| blob | d0448015943f79a5429824f67aa0fb1bccf0a018 |
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 using a validating parser. A
30 validating parser, for example, may be found at
31 http://unicode.org/cldr/data/tools/java/org/unicode/cldr/util/XMLValidator.java
32 Compile it into a class-jar and call it in the
33 i18npool/source/localedata/data/ directory:
34 java -cp <your_path>/XMLValidator.jar org.unicode.cldr.util.XMLValidator your_data.xml
37 A second possiblity is:
39 - temporarily (!) change the DOCTYPE of your file to read (all on one line)
40 <!DOCTYPE Locale SYSTEM "http://cgit.freedesktop.org/libreoffice/core/plain/i18npool/source/localedata/data/locale.dtd">
42 - upload it to the form available at http://www.validome.org/xml/
44 This will validate the file against the HEAD revision of locale.dtd
47 Please test locale data files either in an enable-dbgutil build, which
48 implements some checks and pops up assertion message boxes if the tests
49 fail, or by setting the environment variable OOO_ENABLE_LOCALE_DATA_CHECKS
50 to 'Y' or 'Yes' (or any other string starting with 'Y') or '1' before
51 starting the application, which outputs the same messages to stderr and
52 also works in a product build.
54 Then follow this procedure:
55 1. Create a new spreadsheet document.
56 2. On a cell use context menu -> Format Cells -> Numbers.
57 3. Select the locale in the Language list box => MUST be assertion free.
58 3.a. Assertions are only shown at the very first time a number formatter
59 is created, respectively the first time a specific locale data is
60 used. To repeat steps 1.-3. you'd need to create another spreadsheet
61 document.
63 ===========================================================================
64 NOTE the FormatElement comments further down.
65 ===========================================================================
67 -->
69 <!-- ............................................................... -->
70 <!-- Entities for characters and symbols ........................... -->
72 <!ENTITY % UNOModule
73 'unoid CDATA #IMPLIED'>
75 <!ENTITY % MessageID
76 'msgid CDATA #REQUIRED'>
78 <!ENTITY % RefLocale
79 'ref CDATA #IMPLIED'>
80 <!-- Where given, an element can be inherited from another locale, e.g.
81 ref="en_US" -->
83 <!ENTITY % LIBModule
84 'module CDATA #IMPLIED'>
85 <!-- The locale referred to for the sub categories, implementation detail. -->
89 <!-- See below for the LC_FORMAT element. -->
93 <!-- Locale is made of different sub categories -->
94 <!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)>
96 <!-- Version identifier to prevent mismatching data files being submitted
97 because of older files being copied as templates. The version should be
98 less than or equal to the OOo release, or the OOo release number plus some
99 extension, like "2.0.enhanced", to be able to easily determine the
100 corresponding OOo version. New versions of the DTD with new required
101 elements SHOULD REALLY result in a new versionDTD here and
102 LOCALE_VERSION_DTD in ../LocaleNode.cxx, and *.xml files MUST be adapted
103 then, otherwise building the data or checking it with a validating parser
104 will throw an error.
105 -->
107 <!-- Whether some data elements may be (automatically) updated from the Common
108 Locale Data Repository, see http://cldr.unicode.org/
109 Note: This mechanism currently (2010-02-21) is outdated.
110 -->
112 <!-- Valid number, may designate versioned data -->
119 <!-- LangID must be a valid two or three letter language identifier defined by
120 ISO 639. Use ISO 639-1 two letter code where available, else ISO 639-2 or
121 639-3 three letter code.
122 -->
126 <!-- CountryID must be a valid two letter country identifier defined by ISO 3166. -->
130 <!-- Unused, deprecated, can be generic|unix|win32|macos, best Platform element
131 be absent.
132 -->
135 <!-- Reserved for future use, currently unused. -->
138 <!-- The LC_FORMAT element contains number format codes and may actually appear
139 twice (with the second ocurrence named LC_FORMAT_1). One section is
140 mandatory and MUST contain at least all required format codes
141 formatindex="0" to formatindex="47", MUST NOT contain formatindex 48 and
142 49, and MAY contain other format codes. See below ATTLIST FormatElement
143 formatindex.
