GNU libc (glibc) locale patch page

This page contain information about the GNU libc locale collection, including patches, suggested formates, list of known problems and who to contact regarding the different locales. It is a work in progress. The page maintainer, Petter Reinholdtsen, is working with the GNU libc maintainers to get all the fixes included in the official releases of the source.

References to documentation and standards

It is currently an open question where the format og locales in glibc is documented. This list of references is at least pointing to related docuemnts documenting the format of locales, even if the glibc maintainers might not accept them as authoritative.

• ISO/IEC 9945-2:1993 (not freely available)
• ISO/IEC 9945:2002
• The Open Group Base Specifications Issue 6, IEEE Std 1003.1, 2003 Edition
• ISO/IEC FCD 14652
• ISO/IEC FPDTR 14652:1999(E)
• ISO/IEC TR 14652:2003

To make it easier to compare locales with each other, I recommend using the same order in all locales. Any order will do, so I picked a order used in several locales, and decided to recommend this order:

1. LC_IDENTIFICATION
2. LC_CTYPE
3. LC_COLLATE
4. LC_MONETARY
5. LC_NUMERIC
6. LC_TIME
7. LC_MESSAGES
8. LC_PAPER
9. LC_NAME
10. LC_ADDRESS
11. LC_TELEPHONE
12. LC_MEASUREMENT

Locale related problem reports

From GNU libc gnats database

The GNATS database is dead at the moment (2003-10-31), and has been dead since this summer.

• libc/2254: Insufficient info from the locale tool.
  (patch sent to libc-alpha, partly fixed in CVS 2003-05-21)
• libc/2931: Missing nb_NO locale
  (patch sent to libc-alpha 2003-06-15)
• libc/2954: Should add postal telephone and vehicle codes to locales
  (HTML error in gnats?)
• libc/3606: Update documentation regarding LC_ALL and LANG
• libc/4197: Incorrect hr_HR - currency sumbol position
  (sent email to submitter [bad address] and author)
• libc/4348: Converting charset "eucjp->sjis->iso2022jp" and "eucjp->iso2022jp" give different result.
• libc/4658: 'locale -a' do not list locales in archive
• libc/4938: Segfaults when using locale info before it is initialized
• libc/5013: Incorrect fi_FI - gramatic
  (sent email to submitter and author, patch sent 2003-05-09, patch commited 2003-06-16)
• libc/5018, libc/5019, libc/5028: Incorret uk_UA.
  (sent email to submitter and author)
• libc/5023: Missing locales li_NL, nds_NL and nds_DE
  (sent email to submitter)
• libc/5059: Wrong date format for locale sl_SI
  (sent email to submitter and locale author, approved by author, patch sent to libc-alpha 2003-07-06)

From Debian GNU/Linux BTS

• #111268: Incorrect sv_SE - 13:59 instead of 13.59 for clock time
  (patch sent 2003-05-18)
• #119528: Incorrect ro_RO - rewritten locale
  (sent email to submitter)
• #178513: Incorrect is_IS - Use lower case letters in month and week day names
  (sent email to submitter and author, confirmed by Keld, patch sent to libc-alpha 2003-06-15, patch commited 2003-06-16)
• #194289: New locale ag_ES for Aragonese
  (sent email to submitter, patch commited 2003-06-16 as an_ES)
• #193904: locales.alias even lines are not considered (debian only?)
• #206474: locales: nb_NO should be a real locale, not an alias
  Same as libc/2931
• #208238: locales: wrong charset for et_EE
• #211607: Bad catalan locale currency format
• #214107: locales: en_US.UTF-8 treats [ as a space char
• #214414: Incorrect date format in en_NZ
• #215466: fi_FI locale doesn't contain first_weekday

From RedHat Bugzilla

• #75139: Incorrect week start day for sv_SE (should start on monday)
  (Corrected in CVS 2003-01-14)
• #83276: Norwegian locale in distro
  Same as libc/2931

Patches submitted to the glibc maintainers

If you know of some patches that should be listed here, please send me an email. I want to track all relevant patches, not just the ones submitted by me.

