xref: /netbsd-src/external/gpl2/gettext/dist/gettext-tools/doc/gettext_15.html (revision 946379e7b37692fc43f68eb0d1c10daa0a7f3b6c)
1<HTML>
2<HEAD>
3<!-- This HTML file has been created by texi2html 1.52b
4     from gettext.texi on 27 November 2006 -->
5
6<META HTTP-EQUIV="content-type" CONTENT="text/html; charset=UTF-8">
7<TITLE>GNU gettext utilities - 15  Other Programming Languages</TITLE>
8</HEAD>
9<BODY>
10Go to the <A HREF="gettext_1.html">first</A>, <A HREF="gettext_14.html">previous</A>, <A HREF="gettext_16.html">next</A>, <A HREF="gettext_25.html">last</A> section, <A HREF="gettext_toc.html">table of contents</A>.
11<P><HR><P>
12
13
14<H1><A NAME="SEC229" HREF="gettext_toc.html#TOC229">15  Other Programming Languages</A></H1>
15
16<P>
17While the presentation of <CODE>gettext</CODE> focuses mostly on C and
18implicitly applies to C++ as well, its scope is far broader than that:
19Many programming languages, scripting languages and other textual data
20like GUI resources or package descriptions can make use of the gettext
21approach.
22
23</P>
24
25
26
27<H2><A NAME="SEC230" HREF="gettext_toc.html#TOC230">15.1  The Language Implementor's View</A></H2>
28<P>
29<A NAME="IDX1111"></A>
30<A NAME="IDX1112"></A>
31
32</P>
33<P>
34All programming and scripting languages that have the notion of strings
35are eligible to supporting <CODE>gettext</CODE>.  Supporting <CODE>gettext</CODE>
36means the following:
37
38</P>
39
40<OL>
41<LI>
42
43You should add to the language a syntax for translatable strings.  In
44principle, a function call of <CODE>gettext</CODE> would do, but a shorthand
45syntax helps keeping the legibility of internationalized programs.  For
46example, in C we use the syntax <CODE>_("string")</CODE>, and in GNU awk we use
47the shorthand <CODE>_"string"</CODE>.
48
49<LI>
50
51You should arrange that evaluation of such a translatable string at
52runtime calls the <CODE>gettext</CODE> function, or performs equivalent
53processing.
54
55<LI>
56
57Similarly, you should make the functions <CODE>ngettext</CODE>,
58<CODE>dcgettext</CODE>, <CODE>dcngettext</CODE> available from within the language.
59These functions are less often used, but are nevertheless necessary for
60particular purposes: <CODE>ngettext</CODE> for correct plural handling, and
61<CODE>dcgettext</CODE> and <CODE>dcngettext</CODE> for obeying other locale
62environment variables than <CODE>LC_MESSAGES</CODE>, such as <CODE>LC_TIME</CODE> or
63<CODE>LC_MONETARY</CODE>.  For these latter functions, you need to make the
64<CODE>LC_*</CODE> constants, available in the C header <CODE>&#60;locale.h&#62;</CODE>,
65referenceable from within the language, usually either as enumeration
66values or as strings.
67
68<LI>
69
70You should allow the programmer to designate a message domain, either by
71making the <CODE>textdomain</CODE> function available from within the
72language, or by introducing a magic variable called <CODE>TEXTDOMAIN</CODE>.
73Similarly, you should allow the programmer to designate where to search
74for message catalogs, by providing access to the <CODE>bindtextdomain</CODE>
75function.
76
77<LI>
78
79You should either perform a <CODE>setlocale (LC_ALL, "")</CODE> call during
80the startup of your language runtime, or allow the programmer to do so.
81Remember that gettext will act as a no-op if the <CODE>LC_MESSAGES</CODE> and
82<CODE>LC_CTYPE</CODE> locale facets are not both set.
83
84<LI>
85
86A programmer should have a way to extract translatable strings from a
87program into a PO file.  The GNU <CODE>xgettext</CODE> program is being
88extended to support very different programming languages.  Please
89contact the GNU <CODE>gettext</CODE> maintainers to help them doing this.  If
90the string extractor is best integrated into your language's parser, GNU
91<CODE>xgettext</CODE> can function as a front end to your string extractor.
92
93<LI>
94
95The language's library should have a string formatting facility where
96the arguments of a format string are denoted by a positional number or a
97name.  This is needed because for some languages and some messages with
98more than one substitutable argument, the translation will need to
99output the substituted arguments in different order.  See section <A HREF="gettext_4.html#SEC17">4.6  Special Comments preceding Keywords</A>.
100
101<LI>
102
103If the language has more than one implementation, and not all of the
104implementations use <CODE>gettext</CODE>, but the programs should be portable
105across implementations, you should provide a no-i18n emulation, that
106makes the other implementations accept programs written for yours,
107without actually translating the strings.
108
109<LI>
110
111To help the programmer in the task of marking translatable strings,
112which is sometimes performed using the Emacs PO mode (see section <A HREF="gettext_4.html#SEC16">4.5  Marking Translatable Strings</A>),
113you are welcome to
114contact the GNU <CODE>gettext</CODE> maintainers, so they can add support for
115your language to <TT>&lsquo;po-mode.el&rsquo;</TT>.
116</OL>
117
118<P>
119On the implementation side, three approaches are possible, with
120different effects on portability and copyright:
121
122</P>
123
124<UL>
125<LI>
126
127You may integrate the GNU <CODE>gettext</CODE>'s <TT>&lsquo;intl/&rsquo;</TT> directory in
128your package, as described in section <A HREF="gettext_13.html#SEC196">13  The Maintainer's View</A>.  This allows you to
129have internationalization on all kinds of platforms.  Note that when you
130then distribute your package, it legally falls under the GNU General
131Public License, and the GNU project will be glad about your contribution
132to the Free Software pool.
133
134<LI>
135
136You may link against GNU <CODE>gettext</CODE> functions if they are found in
137the C library.  For example, an autoconf test for <CODE>gettext()</CODE> and
138<CODE>ngettext()</CODE> will detect this situation.  For the moment, this test
139will succeed on GNU systems and not on other platforms.  No severe
140copyright restrictions apply.
141
142<LI>
143
144You may emulate or reimplement the GNU <CODE>gettext</CODE> functionality.
145This has the advantage of full portability and no copyright
146restrictions, but also the drawback that you have to reimplement the GNU
147<CODE>gettext</CODE> features (such as the <CODE>LANGUAGE</CODE> environment
148variable, the locale aliases database, the automatic charset conversion,
149and plural handling).
150</UL>
151
152
153
154<H2><A NAME="SEC231" HREF="gettext_toc.html#TOC231">15.2  The Programmer's View</A></H2>
155
156<P>
157For the programmer, the general procedure is the same as for the C
158language.  The Emacs PO mode marking supports other languages, and the GNU
159<CODE>xgettext</CODE> string extractor recognizes other languages based on the
160file extension or a command-line option.  In some languages,
161<CODE>setlocale</CODE> is not needed because it is already performed by the
162underlying language runtime.
163
164</P>
165
166
167<H2><A NAME="SEC232" HREF="gettext_toc.html#TOC232">15.3  The Translator's View</A></H2>
168
169<P>
170The translator works exactly as in the C language case.  The only
171difference is that when translating format strings, she has to be aware
172of the language's particular syntax for positional arguments in format
173strings.
174
175</P>
176
177
178
179<H3><A NAME="SEC233" HREF="gettext_toc.html#TOC233">15.3.1  C Format Strings</A></H3>
180
181<P>
182C format strings are described in POSIX (IEEE P1003.1 2001), section
183XSH 3 fprintf(),
184<A HREF="http://www.opengroup.org/onlinepubs/007904975/functions/fprintf.html">http://www.opengroup.org/onlinepubs/007904975/functions/fprintf.html</A>.
185See also the fprintf() manual page,
186<A HREF="http://www.linuxvalley.it/encyclopedia/ldp/manpage/man3/printf.3.php">http://www.linuxvalley.it/encyclopedia/ldp/manpage/man3/printf.3.php</A>,
187<A HREF="http://informatik.fh-wuerzburg.de/student/i510/man/printf.html">http://informatik.fh-wuerzburg.de/student/i510/man/printf.html</A>.
188
189</P>
190<P>
191Although format strings with positions that reorder arguments, such as
192
193</P>
194
195<PRE>
196"Only %2$d bytes free on '%1$s'."
197</PRE>
198
199<P>
200which is semantically equivalent to
201
202</P>
203
204<PRE>
205"'%s' has only %d bytes free."
206</PRE>
207
208<P>
209are a POSIX/XSI feature and not specified by ISO C 99, translators can rely
210on this reordering ability: On the few platforms where <CODE>printf()</CODE>,
211<CODE>fprintf()</CODE> etc. don't support this feature natively, <TT>&lsquo;libintl.a&rsquo;</TT>
212or <TT>&lsquo;libintl.so&rsquo;</TT> provides replacement functions, and GNU <CODE>&#60;libintl.h&#62;</CODE>
213activates these replacement functions automatically.
214
215</P>
216<P>
217<A NAME="IDX1113"></A>
218<A NAME="IDX1114"></A>
219As a special feature for Farsi (Persian) and maybe Arabic, translators can
220insert an <SAMP>&lsquo;I&rsquo;</SAMP> flag into numeric format directives.  For example, the
221translation of <CODE>"%d"</CODE> can be <CODE>"%Id"</CODE>.  The effect of this flag,
222on systems with GNU <CODE>libc</CODE>, is that in the output, the ASCII digits are
223replaced with the <SAMP>&lsquo;outdigits&rsquo;</SAMP> defined in the <CODE>LC_CTYPE</CODE> locale
224facet.  On other systems, the <CODE>gettext</CODE> function removes this flag,
225so that it has no effect.
226
227</P>
228<P>
229Note that the programmer should <EM>not</EM> put this flag into the
230untranslated string.  (Putting the <SAMP>&lsquo;I&rsquo;</SAMP> format directive flag into an
231<VAR>msgid</VAR> string would lead to undefined behaviour on platforms without
232glibc when NLS is disabled.)
233
234</P>
235
236
237<H3><A NAME="SEC234" HREF="gettext_toc.html#TOC234">15.3.2  Objective C Format Strings</A></H3>
238
239<P>
240Objective C format strings are like C format strings.  They support an
241additional format directive: "$@", which when executed consumes an argument
242of type <CODE>Object *</CODE>.
243
244</P>
245
246
247<H3><A NAME="SEC235" HREF="gettext_toc.html#TOC235">15.3.3  Shell Format Strings</A></H3>
248
249<P>
250Shell format strings, as supported by GNU gettext and the <SAMP>&lsquo;envsubst&rsquo;</SAMP>
251program, are strings with references to shell variables in the form
252<CODE>$<VAR>variable</VAR></CODE> or <CODE>${<VAR>variable</VAR>}</CODE>.  References of the form
253<CODE>${<VAR>variable</VAR>-<VAR>default</VAR>}</CODE>,
254<CODE>${<VAR>variable</VAR>:-<VAR>default</VAR>}</CODE>,
255<CODE>${<VAR>variable</VAR>=<VAR>default</VAR>}</CODE>,
256<CODE>${<VAR>variable</VAR>:=<VAR>default</VAR>}</CODE>,
257<CODE>${<VAR>variable</VAR>+<VAR>replacement</VAR>}</CODE>,
258<CODE>${<VAR>variable</VAR>:+<VAR>replacement</VAR>}</CODE>,
259<CODE>${<VAR>variable</VAR>?<VAR>ignored</VAR>}</CODE>,
260<CODE>${<VAR>variable</VAR>:?<VAR>ignored</VAR>}</CODE>,
261that would be valid inside shell scripts, are not supported.  The
262<VAR>variable</VAR> names must consist solely of alphanumeric or underscore
263ASCII characters, not start with a digit and be nonempty; otherwise such
264a variable reference is ignored.
265
266</P>
267
268
269<H3><A NAME="SEC236" HREF="gettext_toc.html#TOC236">15.3.4  Python Format Strings</A></H3>
270
271<P>
272Python format strings are described in
273Python Library reference /
2742. Built-in Types, Exceptions and Functions /
2752.2. Built-in Types /
2762.2.6. Sequence Types /
2772.2.6.2. String Formatting Operations.
278<A HREF="http://www.python.org/doc/2.2.1/lib/typesseq-strings.html">http://www.python.org/doc/2.2.1/lib/typesseq-strings.html</A>.
279
280</P>
281
282
283<H3><A NAME="SEC237" HREF="gettext_toc.html#TOC237">15.3.5  Lisp Format Strings</A></H3>
284
285<P>
286Lisp format strings are described in the Common Lisp HyperSpec,
287chapter 22.3 Formatted Output,
288<A HREF="http://www.lisp.org/HyperSpec/Body/sec_22-3.html">http://www.lisp.org/HyperSpec/Body/sec_22-3.html</A>.
289
290</P>
291
292
293<H3><A NAME="SEC238" HREF="gettext_toc.html#TOC238">15.3.6  Emacs Lisp Format Strings</A></H3>
294
295<P>
296Emacs Lisp format strings are documented in the Emacs Lisp reference,
297section Formatting Strings,
298<A HREF="http://www.gnu.org/manual/elisp-manual-21-2.8/html_chapter/elisp_4.html#SEC75">http://www.gnu.org/manual/elisp-manual-21-2.8/html_chapter/elisp_4.html#SEC75</A>.
299Note that as of version 21, XEmacs supports numbered argument specifications
300in format strings while FSF Emacs doesn't.
301
302</P>
303
304
305<H3><A NAME="SEC239" HREF="gettext_toc.html#TOC239">15.3.7  librep Format Strings</A></H3>
306
307<P>
308librep format strings are documented in the librep manual, section
309Formatted Output,
310<A HREF="http://librep.sourceforge.net/librep-manual.html#Formatted%20Output">http://librep.sourceforge.net/librep-manual.html#Formatted%20Output</A>,
311<A HREF="http://www.gwinnup.org/research/docs/librep.html#SEC122">http://www.gwinnup.org/research/docs/librep.html#SEC122</A>.
312
313</P>
314
315
316<H3><A NAME="SEC240" HREF="gettext_toc.html#TOC240">15.3.8  Scheme Format Strings</A></H3>
317
318<P>
319Scheme format strings are documented in the SLIB manual, section
320Format Specification.
321
322</P>
323
324
325<H3><A NAME="SEC241" HREF="gettext_toc.html#TOC241">15.3.9  Smalltalk Format Strings</A></H3>
326
327<P>
328Smalltalk format strings are described in the GNU Smalltalk documentation,
329class <CODE>CharArray</CODE>, methods <SAMP>&lsquo;bindWith:&rsquo;</SAMP> and
330<SAMP>&lsquo;bindWithArguments:&rsquo;</SAMP>.
331<A HREF="http://www.gnu.org/software/smalltalk/gst-manual/gst_68.html#SEC238">http://www.gnu.org/software/smalltalk/gst-manual/gst_68.html#SEC238</A>.
332In summary, a directive starts with <SAMP>&lsquo;%&rsquo;</SAMP> and is followed by <SAMP>&lsquo;%&rsquo;</SAMP>
333or a nonzero digit (<SAMP>&lsquo;1&rsquo;</SAMP> to <SAMP>&lsquo;9&rsquo;</SAMP>).
334
335</P>
336
337
338<H3><A NAME="SEC242" HREF="gettext_toc.html#TOC242">15.3.10  Java Format Strings</A></H3>
339
340<P>
341Java format strings are described in the JDK documentation for class
342<CODE>java.text.MessageFormat</CODE>,
343<A HREF="http://java.sun.com/j2se/1.4/docs/api/java/text/MessageFormat.html">http://java.sun.com/j2se/1.4/docs/api/java/text/MessageFormat.html</A>.
344See also the ICU documentation
345<A HREF="http://oss.software.ibm.com/icu/apiref/classMessageFormat.html">http://oss.software.ibm.com/icu/apiref/classMessageFormat.html</A>.
346
347</P>
348
349
350<H3><A NAME="SEC243" HREF="gettext_toc.html#TOC243">15.3.11  C# Format Strings</A></H3>
351
352<P>
353C# format strings are described in the .NET documentation for class
354<CODE>System.String</CODE> and in
355<A HREF="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpguide/html/cpConFormattingOverview.asp">http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpguide/html/cpConFormattingOverview.asp</A>.
356
357</P>
358
359
360<H3><A NAME="SEC244" HREF="gettext_toc.html#TOC244">15.3.12  awk Format Strings</A></H3>
361
362<P>
363awk format strings are described in the gawk documentation, section
364Printf,
365<A HREF="http://www.gnu.org/manual/gawk/html_node/Printf.html#Printf">http://www.gnu.org/manual/gawk/html_node/Printf.html#Printf</A>.
366
367</P>
368
369
370<H3><A NAME="SEC245" HREF="gettext_toc.html#TOC245">15.3.13  Object Pascal Format Strings</A></H3>
371
372<P>
373Where is this documented?
374
375</P>
376
377
378<H3><A NAME="SEC246" HREF="gettext_toc.html#TOC246">15.3.14  YCP Format Strings</A></H3>
379
380<P>
381YCP sformat strings are described in the libycp documentation
382<A HREF="file:/usr/share/doc/packages/libycp/YCP-builtins.html">file:/usr/share/doc/packages/libycp/YCP-builtins.html</A>.
