11.3 Internationalization

Starting with Samba 3.0, Samba supports Unicode "on the wire," requiring no additional effort on your part to support filenames and other text containing characters in international character sets.

11.3.1 Internationalization Options

Samba 2.2.x has a limited ability to speak foreign tongues : if you need to support filenames containing characters that aren't in standard ASCII, some options that can help you are shown in Table 11-3.

Table 11-3. Internationalization options

Option	Parameters	Function	Default	Scope
client code page	Described in this section	Sets a code page to expect from clients	850	Global
character set	Described in this section	Translates code pages into alternate Unix character sets	None	Global
coding system	Described in this section	Translates code page 932 into an Asian character set	None	Global
valid chars	string (set of characters)	Adds individual characters to a code page	None	Global

11.3.1.1 client code page

The character sets on Windows platforms hark back to the original concept of a code page . These code pages are used by DOS and Windows clients to determine rules for mapping lowercase letters to uppercase letters . Samba can be instructed to use a variety of code pages through the use of the global client code page option to match the corresponding code page in use on the client. This option loads a code page definition file and can take the values specified in Table 11-4.

Table 11-4. Valid code pages with Samba 2.0

Code page	Definition
437	MS-DOS Latin (United States)
737	Windows 95 Greek
850	MS-DOS Latin 1 (Western European)
852	MS-DOS Latin 2 (Eastern European)
861	MS-DOS Icelandic
866	MS-DOS Cyrillic (Russian)
932	MS-DOS Japanese Shift-JIS
936	MS-DOS Simplified Chinese
949	MS-DOS Korean Hangul
950	MS-DOS Traditional Chinese

You can set the client code page as follows :

 [global]     client code page = 852 

The default value of this option is 850, for MS-DOS Latin 1. You can use the make_smbcodepage tool that comes with Samba (by default in /usr/local/samba/bin ) to create your own SMB code pages, in the event that those listed earlier are not sufficient.

11.3.1.2 character set

The global character set option can be used to convert filenames offered through a DOS code page (see the previous section, Section 11.3.1.1) to equivalents that can be represented by Unix character sets other than those in the United States. For example, if you want to convert the Western European MS-DOS character set on the client to a Western European Unix character set on the server, you can use the following in your configuration file:

 [global]     client code page = 850     character set = ISO8859-1 

Note that you must include a client code page option to specify the character set from which you are converting. The valid character sets (and their matching code pages) that Samba accepts are listed in Table 11-5.

Table 11-5. Valid character sets

Character set	Matching code page	Definition
ISO8859-1	850	Western European Unix
ISO8859-2	852	Eastern European Unix
ISO8859-5	866	Russian Cyrillic Unix
ISO8859-7	737	Greek Unix
KOI8-R	866	Alternate Russian Cyrillic Unix

Normally, the character set option is disabled completely.

11.3.1.3 coding system

The coding system option is similar to the character set option. However, its purpose is to determine how to convert a Japanese Shift JIS code page into an appropriate Unix character set. To use this option, the client code page option described previously must be set to page 932 . The valid coding systems that Samba accepts are listed in Table 11-6.

Table 11-6. Valid coding-system parameters

Character set	Definition
SJIS	Standard Shift JIS
JIS8	Eight-bit JIS codes
J8BB	Eight-bit JIS codes
J8BH	Eight-bit JIS codes
J8@B	Eight-bit JIS codes
J8@J	Eight-bit JIS codes
J8@H	Eight-bit JIS codes
JIS7	Seven-bit JIS codes
J7BB	Seven-bit JIS codes
J7BH	Seven-bit JIS codes
J7@B	Seven-bit JIS codes
J7@J	Seven-bit JIS codes
J7@H	Seven-bit JIS codes
JUNET	JUNET codes
JUBB	JUNET codes
JUBH	JUNET codes
JU@B	JUNET codes
JU@J	JUNET codes
JU@H	JUNET codes
EUC	EUC codes
HEX	Three-byte hexadecimal code
CAP	Three-byte hexadecimal code (Columbia AppleTalk Program)

11.3.1.4 valid chars

The valid chars option can be used to add individual characters to a code page. You can use this option as follows:

 valid chars =  valid chars = 0450:0420 0x0A20:0x0A00 valid chars = A:a 

Each character in the list specified should be separated by spaces. If there is a colon between two characters or a numerical equivalent, the data to the left of the colon is considered an uppercase character, while the data to the right is considered the lowercase character. You can represent characters both by literals (if you can type them) and by octal, hexadecimal, or decimal Unicode equivalents.

If you use this option, it must be listed after the client code page to which you wish to add the character.