Pending/rejected

• 2003-04-15 Add sorting test for no_NO
• 2003-05-18 Correct sorting for nn_NO to match no_NO
• 2003-05-18 Norwegian week start on monday, fix no_NO and nn_NO
• 2003-05-18 Change sv_SE to use ':' in clock format (BTS #111268)
• 2003-05-18 Include examples in 'locale --help' (PR libc/2254, rejected)
• 2003-05-29 Generate locales on 'make check' (rejected)
• 2003-05-29 Indent the error output from sort-test.sh
• 2003-06-15 Add nb_NO locale (PR libc/2931)
• 2003-07-06 Correct wrong date format for locale sl_SI (PR libc/5059)
• 2003-08-13 Updated/new locales for sr_CS (Serbia and Montenegro)
• 2003-06-21 Add test for international currency formatting. and 2003-11-23 Add test for international currency formatting (updated)
• 2004-01-01 Change et_EE charset to ISO-8859-15 (rejected)

Accepted/included

• 2003-05-31 Correct email of Denis V. Dmitrienko in uk_UA (commited)
• 2003-05-31 Correct thousand separator for fi_FI (commited)
• 2003-06-15 Correct fi_FI date gramar (PR libc/5013) (commited)
• 2003-06-15 Clean up locale test code (commited)
• 2003-06-15 Use lower case letters in month and week day names in is_IS (commited)
• 2003-06-21 Add missing ETH to LC_COLLATE in no_NO (commited)
• 2003-06-25 Correct an_ES charset and add an_ES.ISO-8859-15 to SUPPORTED (commited)
• 2003-07-05 Updating/adding locale for Ethiopia and Eritrea
• 2003-07-13 Make sort-test handle more then 100 lines, and avoid segfault (commited)
• 2003-09-01 Document why there are two locales for Hebrew
• 2003-09-11 Locales for Uzbekistan (uz_UZ and uz_UZ@Cyrl)
• 2003-11-23 Only check the currency name against 4217, exclude the space
• 2003-11-23 Monday is first work and week day in Finland
• 2003-11-23 Remove unused file localedata/locales/translit_et
• 2004-01-01 Make strfmon() more POSIX-compliant
• 2004-01-01 Report which of the sorting methods that fail
• 2003-11-23 Correct formatting of internatinal currency
• 2003-11-23 Allow quotes in tst-fmon.data, and make output easier to read
• 2003-07-06 Some small locale fixes

Missing locales?

Bruno Haible told me about a list of language codes available from Microsoft, and discovered that some of these are missing as locales in current GNU libc.

Locale code	Language	Region/Country
bnt_TZ	Sutu	Tanzania	(MS calls this one "sx" - not ISO 639) (see also http://www.ethnologue.com/show_language.asp?code=NGU) Correct code st_LS / nso_ZA ?
de_LI	German	Liechtenstein
en_JM	English	Jamaica
en_GD	English	Caribbean
en_BZ	English	Belize
en_TT	English	Trinidad
gd_IE	Gaelic	Ireland
rm_CH	Rhaeto-Romanic	Switzerland
ro_MD	Romanian	Moldavia	(MS calls this one "MO" - not ISO 3166)
ru_MD	Russian	Moldavia	(MS calls this one "MO" - not ISO 3166)
ts_ZA	Tsonga	South Africa	(see also http://www.ethnologue.com/show_language.asp?code=TSO)
tn_BW	Tswana	Botswana	(see also http://www.ethnologue.com/show_language.asp?code=TSW)
ven_ZA	Venda	South Africa	(not sure about ISO 639 code: "ve" or "ven") (see also http://www.ethnologue.com/show_language.asp?code=VEN)
wen_DE	Sorbian	Germany	(MS calls this one "sb" - not ISO 639) (see also http://www.ethnologue.com/show_language.asp?code=WEN)

Mailing list for discussing locales

A mailing list for discussing the GNU libc locales have been created. It is archived on the web. Send an email to libc-locales-subscribe@sources.redhat.com to subscribe.