145 A second LC_FORMAT_1 section may follow containing other format codes. The
146 difference between the two sections is that they are inherited
147 independently if a locale uses the RefLocale mechanism (ref="..."). This
148 may be used to not offer locale dependent format codes to other locales
149 that otherwise inherit the format codes.
151 -->
153 <!-- All FormatElement elements must be given if the RefLocale mechanism is not used! -->
156 <!-- Define placeholder for currency code, usually "[CURRENCY]" -->
158 <!-- Currency code to be used to replace the placeholder, e.g. "[$R-1C09]".
159 Note: The brackets and the leading $ character are mandatory, the
160 hyphen-minus separates the currency symbol from the hexagesimal MS-LCID,
161 letters contained in the LCID have to be in upper case, leading zeros are
162 to be omitted. LCIDs are defined in i18npool/inc/i18npool/lang.h
163 -->
166 <!-- Pattern that defines an input sequence match to be accepted as
167 (abbreviated) date. For example, in en_US locale "M/D" accepts an input of
168 11/23 as CurrentYear-November-23 whereas 11/23/ would not be a date. In
169 de_DE locale "D.M." accepts an input of 23.11. as CurrentYear-November-23
170 whereas 23.11 would not be a date.
172 For each locale one pattern that matches a full date is automatically
173 generated from FormatElement formatIndex="21" and does not need to be
174 defined, for example "M/D/Y" or "D.M.Y", so only if you want abbreviations
175 to be in effect add some patterns.
177 NOTE: use only single letter D,M,Y. Multiple patterns can be defined.
178 -->
189 <!--
190 There may be up to three groups (type="short", type="medium", type="long")
191 for each usage category defined. Each group, if defined, needs excactly one
192 default. The type roughly determines the display string length, for example
193 short, medium, and long date formats.
195 The number formatter determines an ultimate default format of a specific
196 usage category by looking at the medium, long, and short default formats,
197 in that very particular order.
198 -->
199 <!ATTLIST FormatElement usage (FIXED_NUMBER|FRACTION_NUMBER|PERCENT_NUMBER|SCIENTIFIC_NUMBER|CURRENCY|DATE|TIME|DATE_TIME) #REQUIRED >
201 <!--
202 The following FormatElements must follow specific rules:
204 All:
205 The format indices 0..49 are reserved and, for backwards compatibility,
206 MUST be used as stated in
207 offapi/com/sun/star/i18n/NumberFormatIndex.idl.
208 Note that indices 10 ("# ?/?"), 11 ("# ??/??"), 48 (BOOLEAN) and 49 (@
209 Text) are generated internally, as they aren't locale dependent, and
210 must not be used in locale data XML files. All other formats have to be
211 present.
213 Note also that "must be used as stated" does not mean that the format
214 codes must be identical, of course the meaning of a format code should
215 match, for example en_US MM/DD/YY matches de_DE DD.MM.YY
216 Just imagine the same index being used with another locale shouldn't
217 change the meaning of representation significantly.
219 You'll notice differences of non-matching format codes only if
220 documents use the Default language in number formats and either are
221 stored in old SO5 binary file format and loaded on another system where
222 languages/locales aren't the same, or if the default locale is switched
223 under menu Tools.Options.LanguageSettings.Languages.LocaleSetting
224 dialog, which exchanges formats on the fly in the spreadsheet
225 application. Please check it out! So far only very few locale data file
226 we received got that right, especially not in date formats!
228 For easier comparison between locales in future please sort the
229 FormatElements by their formatindex="..." value within a usage group.
230 This isn't necessary to be technically correct and isn't done in many
231 locales yet, but will certainly help.
233 If you want to define yet more formats than the preset and reserved
234 0..49 range that's fine, only make sure those formatindex="..." values
235 are each >=50 and all values are unique within one locale.
