XGETTEXT(1) General Commands Manual XGETTEXT(1)

xgettext — extract gettext call strings from C-language source files (DEVELOPMENT)


xgettext [-j] [-n] [-d default-domain] [-K keyword-spec]... [-p pathname] file...

xgettext -a [-n] [-d default-domain] [-p pathname] [-x exclude-file] file...

The xgettext utility shall automate the creation of portable messages object source files (dot-po files). A dot-po file shall contain copies of string literals that are found in C-language source code in files specified by file operands. The dot-po file can be used as input to the msgfmt(1) utility, to produce a messages object file that can be used by applications.

The xgettext utility shall write msgid argument strings that are passed as string literals in gettext(3), gettext_l(3), ngettext(3), and ngettext_l(3) calls in C-language source code to the default output file; this file shall be named messages.po unless it is changed by the -d option. The xgettext utility shall also write msgid argument strings that are passed as string literals in dcgettext(3), dcgettext_l(3), dcngettext(3), dcngettext_l(3), dgettext(3), dgettext_l(3), dngettext(3), and dngettext_l(3) calls either to the default output file or to the output file domainname where domainname is the first parameter to the call; it is implementation-defined which of those output files is used. A msgid directive shall precede each msgid argument string. For the functions that have a msgid_plural argument, a msgid_plural directive followed by that argument string shall also be written directly after the corresponding msgid directive. A directive or ] directives with an empty string shall be written after the corresponding msgid or msgid_plural directive, respectively. The function names that xgettext searches for can be changed using the -K option.

The first directive in each created dot-po file shall be a directive giving the associated domain name, except that this directive is optional in the default output file.

If the -p pathname option is specified, xgettext shall create the dot-po files in the pathname directory. Otherwise, the dot-po files shall be created in the current working directory.

The msgid values shall be in the same order that the strings are extracted from each file and subsections with duplicate msgid values shall be written to the dot-po files as comment lines.

The xgettext utility shall conform to XBD V1_chap12(7).

The following options shall be supported:

Extract all strings, not just those found in calls to gettext family functions. Only one dot-po file shall be created.

Name the default output file default-domain .po instead of messages.po .
Join messages from C-language source files with existing dot-po files. For each dot-po file that xgettext writes messages to, if the file does not exist, it shall be created. New messages shall be appended but any subsections with duplicate msgid values except the first (including msgid values found in an existing dot-po file) shall either be commented out or omitted in the resulting dot-po file; if omitted, a warning message may be written to standard error. Domain directives in the existing dot-po files shall be ignored; the assumption is that all previous msgid values belong to the same domain. The behavior is unspecified if an existing dot-po file was not created by xgettext or has been modified by another application.

Specify an additional keyword to be looked for:
  • If keyword-spec is an empty string, this shall disable the use of default keywords for the gettext family of functions.
  • If keyword-spec is a C identifier, xgettext shall look for strings in the first argument of each call to the function or macro keyword-spec.
  • If keyword-spec is of the form id: argnum then xgettext shall treat the argnum -th argument of a call to the function or macro id as the msgid argument, where argnum 1 is the first argument.
  • If keyword-spec is of the form id: argnum1, argnum2 then xgettext shall treat strings in the argnum1 -th argument and in the argnum2 -th argument of a call to the function or macro id as the msgid and msgid_plural arguments, respectively.

For all mentioned forms, the application shall ensure that if argnum2 is given, it is not equal to argnum1. All numeric values shall be converted as specified in item 6 in XBD V1_chap12(7).

Add comment lines to the output file indicating pathnames and line numbers in the source files where each extracted string is encountered. These lines shall appear before each msgid directive. Such comments should have the format: 
#: pathname1:linenumber1 [pathname2:linenumber2...]

Create output files in the directory specified by pathname instead of in the current working directory.

Specify a file containing strings that shall not be extracted from the input files. The format of exclude-file is identical to that of a dot-po file. However, only statements containing msgid directives in exclude-file shall be used. All other statements shall be ignored.

