| blob | e348fdc1905a43f5a140d724eed00d05b3a7194a |
2 <!--
3 * This file is part of the LibreOffice project.
4 *
5 * This Source Code Form is subject to the terms of the Mozilla Public
6 * License, v. 2.0. If a copy of the MPL was not distributed with this
7 * file, You can obtain one at http://mozilla.org/MPL/2.0/.
8 *
9 * This file incorporates work covered by the following license notice:
10 *
11 * Licensed to the Apache Software Foundation (ASF) under one or more
12 * contributor license agreements. See the NOTICE file distributed
13 * with this work for additional information regarding copyright
14 * ownership. The ASF licenses this file to you under the Apache
15 * License, Version 2.0 (the "License"); you may not use this file
16 * except in compliance with the License. You may obtain a copy of
17 * the License at http://www.apache.org/licenses/LICENSE-2.0 .
18 -->
20 <!-- ............................................................... -->
21 <!-- Locale data specification DTD ................................. -->
22 <!-- ............................................................... -->
24 <!--
25 ===========================================================================
26 ATTENTION! PLEASE! HEADS UP! IMPORTANT!
27 ===========================================================================
29 Please validate your locale data contribution.
31 A quick validation check can be done if you have xmllint installed and the
32 current locale.dtd file at hand, you can download the locale.dtd file from
33 https://cgit.freedesktop.org/libreoffice/core/plain/i18npool/source/localedata/data/locale.dtd
35 xmllint --dtdvalid locale.dtd --noout your_data.xml
38 Second, using a validating parser. A validating parser, for example, may be found at
39 http://unicode.org/cldr/data/tools/java/org/unicode/cldr/util/XMLValidator.java
40 Compile it into a class-jar and call it in the
41 i18npool/source/localedata/data/ directory:
42 java -cp <your_path>/XMLValidator.jar org.unicode.cldr.util.XMLValidator your_data.xml
45 A third possibility is:
47 - temporarily (!) change the DOCTYPE of your file to read (all on one line)
48 <!DOCTYPE Locale SYSTEM "https://cgit.freedesktop.org/libreoffice/core/plain/i18npool/source/localedata/data/locale.dtd">
50 - upload it to the form available at http://www.validome.org/xml/
52 This will validate the file against the HEAD revision of locale.dtd
55 Please test locale data files either in an enable-dbgutil build, which
56 implements some checks and pops up assertion message boxes if the tests
57 fail, or by setting the environment variable OOO_ENABLE_LOCALE_DATA_CHECKS
58 to 'Y' or 'Yes' (or any other string starting with 'Y') or '1' before
59 starting the application, which outputs the same messages to stderr and
60 also works in a product build.
62 Then follow this procedure:
63 1. Create a new spreadsheet document.
64 2. On a cell use context menu -> Format Cells -> Numbers.
65 3. Select the locale in the Language list box => MUST be assertion free.
66 3.a. Assertions are only shown at the very first time a number formatter
67 is created, respectively the first time a specific locale data is
68 used. To repeat steps 1.-3. you'd need to create another spreadsheet
69 document.
71 ===========================================================================
72 NOTE the FormatElement comments further down.
73 ===========================================================================
75 -->
77 <!-- ............................................................... -->
78 <!-- Entities for characters and symbols ........................... -->
80 <!ENTITY % UNOModule
81 'unoid CDATA #IMPLIED'>
83 <!ENTITY % MessageID
84 'msgid CDATA #REQUIRED'>
86 <!ENTITY % RefLocale
87 'ref CDATA #IMPLIED'>
88 <!-- Where given, an element can be inherited from another locale, e.g.