237 Of usage="FIXED_NUMBER":
238 formatindex="0" MUST be the format containing the 'General' keyword.
239 The keyword itself may be localized, it is good practice though to
240 stick with a wording known from another spreadsheet application for
241 better user experience. Like other format codes it may be prepended
242 with a [NatNum1] modifier if values are to be displayed using native
243 numbering if no specific format was applied. The format must have the
244 default="true" and type="medium" attributes.
246 Of usage="DATE":
247 formatindex="21" is used to edit already existing date data. In order
248 to always edit the full century the long year YYYY code must be used.
249 Furthermore, the format has to be of an editable type, of course, which
250 means parseable. Therefore it should only contain DD, MM, YYYY and date
251 separators, and the YMD default order is determined from the order
252 encountered in this format.
253 TODO: Future versions should make use of an edit="true" attribute
254 instead of relying on this special requirement.
256 formatindex="33" must be ISO 8601 YYYY-MM-DD format code.
258 Of usage="DATE_TIME":
259 formatindex="47" is used to edit already existing combined date/time
260 data. The requirements are the same as for formatindex="21" above.
262 Of usage="TIME":
263 * The FormatElement with default="true" type="medium" is used to edit
264 times and thus must contain all HH and MM and SS codes, e.g. HH:MM:SS
265 * Formatindices 43, 44, 45 are special in the sense that they are
266 programmatically used to automatically display values that meet
267 certain criteria:
268 * Formatindex="43" contains the [HH] format code that displays hour
269 values greater than or equal to 24 (as opposed to a simple HH that
270 displays modulo 24).
271 * Formatindex="44" uses no hour code but 100th seconds and the
272 Time100SecSeparator.
273 * Formatindex="45" uses both [HH] hour code and 100th seconds to be
274 able to edit such time values without loosing information.
276 Of usage="CURRENCY":
277 formatindices 12, 13, 14, 15, 17 with [$xxx-yyy] notation must use the
278 xxx currency symbol that has the attribute
279 usedInCompatibleFormatCodes="true". The hexadecimal yyy LANGID must be
280 properly set. It can be found in project util module tools file
281 tools/inc/lang.hxx. You may verify the proper use of the xxx currency
282 symbol with the AWK script
283 i18npool/source/localedata/data/currency-check.awk, it mustn't display
284 any output. If it does, then there's something wrong.
285 -->
289 <!-- The LC_CALENDAR element defines calendars used with a locale. -->
291 <!-- At least one Calendar element must be given if the RefLocale mechanism is not used! -->
294 <!ELEMENT Calendar (DaysOfWeek, MonthsOfYear, GenitiveMonths*, PartitiveMonths*, Eras, StartDayOfWeek, MinimalDaysInFirstWeek) >
296 <!-- The unoid of a gregorian calendar MUST be lower case "gregorian",
297 calendars MUST match the names defined in the OASIS OpenDocument Format
298 (ODF) 1.2 or later specification. The implementation name registered with
299 the OOo service resgistry MUST match, e.g.
300 com.sun.star.i18n.Calendar_gregorian
301 -->
303 <!-- Exactly one Calendar element has to be the default calendar. -->
306 <!-- All Day elements of a Calendar must be given if the RefLocale mechanism is not used! -->
308 <!-- Sequence of days is important, MUST start with Sunday. -->
311 <!-- Preferably the lower case abbreviated English name like sun for Sunday. -->
313 <!-- The abbreviated day name, e.g. Sun for Sunday. -->
315 <!-- The full day name, e.g. Sunday for Sunday. -->
317 <!-- The narrow day name, e.g. S for Sunday.
318 If not specified, the first letter of the corresponding DefaultFullName is taken.
319 -->
322 <!-- All Month elements of a Calendar must be given if the RefLocale mechanism is not used! -->
324 <!-- Sequence of months is important, MUST start with the first month of a
325 year, e.g. January in a Gregorian calendar.