383In summary, a directive starts with <SAMP>&lsquo;%&rsquo;</SAMP> and is followed by <SAMP>&lsquo;%&rsquo;</SAMP>
384or a nonzero digit (<SAMP>&lsquo;1&rsquo;</SAMP> to <SAMP>&lsquo;9&rsquo;</SAMP>).
385
386</P>
387
388
389<H3><A NAME="SEC247" HREF="gettext_toc.html#TOC247">15.3.15  Tcl Format Strings</A></H3>
390
391<P>
392Tcl format strings are described in the <TT>&lsquo;format.n&rsquo;</TT> manual page,
393<A HREF="http://www.scriptics.com/man/tcl8.3/TclCmd/format.htm">http://www.scriptics.com/man/tcl8.3/TclCmd/format.htm</A>.
394
395</P>
396
397
398<H3><A NAME="SEC248" HREF="gettext_toc.html#TOC248">15.3.16  Perl Format Strings</A></H3>
399
400<P>
401There are two kinds format strings in Perl: those acceptable to the
402Perl built-in function <CODE>printf</CODE>, labelled as <SAMP>&lsquo;perl-format&rsquo;</SAMP>,
403and those acceptable to the <CODE>libintl-perl</CODE> function <CODE>__x</CODE>,
404labelled as <SAMP>&lsquo;perl-brace-format&rsquo;</SAMP>.
405
406</P>
407<P>
408Perl <CODE>printf</CODE> format strings are described in the <CODE>sprintf</CODE>
409section of <SAMP>&lsquo;man perlfunc&rsquo;</SAMP>.
410
411</P>
412<P>
413Perl brace format strings are described in the
414<TT>&lsquo;Locale::TextDomain(3pm)&rsquo;</TT> manual page of the CPAN package
415libintl-perl.  In brief, Perl format uses placeholders put between
416braces (<SAMP>&lsquo;{&rsquo;</SAMP> and <SAMP>&lsquo;}&rsquo;</SAMP>).  The placeholder must have the syntax
417of simple identifiers.
418
419</P>
420
421
422<H3><A NAME="SEC249" HREF="gettext_toc.html#TOC249">15.3.17  PHP Format Strings</A></H3>
423
424<P>
425PHP format strings are described in the documentation of the PHP function
426<CODE>sprintf</CODE>, in <TT>&lsquo;phpdoc/manual/function.sprintf.html&rsquo;</TT> or
427<A HREF="http://www.php.net/manual/en/function.sprintf.php">http://www.php.net/manual/en/function.sprintf.php</A>.
428
429</P>
430
431
432<H3><A NAME="SEC250" HREF="gettext_toc.html#TOC250">15.3.18  GCC internal Format Strings</A></H3>
433
434<P>
435These format strings are used inside the GCC sources.  In such a format
436string, a directive starts with <SAMP>&lsquo;%&rsquo;</SAMP>, is optionally followed by a
437size specifier <SAMP>&lsquo;l&rsquo;</SAMP>, an optional flag <SAMP>&lsquo;+&rsquo;</SAMP>, another optional flag
438<SAMP>&lsquo;#&rsquo;</SAMP>, and is finished by a specifier: <SAMP>&lsquo;%&rsquo;</SAMP> denotes a literal
439percent sign, <SAMP>&lsquo;c&rsquo;</SAMP> denotes a character, <SAMP>&lsquo;s&rsquo;</SAMP> denotes a string,
440<SAMP>&lsquo;i&rsquo;</SAMP> and <SAMP>&lsquo;d&rsquo;</SAMP> denote an integer, <SAMP>&lsquo;o&rsquo;</SAMP>, <SAMP>&lsquo;u&rsquo;</SAMP>, <SAMP>&lsquo;x&rsquo;</SAMP>
441denote an unsigned integer, <SAMP>&lsquo;.*s&rsquo;</SAMP> denotes a string preceded by a
442width specification, <SAMP>&lsquo;H&rsquo;</SAMP> denotes a <SAMP>&lsquo;location_t *&rsquo;</SAMP> pointer,
443<SAMP>&lsquo;D&rsquo;</SAMP> denotes a general declaration, <SAMP>&lsquo;F&rsquo;</SAMP> denotes a function
444declaration, <SAMP>&lsquo;T&rsquo;</SAMP> denotes a type, <SAMP>&lsquo;A&rsquo;</SAMP> denotes a function argument,
445<SAMP>&lsquo;C&rsquo;</SAMP> denotes a tree code, <SAMP>&lsquo;E&rsquo;</SAMP> denotes an expression, <SAMP>&lsquo;L&rsquo;</SAMP>
446denotes a programming language, <SAMP>&lsquo;O&rsquo;</SAMP> denotes a binary operator,
447<SAMP>&lsquo;P&rsquo;</SAMP> denotes a function parameter, <SAMP>&lsquo;Q&rsquo;</SAMP> denotes an assignment
448operator, <SAMP>&lsquo;V&rsquo;</SAMP> denotes a const/volatile qualifier.
449
450</P>
451
452
453<H3><A NAME="SEC251" HREF="gettext_toc.html#TOC251">15.3.19  Qt Format Strings</A></H3>
454
455<P>
456Qt format strings are described in the documentation of the QString class
457<A HREF="file:/usr/lib/qt-3.0.5/doc/html/qstring.html">file:/usr/lib/qt-3.0.5/doc/html/qstring.html</A>.
458In summary, a directive consists of a <SAMP>&lsquo;%&rsquo;</SAMP> followed by a digit. The same
459directive cannot occur more than once in a format string.
460
461</P>
462
463
464<H3><A NAME="SEC252" HREF="gettext_toc.html#TOC252">15.3.20  Boost Format Strings</A></H3>
465
466<P>
467Boost format strings are described in the documentation of the
468<CODE>boost::format</CODE> class, at
469<A HREF="http://www.boost.org/libs/format/doc/format.html">http://www.boost.org/libs/format/doc/format.html</A>.
470In summary, a directive has either the same syntax as in a C format string,
471such as <SAMP>&lsquo;%1$+5d&rsquo;</SAMP>, or may be surrounded by vertical bars, such as
472<SAMP>&lsquo;%|1$+5d|&rsquo;</SAMP> or <SAMP>&lsquo;%|1$+5|&rsquo;</SAMP>, or consists of just an argument number
473between percent signs, such as <SAMP>&lsquo;%1%&rsquo;</SAMP>.
474
475</P>
476
477
478<H2><A NAME="SEC253" HREF="gettext_toc.html#TOC253">15.4  The Maintainer's View</A></H2>
479
480<P>
481For the maintainer, the general procedure differs from the C language
482case in two ways.
483
484</P>
485
486<UL>
487<LI>
488
489For those languages that don't use GNU gettext, the <TT>&lsquo;intl/&rsquo;</TT> directory
490is not needed and can be omitted.  This means that the maintainer calls the
491<CODE>gettextize</CODE> program without the <SAMP>&lsquo;--intl&rsquo;</SAMP> option, and that he
492invokes the <CODE>AM_GNU_GETTEXT</CODE> autoconf macro via
493<SAMP>&lsquo;AM_GNU_GETTEXT([external])&rsquo;</SAMP>.
494
495<LI>
496
497If only a single programming language is used, the <CODE>XGETTEXT_OPTIONS</CODE>
498variable in <TT>&lsquo;po/Makevars&rsquo;</TT> (see section <A HREF="gettext_13.html#SEC203">13.4.3  <TT>&lsquo;Makevars&rsquo;</TT> in <TT>&lsquo;po/&rsquo;</TT></A>) should be adjusted to
499match the <CODE>xgettext</CODE> options for that particular programming language.
500If the package uses more than one programming language with <CODE>gettext</CODE>
501support, it becomes necessary to change the POT file construction rule
502in <TT>&lsquo;po/Makefile.in.in&rsquo;</TT>.  It is recommended to make one <CODE>xgettext</CODE>
503invocation per programming language, each with the options appropriate for
504that language, and to combine the resulting files using <CODE>msgcat</CODE>.
505</UL>
506
507
508
509<H2><A NAME="SEC254" HREF="gettext_toc.html#TOC254">15.5  Individual Programming Languages</A></H2>
510
511
512
513<H3><A NAME="SEC255" HREF="gettext_toc.html#TOC255">15.5.1  C, C++, Objective C</A></H3>
514<P>
515<A NAME="IDX1115"></A>
516
517</P>
518<DL COMPACT>
519
520<DT>RPMs
521<DD>
522gcc, gpp, gobjc, glibc, gettext
523
524<DT>File extension
525<DD>
526For C: <CODE>c</CODE>, <CODE>h</CODE>.
527<BR>For C++: <CODE>C</CODE>, <CODE>c++</CODE>, <CODE>cc</CODE>, <CODE>cxx</CODE>, <CODE>cpp</CODE>, <CODE>hpp</CODE>.
528<BR>For Objective C: <CODE>m</CODE>.
529
530<DT>String syntax
531<DD>
532<CODE>"abc"</CODE>
533
534<DT>gettext shorthand
535<DD>
536<CODE>_("abc")</CODE>
537
538<DT>gettext/ngettext functions
539<DD>
540<CODE>gettext</CODE>, <CODE>dgettext</CODE>, <CODE>dcgettext</CODE>, <CODE>ngettext</CODE>,
541<CODE>dngettext</CODE>, <CODE>dcngettext</CODE>
542
543<DT>textdomain
544<DD>
545<CODE>textdomain</CODE> function
546
547<DT>bindtextdomain
548<DD>
549<CODE>bindtextdomain</CODE> function
550
551<DT>setlocale
552<DD>
553Programmer must call <CODE>setlocale (LC_ALL, "")</CODE>
554
555<DT>Prerequisite
556<DD>
557<CODE>#include &#60;libintl.h&#62;</CODE>
558<BR><CODE>#include &#60;locale.h&#62;</CODE>
559<BR><CODE>#define _(string) gettext (string)</CODE>
560
561<DT>Use or emulate GNU gettext
562<DD>
563Use
564
565<DT>Extractor
566<DD>
567<CODE>xgettext -k_</CODE>
568
569<DT>Formatting with positions
570<DD>
571<CODE>fprintf "%2$d %1$d"</CODE>
572<BR>In C++: <CODE>autosprintf "%2$d %1$d"</CODE>
573(see section ‘Introduction’ in <CITE>GNU autosprintf</CITE>)
574
575<DT>Portability
576<DD>
577autoconf (gettext.m4) and #if ENABLE_NLS
578
579<DT>po-mode marking
580<DD>
581yes
582</DL>
583
584<P>
585The following examples are available in the <TT>&lsquo;examples&rsquo;</TT> directory:
586<CODE>hello-c</CODE>, <CODE>hello-c-gnome</CODE>, <CODE>hello-c++</CODE>, <CODE>hello-c++-qt</CODE>,
587<CODE>hello-c++-kde</CODE>, <CODE>hello-c++-gnome</CODE>, <CODE>hello-c++-wxwidgets</CODE>,
588<CODE>hello-objc</CODE>, <CODE>hello-objc-gnustep</CODE>, <CODE>hello-objc-gnome</CODE>.
589
590</P>
591
592
593<H3><A NAME="SEC256" HREF="gettext_toc.html#TOC256">15.5.2  sh - Shell Script</A></H3>
594<P>
595<A NAME="IDX1116"></A>
596
597</P>
598<DL COMPACT>
599
600<DT>RPMs
601<DD>
602bash, gettext
603
604<DT>File extension
605<DD>
606<CODE>sh</CODE>
607
608<DT>String syntax
609<DD>
610<CODE>"abc"</CODE>, <CODE>'abc'</CODE>, <CODE>abc</CODE>
611
612<DT>gettext shorthand
613<DD>
614<CODE>"`gettext \"abc\"`"</CODE>
615
616<DT>gettext/ngettext functions
617<DD>
618<A NAME="IDX1117"></A>
619<A NAME="IDX1118"></A>
620<CODE>gettext</CODE>, <CODE>ngettext</CODE> programs
621<BR><CODE>eval_gettext</CODE>, <CODE>eval_ngettext</CODE> shell functions
622
623<DT>textdomain
624<DD>
625<A NAME="IDX1119"></A>
626environment variable <CODE>TEXTDOMAIN</CODE>
627
628<DT>bindtextdomain
629<DD>
630<A NAME="IDX1120"></A>
631environment variable <CODE>TEXTDOMAINDIR</CODE>
632
633<DT>setlocale
634<DD>
635automatic
636
637<DT>Prerequisite
638<DD>
639<CODE>. gettext.sh</CODE>
640
641<DT>Use or emulate GNU gettext
642<DD>
643use
644
645<DT>Extractor
646<DD>
647<CODE>xgettext</CODE>
648
649<DT>Formatting with positions
650<DD>
651---
652
653<DT>Portability
654<DD>
655fully portable
656
657<DT>po-mode marking
658<DD>
659---
660</DL>
661
662<P>
663An example is available in the <TT>&lsquo;examples&rsquo;</TT> directory: <CODE>hello-sh</CODE>.
664
665</P>
666
667
668
669<H4><A NAME="SEC257" HREF="gettext_toc.html#TOC257">15.5.2.1  Preparing Shell Scripts for Internationalization</A></H4>
670<P>
671<A NAME="IDX1121"></A>
672
673</P>
674<P>
675Preparing a shell script for internationalization is conceptually similar
676to the steps described in section <A HREF="gettext_4.html#SEC11">4  Preparing Program Sources</A>.  The concrete steps for shell
677scripts are as follows.
678
679</P>
680
681<OL>
682<LI>
683
684Insert the line
685
686
687<PRE>
688. gettext.sh
689</PRE>
690
691near the top of the script.  <CODE>gettext.sh</CODE> is a shell function library
692that provides the functions
693<CODE>eval_gettext</CODE> (see section <A HREF="gettext_15.html#SEC262">15.5.2.6  Invoking the <CODE>eval_gettext</CODE> function</A>) and
694<CODE>eval_ngettext</CODE> (see section <A HREF="gettext_15.html#SEC263">15.5.2.7  Invoking the <CODE>eval_ngettext</CODE> function</A>).
695You have to ensure that <CODE>gettext.sh</CODE> can be found in the <CODE>PATH</CODE>.
696
697<LI>
698
699Set and export the <CODE>TEXTDOMAIN</CODE> and <CODE>TEXTDOMAINDIR</CODE> environment
700variables.  Usually <CODE>TEXTDOMAIN</CODE> is the package or program name, and
701<CODE>TEXTDOMAINDIR</CODE> is the absolute pathname corresponding to
702<CODE>$prefix/share/locale</CODE>, where <CODE>$prefix</CODE> is the installation location.
703
704
705<PRE>
706TEXTDOMAIN=@PACKAGE@
707export TEXTDOMAIN
708TEXTDOMAINDIR=@LOCALEDIR@
709export TEXTDOMAINDIR
710</PRE>
711
712<LI>
713
714Prepare the strings for translation, as described in section <A HREF="gettext_4.html#SEC14">4.3  Preparing Translatable Strings</A>.
715
716<LI>
717
718Simplify translatable strings so that they don't contain command substitution
719(<CODE>"`...`"</CODE> or <CODE>"$(...)"</CODE>), variable access with defaulting (like
720<CODE>${<VAR>variable</VAR>-<VAR>default</VAR>}</CODE>), access to positional arguments
721(like <CODE>$0</CODE>, <CODE>$1</CODE>, ...) or highly volatile shell variables (like
722<CODE>$?</CODE>). This can always be done through simple local code restructuring.
723For example,
724
725
726<PRE>
727echo "Usage: $0 [OPTION] FILE..."
728</PRE>
729
730becomes
731
732
733<PRE>
734program_name=$0
735echo "Usage: $program_name [OPTION] FILE..."
736</PRE>
737
738Similarly,
739
740
741<PRE>
742echo "Remaining files: `ls | wc -l`"
743</PRE>
744
745becomes
746
747
748<PRE>
749filecount="`ls | wc -l`"
750echo "Remaining files: $filecount"
751</PRE>
752
753<LI>
754
755For each translatable string, change the output command <SAMP>&lsquo;echo&rsquo;</SAMP> or
756<SAMP>&lsquo;$echo&rsquo;</SAMP> to <SAMP>&lsquo;gettext&rsquo;</SAMP> (if the string contains no references to
757shell variables) or to <SAMP>&lsquo;eval_gettext&rsquo;</SAMP> (if it refers to shell variables),
758followed by a no-argument <SAMP>&lsquo;echo&rsquo;</SAMP> command (to account for the terminating
759newline). Similarly, for cases with plural handling, replace a conditional
760<SAMP>&lsquo;echo&rsquo;</SAMP> command with an invocation of <SAMP>&lsquo;ngettext&rsquo;</SAMP> or
761<SAMP>&lsquo;eval_ngettext&rsquo;</SAMP>, followed by a no-argument <SAMP>&lsquo;echo&rsquo;</SAMP> command.
