quarter1Abbreviation, quarter2Abbreviation, quarter3Abbreviation, quarter4Abbreviation)>
<!-- order is significant! -->
<!-- trueWord and falseWord are displayed (uppercased) for the BOOLEAN number
     format keyword. "true", "false" -->
<!ELEMENT trueWord (#PCDATA)>
<!ELEMENT falseWord (#PCDATA)>
<!-- quarter1Word,... are displayed for the QQ number format keyword.
     "1st quarter", "2nd quarter", ...; "1er trimestre", "2e trimestre", ... -->
<!ELEMENT quarter1Word (#PCDATA)>
<!ELEMENT quarter2Word (#PCDATA)>
<!ELEMENT quarter3Word (#PCDATA)>
<!ELEMENT quarter4Word (#PCDATA)>
<!-- aboveWord and belowWord are used as reference field content in Writer, it
     is a "physical" object relative position, not "numerical" relative.
     Like section "above" and section "below". -->
<!ELEMENT aboveWord (#PCDATA)>
<!ELEMENT belowWord (#PCDATA)>
<!-- quarter1Abbreviation,... are displayed for the Q number format keyword.
     "Q1" (quarter), "T1" (trimestre), ... -->
<!ELEMENT quarter1Abbreviation (#PCDATA)>
<!ELEMENT quarter2Abbreviation (#PCDATA)>
<!ELEMENT quarter3Abbreviation (#PCDATA)>
<!ELEMENT quarter4Abbreviation (#PCDATA)>


<!--
    In numbering levels, the NumType attribute is a value of the constants
    defined in offapi/com/sun/star/style/NumberingType.idl (
    https://opengrok.libreoffice.org/xref/core/offapi/com/sun/star/style/NumberingType.idl
    )
-->

<!ELEMENT LC_NumberingLevel (NumberingLevel* )>
<!ATTLIST LC_NumberingLevel %RefLocale;>
<!ELEMENT NumberingLevel EMPTY>
<!ATTLIST NumberingLevel Prefix CDATA #REQUIRED>
<!ATTLIST NumberingLevel NumType CDATA #REQUIRED>
<!ATTLIST NumberingLevel Suffix CDATA  #REQUIRED>
<!ATTLIST NumberingLevel Transliteration CDATA #IMPLIED>
<!ATTLIST NumberingLevel NatNum CDATA #IMPLIED>


<!ELEMENT LC_OutLineNumberingLevel (OutlineStyle* )>
<!ATTLIST LC_OutLineNumberingLevel %RefLocale;>
<!ELEMENT OutlineStyle (OutLineNumberingLevel+)>
<!ELEMENT OutLineNumberingLevel EMPTY>
<!ATTLIST OutLineNumberingLevel Prefix CDATA  #REQUIRED >
<!ATTLIST OutLineNumberingLevel NumType CDATA  #REQUIRED >
<!ATTLIST OutLineNumberingLevel Suffix CDATA  #REQUIRED >
<!ATTLIST OutLineNumberingLevel BulletChar CDATA  #REQUIRED >
<!ATTLIST OutLineNumberingLevel BulletFontName CDATA   #REQUIRED>
<!ATTLIST OutLineNumberingLevel ParentNumbering CDATA  #REQUIRED >
<!ATTLIST OutLineNumberingLevel LeftMargin CDATA   #REQUIRED>
<!ATTLIST OutLineNumberingLevel SymbolTextDistance CDATA  #REQUIRED >
<!ATTLIST OutLineNumberingLevel FirstLineOffset CDATA  #REQUIRED >
<!ATTLIST OutLineNumberingLevel Adjust CDATA #IMPLIED >
<!ATTLIST OutLineNumberingLevel Transliteration CDATA #IMPLIED >
<!ATTLIST OutLineNumberingLevel NatNum CDATA #IMPLIED>