A locale is a collection of files, data, and sometimes code, which contain the necessary information to adapt Solaris to a specific geographical market. It is essentially a "bundle", containing information such as the messages displayed to the user, interface component information, for example the size of the buttons; cultural information such as dates and currency formats, and fonts or other writing specific information.

2. What locales are provided with Solaris?

Solaris supports a global set of languages that can be added or removed at any time. Globalization is enhanced through extensive support over 37 languages, 123 locales, and complex text formats for Arabic, Thai and Hebrew, along with support for the development of multilingual applications.

See details for:

Solaris 10 Update Release 2

Solaris 10 UR2 now ships with all the locales listed in the Solaris 10, plus the locales listed at http://developers.sun.com/techtopics/global/products_platforms/solaris/reference/techart/s10-locale-patches.html.

Solaris 10

Solaris 9

Solaris 8

3. How do I add a locale after installing the Base Product?

Solaris 10 added the new localeadm command line utility. It allows you to add support for specific geographic regions. Please take a look at the man page.
For Solaris 7, 8 & 9 it requires that you pkgadd the right packages. For more details, including package lists, please see the Solaris 9, Solaris 8 and Solaris 7 pages.

4. Where can I get specific locales?

There are no downloads for Solaris locales. All the locales are included on one of your CDs or DVD.

5. What is the difference between a partial and a full locale?

A partial locale provides support for date/time format, monetary and numeric format, and collation sequence. A partial locale also defines printable characters, uppercase and lowercase characters, and graphic characters.

A full locale provides the same information plus translated system messages.

6. How do I run a localized environment once it has been installed?

If you're running Gnome, select a locale at login.
If you're running CDE, select a locale at login.

7. Can I use English to install the Japanese Solaris distribution?

Yes. The first window you see when installing Solaris enables you to choose the language you want to use for the installation. See Q3.

8. Can I install Chinese, Japanese, Korean, and European locales on the same Solaris system?

Yes. From Solaris 8, all Chinese, Japanese, Korean, and European locales are included on the base CD with localised messages on a separate "Language" CD. These locales and messages are selected at install time, or added post-install.

9. I installed Chinese packages from languages cd. I then tried "`setenv LANG zh_TW`" but got an error - "`Locale could not be set properly`". What am I doing wrong?

It is most likely the case that one or more required packages were not installed properly. On Solaris 10 please refer to the localeadm utility. Otherwise refer to the locale package pages for Solaris 9, Solaris 8 and Solaris 7.



By-the-way, using Chinese as the installation language is not enough to install all Chinese locales. Users still need to select all of the locales they need from the GEO screen during the installation.
Only zh or zh_TW locale is installed by default when Simplified Chinese or 
Traditional Chinese is used as the installation langauge.

Back to Top
10. What Turkish language support is available for Solaris?
Solaris supports the Turkish language in that you can input, display, and 
print Turkish.
It also supports cultural elements like date/time formats, currency, sorting 
etc. Solaris does not provide Turkish translations for 
system messages and application text (e.g. GUI menus/buttons etc. Turkish 
support has been in Solaris since Solaris 2.6.
Back to Top
11. How do I customize and create my own Solaris locales?
See this techtip.
Back to Top
12. What locales are available on the different versions of Solaris?
This sortable chart lists what locales are available in Solaris 7, 8 and 9.
Back to Top
13. My timezone - TZ - variable is set to ":CST6CDT". Why do Solaris 9 and
Solaris 10 insert a ':' into the TZ variable? Is a colon [:] in the TZ variable POSIX compliant?
The three different ways to set timezone at install are:

 geograpy/region
 GMT offset
 filename (from /usr/share/lib/zoneinfo/*)

If user selects to specify a filename and that filename also qualifies
as a POSIX timezone, then the ":" is prepended to the value stored in
/etc/default/init so that we know that the TZ has been set from a file
as opposed to the POSIX timezone implementation built into libc.



The POSIX standard (since 1990 - POSIX 1003.1:1990, p.152) does support
':' as the first character of the timezone, though it specifies such a
timezone is implementation defined.  So although the behavior is not
specified, ":CST6CDT" is a POSIX-compliant value for TZ.



If your TZ value is from filename and if you wish to have
the TZ value without the ':', you could either edit
the /etc/default/init and remove the character if
the timezone file already exists at Solaris. Alternatively,
specifying "timezone=" value at the sysidcfg(4) file that
can be used during system installation is another way. 


Also refer to the TZ section of environ(5) man page.
Back to Top
14. What is the difference between the en_US.UTF-8 and ja_JP.UTF-8 locales in the Solaris OS?
The two locales are one of the many UTF-8 locales that share
many underlying frameworks and definitions, i.e., locale methods,
output methods, input methods and so on and even LC_CTYPE locale
definition are the same.

