Standardize Localization Tasks The locale functionality is the interface the centos-art.sh script provides to standardize localization tasks inside the working copy. centos-art locale [OPTIONS] [DIRECTORY] The DIRECTORY parameter specifies the directory path, inside the working copy of &TCAR;, where the files you want to process are stored in. This paramter can be provided more than once in order to process more than one directory path in a single command execution. When this parameter is not provided, the current directory path where the command was called from is used instead. The locale functionality accepts the following options: Supress all output messages except error messages. When this option is passed, all confirmation requests are supressed and a possitive answer is assumed for them, just as if the option would have been provided. Assume yes to all confirmation requests. Reduce the list of files to process inside DIRECTORY using REGEX as pattern. You can use this option to control the amount of files you want to locale. The deeper you go into the directory structure the more specific you'll be about the files you want to locale. When you cannot go deeper into the directory structure through DIRECTORY specification, use this option to reduce the list of files therein. Supress all commit and update actions realized over files, before and after the actions themselves had took place over files in the working copy. This option updates both POT and PO files related to source files. Use this option everytime you change translatable strings inside the source files. This option edits the portable object related to source files. When you provide this option, your default text editor is used to open the portable object you, as translator, need to change in order to keep source file messages consistent with their localized versions. In the very specific case of shell scripts localization, this option takes care of updating the machine object (MO) file the shell script requires to displayed translation messages correctly when it is executed. This option unlocalizes source files. When you provide this option, the localization directory related to source files is removed from the working copy in conjunction with all portable objects and machine objects inside it. This option suppresses machine objects creation when shell scripts are localized. The localization process is very tied to the source files we want to provide localized messages for. Inside the working copy of &TCAR; it is possible to localize XML-based files (e.g., SVG and Docbook) and programs written in most popular programming languages (e.g., C, C++, C#, Shell Scripts, Python, Java, GNU awk, PHP, etc.). The localization process initiates by retriving translatable strings from source files. When source files are XML-based files, the only requisite to retrive translatable strings correctly is that they be well-formed. Beyond that, the xml2po command takes care of everything else. When source files are Shell script files, it is necessary that you previously define what strings inside the script are considered as translatable strings in order for xgettext command to retrive them correctly. To define translatable strings inside shell scripts, you need to use either gettext, ngettext, eval_gettext or eval_ngettext command as it is following described: Use the gettext command to display the native language translation of a textual message. MESSAGE="`gettext "There is no entry to create."`" Use the ngettext command to display the native language translation of a textual message whose grammatical form depends on a number. MESSAGE="`ngettext "The following entry will be created" \ "The following entries will be created" \ $COUNT`:" Use the eval_gettext command to display the native language translation of a textual message, performing dollar-substitution on the result. Note that only shell variables mentioned in the message will be dollar-substituted in the result. MESSAGE="`eval_gettext "The location \\\"\\\$LOCATION\\\" is not valid."`" Use the eval_ngettext command to display the native language translation of a textual message whose grammatical form depends on a number, performing dollar-substitution on the result. Note that only shell variables mentioned in messages will be dollar-substituted in the result. MESSAGE="`eval_ngettext "The following entry will be created in \\\$LOCATION" \ "The following entries will be created in \\\$LOCATION" \ $COUNT`:" Once translatable strings are retrived, a portable object template (POT) file is created for storing them. Later, the POT file is used to create a portable object (PO). The portable object is the place where localization itself takes place, it is the file translators edit to localize messages. When translatable strings change inside source files, it is necessary that you update these POT and PO files in order to keep consistency between source file messages and their localized versions. Inside source files, translatable strings are always written in English language. In order to localize translatable strings from English language to another language, you need to be sure the LANG environment variable has been already set to the locale code you want to localize message for or see them printed out before running the locale functionality of centos-art.sh script. Localizing English language to itself is not supported. To have a list of all locale codes you can have localized messages for, run the following command: locale -a | less.