89 ref="en_US" -->
91 <!ENTITY % LIBModule
92 'module CDATA #IMPLIED'>
93 <!-- The locale referred to for the sub categories, implementation detail. -->
97 <!-- See below for the LC_FORMAT element. -->
101 <!-- Locale is made of different sub categories -->
102 <!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)>
104 <!-- Version identifier to prevent mismatching data files being submitted
105 because of older files being copied as templates. The version should be
106 less than or equal to the LibreOffice release, or the LibreOffice release
107 number plus some extension, like "2.0.enhanced", to be able to easily
108 determine the corresponding LibreOffice version. New versions of the DTD
109 with new required elements SHOULD REALLY result in a new versionDTD here
110 and LOCALE_VERSION_DTD in ../LocaleNode.cxx, and *.xml files MUST be
111 adapted then, otherwise building the data or checking it with a validating
112 parser will throw an error.
113 -->
115 <!-- Whether some data elements may be (automatically) updated from the Common
116 Locale Data Repository, see http://cldr.unicode.org/
117 Note: This mechanism currently (2010-02-21) is outdated.
118 -->
120 <!-- Valid number, may designate versioned data -->
127 <!-- LangID must be a valid two or three letter language identifier defined by
128 ISO 639. Use ISO 639-1 two letter code where available, else ISO 639-2 or
129 639-3 three letter code.
131 If the Variant element designates a BCP 47 language tag (because the
132 locale is not expressible as a Language,Country pair, for example contains
133 a script tag) then the LangID value MUST be the code 'qlt' (reserved by
134 ISO 639-3 for private use) to tell the language tag processor that the
135 actual language tag is in the Variant element.
136 -->
140 <!-- CountryID must be a valid two letter country identifier defined by ISO 3166. -->
144 <!-- Unused, deprecated, can be generic|unix|win32|macos, best Platform element
145 be absent.
146 -->
149 <!-- If the LangID element contains the value 'qlt' then the Variant element
150 MUST contain the valid BCP 47 language tag of the locale. If LangID is
151 another ISO 639 code than 'qlt' then the Variant element must be empty or
152 not present.
153 -->
156 <!-- The LC_FORMAT element contains number format codes and may actually appear
157 twice (with the second occurrence named LC_FORMAT_1). One section is
158 mandatory and MUST contain at least all required format codes
159 formatindex="0" to formatindex="47", MUST NOT contain formatindex 48 and
160 49, and MAY contain other format codes. See below ATTLIST FormatElement
161 formatindex.
163 A second LC_FORMAT_1 section may follow containing other format codes. The
164 difference between the two sections is that they are inherited
165 independently if a locale uses the RefLocale mechanism (ref="..."). This
166 may be used to not offer locale dependent format codes to other locales
167 that otherwise inherit the format codes.
169 -->
171 <!-- All FormatElement elements must be given if the RefLocale mechanism is not used! -->
174 <!-- Define placeholder for currency code, usually "[CURRENCY]" -->
176 <!-- Currency code to be used to replace the placeholder, e.g. "[$R-1C09]".
177 Note: The brackets and the leading $ character are mandatory, the
178 hyphen-minus separates the currency symbol from the hexagesimal MS-LCID,
179 letters contained in the LCID have to be in upper case, leading zeros are
180 to be omitted. LCIDs are defined in include/i18nlangtag/lang.h (
181 https://opengrok.libreoffice.org/xref/core/include/i18nlangtag/lang.h )
182 -->
185 <!-- Pattern that defines an input sequence match to be accepted as
186 (abbreviated) date. For example, in en_US locale "M/D" accepts an input of
187 11/23 as CurrentYear-November-23 whereas 11/23/ would not be a date. In
188 de_DE locale "D.M." accepts an input of 23.11. as CurrentYear-November-23
189 whereas 23.11 would not be a date.
191 For each locale one pattern that matches a full date is automatically
192 generated from FormatElement formatIndex="21" and does not need to be
193 defined, for example "M/D/Y" or "D.M.Y".
195 At least one pattern for abbreviated date input must be defined, i.e.
196 contain D and M but not Y.
198 NOTE: use only single letter D,M,Y. Multiple patterns can be defined.
199 -->
210 <!--
211 There may be up to three groups (type="short", type="medium", type="long")
212 for each usage category defined. Each group, if defined, needs exactly one
213 default. The type roughly determines the display string length, for example
214 short, medium, and long date formats.