The following operand shall be supported:

file
A pathname of an input file containing C-language source code. If '-' is specified for an instance of file , the standard input shall be used.

The standard input shall not be used unless a file operand is specified as ‘-’.

The input files specified as file operands shall be C-language source files. The input file specified as the option-argument for the -x option shall be a dot-po file in the format specified as input for the msgfmt(1) utility.

The following environment variables shall affect the execution of xgettext:

Provide a default value for the internationalization variables that are unset or null. (See XBD for the precedence of internationalization variables used to determine the values of locale categories.)
Determine the location of messages objects [ XSI] if NLSPATH is not set or the evaluation of NLSPATH did not lead to a suitable messages object being found.
If set to a non-empty string value, override the values of all the other internationalization variables.
Determine the locale for the interpretation of sequences of bytes of text data as characters (for example, single-byte as opposed to multi-byte characters in arguments and input files).

Determine the locale name used to locate messages objects, and the locale that should be used to affect the format and contents of diagnostic messages written to standard error.
[ XSI] Determine the location of messages objects and message catalogs.

Default.

The standard output shall not be used.

The standard error shall be used for diagnostic messages and may be used for warning messages.

The output files shall be dot-po files in the format specified as input for the msgfmt(1) utility. It is unspecified whether each output file includes a header ( msgid "") before the content derived from the input C-language source files.

None.

The following exit values shall be returned:

Successful completion.
An error occurred.

Default.

Implementations differ as to whether they write all output to the default output file or split the output into separate per-domain files. Portable applications can either ensure that each C-language source file contains calls to gettext family functions for only a single domain, or force all output to be to the default output file by using the -K option to override the default keywords.

Some implementations of xgettext are not able to extract cast strings (unless -a is used), for example casts of literal strings to . Use of a cast is unnecessary anyway, since the prototypes in <libintl.h> already specify this type.

The xgettext utility is not required to handle C preprocessor directives. Therefore if, for example, calls to gettext family functions are wrapped by macros, they might not be found unless the -K option is used to tell xgettext to look for the macro calls.

Example 1

The following example shows how -K can be used to force all output to be to the default output file:

xgettext -K "" -K gettext:1 -K dgettext:2 -K dcgettext:2 \
    -K ngettext:1,2 -K dngettext:2,3 -K dcngettext:2,3 source.c

By overriding the default keywords using the -K option as above, the xgettext utility is directed to ignore the domainname arguments to the dgettext(3), dcgettext(3), dngettext(3), and dcngettext(3) functions. Thus, the utility treats the functions as their respective equivalent without the d prefix, ignoring the domainname argument and writing generated output to the default output file, messages.po. Additional -K options would be needed for the variants of the functions with an suffix if they are used.

Example 2

If the source uses a macro definition such as:

#define i18n gettext

the use of:

xgettext -K i18n:1 source.c

will pick up msgid values from a line such as:

fprintf(stdout, i18n("The value is %s"), value1);

The -K option is based on the -k option of GNU xgettext; the only difference is that GNU's -k takes an optional option-argument whereas -K in this standard has a mandatory option-argument in order to comply with the syntax guidelines.

The standard developers considered including functionality equivalent to the -c, -m, and -M options in existing implementations. However, those letters could not be used as the syntax differed between implementations. The usual solution of adding an uppercase equivalent of lowercase options with the standard syntax instead was not possible, for obvious reasons for -m and -M, and as -C was already in use for another purpose in one implementation.

The -s option is not included as it has been deprecated in at least one implementation because it has been found to deprive translators of valuable context.

If this utility is directed to create a new directory entry that contains any bytes that have the encoded value of a <newline> character, implementations are encouraged to treat this as an error. A future version of this standard may require implementations to treat this as an error.

A future version of this standard may change the description of the -n option to mandate the given comment format (by using "shall" instead of "should").

gettext(1), msgfmt(1)

XBD V1_chap08(7), V1_chap12(7)

XSH gettext(3)

First released in Issue 8.

June 14, 2024 posix.fail