762
763When doing this, you also need to add an extra backslash before the dollar
764sign in references to shell variables, so that the <SAMP>&lsquo;eval_gettext&rsquo;</SAMP>
765function receives the translatable string before the variable values are
766substituted into it. For example,
767
768
769<PRE>
770echo "Remaining files: $filecount"
771</PRE>
772
773becomes
774
775
776<PRE>
777eval_gettext "Remaining files: \$filecount"; echo
778</PRE>
779
780If the output command is not <SAMP>&lsquo;echo&rsquo;</SAMP>, you can make it use <SAMP>&lsquo;echo&rsquo;</SAMP>
781nevertheless, through the use of backquotes. However, note that inside
782backquotes, backslashes must be doubled to be effective (because the
783backquoting eats one level of backslashes). For example, assuming that
784<SAMP>&lsquo;error&rsquo;</SAMP> is a shell function that signals an error,
785
786
787<PRE>
788error "file not found: $filename"
789</PRE>
790
791is first transformed into
792
793
794<PRE>
795error "`echo \"file not found: \$filename\"`"
796</PRE>
797
798which then becomes
799
800
801<PRE>
802error "`eval_gettext \"file not found: \\\$filename\"`"
803</PRE>
804
805</OL>
806
807
808
809<H4><A NAME="SEC258" HREF="gettext_toc.html#TOC258">15.5.2.2  Contents of <CODE>gettext.sh</CODE></A></H4>
810
811<P>
812<CODE>gettext.sh</CODE>, contained in the run-time package of GNU gettext, provides
813the following:
814
815</P>
816
817<UL>
818<LI>$echo
819
820The variable <CODE>echo</CODE> is set to a command that outputs its first argument
821and a newline, without interpreting backslashes in the argument string.
822
823<LI>eval_gettext
824
825See section <A HREF="gettext_15.html#SEC262">15.5.2.6  Invoking the <CODE>eval_gettext</CODE> function</A>.
826
827<LI>eval_ngettext
828
829See section <A HREF="gettext_15.html#SEC263">15.5.2.7  Invoking the <CODE>eval_ngettext</CODE> function</A>.
830</UL>
831
832
833
834<H4><A NAME="SEC259" HREF="gettext_toc.html#TOC259">15.5.2.3  Invoking the <CODE>gettext</CODE> program</A></H4>
835
836<P>
837<A NAME="IDX1122"></A>
838<A NAME="IDX1123"></A>
839
840<PRE>
841gettext [<VAR>option</VAR>] [[<VAR>textdomain</VAR>] <VAR>msgid</VAR>]
842gettext [<VAR>option</VAR>] -s [<VAR>msgid</VAR>]...
843</PRE>
844
845<P>
846<A NAME="IDX1124"></A>
847The <CODE>gettext</CODE> program displays the native language translation of a
848textual message.
849
850</P>
851<P>
852<STRONG>Arguments</STRONG>
853
854</P>
855<DL COMPACT>
856
857<DT><SAMP>&lsquo;-d <VAR>textdomain</VAR>&rsquo;</SAMP>
858<DD>
859<DT><SAMP>&lsquo;--domain=<VAR>textdomain</VAR>&rsquo;</SAMP>
860<DD>
861<A NAME="IDX1125"></A>
862<A NAME="IDX1126"></A>
863Retrieve translated messages from <VAR>textdomain</VAR>.  Usually a <VAR>textdomain</VAR>
864corresponds to a package, a program, or a module of a program.
865
866<DT><SAMP>&lsquo;-e&rsquo;</SAMP>
867<DD>
868<A NAME="IDX1127"></A>
869Enable expansion of some escape sequences.  This option is for compatibility
870with the <SAMP>&lsquo;echo&rsquo;</SAMP> program or shell built-in.  The escape sequences
871<SAMP>&lsquo;\a&rsquo;</SAMP>, <SAMP>&lsquo;\b&rsquo;</SAMP>, <SAMP>&lsquo;\c&rsquo;</SAMP>, <SAMP>&lsquo;\f&rsquo;</SAMP>, <SAMP>&lsquo;\n&rsquo;</SAMP>, <SAMP>&lsquo;\r&rsquo;</SAMP>, <SAMP>&lsquo;\t&rsquo;</SAMP>,
872<SAMP>&lsquo;\v&rsquo;</SAMP>, <SAMP>&lsquo;\\&rsquo;</SAMP>, and <SAMP>&lsquo;\&rsquo;</SAMP> followed by one to three octal digits, are
873interpreted like the System V <SAMP>&lsquo;echo&rsquo;</SAMP> program did.
874
875<DT><SAMP>&lsquo;-E&rsquo;</SAMP>
876<DD>
877<A NAME="IDX1128"></A>
878This option is only for compatibility with the <SAMP>&lsquo;echo&rsquo;</SAMP> program or shell
879built-in.  It has no effect.
880
881<DT><SAMP>&lsquo;-h&rsquo;</SAMP>
882<DD>
883<DT><SAMP>&lsquo;--help&rsquo;</SAMP>
884<DD>
885<A NAME="IDX1129"></A>
886<A NAME="IDX1130"></A>
887Display this help and exit.
888
889<DT><SAMP>&lsquo;-n&rsquo;</SAMP>
890<DD>
891<A NAME="IDX1131"></A>
892Suppress trailing newline.  By default, <CODE>gettext</CODE> adds a newline to
893the output.
894
895<DT><SAMP>&lsquo;-V&rsquo;</SAMP>
896<DD>
897<DT><SAMP>&lsquo;--version&rsquo;</SAMP>
898<DD>
899<A NAME="IDX1132"></A>
900<A NAME="IDX1133"></A>
901Output version information and exit.
902
903<DT><SAMP>&lsquo;[<VAR>textdomain</VAR>] <VAR>msgid</VAR>&rsquo;</SAMP>
904<DD>
905Retrieve translated message corresponding to <VAR>msgid</VAR> from <VAR>textdomain</VAR>.
906
907</DL>
908
909<P>
910If the <VAR>textdomain</VAR> parameter is not given, the domain is determined from
911the environment variable <CODE>TEXTDOMAIN</CODE>.  If the message catalog is not
912found in the regular directory, another location can be specified with the
913environment variable <CODE>TEXTDOMAINDIR</CODE>.
914
915</P>
916<P>
917When used with the <CODE>-s</CODE> option the program behaves like the <SAMP>&lsquo;echo&rsquo;</SAMP>
918command.  But it does not simply copy its arguments to stdout.  Instead those
919messages found in the selected catalog are translated.
920
921</P>
922
923
924<H4><A NAME="SEC260" HREF="gettext_toc.html#TOC260">15.5.2.4  Invoking the <CODE>ngettext</CODE> program</A></H4>
925
926<P>
927<A NAME="IDX1134"></A>
928<A NAME="IDX1135"></A>
929
930<PRE>
931ngettext [<VAR>option</VAR>] [<VAR>textdomain</VAR>] <VAR>msgid</VAR> <VAR>msgid-plural</VAR> <VAR>count</VAR>
932</PRE>
933
934<P>
935<A NAME="IDX1136"></A>
936The <CODE>ngettext</CODE> program displays the native language translation of a
937textual message whose grammatical form depends on a number.
938
939</P>
940<P>
941<STRONG>Arguments</STRONG>
942
943</P>
944<DL COMPACT>
945
946<DT><SAMP>&lsquo;-d <VAR>textdomain</VAR>&rsquo;</SAMP>
947<DD>
948<DT><SAMP>&lsquo;--domain=<VAR>textdomain</VAR>&rsquo;</SAMP>
949<DD>
950<A NAME="IDX1137"></A>
951<A NAME="IDX1138"></A>
952Retrieve translated messages from <VAR>textdomain</VAR>.  Usually a <VAR>textdomain</VAR>
953corresponds to a package, a program, or a module of a program.
954
955<DT><SAMP>&lsquo;-e&rsquo;</SAMP>
956<DD>
957<A NAME="IDX1139"></A>
958Enable expansion of some escape sequences.  This option is for compatibility
959with the <SAMP>&lsquo;gettext&rsquo;</SAMP> program.  The escape sequences
960<SAMP>&lsquo;\a&rsquo;</SAMP>, <SAMP>&lsquo;\b&rsquo;</SAMP>, <SAMP>&lsquo;\c&rsquo;</SAMP>, <SAMP>&lsquo;\f&rsquo;</SAMP>, <SAMP>&lsquo;\n&rsquo;</SAMP>, <SAMP>&lsquo;\r&rsquo;</SAMP>, <SAMP>&lsquo;\t&rsquo;</SAMP>,
961<SAMP>&lsquo;\v&rsquo;</SAMP>, <SAMP>&lsquo;\\&rsquo;</SAMP>, and <SAMP>&lsquo;\&rsquo;</SAMP> followed by one to three octal digits, are
962interpreted like the System V <SAMP>&lsquo;echo&rsquo;</SAMP> program did.
963
964<DT><SAMP>&lsquo;-E&rsquo;</SAMP>
965<DD>
966<A NAME="IDX1140"></A>
967This option is only for compatibility with the <SAMP>&lsquo;gettext&rsquo;</SAMP> program.  It has
968no effect.
969
970<DT><SAMP>&lsquo;-h&rsquo;</SAMP>
971<DD>
972<DT><SAMP>&lsquo;--help&rsquo;</SAMP>
973<DD>
974<A NAME="IDX1141"></A>
975<A NAME="IDX1142"></A>
976Display this help and exit.
977
978<DT><SAMP>&lsquo;-V&rsquo;</SAMP>
979<DD>
980<DT><SAMP>&lsquo;--version&rsquo;</SAMP>
981<DD>
982<A NAME="IDX1143"></A>
983<A NAME="IDX1144"></A>
984Output version information and exit.
985
986<DT><SAMP>&lsquo;<VAR>textdomain</VAR>&rsquo;</SAMP>
987<DD>
988Retrieve translated message from <VAR>textdomain</VAR>.
989
990<DT><SAMP>&lsquo;<VAR>msgid</VAR> <VAR>msgid-plural</VAR>&rsquo;</SAMP>
991<DD>
992Translate <VAR>msgid</VAR> (English singular) / <VAR>msgid-plural</VAR> (English plural).
993
994<DT><SAMP>&lsquo;<VAR>count</VAR>&rsquo;</SAMP>
995<DD>
996Choose singular/plural form based on this value.
997
998</DL>
999
1000<P>
1001If the <VAR>textdomain</VAR> parameter is not given, the domain is determined from
1002the environment variable <CODE>TEXTDOMAIN</CODE>.  If the message catalog is not
1003found in the regular directory, another location can be specified with the
1004environment variable <CODE>TEXTDOMAINDIR</CODE>.
1005
1006</P>
1007
1008
1009<H4><A NAME="SEC261" HREF="gettext_toc.html#TOC261">15.5.2.5  Invoking the <CODE>envsubst</CODE> program</A></H4>
1010
1011<P>
1012<A NAME="IDX1145"></A>
1013<A NAME="IDX1146"></A>
1014
1015<PRE>
1016envsubst [<VAR>option</VAR>] [<VAR>shell-format</VAR>]
1017</PRE>
1018
1019<P>
1020<A NAME="IDX1147"></A>
1021<A NAME="IDX1148"></A>
1022<A NAME="IDX1149"></A>
1023The <CODE>envsubst</CODE> program substitutes the values of environment variables.
1024
1025</P>
1026<P>
1027<STRONG>Operation mode</STRONG>
1028
1029</P>
1030<DL COMPACT>
1031
1032<DT><SAMP>&lsquo;-v&rsquo;</SAMP>
1033<DD>
1034<DT><SAMP>&lsquo;--variables&rsquo;</SAMP>
1035<DD>
1036<A NAME="IDX1150"></A>
1037<A NAME="IDX1151"></A>
1038Output the variables occurring in <VAR>shell-format</VAR>.
1039
1040</DL>
1041
1042<P>
1043<STRONG>Informative output</STRONG>
1044
1045</P>
1046<DL COMPACT>
1047
1048<DT><SAMP>&lsquo;-h&rsquo;</SAMP>
1049<DD>
1050<DT><SAMP>&lsquo;--help&rsquo;</SAMP>
1051<DD>
1052<A NAME="IDX1152"></A>
1053<A NAME="IDX1153"></A>
1054Display this help and exit.
1055
1056<DT><SAMP>&lsquo;-V&rsquo;</SAMP>
1057<DD>
1058<DT><SAMP>&lsquo;--version&rsquo;</SAMP>
1059<DD>
1060<A NAME="IDX1154"></A>
1061<A NAME="IDX1155"></A>
1062Output version information and exit.
1063
1064</DL>
1065
1066<P>
1067In normal operation mode, standard input is copied to standard output,
1068with references to environment variables of the form <CODE>$VARIABLE</CODE> or
1069<CODE>${VARIABLE}</CODE> being replaced with the corresponding values.  If a
1070<VAR>shell-format</VAR> is given, only those environment variables that are
1071referenced in <VAR>shell-format</VAR> are substituted; otherwise all environment
1072variables references occurring in standard input are substituted.
1073
1074</P>
1075<P>
1076These substitutions are a subset of the substitutions that a shell performs
1077on unquoted and double-quoted strings.  Other kinds of substitutions done
1078by a shell, such as <CODE>${<VAR>variable</VAR>-<VAR>default</VAR>}</CODE> or
1079<CODE>$(<VAR>command-list</VAR>)</CODE> or <CODE>`<VAR>command-list</VAR>`</CODE>, are not performed
1080by the <CODE>envsubst</CODE> program, due to security reasons.
1081
1082</P>
1083<P>
1084When <CODE>--variables</CODE> is used, standard input is ignored, and the output
1085consists of the environment variables that are referenced in
1086<VAR>shell-format</VAR>, one per line.
1087
1088</P>
1089
1090
1091<H4><A NAME="SEC262" HREF="gettext_toc.html#TOC262">15.5.2.6  Invoking the <CODE>eval_gettext</CODE> function</A></H4>
1092
1093<P>
1094<A NAME="IDX1156"></A>
1095
1096<PRE>
1097eval_gettext <VAR>msgid</VAR>
1098</PRE>
1099
1100<P>
1101<A NAME="IDX1157"></A>
1102This function outputs the native language translation of a textual message,
1103performing dollar-substitution on the result.  Note that only shell variables
1104mentioned in <VAR>msgid</VAR> will be dollar-substituted in the result.
1105
1106</P>
1107
1108
1109<H4><A NAME="SEC263" HREF="gettext_toc.html#TOC263">15.5.2.7  Invoking the <CODE>eval_ngettext</CODE> function</A></H4>
1110
1111<P>
1112<A NAME="IDX1158"></A>
1113
1114<PRE>
1115eval_ngettext <VAR>msgid</VAR> <VAR>msgid-plural</VAR> <VAR>count</VAR>
1116</PRE>
1117
1118<P>
1119<A NAME="IDX1159"></A>
1120This function outputs the native language translation of a textual message
1121whose grammatical form depends on a number, performing dollar-substitution
1122on the result.  Note that only shell variables mentioned in <VAR>msgid</VAR> or
1123<VAR>msgid-plural</VAR> will be dollar-substituted in the result.
1124
1125</P>
1126
1127
1128<H3><A NAME="SEC264" HREF="gettext_toc.html#TOC264">15.5.3  bash - Bourne-Again Shell Script</A></H3>
1129<P>
1130<A NAME="IDX1160"></A>
1131
1132</P>
1133<P>
1134GNU <CODE>bash</CODE> 2.0 or newer has a special shorthand for translating a
1135string and substituting variable values in it: <CODE>$"msgid"</CODE>.  But
1136the use of this construct is <STRONG>discouraged</STRONG>, due to the security
1137holes it opens and due to its portability problems.
1138
1139</P>
1140<P>
1141The security holes of <CODE>$"..."</CODE> come from the fact that after looking up
1142the translation of the string, <CODE>bash</CODE> processes it like it processes
1143any double-quoted string: dollar and backquote processing, like <SAMP>&lsquo;eval&rsquo;</SAMP>
1144does.
1145
1146</P>
1147
1148<OL>
1149<LI>
1150
1151In a locale whose encoding is one of BIG5, BIG5-HKSCS, GBK, GB18030, SHIFT_JIS,
1152JOHAB, some double-byte characters have a second byte whose value is
1153<CODE>0x60</CODE>.  For example, the byte sequence <CODE>\xe0\x60</CODE> is a single
1154character in these locales.  Many versions of <CODE>bash</CODE> (all versions
1155up to bash-2.05, and newer versions on platforms without <CODE>mbsrtowcs()</CODE>
1156function) don't know about character boundaries and see a backquote character
1157where there is only a particular Chinese character.  Thus it can start
1158executing part of the translation as a command list.  This situation can occur
1159even without the translator being aware of it: if the translator provides
1160translations in the UTF-8 encoding, it is the <CODE>gettext()</CODE> function which
1161will, during its conversion from the translator's encoding to the user's
1162locale's encoding, produce the dangerous <CODE>\x60</CODE> bytes.
1163
1164<LI>
1165
1166A translator could - voluntarily or inadvertently - use backquotes
1167<CODE>"`...`"</CODE> or dollar-parentheses <CODE>"$(...)"</CODE> in her translations.
1168The enclosed strings would be executed as command lists by the shell.
1169</OL>
1170
1171<P>
1172The portability problem is that <CODE>bash</CODE> must be built with
1173internationalization support; this is normally not the case on systems
1174that don't have the <CODE>gettext()</CODE> function in libc.
