Standard C Language Internationalization Programming

Standard C Language Internationalization Programming
Messaging

An international program should provide a user interface in the user's native language. Language C provides the message catalog, a mechanism to select messages dynamically depending on the current locale. You can use the message catalog as follows:

Create a translated message catalog
- Pick up the English string literals from source codes and put them into a message text file (.msg).
- Translate the string literals in the message text file (into Japanese, for example).
- Generate a message catalog file (.cat) from the message text file with 'gencat' command.
- Install the message catalog file in one of the system message catalog directories provided for each locale (/usr/share/locale/<locale-name>/LC_MESSAGES/, for example).
Use the message catalog from the program
- Open the message catalog by calling 'catopen()'. A message catalog file of the current locale in the system message catalog directories is searched with LC_MESSAGES locale category.
- Get a translated message from the opened message catalog by calling 'catgets()'.

The following are the standard C library functions related to message catalog:

------------------------------------------------------
Function	Description
------------------------------------------------------
catopen()	Opens a message catalog
catgets()	Gets a message from a message catalog
catclose()	Closes a message catalog
------------------------------------------------------

Most string literals should be translated into a natural language but some string literals, such as pathnames, are common for all languages. A string literal may not have to be translated if it is an argument of a certain function. Generally speaking, you may not have to translate string literals such as the following:

a single ASCII word
a value of #include and #define macros
a value of a variable whose name contains "sql" (maybe SQL statement)
an argument of 'catopen()', 'catgets()', and 'strfmon()'

World Wide Navi regards the messages which do not have to be translated. This setting can be changed by the DB customizing functionality.

*** Message Catalog Details ***

The format of the .msg files is described below.

wwnavi_cat.msg

$ $NLS-WWNAVI 2008-05-24T16:48:54+0900
$ 
$quote "

$set 1
1 "Failed to create fontset\n" <-- *Message ID and Message
2 "Hello, how are you? This is sample for catgets. (Push <enter>.)"

The locations of the .cat files depend on the environment variable NLSPATH.

For example, run the program 'Test' according to the settings below.

export NLSPATH=./%L/%N

export LANG=en_US.UTF-8
./Test

export LANG=ja_JP.UTF-8
./Test

The location of the .cat files will be the following:

./en_US.UTF-8/wwnavi_cat.cat <-- *English message catalog
./ja_JP.UTF-8/wwnavi_cat.cat <-- *Japanese message catalog

Loading messages in the source codes is described below.

setlocale (LC_ALL, ""); <-- *Described in the main function
nl_catd wwnavi_cd = catopen ("wwnavi_cat.cat", NL_CAT_LOCALE); <-- *Described in the main function 
... 
printf(catgets(wwnavi_cd, 1, 1, "Failed to create fontset\n")); <-- *Loading messages

This process can be checked with 'catgets' externalization of the sample 'linux_catgets' in World Wide Navi.

Go to Internationalization Programming Top

Standard C Language Internationalization Programming Messaging

Standard C Language Internationalization Programming
Messaging