326 -->
329 <!-- Possessive genitive case month names, for example in Slavic locales. The
330 element is optional, but if present all Month elements of a Calendar must
331 be given if the RefLocale mechanism is not used! If not present,
332 MonthsOfYear names will be used. -->
334 <!-- Sequence of months is important, MUST start with the first month of a
335 year, e.g. January in a Gregorian calendar.
336 -->
339 <!-- Partitive case month names, for example in Finnish locales. The
340 element is optional, but if present all Month elements of a Calendar must
341 be given if the RefLocale mechanism is not used! If not present,
342 GenitiveMonths names will be used. -->
344 <!-- Sequence of months is important, MUST start with the first month of a
345 year, e.g. January in a Gregorian calendar.
346 -->
350 <!-- Preferably the lower case abbreviated English name like jan for January. -->
353 <!-- All Era elements of a Calendar must be given if the RefLocale mechanism is not used! -->
355 <!-- The eras MUST be in chronological order, e.g. first BC then AC. -->
358 <!-- If a calendar has special eras (like zh_TW ROC or ja_JP Gengou calendar)
359 and a date before those eras is undefined, a leading (first) dummy era
360 with EraID="Dummy" has to be defined to enable the number formatter to
361 fall back to a Gregorian calendar for those date values if the XCalendar
362 implementation returns an era value of 0.
363 -->
366 <!-- MUST exactly match (case significant!) one of the DayID of DaysOfWeek -->
369 <!-- The number of days of a week that must reside in the beginning of a year
370 to make a week the first week of the year. For example, a value of 4 means
371 that at least 4 days of a week must be in the new year. So if the week
372 starts on Monday, the first week of a year will be the week where Thursday
373 is in the new year.
374 -->
377 <!-- The LC_CURRENCY element defines currencies used with a locale. -->
379 <!-- At least one Currency element must be given if the RefLocale mechanism is not used! -->
383 <!-- Exactly one Currency element has to be the default currency. -->
385 <!-- If this currency is the one used in compatible number format codes with
386 <member>FormatElement::formatIndex</member> values in the range 12..17.
387 Those format codes are used to generate some old style currency format
388 codes for compatibility with StarOffice5 and StarOffice4.
389 Every locale data file MUST contain exactly one currency having this set to "true",
390 and that currency MUST be used in format codes 12..17.
391 For European countries using the EURo it MUST be the old currency, for example, DM.
392 -->
394 <!-- If this Currency element exists only to be able to correctly load legacy
395 documents and is not selectable in the UI otherwise. Defaults to "false"
396 if not specified. If this attribute is "true", 'default' and
397 'usedInCompatibleFormatCodes' must both be "false".
399 Currency elements must not be changed to contain only a different
400 CurrencySymbol element without changing the CurrencyID and BankSymbol
401 elements, instead the entire Currency element must be duplicated, the old
402 element needs this 'legacyOnly' attribute be added and 'default' and
403 'usedInCompatibleFormatCodes' attributes must be set to "false", and only
404 in the new duplicated Currency element the CurrencySymbol element be
405 changed.
406 -->
408 <!-- The ISO 4217 three letter currency code, e.g. USD or EUR. -->
410 <!-- The currency symbol, e.g. $ or €. -->
412 <!-- The ISO 4217 three letter currency code, e.g. USD or EUR. -->
414 <!-- The native currency name, e.g. Dollar or Euro. -->
416 <!-- Number of decimal places used with the currency, usually 2 or 0, e.g. 2
417 for cents.
418 -->
422 <!-- All elements must be given if the RefLocale mechanism is not used! -->
426 <!ELEMENT Separators (DateSeparator, ThousandSeparator, DecimalSeparator, TimeSeparator, Time100SecSeparator, ListSeparator, LongDateDayOfWeekSeparator, LongDateDaySeparator, LongDateMonthSeparator, LongDateYearSeparator)>
450 <!-- All elements must be given if the RefLocale mechanism is not used! -->
452 <!--
453 Optional ICU tailoring.
455 See Collation Customization in ICU User Guide for syntax,
456 http://www.icu-project.org/userguide/Collate_Customization.html
458 There are two ways to add language specific tailoring in OpenOffice.org.
459 For small tailoring, you can directly add it in locale data here. For
460 large tailoring, it is suggested to put the data under collator/data, so
461 it will be compiled to a binary format in build time and improve performance
462 in run time.