1175
1176</P>
1177
1178
1179<H3><A NAME="SEC265" HREF="gettext_toc.html#TOC265">15.5.4  Python</A></H3>
1180<P>
1181<A NAME="IDX1161"></A>
1182
1183</P>
1184<DL COMPACT>
1185
1186<DT>RPMs
1187<DD>
1188python
1189
1190<DT>File extension
1191<DD>
1192<CODE>py</CODE>
1193
1194<DT>String syntax
1195<DD>
1196<CODE>'abc'</CODE>, <CODE>u'abc'</CODE>, <CODE>r'abc'</CODE>, <CODE>ur'abc'</CODE>,
1197<BR><CODE>"abc"</CODE>, <CODE>u"abc"</CODE>, <CODE>r"abc"</CODE>, <CODE>ur"abc"</CODE>,
1198<BR><CODE>”'abc”'</CODE>, <CODE>u”'abc”'</CODE>, <CODE>r”'abc”'</CODE>, <CODE>ur”'abc”'</CODE>,
1199<BR><CODE>"""abc"""</CODE>, <CODE>u"""abc"""</CODE>, <CODE>r"""abc"""</CODE>, <CODE>ur"""abc"""</CODE>
1200
1201<DT>gettext shorthand
1202<DD>
1203<CODE>_('abc')</CODE> etc.
1204
1205<DT>gettext/ngettext functions
1206<DD>
1207<CODE>gettext.gettext</CODE>, <CODE>gettext.dgettext</CODE>,
1208<CODE>gettext.ngettext</CODE>, <CODE>gettext.dngettext</CODE>,
1209also <CODE>ugettext</CODE>, <CODE>ungettext</CODE>
1210
1211<DT>textdomain
1212<DD>
1213<CODE>gettext.textdomain</CODE> function, or
1214<CODE>gettext.install(<VAR>domain</VAR>)</CODE> function
1215
1216<DT>bindtextdomain
1217<DD>
1218<CODE>gettext.bindtextdomain</CODE> function, or
1219<CODE>gettext.install(<VAR>domain</VAR>,<VAR>localedir</VAR>)</CODE> function
1220
1221<DT>setlocale
1222<DD>
1223not used by the gettext emulation
1224
1225<DT>Prerequisite
1226<DD>
1227<CODE>import gettext</CODE>
1228
1229<DT>Use or emulate GNU gettext
1230<DD>
1231emulate
1232
1233<DT>Extractor
1234<DD>
1235<CODE>xgettext</CODE>
1236
1237<DT>Formatting with positions
1238<DD>
1239<CODE>'...%(ident)d...' % { 'ident': value }</CODE>
1240
1241<DT>Portability
1242<DD>
1243fully portable
1244
1245<DT>po-mode marking
1246<DD>
1247---
1248</DL>
1249
1250<P>
1251An example is available in the <TT>&lsquo;examples&rsquo;</TT> directory: <CODE>hello-python</CODE>.
1252
1253</P>
1254
1255
1256<H3><A NAME="SEC266" HREF="gettext_toc.html#TOC266">15.5.5  GNU clisp - Common Lisp</A></H3>
1257<P>
1258<A NAME="IDX1162"></A>
1259<A NAME="IDX1163"></A>
1260<A NAME="IDX1164"></A>
1261
1262</P>
1263<DL COMPACT>
1264
1265<DT>RPMs
1266<DD>
1267clisp 2.28 or newer
1268
1269<DT>File extension
1270<DD>
1271<CODE>lisp</CODE>
1272
1273<DT>String syntax
1274<DD>
1275<CODE>"abc"</CODE>
1276
1277<DT>gettext shorthand
1278<DD>
1279<CODE>(_ "abc")</CODE>, <CODE>(ENGLISH "abc")</CODE>
1280
1281<DT>gettext/ngettext functions
1282<DD>
1283<CODE>i18n:gettext</CODE>, <CODE>i18n:ngettext</CODE>
1284
1285<DT>textdomain
1286<DD>
1287<CODE>i18n:textdomain</CODE>
1288
1289<DT>bindtextdomain
1290<DD>
1291<CODE>i18n:textdomaindir</CODE>
1292
1293<DT>setlocale
1294<DD>
1295automatic
1296
1297<DT>Prerequisite
1298<DD>
1299---
1300
1301<DT>Use or emulate GNU gettext
1302<DD>
1303use
1304
1305<DT>Extractor
1306<DD>
1307<CODE>xgettext -k_ -kENGLISH</CODE>
1308
1309<DT>Formatting with positions
1310<DD>
1311<CODE>format "~1@*~D ~0@*~D"</CODE>
1312
1313<DT>Portability
1314<DD>
1315On platforms without gettext, no translation.
1316
1317<DT>po-mode marking
1318<DD>
1319---
1320</DL>
1321
1322<P>
1323An example is available in the <TT>&lsquo;examples&rsquo;</TT> directory: <CODE>hello-clisp</CODE>.
1324
1325</P>
1326
1327
1328<H3><A NAME="SEC267" HREF="gettext_toc.html#TOC267">15.5.6  GNU clisp C sources</A></H3>
1329<P>
1330<A NAME="IDX1165"></A>
1331
1332</P>
1333<DL COMPACT>
1334
1335<DT>RPMs
1336<DD>
1337clisp
1338
1339<DT>File extension
1340<DD>
1341<CODE>d</CODE>
1342
1343<DT>String syntax
1344<DD>
1345<CODE>"abc"</CODE>
1346
1347<DT>gettext shorthand
1348<DD>
1349<CODE>ENGLISH ? "abc" : ""</CODE>
1350<BR><CODE>GETTEXT("abc")</CODE>
1351<BR><CODE>GETTEXTL("abc")</CODE>
1352
1353<DT>gettext/ngettext functions
1354<DD>
1355<CODE>clgettext</CODE>, <CODE>clgettextl</CODE>
1356
1357<DT>textdomain
1358<DD>
1359---
1360
1361<DT>bindtextdomain
1362<DD>
1363---
1364
1365<DT>setlocale
1366<DD>
1367automatic
1368
1369<DT>Prerequisite
1370<DD>
1371<CODE>#include "lispbibl.c"</CODE>
1372
1373<DT>Use or emulate GNU gettext
1374<DD>
1375use
1376
1377<DT>Extractor
1378<DD>
1379<CODE>clisp-xgettext</CODE>
1380
1381<DT>Formatting with positions
1382<DD>
1383<CODE>fprintf "%2$d %1$d"</CODE>
1384
1385<DT>Portability
1386<DD>
1387On platforms without gettext, no translation.
1388
1389<DT>po-mode marking
1390<DD>
1391---
1392</DL>
1393
1394
1395
1396<H3><A NAME="SEC268" HREF="gettext_toc.html#TOC268">15.5.7  Emacs Lisp</A></H3>
1397<P>
1398<A NAME="IDX1166"></A>
1399
1400</P>
1401<DL COMPACT>
1402
1403<DT>RPMs
1404<DD>
1405emacs, xemacs
1406
1407<DT>File extension
1408<DD>
1409<CODE>el</CODE>
1410
1411<DT>String syntax
1412<DD>
1413<CODE>"abc"</CODE>
1414
1415<DT>gettext shorthand
1416<DD>
1417<CODE>(_"abc")</CODE>
1418
1419<DT>gettext/ngettext functions
1420<DD>
1421<CODE>gettext</CODE>, <CODE>dgettext</CODE> (xemacs only)
1422
1423<DT>textdomain
1424<DD>
1425<CODE>domain</CODE> special form (xemacs only)
1426
1427<DT>bindtextdomain
1428<DD>
1429<CODE>bind-text-domain</CODE> function (xemacs only)
1430
1431<DT>setlocale
1432<DD>
1433automatic
1434
1435<DT>Prerequisite
1436<DD>
1437---
1438
1439<DT>Use or emulate GNU gettext
1440<DD>
1441use
1442
1443<DT>Extractor
1444<DD>
1445<CODE>xgettext</CODE>
1446
1447<DT>Formatting with positions
1448<DD>
1449<CODE>format "%2$d %1$d"</CODE>
1450
1451<DT>Portability
1452<DD>
1453Only XEmacs.  Without <CODE>I18N3</CODE> defined at build time, no translation.
1454
1455<DT>po-mode marking
1456<DD>
1457---
1458</DL>
1459
1460
1461
1462<H3><A NAME="SEC269" HREF="gettext_toc.html#TOC269">15.5.8  librep</A></H3>
1463<P>
1464<A NAME="IDX1167"></A>
1465
1466</P>
1467<DL COMPACT>
1468
1469<DT>RPMs
1470<DD>
1471librep 0.15.3 or newer
1472
1473<DT>File extension
1474<DD>
1475<CODE>jl</CODE>
1476
1477<DT>String syntax
1478<DD>
1479<CODE>"abc"</CODE>
1480
1481<DT>gettext shorthand
1482<DD>
1483<CODE>(_"abc")</CODE>
1484
1485<DT>gettext/ngettext functions
1486<DD>
1487<CODE>gettext</CODE>
1488
1489<DT>textdomain
1490<DD>
1491<CODE>textdomain</CODE> function
1492
1493<DT>bindtextdomain
1494<DD>
1495<CODE>bindtextdomain</CODE> function
1496
1497<DT>setlocale
1498<DD>
1499---
1500
1501<DT>Prerequisite
1502<DD>
1503<CODE>(require 'rep.i18n.gettext)</CODE>
1504
1505<DT>Use or emulate GNU gettext
1506<DD>
1507use
1508
1509<DT>Extractor
1510<DD>
1511<CODE>xgettext</CODE>
1512
1513<DT>Formatting with positions
1514<DD>
1515<CODE>format "%2$d %1$d"</CODE>
1516
1517<DT>Portability
1518<DD>
1519On platforms without gettext, no translation.
1520
1521<DT>po-mode marking
1522<DD>
1523---
1524</DL>
1525
1526<P>
1527An example is available in the <TT>&lsquo;examples&rsquo;</TT> directory: <CODE>hello-librep</CODE>.
1528
1529</P>
1530
1531
1532<H3><A NAME="SEC270" HREF="gettext_toc.html#TOC270">15.5.9  GNU guile - Scheme</A></H3>
1533<P>
1534<A NAME="IDX1168"></A>
1535<A NAME="IDX1169"></A>
1536
1537</P>
1538<DL COMPACT>
1539
1540<DT>RPMs
1541<DD>
1542guile
1543
1544<DT>File extension
1545<DD>
1546<CODE>scm</CODE>
1547
1548<DT>String syntax
1549<DD>
1550<CODE>"abc"</CODE>
1551
1552<DT>gettext shorthand
1553<DD>
1554<CODE>(_ "abc")</CODE>
1555
1556<DT>gettext/ngettext functions
1557<DD>
1558<CODE>gettext</CODE>, <CODE>ngettext</CODE>
1559
1560<DT>textdomain
1561<DD>
1562<CODE>textdomain</CODE>
1563
1564<DT>bindtextdomain
1565<DD>
1566<CODE>bindtextdomain</CODE>
1567
1568<DT>setlocale
1569<DD>
1570<CODE>(catch #t (lambda () (setlocale LC_ALL "")) (lambda args #f))</CODE>
1571
1572<DT>Prerequisite
1573<DD>
1574<CODE>(use-modules (ice-9 format))</CODE>
1575
1576<DT>Use or emulate GNU gettext
1577<DD>
1578use
1579
1580<DT>Extractor
1581<DD>
1582<CODE>xgettext -k_</CODE>
1583
1584<DT>Formatting with positions
1585<DD>
1586---
1587
1588<DT>Portability
1589<DD>
1590On platforms without gettext, no translation.
1591
1592<DT>po-mode marking
1593<DD>
1594---
1595</DL>
1596
1597<P>
1598An example is available in the <TT>&lsquo;examples&rsquo;</TT> directory: <CODE>hello-guile</CODE>.
1599
1600</P>
1601
1602
1603<H3><A NAME="SEC271" HREF="gettext_toc.html#TOC271">15.5.10  GNU Smalltalk</A></H3>
1604<P>
1605<A NAME="IDX1170"></A>
1606
1607</P>
1608<DL COMPACT>
1609
1610<DT>RPMs
1611<DD>
1612smalltalk
1613
1614<DT>File extension
1615<DD>
1616<CODE>st</CODE>
1617
1618<DT>String syntax
1619<DD>
1620<CODE>'abc'</CODE>
1621
1622<DT>gettext shorthand
1623<DD>
1624<CODE>NLS ? 'abc'</CODE>
1625
1626<DT>gettext/ngettext functions
1627<DD>
1628<CODE>LcMessagesDomain&#62;&#62;#at:</CODE>, <CODE>LcMessagesDomain&#62;&#62;#at:plural:with:</CODE>
1629
1630<DT>textdomain
1631<DD>
1632<CODE>LcMessages&#62;&#62;#domain:localeDirectory:</CODE> (returns a <CODE>LcMessagesDomain</CODE>
1633object).<BR>
1634Example: <CODE>I18N Locale default messages domain: 'gettext' localeDirectory: /usr/local/share/locale'</CODE>
1635
1636<DT>bindtextdomain
1637<DD>
1638<CODE>LcMessages&#62;&#62;#domain:localeDirectory:</CODE>, see above.
1639
1640<DT>setlocale
1641<DD>
1642Automatic if you use <CODE>I18N Locale default</CODE>.
1643
1644<DT>Prerequisite
1645<DD>
1646<CODE>PackageLoader fileInPackage: 'I18N'!</CODE>
1647
1648<DT>Use or emulate GNU gettext
1649<DD>
1650emulate
1651
1652<DT>Extractor
1653<DD>
1654<CODE>xgettext</CODE>
1655
1656<DT>Formatting with positions
1657<DD>
1658<CODE>'%1 %2' bindWith: 'Hello' with: 'world'</CODE>
1659
1660<DT>Portability
1661<DD>
1662fully portable
1663
1664<DT>po-mode marking
1665<DD>
1666---
1667</DL>
1668
1669<P>
1670An example is available in the <TT>&lsquo;examples&rsquo;</TT> directory:
1671<CODE>hello-smalltalk</CODE>.
1672
1673</P>
1674
1675
1676<H3><A NAME="SEC272" HREF="gettext_toc.html#TOC272">15.5.11  Java</A></H3>
1677<P>
1678<A NAME="IDX1171"></A>
1679
1680</P>
1681<DL COMPACT>
1682
1683<DT>RPMs
1684<DD>
1685java, java2
1686
1687<DT>File extension
1688<DD>
1689<CODE>java</CODE>
1690
1691<DT>String syntax
1692<DD>
1693"abc"
1694
1695<DT>gettext shorthand
1696<DD>
1697_("abc")
1698
1699<DT>gettext/ngettext functions
1700<DD>
1701<CODE>GettextResource.gettext</CODE>, <CODE>GettextResource.ngettext</CODE>
1702
1703<DT>textdomain
1704<DD>
1705---, use <CODE>ResourceBundle.getResource</CODE> instead
1706
1707<DT>bindtextdomain
1708<DD>
1709---, use CLASSPATH instead
1710
1711<DT>setlocale
1712<DD>
1713automatic
1714
1715<DT>Prerequisite
1716<DD>
1717---
1718
1719<DT>Use or emulate GNU gettext
1720<DD>
1721---, uses a Java specific message catalog format
1722
1723<DT>Extractor
1724<DD>
1725<CODE>xgettext -k_</CODE>
1726
1727<DT>Formatting with positions
1728<DD>
1729<CODE>MessageFormat.format "{1,number} {0,number}"</CODE>
1730
1731<DT>Portability
1732<DD>
1733fully portable
1734
1735<DT>po-mode marking
1736<DD>
1737---
1738</DL>
1739
1740<P>
1741Before marking strings as internationalizable, uses of the string
1742concatenation operator need to be converted to <CODE>MessageFormat</CODE>
1743applications.  For example, <CODE>"file "+filename+" not found"</CODE> becomes
1744<CODE>MessageFormat.format("file {0} not found", new Object[] { filename })</CODE>.
1745Only after this is done, can the strings be marked and extracted.
1746
1747</P>
1748<P>
1749GNU gettext uses the native Java internationalization mechanism, namely
1750<CODE>ResourceBundle</CODE>s.  There are two formats of <CODE>ResourceBundle</CODE>s:
1751<CODE>.properties</CODE> files and <CODE>.class</CODE> files.  The <CODE>.properties</CODE>
1752format is a text file which the translators can directly edit, like PO
1753files, but which doesn't support plural forms.  Whereas the <CODE>.class</CODE>
1754format is compiled from <CODE>.java</CODE> source code and can support plural
1755forms (provided it is accessed through an appropriate API, see below).
1756
1757</P>
1758<P>
1759To convert a PO file to a <CODE>.properties</CODE> file, the <CODE>msgcat</CODE>
1760program can be used with the option <CODE>--properties-output</CODE>.  To convert
1761a <CODE>.properties</CODE> file back to a PO file, the <CODE>msgcat</CODE> program
1762can be used with the option <CODE>--properties-input</CODE>.  All the tools
1763that manipulate PO files can work with <CODE>.properties</CODE> files as well,
1764if given the <CODE>--properties-input</CODE> and/or <CODE>--properties-output</CODE>
1765option.