216 The number formatter determines an ultimate default format of a specific
217 usage category by looking at the medium, long, and short default formats,
218 in that very particular order.
219 -->
220 <!ATTLIST FormatElement usage (FIXED_NUMBER|FRACTION_NUMBER|PERCENT_NUMBER|SCIENTIFIC_NUMBER|CURRENCY|DATE|TIME|DATE_TIME) #REQUIRED >
222 <!--
223 The following FormatElements must follow specific rules:
225 All:
226 The format indices 0..59 are reserved and, for backwards compatibility,
227 indices 0..49 MUST be used as stated in
228 offapi/com/sun/star/i18n/NumberFormatIndex.idl (
229 https://opengrok.libreoffice.org/xref/core/offapi/com/sun/star/i18n/NumberFormatIndex.idl )
230 Note that indices 10 ("# ?/?"), 11 ("# ??/??"), 48 (BOOLEAN) and 49 (@
231 Text) are generated internally, as they aren't locale dependent, and
232 must not be used in locale data XML files. All other formats have to be
233 present.
235 Note also that "must be used as stated" does not mean that the format
236 codes must be identical, of course the meaning of a format code should
237 match, for example en_US MM/DD/YY matches de_DE DD.MM.YY
238 Just imagine the same index being used with another locale shouldn't
239 change the meaning of representation significantly.
241 You'll notice differences of non-matching format codes only if
242 documents use the Default language in number formats and either are
243 stored in old SO5 binary file format and loaded on another system where
244 languages/locales aren't the same, or if the default locale is switched
245 under menu Tools.Options.LanguageSettings.Languages.LocaleSetting
246 dialog, which exchanges formats on the fly in the spreadsheet
247 application. Please check it out! So far only very few locale data file
248 we received got that right, especially not in date formats!
250 For easier comparison between locales in future please sort the
251 FormatElements by their formatindex="..." value within a usage group.
252 This isn't necessary to be technically correct and isn't done in many
253 locales yet, but will certainly help.
255 If you want to define yet more formats than the preset and reserved
256 0..49 range that's fine, only make sure those formatindex="..." values
257 are each >=60 and all values are unique within one locale.
259 Of usage="FIXED_NUMBER":
260 formatindex="0" MUST be the format containing the 'General' keyword.
261 The keyword itself may be localized, it is good practice though to
262 stick with a wording known from another spreadsheet application for
263 better user experience. Like other format codes it may be prepended
264 with a [NatNum1] modifier if values are to be displayed using native
265 numbering if no specific format was applied. The format must have the
266 default="true" and type="medium" attributes.
268 Of usage="DATE":
269 formatindex="21" is used to edit already existing date data. In order
270 to always edit the full century the long year YYYY code must be used.
271 Furthermore, the format has to be of an editable type, of course, which
272 means parseable. Therefore it should only contain DD, MM, YYYY and date
273 separators, and the YMD default order is determined from the order
274 encountered in this format.
275 TODO: Future versions should make use of an edit="true" attribute
276 instead of relying on this special requirement.
278 formatindex="33" must be ISO 8601 YYYY-MM-DD format code.
280 Of usage="DATE_TIME":
281 formatindex="47" is used to edit already existing combined date/time
282 data. The requirements are the same as for formatindex="21" above.
284 Of usage="TIME":
285 * The FormatElement with default="true" type="medium" is used to edit
286 times and thus must contain all HH and MM and SS codes, e.g. HH:MM:SS
287 * Formatindices 43, 44, 45 are special in the sense that they are
288 programmatically used to automatically display values that meet
289 certain criteria:
290 * Formatindex="43" contains the [HH] format code that displays hour
291 values greater than or equal to 24 (as opposed to a simple HH that
292 displays modulo 24).
293 * Formatindex="44" uses no hour code but 100th seconds and the
294 Time100SecSeparator.
295 * Formatindex="45" uses both [HH] hour code and 100th seconds to be
296 able to edit such time values without losing information.
