How-To internationalize shell scripts with gettext in 3 easy steps!

1- MODIFICATION OF THE ORIGINAL SCRIPT:

(We shall take excerpts from the script keyboardconfig as an example.)

1-a- Add the following lines after the usual introductory comments
but before your actual code starts:

	# Gettext internationalization
	export TEXTDOMAIN="name_of_the_program"
	export TEXTDOMAINDIR="destination_path_to_the_compiled_translations"
	. gettext.sh

Example...

-> Before modification:
	
	# Version 4.3 - 28/07/2009
	#  * replaced xorg.conf by fdi policy rule

	# Take a look at "Xdialog" and use it instead of "dialog" in case X is running
	if [[ "$DISPLAY" && "$(which Xdialog 2>&1 | grep -v "which: no")" ]]; then

	
	
-> After modification:
	
	# Version 4.3 - 28/07/2009
	#  * replaced xorg.conf by fdi policy rule

	# Gettext internationalization
	export TEXTDOMAIN="keyboardconfig"
	export TEXTDOMAINDIR="/usr/share/locale"
	. gettext.sh

	# Take a look at "Xdialog" and use it instead of "dialog" in case X is running
	if [[ "$DISPLAY" && "$(which Xdialog 2>&1 | grep -v "which: no")" ]]; then
	
	
1-b- Put the following tags around the strings that need to be translated:

	`eval_gettext 'string_to_be_translated'`

Exemple...

-> Before modification:
	
	if [ "$xflag" = "yes" ] ; then
	answer="$(eval $dialog \
	--stdout \
	--title \"Keyboard configuration\" \
	--default-item \"$currentpathkeymap\" \
	--cancel-label \"Exit\" \
	--icon \"keyboardconfig\" \
	--check \"numlock\" \"on\" \
	--menu \
	\"\\n        Please select your prefered keyboard map:\" 20 75 11 "$list" )"	
	else

-> After Modification:
	
	if [ "$xflag" = "yes" ] ; then
	answer="$(eval $dialog \
	--stdout \
	--title \"`eval_gettext 'Keyboard configuration'`\" \
	--default-item \"$currentpathkeymap\" \
	--cancel-label \"`eval_gettext 'Exit'`\" \
	--icon \"keyboardconfig\" \
	--check \"numlock\" \"on\" \
	--menu \
	\"\\n        `eval_gettext 'Please select your prefered keyboard map:'`\" 20 75 11 "$list" )"
	else


2- TRANSLATION PROCESS:

2-a- We first need to extract the strings to be translated from the shell script to a model file 
that will be distributed to the translators. To do this, run the following command from the working 
directory where the shell script is located:

		xgettext --from-code=utf-8 -L shell -o name_of_shell_script.pot name_of_shell_script
		
Example: xgettext --from-code=utf-8 -L shell -o keyboardconfig.pot keyboardconfig 

					
2-b- Each translator will use the model file to generate his own translation work file by running the 
following command from the directory where the .pot file (our translation model) is located. It is 
assumed that the environment locale of his operating system is the same as the locale he wants to
translate into (or else he should consult the man page of msginit):

	msginit -i name_of_shell_script.pot -o name_of_shell_script-name_of_locale.po
	
Example: msginit -i keyboardconfig.pot -o keyboardconfig-fr.po

2-c- Looking into the file is self explanatory, using a text editor of his choice
each translator translate the strings in his language po file.

Example...

	Before:
	
	#: keyboardconfig:93 keyboardconfig:103
	#, sh-format
	msgid "Keyboard configuration"
	msgstr ""
		
	After:
	
	#: keyboardconfig:93 keyboardconfig:103
	#, sh-format
	msgid "Keyboard configuration"
	msgstr "Configuration du clavier"
		
		
3- EXTENDING THE TRANSLATIONS TO THE SHELL SCRIPT:

3-a- When the maintainer of the shell script receives back a completed po file from a 
translator, he first compiles it by running the following command (from the working 
directory where the translated po file is located):

msgfmt name_of_shell_script-name_of_locale.po -o name_of_shell_script.mo

Example: msgfmt keyboardconfig-fr.po -o keyboardconfig.mo

3-b- Then he places the mo file he generated into its appropriate locale directory:

	/usr/share/locale/name_of_locale/LC_MESSAGES/
	
Example: /usr/share/locale/fr/LC_MESSAGES/

That's all! 

NOTES:

Maintenance of the shell script & future translations:

	If later on, the shell script is modified, its maintainer will issue a new 
	pot (model) file as in point 2a which he can send back to the translators.

	Each translator will then have to run the following command in order to generate 
	a new po file for their language that will include their previous translation work
	as well as the new strings net yet translated. Both the new pot file & the last
	translated po file should be located in the directory where this command is issued.


	msgmerge -U name_of_shell_script-name_of_locale.po name_of_shell_script.pot

	Exemple: msgmerge -U keyboardconfig-fr.po keyboardconfig.pot

	And then redo points 2c, 3a & 3b.