1766
1767</P>
1768<P>
1769To convert a PO file to a ResourceBundle class, the <CODE>msgfmt</CODE> program
1770can be used with the option <CODE>--java</CODE> or <CODE>--java2</CODE>.  To convert a
1771ResourceBundle back to a PO file, the <CODE>msgunfmt</CODE> program can be used
1772with the option <CODE>--java</CODE>.
1773
1774</P>
1775<P>
1776Two different programmatic APIs can be used to access ResourceBundles.
1777Note that both APIs work with all kinds of ResourceBundles, whether
1778GNU gettext generated classes, or other <CODE>.class</CODE> or <CODE>.properties</CODE>
1779files.
1780
1781</P>
1782
1783<OL>
1784<LI>
1785
1786The <CODE>java.util.ResourceBundle</CODE> API.
1787
1788In particular, its <CODE>getString</CODE> function returns a string translation.
1789Note that a missing translation yields a <CODE>MissingResourceException</CODE>.
1790
1791This has the advantage of being the standard API.  And it does not require
1792any additional libraries, only the <CODE>msgcat</CODE> generated <CODE>.properties</CODE>
1793files or the <CODE>msgfmt</CODE> generated <CODE>.class</CODE> files.  But it cannot do
1794plural handling, even if the resource was generated by <CODE>msgfmt</CODE> from
1795a PO file with plural handling.
1796
1797<LI>
1798
1799The <CODE>gnu.gettext.GettextResource</CODE> API.
1800
1801Reference documentation in Javadoc 1.1 style format
1802is in the <A HREF="javadoc1/tree.html">javadoc1 directory</A> and
1803in Javadoc 2 style format
1804in the <A HREF="javadoc2/index.html">javadoc2 directory</A>.
1805
1806Its <CODE>gettext</CODE> function returns a string translation.  Note that when
1807a translation is missing, the <VAR>msgid</VAR> argument is returned unchanged.
1808
1809This has the advantage of having the <CODE>ngettext</CODE> function for plural
1810handling.
1811
1812<A NAME="IDX1172"></A>
1813To use this API, one needs the <CODE>libintl.jar</CODE> file which is part of
1814the GNU gettext package and distributed under the LGPL.
1815</OL>
1816
1817<P>
1818Three examples, using the second API, are available in the <TT>&lsquo;examples&rsquo;</TT>
1819directory: <CODE>hello-java</CODE>, <CODE>hello-java-awt</CODE>, <CODE>hello-java-swing</CODE>.
1820
1821</P>
1822<P>
1823Now, to make use of the API and define a shorthand for <SAMP>&lsquo;getString&rsquo;</SAMP>,
1824there are three idioms that you can choose from:
1825
1826</P>
1827
1828<UL>
1829<LI>
1830
1831(This one assumes Java 1.5 or newer.)
1832In a unique class of your project, say <SAMP>&lsquo;Util&rsquo;</SAMP>, define a static variable
1833holding the <CODE>ResourceBundle</CODE> instance and the shorthand:
1834
1835
1836<PRE>
1837private static ResourceBundle myResources =
1838  ResourceBundle.getBundle("domain-name");
1839public static String _(String s) {
1840  return myResources.getString(s);
1841}
1842</PRE>
1843
1844All classes containing internationalized strings then contain
1845
1846
1847<PRE>
1848import static Util._;
1849</PRE>
1850
1851and the shorthand is used like this:
1852
1853
1854<PRE>
1855System.out.println(_("Operation completed."));
1856</PRE>
1857
1858<LI>
1859
1860In a unique class of your project, say <SAMP>&lsquo;Util&rsquo;</SAMP>, define a static variable
1861holding the <CODE>ResourceBundle</CODE> instance:
1862
1863
1864<PRE>
1865public static ResourceBundle myResources =
1866  ResourceBundle.getBundle("domain-name");
1867</PRE>
1868
1869All classes containing internationalized strings then contain
1870
1871
1872<PRE>
1873private static ResourceBundle res = Util.myResources;
1874private static String _(String s) { return res.getString(s); }
1875</PRE>
1876
1877and the shorthand is used like this:
1878
1879
1880<PRE>
1881System.out.println(_("Operation completed."));
1882</PRE>
1883
1884<LI>
1885
1886You add a class with a very short name, say <SAMP>&lsquo;S&rsquo;</SAMP>, containing just the
1887definition of the resource bundle and of the shorthand:
1888
1889
1890<PRE>
1891public class S {
1892  public static ResourceBundle myResources =
1893    ResourceBundle.getBundle("domain-name");
1894  public static String _(String s) {
1895    return myResources.getString(s);
1896  }
1897}
1898</PRE>
1899
1900and the shorthand is used like this:
1901
1902
1903<PRE>
1904System.out.println(S._("Operation completed."));
1905</PRE>
1906
1907</UL>
1908
1909<P>
1910Which of the three idioms you choose, will depend on whether your project
1911requires portability to Java versions prior to Java 1.5 and, if so, whether
1912copying two lines of codes into every class is more acceptable in your project
1913than a class with a single-letter name.
1914
1915</P>
1916
1917
1918<H3><A NAME="SEC273" HREF="gettext_toc.html#TOC273">15.5.12  C#</A></H3>
1919<P>
1920<A NAME="IDX1173"></A>
1921
1922</P>
1923<DL COMPACT>
1924
1925<DT>RPMs
1926<DD>
1927pnet, pnetlib 0.6.2 or newer, or mono 0.29 or newer
1928
1929<DT>File extension
1930<DD>
1931<CODE>cs</CODE>
1932
1933<DT>String syntax
1934<DD>
1935<CODE>"abc"</CODE>, <CODE>@"abc"</CODE>
1936
1937<DT>gettext shorthand
1938<DD>
1939_("abc")
1940
1941<DT>gettext/ngettext functions
1942<DD>
1943<CODE>GettextResourceManager.GetString</CODE>,
1944<CODE>GettextResourceManager.GetPluralString</CODE>
1945
1946<DT>textdomain
1947<DD>
1948<CODE>new GettextResourceManager(domain)</CODE>
1949
1950<DT>bindtextdomain
1951<DD>
1952---, compiled message catalogs are located in subdirectories of the directory
1953containing the executable
1954
1955<DT>setlocale
1956<DD>
1957automatic
1958
1959<DT>Prerequisite
1960<DD>
1961---
1962
1963<DT>Use or emulate GNU gettext
1964<DD>
1965---, uses a C# specific message catalog format
1966
1967<DT>Extractor
1968<DD>
1969<CODE>xgettext -k_</CODE>
1970
1971<DT>Formatting with positions
1972<DD>
1973<CODE>String.Format "{1} {0}"</CODE>
1974
1975<DT>Portability
1976<DD>
1977fully portable
1978
1979<DT>po-mode marking
1980<DD>
1981---
1982</DL>
1983
1984<P>
1985Before marking strings as internationalizable, uses of the string
1986concatenation operator need to be converted to <CODE>String.Format</CODE>
1987invocations.  For example, <CODE>"file "+filename+" not found"</CODE> becomes
1988<CODE>String.Format("file {0} not found", filename)</CODE>.
1989Only after this is done, can the strings be marked and extracted.
1990
1991</P>
1992<P>
1993GNU gettext uses the native C#/.NET internationalization mechanism, namely
1994the classes <CODE>ResourceManager</CODE> and <CODE>ResourceSet</CODE>.  Applications
1995use the <CODE>ResourceManager</CODE> methods to retrieve the native language
1996translation of strings.  An instance of <CODE>ResourceSet</CODE> is the in-memory
1997representation of a message catalog file.  The <CODE>ResourceManager</CODE> loads
1998and accesses <CODE>ResourceSet</CODE> instances as needed to look up the
1999translations.
2000
2001</P>
2002<P>
2003There are two formats of <CODE>ResourceSet</CODE>s that can be directly loaded by
2004the C# runtime: <CODE>.resources</CODE> files and <CODE>.dll</CODE> files.
2005
2006</P>
2007
2008<UL>
2009<LI>
2010
2011The <CODE>.resources</CODE> format is a binary file usually generated through the
2012<CODE>resgen</CODE> or <CODE>monoresgen</CODE> utility, but which doesn't support plural
2013forms.  <CODE>.resources</CODE> files can also be embedded in .NET <CODE>.exe</CODE> files.
2014This only affects whether a file system access is performed to load the message
2015catalog; it doesn't affect the contents of the message catalog.
2016
2017<LI>
2018
2019On the other hand, the <CODE>.dll</CODE> format is a binary file that is compiled
2020from <CODE>.cs</CODE> source code and can support plural forms (provided it is
2021accessed through the GNU gettext API, see below).
2022</UL>
2023
2024<P>
2025Note that these .NET <CODE>.dll</CODE> and <CODE>.exe</CODE> files are not tied to a
2026particular platform; their file format and GNU gettext for C# can be used
2027on any platform.
2028
2029</P>
2030<P>
2031To convert a PO file to a <CODE>.resources</CODE> file, the <CODE>msgfmt</CODE> program
2032can be used with the option <SAMP>&lsquo;--csharp-resources&rsquo;</SAMP>.  To convert a
2033<CODE>.resources</CODE> file back to a PO file, the <CODE>msgunfmt</CODE> program can be
2034used with the option <SAMP>&lsquo;--csharp-resources&rsquo;</SAMP>.  You can also, in some cases,
2035use the <CODE>resgen</CODE> program (from the <CODE>pnet</CODE> package) or the
2036<CODE>monoresgen</CODE> program (from the <CODE>mono</CODE>/<CODE>mcs</CODE> package).  These
2037programs can also convert a <CODE>.resources</CODE> file back to a PO file.  But
2038beware: as of this writing (January 2004), the <CODE>monoresgen</CODE> converter is
2039quite buggy and the <CODE>resgen</CODE> converter ignores the encoding of the PO
2040files.
2041
2042</P>
2043<P>
2044To convert a PO file to a <CODE>.dll</CODE> file, the <CODE>msgfmt</CODE> program can be
2045used with the option <CODE>--csharp</CODE>.  The result will be a <CODE>.dll</CODE> file
2046containing a subclass of <CODE>GettextResourceSet</CODE>, which itself is a subclass
2047of <CODE>ResourceSet</CODE>.  To convert a <CODE>.dll</CODE> file containing a
2048<CODE>GettextResourceSet</CODE> subclass back to a PO file, the <CODE>msgunfmt</CODE>
2049program can be used with the option <CODE>--csharp</CODE>.
2050
2051</P>
2052<P>
2053The advantages of the <CODE>.dll</CODE> format over the <CODE>.resources</CODE> format
2054are:
2055
2056</P>
2057
2058<OL>
2059<LI>
2060
2061Freedom to localize: Users can add their own translations to an application
2062after it has been built and distributed.  Whereas when the programmer uses
2063a <CODE>ResourceManager</CODE> constructor provided by the system, the set of
2064<CODE>.resources</CODE> files for an application must be specified when the
2065application is built and cannot be extended afterwards.
2066
2067<LI>
2068
2069Plural handling: A message catalog in <CODE>.dll</CODE> format supports the plural
2070handling function <CODE>GetPluralString</CODE>.  Whereas <CODE>.resources</CODE> files can
2071only contain data and only support lookups that depend on a single string.
2072
2073<LI>
2074
2075The <CODE>GettextResourceManager</CODE> that loads the message catalogs in
2076<CODE>.dll</CODE> format also provides for inheritance on a per-message basis.
2077For example, in Austrian (<CODE>de_AT</CODE>) locale, translations from the German
2078(<CODE>de</CODE>) message catalog will be used for messages not found in the
2079Austrian message catalog.  This has the consequence that the Austrian
2080translators need only translate those few messages for which the translation
2081into Austrian differs from the German one.  Whereas when working with
2082<CODE>.resources</CODE> files, each message catalog must provide the translations
2083of all messages by itself.
2084
2085<LI>
2086
2087The <CODE>GettextResourceManager</CODE> that loads the message catalogs in
2088<CODE>.dll</CODE> format also provides for a fallback: The English <VAR>msgid</VAR> is
2089returned when no translation can be found.  Whereas when working with
2090<CODE>.resources</CODE> files, a language-neutral <CODE>.resources</CODE> file must
2091explicitly be provided as a fallback.
2092</OL>
2093
2094<P>
2095On the side of the programmatic APIs, the programmer can use either the
2096standard <CODE>ResourceManager</CODE> API and the GNU <CODE>GettextResourceManager</CODE>
2097API.  The latter is an extension of the former, because
2098<CODE>GettextResourceManager</CODE> is a subclass of <CODE>ResourceManager</CODE>.
2099
2100</P>
2101
2102<OL>
2103<LI>
2104
2105The <CODE>System.Resources.ResourceManager</CODE> API.
2106
2107This API works with resources in <CODE>.resources</CODE> format.
2108
2109The creation of the <CODE>ResourceManager</CODE> is done through
2110
2111<PRE>
2112  new ResourceManager(domainname, Assembly.GetExecutingAssembly())
2113</PRE>
2114
2115
2116The <CODE>GetString</CODE> function returns a string's translation.  Note that this
2117function returns null when a translation is missing (i.e. not even found in
2118the fallback resource file).
2119
2120<LI>
2121
2122The <CODE>GNU.Gettext.GettextResourceManager</CODE> API.
2123
2124This API works with resources in <CODE>.dll</CODE> format.
2125
2126Reference documentation is in the
2127<A HREF="csharpdoc/index.html">csharpdoc directory</A>.
2128
2129The creation of the <CODE>ResourceManager</CODE> is done through
2130
2131<PRE>
2132  new GettextResourceManager(domainname)
2133</PRE>
2134
2135The <CODE>GetString</CODE> function returns a string's translation.  Note that when
2136a translation is missing, the <VAR>msgid</VAR> argument is returned unchanged.
2137
2138The <CODE>GetPluralString</CODE> function returns a string translation with plural
2139handling, like the <CODE>ngettext</CODE> function in C.
2140
2141<A NAME="IDX1174"></A>
2142To use this API, one needs the <CODE>GNU.Gettext.dll</CODE> file which is part of
2143the GNU gettext package and distributed under the LGPL.
2144</OL>
2145
2146<P>
2147You can also mix both approaches: use the
2148<CODE>GNU.Gettext.GettextResourceManager</CODE> constructor, but otherwise use
2149only the <CODE>ResourceManager</CODE> type and only the <CODE>GetString</CODE> method.
2150This is appropriate when you want to profit from the tools for PO files,
2151but don't want to change an existing source code that uses
2152<CODE>ResourceManager</CODE> and don't (yet) need the <CODE>GetPluralString</CODE> method.
2153
2154</P>
2155<P>
2156Two examples, using the second API, are available in the <TT>&lsquo;examples&rsquo;</TT>
2157directory: <CODE>hello-csharp</CODE>, <CODE>hello-csharp-forms</CODE>.
2158
2159</P>
2160<P>
2161Now, to make use of the API and define a shorthand for <SAMP>&lsquo;GetString&rsquo;</SAMP>,
2162there are two idioms that you can choose from:
2163
2164</P>
2165
2166<UL>
2167<LI>
2168
2169In a unique class of your project, say <SAMP>&lsquo;Util&rsquo;</SAMP>, define a static variable
2170holding the <CODE>ResourceManager</CODE> instance:
2171
2172
2173<PRE>
2174public static GettextResourceManager MyResourceManager =
2175  new GettextResourceManager("domain-name");
2176</PRE>
2177
2178All classes containing internationalized strings then contain
2179
2180
2181<PRE>
2182private static GettextResourceManager Res = Util.MyResourceManager;
2183private static String _(String s) { return Res.GetString(s); }
2184</PRE>
2185
2186and the shorthand is used like this:
2187
2188
2189<PRE>
2190Console.WriteLine(_("Operation completed."));
2191</PRE>
2192
2193<LI>
2194
2195You add a class with a very short name, say <SAMP>&lsquo;S&rsquo;</SAMP>, containing just the
2196definition of the resource manager and of the shorthand:
2197
2198
2199<PRE>
2200public class S {
2201  public static GettextResourceManager MyResourceManager =
2202    new GettextResourceManager("domain-name");
2203  public static String _(String s) {
2204     return MyResourceManager.GetString(s);
2205  }
2206}
2207</PRE>
2208
2209and the shorthand is used like this:
2210
2211
2212<PRE>
2213Console.WriteLine(S._("Operation completed."));
2214</PRE>
2215
2216</UL>
2217
2218<P>
2219Which of the two idioms you choose, will depend on whether copying two lines
2220of codes into every class is more acceptable in your project than a class
2221with a single-letter name.