298 Of usage="CURRENCY":
299 formatindices 12, 13, 14, 15, 17 with [$xxx-yyy] notation must use the
300 xxx currency symbol that has the attribute
301 usedInCompatibleFormatCodes="true". The hexadecimal yyy LANGID must be
302 properly set. It can be found in the file include/i18nlangtag/lang.h (
303 https://opengrok.libreoffice.org/xref/core/include/i18nlangtag/lang.h )
304 You may verify the proper use of the xxx currency symbol with the AWK
305 script i18npool/source/localedata/data/currency-check.awk, it mustn't
306 display any output. If it does, then there's something wrong.
308 -->
312 <!-- The LC_CALENDAR element defines calendars used with a locale. -->
314 <!-- At least one Calendar element must be given if the RefLocale mechanism is not used! -->
317 <!ELEMENT Calendar (DaysOfWeek, MonthsOfYear, GenitiveMonths*, PartitiveMonths*, Eras, StartDayOfWeek, MinimalDaysInFirstWeek) >
319 <!-- The unoid of a gregorian calendar MUST be lower case "gregorian",
320 calendars MUST match the names defined in the OASIS OpenDocument Format
321 (ODF) 1.2 or later specification. The implementation name registered with
322 the LibreOffice service registry MUST match, e.g.
323 com.sun.star.i18n.Calendar_gregorian
324 -->
326 <!-- Exactly one Calendar element has to be the default calendar. -->
329 <!-- All Day elements of a Calendar must be given if the RefLocale mechanism is not used! -->
331 <!-- Sequence of days is important, MUST start with Sunday. -->
334 <!-- Preferably the lower case abbreviated English name like sun for Sunday. -->
336 <!-- The abbreviated day name, e.g. Sun for Sunday. -->
338 <!-- The full day name, e.g. Sunday for Sunday. -->
340 <!-- The narrow day name, e.g. S for Sunday.
341 If not specified, the first letter of the corresponding DefaultFullName is taken.
342 -->
345 <!-- Nominative month names.
346 All Month elements of a Calendar must be given if the RefLocale mechanism is not used!
347 -->
349 <!-- Sequence of months is important, MUST start with the first month of a
350 year, e.g. January in a Gregorian calendar.
351 -->
354 <!-- Possessive genitive case month names, for example in Slavic locales. The
355 element is optional, but if present all Month elements of a Calendar must
356 be given if the RefLocale mechanism is not used! If not specified,
357 the MonthsOfYear names will be used in the context of the number
358 formatter's genitive case. -->
360 <!-- Sequence of months is important, MUST start with the first month of a
361 year, e.g. January in a Gregorian calendar.
362 -->
365 <!-- Partitive case month names, for example in Finnish locales. The
366 element is optional, but if present all Month elements of a Calendar must
367 be given if the RefLocale mechanism is not used! If not specified,
368 GenitiveMonths names will be used, or if those are not specified then
369 MonthsOfYear, in the context of the number formatter's partitive case. -->
371 <!-- Sequence of months is important, MUST start with the first month of a
372 year, e.g. January in a Gregorian calendar.
373 -->
375 <!-- Rules for use of nominative / genitive / partitive case month names in
376 number formatter when encountering MMM or MMMM:
378 * MMM or MMMM immediately preceded or followed by a literal character
379 other than space => nominative month name (noun), for Excel and
380 backwards compatibility such as Finnish MMMM"ta"
381 * no day of month (D or DD) present in format code => nominative name
382 * day of month (D or DD) after MMM or MMMM => genitive name
383 * no genitive names defined => nominative name
384 * day of month (D or DD) before MMM or MMMM => partitive name
385 * no partitive names defined => genitive name
386 * no genitive names defined => nominative name
388 NOTE:
390 If only <MonthsOfYear> and <PartitiveMonths> are specified but not
391 <GenitiveMonths>, then for MMM(M) D(D) formats the <MonthsOfYear>
392 nominative name is displayed. Only for D(D) MMM(M) formats the
393 <PartitiveMonths> name is displayed.