Also, most notably, both share the same file and process/wchar_t code
representations, and so text data are completely interchangeable between
the two locales without any need of code conversions.



All other locale definitions are however different, for instance, they
both have different LC_COLLATE, LC_MONETARY, LC_NUMERIC etc. definitions.
The ja_JP.UTF-8 also has different and additional font charsets
supported since Japanese users need more JIS X font charsets such as JIS
X 0212.

The ja_JP.UTF-8 locale provides fully translated Japanese messages in
GUI, CUI, man pages and so on (well, on most of them).
The ja_JP.UTF-8 locale also provides its own libraries and utilities
that Japanese customers are expecting to have in their locale environment.
Back to Top
15. The gethostname and sethostname calls take a char* as first
argument. Does either of these functions support non-ASCII characters?
Hostnames and Internet style domain names must stay in
the subset of 7-bit ASCII as specified in RFC952, 1035, 1122, and 1123.
http://www.ietf.org/rfc/rfc952.txt.


If you though wish to support non-ASCII characters at the host or
domain names with your program, you can use IDN conversion functions
such as idn_encodename(3EXT), idn_decodename(3EXT)
to convert between non-ASCII names and ACE (ASCII Compatible Encoding)
names and use the ACE names at the system files, networking interfaces,
and so on where the 7-bit ASCII names are required. 

In other words, your application is required to do the conversions between non-ASCII
names and ACE names before using the ACE names at the various places
where host and domain names are used in the system. (Of course, users
would expect to see non-ASCII names when they are dealing with your
program and so this conversion should be supported transparently by
your program.)



For more details, please see the idn_encodename(3EXT) man page at
S9 Update 6 or later and S10 or later systems and also
"Internationalized Domain Name (IDN) Support" section of the S10
International Language Environments Guide at the following:


http://docs.sun.com/app/docs/doc/817-2521/6mi67tj3q?a=view


We also added idnconv(1) utility in addition to iconv code conversion at
S10 by the way. Please also see RFC3490 for more info on IDN.
Back to Top
16. The /etc/rc2.d/S95IIim script starts the htt daemon on port 9010. If my
default locale is C or en_US.ISO8859-1, will I be missing any functionality if I don't start this daemon?
No, you won't miss anything with C or en_US.ISO8859-1.
Back to Top
17. I need to be able to view a UTF-8 encoded .csv file with Chinese
characters from an Oracle database. Right now all the Chinese characters
are garbled.
 You could login to any of the UTF-8 locales - for example zh_CN.UTF-8, and
then open the file with vi or any other editor.

Alternatively, you could just open the .csv file in a browser, and
change the encoding to UTF-8 [View/Character Encoding/UTF-8 in Mozilla].
This should work for modern browsers - with the right fonts - regardless
of the locale you are actually running in. See here.
Back to Top
18. How do I install the en_CA.ISO8859-1 locale on Solaris 10?
See the above question, or the localeadm man page.
Back to Top
19. In Solaris 10, can I add locales at install time with both the CUI and 
GUI install interfaces - or do I have to use localeadm post install?
Locales can be added at install time via the CUI or GUI.
One also could do a JumpStart based installation and specify all 
necessary locales to be installed in the JumpStart profile file.
Back to Top
20. Is there a Dutch UTF-8 locale in Solaris 9?
Dutch UTF-8 support [as well as UTF-8 support for many other European languages] for Solaris 9 is available as a sparc patch. The patch ID is 117100-04. It can be downloaded from SunSolve.
Back to Top
21. Why does sort fail in a UTF-8 locale, but works fine in the C locale?
The file that you are sorting probably has Latin-1 characters. These
bytes are illegal (and considered as harmful due to security reason)
characters in UTF-8 locales and thus during the collation, such illegal
characters were detected. This somehow mapped to EMSG_ERROR error
condition at the sort(1).


The solution is to convert your file to UTF-8 using iconv(1) prior to 
the sorting. Try this:



# iconv -f 8859-1 -t UTF-8 my-file.txt



Then run sort as normal.
Back to Top
22. I can't properly see Hebrew characters, I see only /347/322/251. Is there any problem in installation? 
The first thing to check is if you have the Hebrew
localization packages properly installed and what locale is being
used on your desktop. In C locale where only ASCII characters
are allowed, all Hebrew characters, esp., in OS commands, will show up
as escaped octal character sequences as you explained and showed.



There are several things to be done to properly see Hebrew
characters:
Please install all Hebrew localization packages. If you would like to
use Hebrew script in any of the Unicode/UTF-8 locales, please also make
sure to install the corresponding UTF-8 locale packages. View the package lists 
here.
Once these locales have been properly installed, you can login to CDE in either an 'he' 
(ISO8859-8) or, say, en_US.UTF-8 locale - depending on your preferred codeset.
 Pretty much all applications will be able to handle (input and output)
Hebrew characters after the CDE Login.