2222
2223</P>
2224
2225
2226<H3><A NAME="SEC274" HREF="gettext_toc.html#TOC274">15.5.13  GNU awk</A></H3>
2227<P>
2228<A NAME="IDX1175"></A>
2229<A NAME="IDX1176"></A>
2230
2231</P>
2232<DL COMPACT>
2233
2234<DT>RPMs
2235<DD>
2236gawk 3.1 or newer
2237
2238<DT>File extension
2239<DD>
2240<CODE>awk</CODE>
2241
2242<DT>String syntax
2243<DD>
2244<CODE>"abc"</CODE>
2245
2246<DT>gettext shorthand
2247<DD>
2248<CODE>_"abc"</CODE>
2249
2250<DT>gettext/ngettext functions
2251<DD>
2252<CODE>dcgettext</CODE>, missing <CODE>dcngettext</CODE> in gawk-3.1.0
2253
2254<DT>textdomain
2255<DD>
2256<CODE>TEXTDOMAIN</CODE> variable
2257
2258<DT>bindtextdomain
2259<DD>
2260<CODE>bindtextdomain</CODE> function
2261
2262<DT>setlocale
2263<DD>
2264automatic, but missing <CODE>setlocale (LC_MESSAGES, "")</CODE> in gawk-3.1.0
2265
2266<DT>Prerequisite
2267<DD>
2268---
2269
2270<DT>Use or emulate GNU gettext
2271<DD>
2272use
2273
2274<DT>Extractor
2275<DD>
2276<CODE>xgettext</CODE>
2277
2278<DT>Formatting with positions
2279<DD>
2280<CODE>printf "%2$d %1$d"</CODE> (GNU awk only)
2281
2282<DT>Portability
2283<DD>
2284On platforms without gettext, no translation.  On non-GNU awks, you must
2285define <CODE>dcgettext</CODE>, <CODE>dcngettext</CODE> and <CODE>bindtextdomain</CODE>
2286yourself.
2287
2288<DT>po-mode marking
2289<DD>
2290---
2291</DL>
2292
2293<P>
2294An example is available in the <TT>&lsquo;examples&rsquo;</TT> directory: <CODE>hello-gawk</CODE>.
2295
2296</P>
2297
2298
2299<H3><A NAME="SEC275" HREF="gettext_toc.html#TOC275">15.5.14  Pascal - Free Pascal Compiler</A></H3>
2300<P>
2301<A NAME="IDX1177"></A>
2302<A NAME="IDX1178"></A>
2303<A NAME="IDX1179"></A>
2304
2305</P>
2306<DL COMPACT>
2307
2308<DT>RPMs
2309<DD>
2310fpk
2311
2312<DT>File extension
2313<DD>
2314<CODE>pp</CODE>, <CODE>pas</CODE>
2315
2316<DT>String syntax
2317<DD>
2318<CODE>'abc'</CODE>
2319
2320<DT>gettext shorthand
2321<DD>
2322automatic
2323
2324<DT>gettext/ngettext functions
2325<DD>
2326---, use <CODE>ResourceString</CODE> data type instead
2327
2328<DT>textdomain
2329<DD>
2330---, use <CODE>TranslateResourceStrings</CODE> function instead
2331
2332<DT>bindtextdomain
2333<DD>
2334---, use <CODE>TranslateResourceStrings</CODE> function instead
2335
2336<DT>setlocale
2337<DD>
2338automatic, but uses only LANG, not LC_MESSAGES or LC_ALL
2339
2340<DT>Prerequisite
2341<DD>
2342<CODE>{$mode delphi}</CODE> or <CODE>{$mode objfpc}</CODE><BR><CODE>uses gettext;</CODE>
2343
2344<DT>Use or emulate GNU gettext
2345<DD>
2346emulate partially
2347
2348<DT>Extractor
2349<DD>
2350<CODE>ppc386</CODE> followed by <CODE>xgettext</CODE> or <CODE>rstconv</CODE>
2351
2352<DT>Formatting with positions
2353<DD>
2354<CODE>uses sysutils;</CODE><BR><CODE>format "%1:d %0:d"</CODE>
2355
2356<DT>Portability
2357<DD>
2358?
2359
2360<DT>po-mode marking
2361<DD>
2362---
2363</DL>
2364
2365<P>
2366The Pascal compiler has special support for the <CODE>ResourceString</CODE> data
2367type.  It generates a <CODE>.rst</CODE> file.  This is then converted to a
2368<CODE>.pot</CODE> file by use of <CODE>xgettext</CODE> or <CODE>rstconv</CODE>.  At runtime,
2369a <CODE>.mo</CODE> file corresponding to translations of this <CODE>.pot</CODE> file
2370can be loaded using the <CODE>TranslateResourceStrings</CODE> function in the
2371<CODE>gettext</CODE> unit.
2372
2373</P>
2374<P>
2375An example is available in the <TT>&lsquo;examples&rsquo;</TT> directory: <CODE>hello-pascal</CODE>.
2376
2377</P>
2378
2379
2380<H3><A NAME="SEC276" HREF="gettext_toc.html#TOC276">15.5.15  wxWidgets library</A></H3>
2381<P>
2382<A NAME="IDX1180"></A>
2383
2384</P>
2385<DL COMPACT>
2386
2387<DT>RPMs
2388<DD>
2389wxGTK, gettext
2390
2391<DT>File extension
2392<DD>
2393<CODE>cpp</CODE>
2394
2395<DT>String syntax
2396<DD>
2397<CODE>"abc"</CODE>
2398
2399<DT>gettext shorthand
2400<DD>
2401<CODE>_("abc")</CODE>
2402
2403<DT>gettext/ngettext functions
2404<DD>
2405<CODE>wxLocale::GetString</CODE>, <CODE>wxGetTranslation</CODE>
2406
2407<DT>textdomain
2408<DD>
2409<CODE>wxLocale::AddCatalog</CODE>
2410
2411<DT>bindtextdomain
2412<DD>
2413<CODE>wxLocale::AddCatalogLookupPathPrefix</CODE>
2414
2415<DT>setlocale
2416<DD>
2417<CODE>wxLocale::Init</CODE>, <CODE>wxSetLocale</CODE>
2418
2419<DT>Prerequisite
2420<DD>
2421<CODE>#include &#60;wx/intl.h&#62;</CODE>
2422
2423<DT>Use or emulate GNU gettext
2424<DD>
2425emulate, see <CODE>include/wx/intl.h</CODE> and <CODE>src/common/intl.cpp</CODE>
2426
2427<DT>Extractor
2428<DD>
2429<CODE>xgettext</CODE>
2430
2431<DT>Formatting with positions
2432<DD>
2433wxString::Format supports positions if and only if the system has
2434<CODE>wprintf()</CODE>, <CODE>vswprintf()</CODE> functions and they support positions
2435according to POSIX.
2436
2437<DT>Portability
2438<DD>
2439fully portable
2440
2441<DT>po-mode marking
2442<DD>
2443yes
2444</DL>
2445
2446
2447
2448<H3><A NAME="SEC277" HREF="gettext_toc.html#TOC277">15.5.16  YCP - YaST2 scripting language</A></H3>
2449<P>
2450<A NAME="IDX1181"></A>
2451<A NAME="IDX1182"></A>
2452
2453</P>
2454<DL COMPACT>
2455
2456<DT>RPMs
2457<DD>
2458libycp, libycp-devel, yast2-core, yast2-core-devel
2459
2460<DT>File extension
2461<DD>
2462<CODE>ycp</CODE>
2463
2464<DT>String syntax
2465<DD>
2466<CODE>"abc"</CODE>
2467
2468<DT>gettext shorthand
2469<DD>
2470<CODE>_("abc")</CODE>
2471
2472<DT>gettext/ngettext functions
2473<DD>
2474<CODE>_()</CODE> with 1 or 3 arguments
2475
2476<DT>textdomain
2477<DD>
2478<CODE>textdomain</CODE> statement
2479
2480<DT>bindtextdomain
2481<DD>
2482---
2483
2484<DT>setlocale
2485<DD>
2486---
2487
2488<DT>Prerequisite
2489<DD>
2490---
2491
2492<DT>Use or emulate GNU gettext
2493<DD>
2494use
2495
2496<DT>Extractor
2497<DD>
2498<CODE>xgettext</CODE>
2499
2500<DT>Formatting with positions
2501<DD>
2502<CODE>sformat "%2 %1"</CODE>
2503
2504<DT>Portability
2505<DD>
2506fully portable
2507
2508<DT>po-mode marking
2509<DD>
2510---
2511</DL>
2512
2513<P>
2514An example is available in the <TT>&lsquo;examples&rsquo;</TT> directory: <CODE>hello-ycp</CODE>.
2515
2516</P>
2517
2518
2519<H3><A NAME="SEC278" HREF="gettext_toc.html#TOC278">15.5.17  Tcl - Tk's scripting language</A></H3>
2520<P>
2521<A NAME="IDX1183"></A>
2522<A NAME="IDX1184"></A>
2523
2524</P>
2525<DL COMPACT>
2526
2527<DT>RPMs
2528<DD>
2529tcl
2530
2531<DT>File extension
2532<DD>
2533<CODE>tcl</CODE>
2534
2535<DT>String syntax
2536<DD>
2537<CODE>"abc"</CODE>
2538
2539<DT>gettext shorthand
2540<DD>
2541<CODE>[_ "abc"]</CODE>
2542
2543<DT>gettext/ngettext functions
2544<DD>
2545<CODE>::msgcat::mc</CODE>
2546
2547<DT>textdomain
2548<DD>
2549---
2550
2551<DT>bindtextdomain
2552<DD>
2553---, use <CODE>::msgcat::mcload</CODE> instead
2554
2555<DT>setlocale
2556<DD>
2557automatic, uses LANG, but ignores LC_MESSAGES and LC_ALL
2558
2559<DT>Prerequisite
2560<DD>
2561<CODE>package require msgcat</CODE>
2562<BR><CODE>proc _ {s} {return [::msgcat::mc $s]}</CODE>
2563
2564<DT>Use or emulate GNU gettext
2565<DD>
2566---, uses a Tcl specific message catalog format
2567
2568<DT>Extractor
2569<DD>
2570<CODE>xgettext -k_</CODE>
2571
2572<DT>Formatting with positions
2573<DD>
2574<CODE>format "%2\$d %1\$d"</CODE>
2575
2576<DT>Portability
2577<DD>
2578fully portable
2579
2580<DT>po-mode marking
2581<DD>
2582---
2583</DL>
2584
2585<P>
2586Two examples are available in the <TT>&lsquo;examples&rsquo;</TT> directory:
2587<CODE>hello-tcl</CODE>, <CODE>hello-tcl-tk</CODE>.
2588
2589</P>
2590<P>
2591Before marking strings as internationalizable, substitutions of variables
2592into the string need to be converted to <CODE>format</CODE> applications.  For
2593example, <CODE>"file $filename not found"</CODE> becomes
2594<CODE>[format "file %s not found" $filename]</CODE>.
2595Only after this is done, can the strings be marked and extracted.
2596After marking, this example becomes
2597<CODE>[format [_ "file %s not found"] $filename]</CODE> or
2598<CODE>[msgcat::mc "file %s not found" $filename]</CODE>.  Note that the
2599<CODE>msgcat::mc</CODE> function implicitly calls <CODE>format</CODE> when more than one
2600argument is given.
2601
2602</P>
2603
2604
2605<H3><A NAME="SEC279" HREF="gettext_toc.html#TOC279">15.5.18  Perl</A></H3>
2606<P>
2607<A NAME="IDX1185"></A>
2608
2609</P>
2610<DL COMPACT>
2611
2612<DT>RPMs
2613<DD>
2614perl
2615
2616<DT>File extension
2617<DD>
2618<CODE>pl</CODE>, <CODE>PL</CODE>, <CODE>pm</CODE>, <CODE>cgi</CODE>
2619
2620<DT>String syntax
2621<DD>
2622
2623<UL>
2624
2625<LI><CODE>"abc"</CODE>
2626
2627<LI><CODE>'abc'</CODE>
2628
2629<LI><CODE>qq (abc)</CODE>
2630
2631<LI><CODE>q (abc)</CODE>
2632
2633<LI><CODE>qr /abc/</CODE>
2634
2635<LI><CODE>qx (/bin/date)</CODE>
2636
2637<LI><CODE>/pattern match/</CODE>
2638
2639<LI><CODE>?pattern match?</CODE>
2640
2641<LI><CODE>s/substitution/operators/</CODE>
2642
2643<LI><CODE>$tied_hash{"message"}</CODE>
2644
2645<LI><CODE>$tied_hash_reference-&#62;{"message"}</CODE>
2646
2647<LI>etc., issue the command <SAMP>&lsquo;man perlsyn&rsquo;</SAMP> for details
2648
2649</UL>
2650
2651<DT>gettext shorthand
2652<DD>
2653<CODE>__</CODE> (double underscore)
2654
2655<DT>gettext/ngettext functions
2656<DD>
2657<CODE>gettext</CODE>, <CODE>dgettext</CODE>, <CODE>dcgettext</CODE>, <CODE>ngettext</CODE>,
2658<CODE>dngettext</CODE>, <CODE>dcngettext</CODE>
2659
2660<DT>textdomain
2661<DD>
2662<CODE>textdomain</CODE> function
2663
2664<DT>bindtextdomain
2665<DD>
2666<CODE>bindtextdomain</CODE> function
2667
2668<DT>bind_textdomain_codeset
2669<DD>
2670<CODE>bind_textdomain_codeset</CODE> function
2671
2672<DT>setlocale
2673<DD>
2674Use <CODE>setlocale (LC_ALL, "");</CODE>
2675
2676<DT>Prerequisite
2677<DD>
2678<CODE>use POSIX;</CODE>
2679<BR><CODE>use Locale::TextDomain;</CODE> (included in the package libintl-perl
2680which is available on the Comprehensive Perl Archive Network CPAN,
2681http://www.cpan.org/).
2682
2683<DT>Use or emulate GNU gettext
2684<DD>
2685platform dependent: gettext_pp emulates, gettext_xs uses GNU gettext
2686
2687<DT>Extractor
2688<DD>
2689<CODE>xgettext -k__ -k\$__ -k%__ -k__x -k__n:1,2 -k__nx:1,2 -k__xn:1,2 -kN__ -k</CODE>
2690
2691<DT>Formatting with positions
2692<DD>
2693Both kinds of format strings support formatting with positions.
2694<BR><CODE>printf "%2\$d %1\$d", ...</CODE> (requires Perl 5.8.0 or newer)
2695<BR><CODE>__expand("[new] replaces [old]", old =&#62; $oldvalue, new =&#62; $newvalue)</CODE>
2696
2697<DT>Portability
2698<DD>
2699The <CODE>libintl-perl</CODE> package is platform independent but is not
2700part of the Perl core.  The programmer is responsible for
2701providing a dummy implementation of the required functions if the
2702package is not installed on the target system.
2703
2704<DT>po-mode marking
2705<DD>
2706---
2707
2708<DT>Documentation
2709<DD>
2710Included in <CODE>libintl-perl</CODE>, available on CPAN
2711(http://www.cpan.org/).
2712
2713</DL>
2714
2715<P>
2716An example is available in the <TT>&lsquo;examples&rsquo;</TT> directory: <CODE>hello-perl</CODE>.
2717
2718</P>
2719<P>
2720<A NAME="IDX1186"></A>
2721
2722</P>
2723<P>
2724The <CODE>xgettext</CODE> parser backend for Perl differs significantly from
2725the parser backends for other programming languages, just as Perl
2726itself differs significantly from other programming languages.  The
2727Perl parser backend offers many more string marking facilities than
2728the other backends but it also has some Perl specific limitations, the
2729worst probably being its imperfectness.
2730
2731</P>
2732
2733
2734
2735<H4><A NAME="SEC280" HREF="gettext_toc.html#TOC280">15.5.18.1  General Problems Parsing Perl Code</A></H4>
2736
2737<P>
2738It is often heard that only Perl can parse Perl.  This is not true.
2739Perl cannot be <EM>parsed</EM> at all, it can only be <EM>executed</EM>.
2740Perl has various built-in ambiguities that can only be resolved at runtime.
2741
2742</P>
2743<P>
2744The following example may illustrate one common problem:
2745
2746</P>
2747
2748<PRE>
2749print gettext "Hello World!";
2750</PRE>
2751
2752<P>
2753Although this example looks like a bullet-proof case of a function
2754invocation, it is not:
2755
2756</P>
2757
2758<PRE>
2759open gettext, "&#62;testfile" or die;
2760print gettext "Hello world!"
2761</PRE>
2762
2763<P>
2764In this context, the string <CODE>gettext</CODE> looks more like a
2765file handle.  But not necessarily:
2766
2767</P>
2768
2769<PRE>
2770use Locale::Messages qw (:libintl_h);
2771open gettext "&#62;testfile" or die;
2772print gettext "Hello world!";
2773</PRE>
2774
2775<P>
2776Now, the file is probably syntactically incorrect, provided that the module
2777<CODE>Locale::Messages</CODE> found first in the Perl include path exports a
2778function <CODE>gettext</CODE>.  But what if the module
2779<CODE>Locale::Messages</CODE> really looks like this?
2780
2781</P>
2782
2783<PRE>
2784use vars qw (*gettext);
2785
27861;
2787</PRE>
2788
2789<P>
2790In this case, the string <CODE>gettext</CODE> will be interpreted as a file
2791handle again, and the above example will create a file <TT>&lsquo;testfile&rsquo;</TT>
2792and write the string “Hello world!” into it.  Even advanced
2793control flow analysis will not really help:
2794
2795</P>
2796
2797<PRE>
2798if (0.5 &#60; rand) {
2799   eval "use Sane";
2800} else {
2801   eval "use InSane";
2802}
2803print gettext "Hello world!";
2804</PRE>
2805
2806<P>
2807If the module <CODE>Sane</CODE> exports a function <CODE>gettext</CODE> that does
2808what we expect, and the module <CODE>InSane</CODE> opens a file for writing
2809and associates the <EM>handle</EM> <CODE>gettext</CODE> with this output
2810stream, we are clueless again about what will happen at runtime.  It is
2811completely unpredictable.  The truth is that Perl has so many ways to
2812fill its symbol table at runtime that it is impossible to interpret a
2813particular piece of code without executing it.