395 If only for MMM(M) D(D) formats the <GenitiveMonths> are to be displayed
396 but nominative names for D(D) MMM(M), then specify <PartitiveMonths>
397 identical to <MonthsOfYear>, do not omit it as otherwise it would inherit
398 from <GenitiveMonths> again.
400 -->
404 <!-- Preferably the lower case abbreviated English name like jan for January. -->
407 <!-- All Era elements of a Calendar must be given if the RefLocale mechanism is not used! -->
409 <!-- The eras MUST be in chronological order, e.g. first BC then AC. -->
412 <!-- If a calendar has special eras (like zh_TW ROC or ja_JP Gengou calendar)
413 and a date before those eras is undefined, a leading (first) dummy era
414 with EraID="Dummy" has to be defined to enable the number formatter to
415 fall back to a Gregorian calendar for those date values if the XCalendar
416 implementation returns an era value of 0.
417 -->
420 <!-- MUST exactly match (case significant!) one of the DayID of DaysOfWeek -->
423 <!-- The number of days of a week that must reside in the beginning of a year
424 to make a week the first week of the year. For example, a value of 4 means
425 that at least 4 days of a week must be in the new year. So if the week
426 starts on Monday, the first week of a year will be the week where Thursday
427 is in the new year.
428 -->
431 <!-- The LC_CURRENCY element defines currencies used with a locale. -->
433 <!-- At least one Currency element must be given if the RefLocale mechanism is not used! -->
437 <!-- Exactly one Currency element has to be the default currency. -->
439 <!-- If this currency is the one used in compatible number format codes with
440 <member>FormatElement::formatIndex</member> values in the range 12..17.
441 Those format codes are used to generate some old style currency format
442 codes for compatibility with StarOffice5 and StarOffice4.
443 Every locale data file MUST contain exactly one currency having this set to "true",
444 and that currency MUST be used in format codes 12..17.
445 For European countries using EUR it MUST be the old currency, for example, DM.
446 -->
448 <!-- If this Currency element exists only to be able to correctly load legacy
449 documents and is not selectable in the UI otherwise. Defaults to "false"
450 if not specified. If this attribute is "true", 'default' and
451 'usedInCompatibleFormatCodes' must both be "false".
453 Currency elements must not be changed to contain only a different
454 CurrencySymbol element without changing the CurrencyID and BankSymbol
455 elements, instead the entire Currency element must be duplicated, the old
456 element needs this 'legacyOnly' attribute be added and 'default' and
457 'usedInCompatibleFormatCodes' attributes must be set to "false", and only
458 in the new duplicated Currency element the CurrencySymbol element be
459 changed.
460 -->
462 <!-- The ISO 4217 three letter currency code, e.g. USD or EUR. -->
464 <!-- The currency symbol, e.g. $ or €. -->
466 <!-- The ISO 4217 three letter currency code, e.g. USD or EUR. -->
468 <!-- The native currency name, e.g. Dollar or Euro. -->
470 <!-- Number of decimal places used with the currency, usually 2 or 0, e.g. 2
471 for cents.
472 -->
476 <!-- All elements must be given if the RefLocale mechanism is not used! -->
480 <!ELEMENT Separators (DateSeparator, ThousandSeparator, DecimalSeparator, DecimalSeparatorAlternative?, TimeSeparator, Time100SecSeparator, ListSeparator, LongDateDayOfWeekSeparator, LongDateDaySeparator, LongDateMonthSeparator, LongDateYearSeparator)>
505 <!-- All elements must be given if the RefLocale mechanism is not used! -->
507 <!--
508 Optional ICU tailoring.
510 See Collation Customization in ICU User Guide for syntax,
511 http://www.icu-project.org/userguide/Collate_Customization.html
513 There are two ways to add language specific tailoring in LibreOffice.
514 For small tailoring, you can directly add it in locale data here. For
515 large tailoring, it is suggested to put the data under collator/data, so
516 it will be compiled to a binary format in build time and improve performance
517 in run time.