464 -->
473 <!-- All elements must be given if the RefLocale mechanism is not used! -->
481 <!--
482 The IndexKey element is optional, but should be given if the locale
483 requires a specific sort order in Writer's index tables or entries are to
484 be combined under keys.
486 Index key for the algorithm and language, like >A-Z< for English => A, B,
487 C, ..., Y, Z. The letters specify under which key an entry goes and the
488 order the keys are sorted. Keys may be reordered or letters inserted to
489 form a specific order, for example (ve_ZA) >A-D Ḓ E-L Ḽ M N Ṋ Ṅ O-T Ṱ U-Z<
490 Entries that don't match a defined key are appended to the index list. Used
491 in Writer textprocessor.
493 The initial data was setup according to the ICU collation chart at
494 http://oss.software.ibm.com/icu/charts/collation/
495 Note: ICU site was moved to http://www.icu-project.org/ and as ICU per
496 default uses CLDR, collation charts are available at
497 http://unicode.org/cldr/comparison_charts.html#Collation
499 Possible notations in the syntax of the IndexKey element are:
501 '-' (dash): Ellipsis, all letters elided by the ellipsis are included as
502 index keys in alphabetic order. For example, 'A-Z' includes all ASCII
503 letters A to Z.
505 '[]' (square brackets): all letters included in square brackets are
506 skipping letters. It is used for CTL languages, for example in Thai
507 (th_TH), to skip prefix vowels or signs. For example, if ["] double quote
508 is defined as skipping letter, index item '"Index"' will be under 'I', not
509 '"'.
511 '{}' (curly brackets): define multiple letters index key, for example
512 '{Cs}' is one of the index keys for Hungarian (hu_HU).
514 '()' (parentheses): define optional description for index key. If defined,
515 the description will be shown as index key title, instead of the index key
516 itself. For example, 'E(E, É)' in Hungarian adds entries with 'E' to the
517 description 'E, É'.
519 -->
526 <!--
527 The Unicode script types are those of
528 offapi/com/sun/star/i18n/UnicodeScript.idl, they define the code range for
529 the language.
530 -->
533 <!--
534 The FollowPageWord entries were originally hard-coded in
535 ../../indexentry/indexentrysupplier.cxx, most locales used the English
536 ``p.'' and ``pp.'', valid data should be provided by native speakers.
537 These words or abbreviations are used in the Writer's index table. The
538 first FollowPageWord element is the abbreviation for "page" (p.), the
539 second FollowPageWord element the abbreviation for "page and following
540 pages" (pp.).
541 -->
552 <!ELEMENT ForbiddenCharacters (ForbiddenLineBeginCharacters, ForbiddenLineEndCharacters, LineBreakHangingCharacters)>
556 <!-- optional break iterator rules for the languages.
557 if defined, 5 rules should be all listed, each of them can be blank and default will be used.
558 order of the rules is significant!
559 -->
560 <!ELEMENT BreakIteratorRules (EditMode, DictionaryMode, WordCountMode, CharacterMode, LineMode)>
567 <!ELEMENT ReservedWords (trueWord, falseWord, quarter1Word, quarter2Word, quarter3Word, quarter4Word, aboveWord, belowWord, quarter1Abbreviation, quarter2Abbreviation, quarter3Abbreviation, quarter4Abbreviation)>
568 <!-- order is significant! -->
577 <!-- aboveWord and belowWord are used as reference field content in Writer, it
578 is a "physical" object relative position, not "numerical" relative.
579 -->
586 <!--
587 In numbering levels, the NumType attribute is a value of the constants
588 defined in offapi/com/sun/star/style/NumberingType.idl
589 -->