2814
2815</P>
2816<P>
2817Of course, <CODE>xgettext</CODE> will not execute your Perl sources while
2818scanning for translatable strings, but rather use heuristics in order
2819to guess what you meant.
2820
2821</P>
2822<P>
2823Another problem is the ambiguity of the slash and the question mark.
2824Their interpretation depends on the context:
2825
2826</P>
2827
2828<PRE>
2829# A pattern match.
2830print "OK\n" if /foobar/;
2831
2832# A division.
2833print 1 / 2;
2834
2835# Another pattern match.
2836print "OK\n" if ?foobar?;
2837
2838# Conditional.
2839print $x ? "foo" : "bar";
2840</PRE>
2841
2842<P>
2843The slash may either act as the division operator or introduce a
2844pattern match, whereas the question mark may act as the ternary
2845conditional operator or as a pattern match, too.  Other programming
2846languages like <CODE>awk</CODE> present similar problems, but the consequences of a
2847misinterpretation are particularly nasty with Perl sources.  In <CODE>awk</CODE>
2848for instance, a statement can never exceed one line and the parser
2849can recover from a parsing error at the next newline and interpret
2850the rest of the input stream correctly.  Perl is different, as a
2851pattern match is terminated by the next appearance of the delimiter
2852(the slash or the question mark) in the input stream, regardless of
2853the semantic context.  If a slash is really a division sign but
2854mis-interpreted as a pattern match, the rest of the input file is most
2855probably parsed incorrectly.
2856
2857</P>
2858<P>
2859If you find that <CODE>xgettext</CODE> fails to extract strings from
2860portions of your sources, you should therefore look out for slashes
2861and/or question marks preceding these sections.  You may have come
2862across a bug in <CODE>xgettext</CODE>'s Perl parser (and of course you
2863should report that bug).  In the meantime you should consider to
2864reformulate your code in a manner less challenging to <CODE>xgettext</CODE>.
2865
2866</P>
2867
2868
2869<H4><A NAME="SEC281" HREF="gettext_toc.html#TOC281">15.5.18.2  Which keywords will xgettext look for?</A></H4>
2870<P>
2871<A NAME="IDX1187"></A>
2872
2873</P>
2874<P>
2875Unless you instruct <CODE>xgettext</CODE> otherwise by invoking it with one
2876of the options <CODE>--keyword</CODE> or <CODE>-k</CODE>, it will recognize the
2877following keywords in your Perl sources:
2878
2879</P>
2880
2881<UL>
2882
2883<LI><CODE>gettext</CODE>
2884
2885<LI><CODE>dgettext</CODE>
2886
2887<LI><CODE>dcgettext</CODE>
2888
2889<LI><CODE>ngettext:1,2</CODE>
2890
2891The first (singular) and the second (plural) argument will be
2892extracted.
2893
2894<LI><CODE>dngettext:1,2</CODE>
2895
2896The first (singular) and the second (plural) argument will be
2897extracted.
2898
2899<LI><CODE>dcngettext:1,2</CODE>
2900
2901The first (singular) and the second (plural) argument will be
2902extracted.
2903
2904<LI><CODE>gettext_noop</CODE>
2905
2906<LI><CODE>%gettext</CODE>
2907
2908The keys of lookups into the hash <CODE>%gettext</CODE> will be extracted.
2909
2910<LI><CODE>$gettext</CODE>
2911
2912The keys of lookups into the hash reference <CODE>$gettext</CODE> will be extracted.
2913
2914</UL>
2915
2916
2917
2918<H4><A NAME="SEC282" HREF="gettext_toc.html#TOC282">15.5.18.3  How to Extract Hash Keys</A></H4>
2919<P>
2920<A NAME="IDX1188"></A>
2921
2922</P>
2923<P>
2924Translating messages at runtime is normally performed by looking up the
2925original string in the translation database and returning the
2926translated version.  The “natural” Perl implementation is a hash
2927lookup, and, of course, <CODE>xgettext</CODE> supports such practice.
2928
2929</P>
2930
2931<PRE>
2932print __"Hello world!";
2933print $__{"Hello world!"};
2934print $__-&#62;{"Hello world!"};
2935print $$__{"Hello world!"};
2936</PRE>
2937
2938<P>
2939The above four lines all do the same thing.  The Perl module
2940<CODE>Locale::TextDomain</CODE> exports by default a hash <CODE>%__</CODE> that
2941is tied to the function <CODE>__()</CODE>.  It also exports a reference
2942<CODE>$__</CODE> to <CODE>%__</CODE>.
2943
2944</P>
2945<P>
2946If an argument to the <CODE>xgettext</CODE> option <CODE>--keyword</CODE>,
2947resp. <CODE>-k</CODE> starts with a percent sign, the rest of the keyword is
2948interpreted as the name of a hash.  If it starts with a dollar
2949sign, the rest of the keyword is interpreted as a reference to a
2950hash.
2951
2952</P>
2953<P>
2954Note that you can omit the quotation marks (single or double) around
2955the hash key (almost) whenever Perl itself allows it:
2956
2957</P>
2958
2959<PRE>
2960print $gettext{Error};
2961</PRE>
2962
2963<P>
2964The exact rule is: You can omit the surrounding quotes, when the hash
2965key is a valid C (!) identifier, i.e. when it starts with an
2966underscore or an ASCII letter and is followed by an arbitrary number
2967of underscores, ASCII letters or digits.  Other Unicode characters
2968are <EM>not</EM> allowed, regardless of the <CODE>use utf8</CODE> pragma.
2969
2970</P>
2971
2972
2973<H4><A NAME="SEC283" HREF="gettext_toc.html#TOC283">15.5.18.4  What are Strings And Quote-like Expressions?</A></H4>
2974<P>
2975<A NAME="IDX1189"></A>
2976
2977</P>
2978<P>
2979Perl offers a plethora of different string constructs.  Those that can
2980be used either as arguments to functions or inside braces for hash
2981lookups are generally supported by <CODE>xgettext</CODE>.
2982
2983</P>
2984
2985<UL>
2986<LI><STRONG>double-quoted strings</STRONG>
2987
2988<BR>
2989
2990<PRE>
2991print gettext "Hello World!";
2992</PRE>
2993
2994<LI><STRONG>single-quoted strings</STRONG>
2995
2996<BR>
2997
2998<PRE>
2999print gettext 'Hello World!';
3000</PRE>
3001
3002<LI><STRONG>the operator qq</STRONG>
3003
3004<BR>
3005
3006<PRE>
3007print gettext qq |Hello World!|;
3008print gettext qq &#60;E-mail: &#60;guido\@imperia.net&#62;&#62;;
3009</PRE>
3010
3011The operator <CODE>qq</CODE> is fully supported.  You can use arbitrary
3012delimiters, including the four bracketing delimiters (round, angle,
3013square, curly) that nest.
3014
3015<LI><STRONG>the operator q</STRONG>
3016
3017<BR>
3018
3019<PRE>
3020print gettext q |Hello World!|;
3021print gettext q &#60;E-mail: &#60;guido@imperia.net&#62;&#62;;
3022</PRE>
3023
3024The operator <CODE>q</CODE> is fully supported.  You can use arbitrary
3025delimiters, including the four bracketing delimiters (round, angle,
3026square, curly) that nest.
3027
3028<LI><STRONG>the operator qx</STRONG>
3029
3030<BR>
3031
3032<PRE>
3033print gettext qx ;LANGUAGE=C /bin/date;
3034print gettext qx [/usr/bin/ls | grep '^[A-Z]*'];
3035</PRE>
3036
3037The operator <CODE>qx</CODE> is fully supported.  You can use arbitrary
3038delimiters, including the four bracketing delimiters (round, angle,
3039square, curly) that nest.
3040
3041The example is actually a useless use of <CODE>gettext</CODE>.  It will
3042invoke the <CODE>gettext</CODE> function on the output of the command
3043specified with the <CODE>qx</CODE> operator.  The feature was included
3044in order to make the interface consistent (the parser will extract
3045all strings and quote-like expressions).
3046
3047<LI><STRONG>here documents</STRONG>
3048
3049<BR>
3050
3051<PRE>
3052print gettext &#60;&#60;'EOF';
3053program not found in $PATH
3054EOF
3055
3056print ngettext &#60;&#60;EOF, &#60;&#60;"EOF";
3057one file deleted
3058EOF
3059several files deleted
3060EOF
3061</PRE>
3062
3063Here-documents are recognized.  If the delimiter is enclosed in single
3064quotes, the string is not interpolated.  If it is enclosed in double
3065quotes or has no quotes at all, the string is interpolated.
3066
3067Delimiters that start with a digit are not supported!
3068
3069</UL>
3070
3071
3072
3073<H4><A NAME="SEC284" HREF="gettext_toc.html#TOC284">15.5.18.5  Invalid Uses Of String Interpolation</A></H4>
3074<P>
3075<A NAME="IDX1190"></A>
3076
3077</P>
3078<P>
3079Perl is capable of interpolating variables into strings.  This offers
3080some nice features in localized programs but can also lead to
3081problems.
3082
3083</P>
3084<P>
3085A common error is a construct like the following:
3086
3087</P>
3088
3089<PRE>
3090print gettext "This is the program $0!\n";
3091</PRE>
3092
3093<P>
3094Perl will interpolate at runtime the value of the variable <CODE>$0</CODE>
3095into the argument of the <CODE>gettext()</CODE> function.  Hence, this
3096argument is not a string constant but a variable argument (<CODE>$0</CODE>
3097is a global variable that holds the name of the Perl script being
3098executed).  The interpolation is performed by Perl before the string
3099argument is passed to <CODE>gettext()</CODE> and will therefore depend on
3100the name of the script which can only be determined at runtime.
3101Consequently, it is almost impossible that a translation can be looked
3102up at runtime (except if, by accident, the interpolated string is found
3103in the message catalog).
3104
3105</P>
3106<P>
3107The <CODE>xgettext</CODE> program will therefore terminate parsing with a fatal
3108error if it encounters a variable inside of an extracted string.  In
3109general, this will happen for all kinds of string interpolations that
3110cannot be safely performed at compile time.  If you absolutely know
3111what you are doing, you can always circumvent this behavior:
3112
3113</P>
3114
3115<PRE>
3116my $know_what_i_am_doing = "This is program $0!\n";
3117print gettext $know_what_i_am_doing;
3118</PRE>
3119
3120<P>
3121Since the parser only recognizes strings and quote-like expressions,
3122but not variables or other terms, the above construct will be
3123accepted.  You will have to find another way, however, to let your
3124original string make it into your message catalog.
3125
3126</P>
3127<P>
3128If invoked with the option <CODE>--extract-all</CODE>, resp. <CODE>-a</CODE>,
3129variable interpolation will be accepted.  Rationale: You will
3130generally use this option in order to prepare your sources for
3131internationalization.
3132
3133</P>
3134<P>
3135Please see the manual page <SAMP>&lsquo;man perlop&rsquo;</SAMP> for details of strings and
3136quote-like expressions that are subject to interpolation and those
3137that are not.  Safe interpolations (that will not lead to a fatal
3138error) are:
3139
3140</P>
3141
3142<UL>
3143
3144<LI>the escape sequences <CODE>\t</CODE> (tab, HT, TAB), <CODE>\n</CODE>
3145
3146(newline, NL), <CODE>\r</CODE> (return, CR), <CODE>\f</CODE> (form feed, FF),
3147<CODE>\b</CODE> (backspace, BS), <CODE>\a</CODE> (alarm, bell, BEL), and <CODE>\e</CODE>
3148(escape, ESC).
3149
3150<LI>octal chars, like <CODE>\033</CODE>
3151
3152<BR>
3153Note that octal escapes in the range of 400-777 are translated into a
3154UTF-8 representation, regardless of the presence of the <CODE>use utf8</CODE> pragma.
3155
3156<LI>hex chars, like <CODE>\x1b</CODE>
3157
3158<LI>wide hex chars, like <CODE>\x{263a}</CODE>
3159
3160<BR>
3161Note that this escape is translated into a UTF-8 representation,
3162regardless of the presence of the <CODE>use utf8</CODE> pragma.
3163
3164<LI>control chars, like <CODE>\c[</CODE> (CTRL-[)
3165
3166<LI>named Unicode chars, like <CODE>\N{LATIN CAPITAL LETTER C WITH CEDILLA}</CODE>
3167
3168<BR>
3169Note that this escape is translated into a UTF-8 representation,
3170regardless of the presence of the <CODE>use utf8</CODE> pragma.
3171</UL>
3172
3173<P>
3174The following escapes are considered partially safe:
3175
3176</P>
3177
3178<UL>
3179
3180<LI><CODE>\l</CODE> lowercase next char
3181
3182<LI><CODE>\u</CODE> uppercase next char
3183
3184<LI><CODE>\L</CODE> lowercase till \E
3185
3186<LI><CODE>\U</CODE> uppercase till \E
3187
3188<LI><CODE>\E</CODE> end case modification
3189
3190<LI><CODE>\Q</CODE> quote non-word characters till \E
3191
3192</UL>
3193
3194<P>
3195These escapes are only considered safe if the string consists of
3196ASCII characters only.  Translation of characters outside the range
3197defined by ASCII is locale-dependent and can actually only be performed
3198at runtime; <CODE>xgettext</CODE> doesn't do these locale-dependent translations
3199at extraction time.
3200
3201</P>
3202<P>
3203Except for the modifier <CODE>\Q</CODE>, these translations, albeit valid,
3204are generally useless and only obfuscate your sources.  If a
3205translation can be safely performed at compile time you can just as
3206well write what you mean.
3207
3208</P>
3209
3210
3211<H4><A NAME="SEC285" HREF="gettext_toc.html#TOC285">15.5.18.6  Valid Uses Of String Interpolation</A></H4>
3212<P>
3213<A NAME="IDX1191"></A>
3214
3215</P>
3216<P>
3217Perl is often used to generate sources for other programming languages
3218or arbitrary file formats.  Web applications that output HTML code
3219make a prominent example for such usage.
3220
3221</P>
3222<P>
3223You will often come across situations where you want to intersperse
3224code written in the target (programming) language with translatable
3225messages, like in the following HTML example:
3226
3227</P>
3228
3229<PRE>
3230print gettext &#60;&#60;EOF;
3231&#60;h1&#62;My Homepage&#60;/h1&#62;
3232&#60;script language="JavaScript"&#62;&#60;!--
3233for (i = 0; i &#60; 100; ++i) {
3234    alert ("Thank you so much for visiting my homepage!");
3235}
3236//--&#62;&#60;/script&#62;
3237EOF
3238</PRE>
3239
3240<P>
3241The parser will extract the entire here document, and it will appear
3242entirely in the resulting PO file, including the JavaScript snippet
3243embedded in the HTML code.  If you exaggerate with constructs like
3244the above, you will run the risk that the translators of your package
3245will look out for a less challenging project.  You should consider an
3246alternative expression here:
3247
3248</P>
3249
3250<PRE>
3251print &#60;&#60;EOF;
3252&#60;h1&#62;$gettext{"My Homepage"}&#60;/h1&#62;
3253&#60;script language="JavaScript"&#62;&#60;!--
3254for (i = 0; i &#60; 100; ++i) {
3255    alert ("$gettext{'Thank you so much for visiting my homepage!'}");
3256}
3257//--&#62;&#60;/script&#62;
3258EOF
3259</PRE>
3260
3261<P>
3262Only the translatable portions of the code will be extracted here, and
3263the resulting PO file will begrudgingly improve in terms of readability.
3264
3265</P>
3266<P>
3267You can interpolate hash lookups in all strings or quote-like
3268expressions that are subject to interpolation (see the manual page
3269<SAMP>&lsquo;man perlop&rsquo;</SAMP> for details).  Double interpolation is invalid, however:
3270
3271</P>
3272
3273<PRE>
3274# TRANSLATORS: Replace "the earth" with the name of your planet.
3275print gettext qq{Welcome to $gettext-&#62;{"the earth"}};
3276</PRE>
3277
3278<P>
3279The <CODE>qq</CODE>-quoted string is recognized as an argument to <CODE>xgettext</CODE> in
3280the first place, and checked for invalid variable interpolation.  The
3281dollar sign of hash-dereferencing will therefore terminate the parser
3282with an “invalid interpolation” error.
3283
3284</P>
3285<P>
3286It is valid to interpolate hash lookups in regular expressions:
3287
3288</P>
3289
3290<PRE>
3291if ($var =~ /$gettext{"the earth"}/) {
3292   print gettext "Match!\n";
3293}
3294s/$gettext{"U. S. A."}/$gettext{"U. S. A."} $gettext{"(dial +0)"}/g;
3295</PRE>
3296
3297
3298
3299<H4><A NAME="SEC286" HREF="gettext_toc.html#TOC286">15.5.18.7  When To Use Parentheses</A></H4>
3300<P>
3301<A NAME="IDX1192"></A>
3302
3303</P>
3304<P>
3305In Perl, parentheses around function arguments are mostly optional.