For more info on UTF-8 locale and Hebrew input, please see the Overview of
en_US.UTF-8 Locale Support chapter at the S8 International Language
Environments Guide at the following URL 
http://docs.sun.com/app/docs/doc/806-0169

Back to Top
23. We are trying to view a file in Japanese language on the Solaris 8 platform. We're not 
sure of its encoding, but have converted it to UTF-8 with:

iconv -f  eucJP -t UTF-8  ja-file.txt
iconv -f  PCK -t UTF-8  ja-file.txt


However it produdes errors like iconv: conversion error and Conversion error detected.

The iconv is failing because, I believe, your source file is not encoded as either eucJP or PCK. Or, 
another possibility is that there might be an error code in the source file even though it is PCK or eucJP file.


That said, if you only want to *view* the Japanese file then you don't
necessarily have to convert it to UTF-8. Open it in a browser [browser must have
correct font support], and follow the steps in the second part of this 
article.

Using the new Solaris 10 auto_ef utility is the best way to determine the
encoding of a file. If you don't have Solaris 10 - you can download it from
http://www.sun.com/software/solaris/.
Back to Top
24. How do I reset the default locale from ssh and reboot with the new locale?
To change the default locale of a machine, you need to set environment variables in the /etc/default/init file.


See http://docs.sun.com/app/docs/doc/816-0677/6m758jcqf?a=view for an example of how it is done.
If the LC_* variables are not set in this file, then the default locale is the 'C' locale.


To reboot the system from the ssh shell, as root, simply type 'reboot'.

Back to Top
25. Do you offer a locale which conforms to ISO8601?
No we don't offer a locale which conforms to ISO8601. I'm afraid there is no current plan to introduce an ISO8601 type locale variant so for the moment you have at least 3 options:
Build your own with localedef.
Use the strftime () function to format the date if they need the date inside an app written in C/C++/perl etc.
Use "date +%Y-%m-%d" if it's a shell script.

Back to Top
26. How do I know which locales are installed on Solaris?
You simply need to run:


# locale -a

Note: Some locales are partial, others are full. A partial locale has the currency, date/time format etc. - like the Polish locale: pl_PL.UTF-8, while a full locale has that plus full translations for the user interface. - say the Spanish es_ES.UTF-8 locale.  


For adding locale support see http://developers.sun.com/dev/gadc/techtips/sol9_locale_pkgs/index.html

Back to Top
27. Is UTF-16 a valid and supported font set in Solaris 8 and 9?
Java 1.5 supports UTF-16 as its primary codeset. Solaris support UTF-16 only through iconv code conversions.


All major TrueType (and so thus OpenType) fonts themselves are encoded by using UTF-16 but only BMP portion usually, i.e., UCS-2. Solaris maps such and any other font encodings into various X11 encodings such as ISO8859-1, eucJP, Big5, Shift_JIS, KOI8-R, UTF-8 and so on by using various internal mechanisms that are transparent to users and fonts.


We do not support UTF-16 as an X11 fontset. Unicode locales in Solaris however supports an X11 fontset (UTF-8) to support Unicode/UTF-8 characters; that fontset can be used to display Unicode characters in X11 applications.


We also do not have any plan to support UTF-16 fontset directly since that's not really possible due to for instance, UTF-16 encoding alone doesn't contain enough characters required for rendering of context-dependent shaping and bidirectional scripts. I.e., Unicode defines logical characters but not all presentation characters and some scripts/languages require many more presentation characters than what the Unicode defines; it is up to each GUI components of platforms on how to map such logical characters to presentation characters after shaping and ordering.

Back to Top
28. What locales are shipped with Solaris 10 6/06?
Solaris 10 6/06 (Update Release 2) ships many new UTF-8 locales. See here for the complete list. If you have an earlier release of Solaris 10, and want to add those locales that have only been shipped with Solaris 10 UR2, then please see the Solaris 10 Locale Patches page.
Back to Top
29. What are the specific packages required for Solaris 10 locales?
Answer
Back to Top
30. I noticed some of my locale data has changed in Solaris 10 8/07, what has happened?
This change, which is noted in the Solaris 10 8/07 "What's New", was made in order to align our locale data with the Common Locale Data Repository (CLDR). This improves the quality and consistency of our locale data.


The CLDR is the industry standard for locale data.



In order to guard against changes to locale data adversely affecting programs, it is recommended that such programs should not depend  on the format of the default output of the date(1) command
as it may change, but instead should explicitly format the date by following the rules described in the man pages of date(1) and strftime(3C). For example :

# env LC_ALL=en_US.ISO8859-1 date +"%a %b %d %H:%M:%S %Z %Y"



will output :

Wed Sep 26 10:57:31 IST 2007