519 -->
528 <!-- All elements must be given if the RefLocale mechanism is not used! -->
536 <!--
537 The IndexKey element is optional, but should be given if the locale
538 requires a specific sort order in Writer's index tables or entries are to
539 be combined under keys.
541 Index key for the algorithm and language, like >A-Z< for English => A, B,
542 C, ..., Y, Z. The letters specify under which key an entry goes and the
543 order the keys are sorted. Keys may be reordered or letters inserted to
544 form a specific order, for example (ve_ZA) >A-D Ḓ E-L Ḽ M N Ṋ Ṅ O-T Ṱ U-Z<
545 Entries that don't match a defined key are appended to the index list. Used
546 in Writer textprocessor.
548 The initial data was setup according to the ICU collation chart at
549 http://oss.software.ibm.com/icu/charts/collation/
550 Note: ICU site was moved to http://www.icu-project.org/ and as ICU per
551 default uses CLDR, collation charts are available at
552 http://unicode.org/cldr/comparison_charts.html#Collation
554 Possible notations in the syntax of the IndexKey element are:
556 '-' (dash): Ellipsis, all letters elided by the ellipsis are included as
557 index keys in alphabetic order. For example, 'A-Z' includes all ASCII
558 letters A to Z.
560 '[]' (square brackets): all letters included in square brackets are
561 skipping letters. It is used for CTL languages, for example in Thai
562 (th_TH), to skip prefix vowels or signs. For example, if ["] double quote
563 is defined as skipping letter, index item '"Index"' will be under 'I', not
564 '"'.
566 '{}' (curly brackets): define multiple letters index key, for example
567 '{Cs}' is one of the index keys for Hungarian (hu_HU).
569 '()' (parentheses): define optional description for index key. If defined,
570 the description will be shown as index key title, instead of the index key
571 itself. For example, 'E(E, É)' in Hungarian adds entries with 'E' to the
572 description 'E, É'.
574 -->
581 <!--
582 The Unicode script types are those of
583 offapi/com/sun/star/i18n/UnicodeScript.idl, they define the code range for
584 the language.
585 -->
588 <!--
589 The FollowPageWord entries were originally hard-coded in
590 ../../indexentry/indexentrysupplier.cxx, most locales used the English
591 ``p.'' and ``pp.'', valid data should be provided by native speakers.
592 These words or abbreviations are used in the Writer's index table. The
593 first FollowPageWord element is the abbreviation for "page" (p.), the
594 second FollowPageWord element the abbreviation for "page and following
595 pages" (pp.).
596 -->
607 <!ELEMENT ForbiddenCharacters (ForbiddenLineBeginCharacters, ForbiddenLineEndCharacters, LineBreakHangingCharacters)>
611 <!-- optional break iterator rules for the languages.
612 if defined, 5 rules should be all listed, each of them can be blank and default will be used.
613 order of the rules is significant!
614 -->
615 <!ELEMENT BreakIteratorRules (EditMode, DictionaryMode, WordCountMode, CharacterMode, LineMode)>
622 <!ELEMENT ReservedWords (trueWord, falseWord, quarter1Word, quarter2Word, quarter3Word, quarter4Word, aboveWord, belowWord, quarter1Abbreviation, quarter2Abbreviation, quarter3Abbreviation, quarter4Abbreviation)>
623 <!-- order is significant! -->
632 <!-- aboveWord and belowWord are used as reference field content in Writer, it
633 is a "physical" object relative position, not "numerical" relative.
634 -->
641 <!--
642 In numbering levels, the NumType attribute is a value of the constants
643 defined in offapi/com/sun/star/style/NumberingType.idl (
644 https://opengrok.libreoffice.org/xref/core/offapi/com/sun/star/style/NumberingType.idl
645 )
646 -->