3306<CODE>xgettext</CODE> will always assume that all
3307recognized keywords (except for hashes and hash references) are names
3308of properly prototyped functions, and will (hopefully) only require
3309parentheses where Perl itself requires them.  All constructs in the
3310following example are therefore ok to use:
3311
3312</P>
3313
3314<PRE>
3315print gettext ("Hello World!\n");
3316print gettext "Hello World!\n";
3317print dgettext ($package =&#62; "Hello World!\n");
3318print dgettext $package, "Hello World!\n";
3319
3320# The "fat comma" =&#62; turns the left-hand side argument into a
3321# single-quoted string!
3322print dgettext smellovision =&#62; "Hello World!\n";
3323
3324# The following assignment only works with prototyped functions.
3325# Otherwise, the functions will act as "greedy" list operators and
3326# eat up all following arguments.
3327my $anonymous_hash = {
3328   planet =&#62; gettext "earth",
3329   cakes =&#62; ngettext "one cake", "several cakes", $n,
3330   still =&#62; $works,
3331};
3332# The same without fat comma:
3333my $other_hash = {
3334   'planet', gettext "earth",
3335   'cakes', ngettext "one cake", "several cakes", $n,
3336   'still', $works,
3337};
3338
3339# Parentheses are only significant for the first argument.
3340print dngettext 'package', ("one cake", "several cakes", $n), $discarded;
3341</PRE>
3342
3343
3344
3345<H4><A NAME="SEC287" HREF="gettext_toc.html#TOC287">15.5.18.8  How To Grok with Long Lines</A></H4>
3346<P>
3347<A NAME="IDX1193"></A>
3348
3349</P>
3350<P>
3351The necessity of long messages can often lead to a cumbersome or
3352unreadable coding style.  Perl has several options that may prevent
3353you from writing unreadable code, and
3354<CODE>xgettext</CODE> does its best to do likewise.  This is where the dot
3355operator (the string concatenation operator) may come in handy:
3356
3357</P>
3358
3359<PRE>
3360print gettext ("This is a very long"
3361               . " message that is still"
3362               . " readable, because"
3363               . " it is split into"
3364               . " multiple lines.\n");
3365</PRE>
3366
3367<P>
3368Perl is smart enough to concatenate these constant string fragments
3369into one long string at compile time, and so is
3370<CODE>xgettext</CODE>.  You will only find one long message in the resulting
3371POT file.
3372
3373</P>
3374<P>
3375Note that the future Perl 6 will probably use the underscore
3376(<SAMP>&lsquo;_&rsquo;</SAMP>) as the string concatenation operator, and the dot
3377(<SAMP>&lsquo;.&rsquo;</SAMP>) for dereferencing.  This new syntax is not yet supported by
3378<CODE>xgettext</CODE>.
3379
3380</P>
3381<P>
3382If embedded newline characters are not an issue, or even desired, you
3383may also insert newline characters inside quoted strings wherever you
3384feel like it:
3385
3386</P>
3387
3388<PRE>
3389print gettext ("&#60;em&#62;In HTML output
3390embedded newlines are generally no
3391problem, since adjacent whitespace
3392is always rendered into a single
3393space character.&#60;/em&#62;");
3394</PRE>
3395
3396<P>
3397You may also consider to use here documents:
3398
3399</P>
3400
3401<PRE>
3402print gettext &#60;&#60;EOF;
3403&#60;em&#62;In HTML output
3404embedded newlines are generally no
3405problem, since adjacent whitespace
3406is always rendered into a single
3407space character.&#60;/em&#62;
3408EOF
3409</PRE>
3410
3411<P>
3412Please do not forget that the line breaks are real, i.e. they
3413translate into newline characters that will consequently show up in
3414the resulting POT file.
3415
3416</P>
3417
3418
3419<H4><A NAME="SEC288" HREF="gettext_toc.html#TOC288">15.5.18.9  Bugs, Pitfalls, And Things That Do Not Work</A></H4>
3420<P>
3421<A NAME="IDX1194"></A>
3422
3423</P>
3424<P>
3425The foregoing sections should have proven that
3426<CODE>xgettext</CODE> is quite smart in extracting translatable strings from
3427Perl sources.  Yet, some more or less exotic constructs that could be
3428expected to work, actually do not work.
3429
3430</P>
3431<P>
3432One of the more relevant limitations can be found in the
3433implementation of variable interpolation inside quoted strings.  Only
3434simple hash lookups can be used there:
3435
3436</P>
3437
3438<PRE>
3439print &#60;&#60;EOF;
3440$gettext{"The dot operator"
3441          . " does not work"
3442          . "here!"}
3443Likewise, you cannot @{[ gettext ("interpolate function calls") ]}
3444inside quoted strings or quote-like expressions.
3445EOF
3446</PRE>
3447
3448<P>
3449This is valid Perl code and will actually trigger invocations of the
3450<CODE>gettext</CODE> function at runtime.  Yet, the Perl parser in
3451<CODE>xgettext</CODE> will fail to recognize the strings.  A less obvious
3452example can be found in the interpolation of regular expressions:
3453
3454</P>
3455
3456<PRE>
3457s/&#60;!--START_OF_WEEK--&#62;/gettext ("Sunday")/e;
3458</PRE>
3459
3460<P>
3461The modifier <CODE>e</CODE> will cause the substitution to be interpreted as
3462an evaluable statement.  Consequently, at runtime the function
3463<CODE>gettext()</CODE> is called, but again, the parser fails to extract the
3464string “Sunday”.  Use a temporary variable as a simple workaround if
3465you really happen to need this feature:
3466
3467</P>
3468
3469<PRE>
3470my $sunday = gettext "Sunday";
3471s/&#60;!--START_OF_WEEK--&#62;/$sunday/;
3472</PRE>
3473
3474<P>
3475Hash slices would also be handy but are not recognized:
3476
3477</P>
3478
3479<PRE>
3480my @weekdays = @gettext{'Sunday', 'Monday', 'Tuesday', 'Wednesday',
3481                        'Thursday', 'Friday', 'Saturday'};
3482# Or even:
3483@weekdays = @gettext{qw (Sunday Monday Tuesday Wednesday Thursday
3484                         Friday Saturday) };
3485</PRE>
3486
3487<P>
3488This is perfectly valid usage of the tied hash <CODE>%gettext</CODE> but the
3489strings are not recognized and therefore will not be extracted.
3490
3491</P>
3492<P>
3493Another caveat of the current version is its rudimentary support for
3494non-ASCII characters in identifiers.  You may encounter serious
3495problems if you use identifiers with characters outside the range of
3496'A'-'Z', 'a'-'z', '0'-'9' and the underscore '_'.
3497
3498</P>
3499<P>
3500Maybe some of these missing features will be implemented in future
3501versions, but since you can always make do without them at minimal effort,
3502these todos have very low priority.
3503
3504</P>
3505<P>
3506A nasty problem are brace format strings that already contain braces
3507as part of the normal text, for example the usage strings typically
3508encountered in programs:
3509
3510</P>
3511
3512<PRE>
3513die "usage: $0 {OPTIONS} FILENAME...\n";
3514</PRE>
3515
3516<P>
3517If you want to internationalize this code with Perl brace format strings,
3518you will run into a problem:
3519
3520</P>
3521
3522<PRE>
3523die __x ("usage: {program} {OPTIONS} FILENAME...\n", program =&#62; $0);
3524</PRE>
3525
3526<P>
3527Whereas <SAMP>&lsquo;{program}&rsquo;</SAMP> is a placeholder, <SAMP>&lsquo;{OPTIONS}&rsquo;</SAMP>
3528is not and should probably be translated. Yet, there is no way to teach
3529the Perl parser in <CODE>xgettext</CODE> to recognize the first one, and leave
3530the other one alone.
3531
3532</P>
3533<P>
3534There are two possible work-arounds for this problem.  If you are
3535sure that your program will run under Perl 5.8.0 or newer (these
3536Perl versions handle positional parameters in <CODE>printf()</CODE>) or
3537if you are sure that the translator will not have to reorder the arguments
3538in her translation -- for example if you have only one brace placeholder
3539in your string, or if it describes a syntax, like in this one --, you can
3540mark the string as <CODE>no-perl-brace-format</CODE> and use <CODE>printf()</CODE>:
3541
3542</P>
3543
3544<PRE>
3545# xgettext: no-perl-brace-format
3546die sprintf ("usage: %s {OPTIONS} FILENAME...\n", $0);
3547</PRE>
3548
3549<P>
3550If you want to use the more portable Perl brace format, you will have to do
3551put placeholders in place of the literal braces:
3552
3553</P>
3554
3555<PRE>
3556die __x ("usage: {program} {[}OPTIONS{]} FILENAME...\n",
3557         program =&#62; $0, '[' =&#62; '{', ']' =&#62; '}');
3558</PRE>
3559
3560<P>
3561Perl brace format strings know no escaping mechanism.  No matter how this
3562escaping mechanism looked like, it would either give the programmer a
3563hard time, make translating Perl brace format strings heavy-going, or
3564result in a performance penalty at runtime, when the format directives
3565get executed.  Most of the time you will happily get along with
3566<CODE>printf()</CODE> for this special case.
3567
3568</P>
3569
3570
3571<H3><A NAME="SEC289" HREF="gettext_toc.html#TOC289">15.5.19  PHP Hypertext Preprocessor</A></H3>
3572<P>
3573<A NAME="IDX1195"></A>
3574
3575</P>
3576<DL COMPACT>
3577
3578<DT>RPMs
3579<DD>
3580mod_php4, mod_php4-core, phpdoc
3581
3582<DT>File extension
3583<DD>
3584<CODE>php</CODE>, <CODE>php3</CODE>, <CODE>php4</CODE>
3585
3586<DT>String syntax
3587<DD>
3588<CODE>"abc"</CODE>, <CODE>'abc'</CODE>
3589
3590<DT>gettext shorthand
3591<DD>
3592<CODE>_("abc")</CODE>
3593
3594<DT>gettext/ngettext functions
3595<DD>
3596<CODE>gettext</CODE>, <CODE>dgettext</CODE>, <CODE>dcgettext</CODE>; starting with PHP 4.2.0
3597also <CODE>ngettext</CODE>, <CODE>dngettext</CODE>, <CODE>dcngettext</CODE>
3598
3599<DT>textdomain
3600<DD>
3601<CODE>textdomain</CODE> function
3602
3603<DT>bindtextdomain
3604<DD>
3605<CODE>bindtextdomain</CODE> function
3606
3607<DT>setlocale
3608<DD>
3609Programmer must call <CODE>setlocale (LC_ALL, "")</CODE>
3610
3611<DT>Prerequisite
3612<DD>
3613---
3614
3615<DT>Use or emulate GNU gettext
3616<DD>
3617use
3618
3619<DT>Extractor
3620<DD>
3621<CODE>xgettext</CODE>
3622
3623<DT>Formatting with positions
3624<DD>
3625<CODE>printf "%2\$d %1\$d"</CODE>
3626
3627<DT>Portability
3628<DD>
3629On platforms without gettext, the functions are not available.
3630
3631<DT>po-mode marking
3632<DD>
3633---
3634</DL>
3635
3636<P>
3637An example is available in the <TT>&lsquo;examples&rsquo;</TT> directory: <CODE>hello-php</CODE>.
3638
3639</P>
3640
3641
3642<H3><A NAME="SEC290" HREF="gettext_toc.html#TOC290">15.5.20  Pike</A></H3>
3643<P>
3644<A NAME="IDX1196"></A>
3645
3646</P>
3647<DL COMPACT>
3648
3649<DT>RPMs
3650<DD>
3651roxen
3652
3653<DT>File extension
3654<DD>
3655<CODE>pike</CODE>
3656
3657<DT>String syntax
3658<DD>
3659<CODE>"abc"</CODE>
3660
3661<DT>gettext shorthand
3662<DD>
3663---
3664
3665<DT>gettext/ngettext functions
3666<DD>
3667<CODE>gettext</CODE>, <CODE>dgettext</CODE>, <CODE>dcgettext</CODE>
3668
3669<DT>textdomain
3670<DD>
3671<CODE>textdomain</CODE> function
3672
3673<DT>bindtextdomain
3674<DD>
3675<CODE>bindtextdomain</CODE> function
3676
3677<DT>setlocale
3678<DD>
3679<CODE>setlocale</CODE> function
3680
3681<DT>Prerequisite
3682<DD>
3683<CODE>import Locale.Gettext;</CODE>
3684
3685<DT>Use or emulate GNU gettext
3686<DD>
3687use
3688
3689<DT>Extractor
3690<DD>
3691---
3692
3693<DT>Formatting with positions
3694<DD>
3695---
3696
3697<DT>Portability
3698<DD>
3699On platforms without gettext, the functions are not available.
3700
3701<DT>po-mode marking
3702<DD>
3703---
3704</DL>
3705
3706
3707
3708<H3><A NAME="SEC291" HREF="gettext_toc.html#TOC291">15.5.21  GNU Compiler Collection sources</A></H3>
3709<P>
3710<A NAME="IDX1197"></A>
3711
3712</P>
3713<DL COMPACT>
3714
3715<DT>RPMs
3716<DD>
3717gcc
3718
3719<DT>File extension
3720<DD>
3721<CODE>c</CODE>, <CODE>h</CODE>.
3722
3723<DT>String syntax
3724<DD>
3725<CODE>"abc"</CODE>
3726
3727<DT>gettext shorthand
3728<DD>
3729<CODE>_("abc")</CODE>
3730
3731<DT>gettext/ngettext functions
3732<DD>
3733<CODE>gettext</CODE>, <CODE>dgettext</CODE>, <CODE>dcgettext</CODE>, <CODE>ngettext</CODE>,
3734<CODE>dngettext</CODE>, <CODE>dcngettext</CODE>
3735
3736<DT>textdomain
3737<DD>
3738<CODE>textdomain</CODE> function
3739
3740<DT>bindtextdomain
3741<DD>
3742<CODE>bindtextdomain</CODE> function
3743
3744<DT>setlocale
3745<DD>
3746Programmer must call <CODE>setlocale (LC_ALL, "")</CODE>
3747
3748<DT>Prerequisite
3749<DD>
3750<CODE>#include "intl.h"</CODE>
3751
3752<DT>Use or emulate GNU gettext
3753<DD>
3754Use
3755
3756<DT>Extractor
3757<DD>
3758<CODE>xgettext -k_</CODE>
3759
3760<DT>Formatting with positions
3761<DD>
3762---
3763
3764<DT>Portability
3765<DD>
3766Uses autoconf macros
3767
3768<DT>po-mode marking
3769<DD>
3770yes
3771</DL>
3772
3773
3774
3775<H2><A NAME="SEC292" HREF="gettext_toc.html#TOC292">15.6  Internationalizable Data</A></H2>
3776
3777<P>
3778Here is a list of other data formats which can be internationalized
3779using GNU gettext.
3780
3781</P>
3782
3783
3784
3785<H3><A NAME="SEC293" HREF="gettext_toc.html#TOC293">15.6.1  POT - Portable Object Template</A></H3>
3786
3787<DL COMPACT>
3788
3789<DT>RPMs
3790<DD>
3791gettext
3792
3793<DT>File extension
3794<DD>
3795<CODE>pot</CODE>, <CODE>po</CODE>
3796
3797<DT>Extractor
3798<DD>
3799<CODE>xgettext</CODE>
3800</DL>
3801
3802
3803
3804<H3><A NAME="SEC294" HREF="gettext_toc.html#TOC294">15.6.2  Resource String Table</A></H3>
3805<P>
3806<A NAME="IDX1198"></A>
3807
3808</P>
3809<DL COMPACT>
3810
3811<DT>RPMs
3812<DD>
3813fpk
3814
3815<DT>File extension
3816<DD>
3817<CODE>rst</CODE>
3818
3819<DT>Extractor
3820<DD>
3821<CODE>xgettext</CODE>, <CODE>rstconv</CODE>
3822</DL>
3823
3824
3825
3826<H3><A NAME="SEC295" HREF="gettext_toc.html#TOC295">15.6.3  Glade - GNOME user interface description</A></H3>
3827
3828<DL COMPACT>
3829
3830<DT>RPMs
3831<DD>
3832glade, libglade, glade2, libglade2, intltool
3833
3834<DT>File extension
3835<DD>
3836<CODE>glade</CODE>, <CODE>glade2</CODE>
3837
3838<DT>Extractor
3839<DD>
3840<CODE>xgettext</CODE>, <CODE>libglade-xgettext</CODE>, <CODE>xml-i18n-extract</CODE>, <CODE>intltool-extract</CODE>
3841</DL>
3842
3843<P><HR><P>
3844Go to the <A HREF="gettext_1.html">first</A>, <A HREF="gettext_14.html">previous</A>, <A HREF="gettext_16.html">next</A>, <A HREF="gettext_25.html">last</A> section, <A HREF="gettext_toc.html">table of contents</A>.
3845</BODY>
3846</HTML>
3847