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 - 4 Preparing Program Sources</TITLE> 8</HEAD> 9<BODY> 10Go to the <A HREF="gettext_1.html">first</A>, <A HREF="gettext_3.html">previous</A>, <A HREF="gettext_5.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="SEC11" HREF="gettext_toc.html#TOC11">4 Preparing Program Sources</A></H1> 15<P> 16<A NAME="IDX102"></A> 17 18</P> 19 20<P> 21For the programmer, changes to the C source code fall into three 22categories. First, you have to make the localization functions 23known to all modules needing message translation. Second, you should 24properly trigger the operation of GNU <CODE>gettext</CODE> when the program 25initializes, usually from the <CODE>main</CODE> function. Last, you should 26identify, adjust and mark all constant strings in your program 27needing translation. 28 29</P> 30 31 32 33<H2><A NAME="SEC12" HREF="gettext_toc.html#TOC12">4.1 Importing the <CODE>gettext</CODE> declaration</A></H2> 34 35<P> 36Presuming that your set of programs, or package, has been adjusted 37so all needed GNU <CODE>gettext</CODE> files are available, and your 38<TT>‘Makefile’</TT> files are adjusted (see section <A HREF="gettext_13.html#SEC196">13 The Maintainer's View</A>), each C module 39having translated C strings should contain the line: 40 41</P> 42<P> 43<A NAME="IDX103"></A> 44 45<PRE> 46#include <libintl.h> 47</PRE> 48 49<P> 50Similarly, each C module containing <CODE>printf()</CODE>/<CODE>fprintf()</CODE>/... 51calls with a format string that could be a translated C string (even if 52the C string comes from a different C module) should contain the line: 53 54</P> 55 56<PRE> 57#include <libintl.h> 58</PRE> 59 60 61 62<H2><A NAME="SEC13" HREF="gettext_toc.html#TOC13">4.2 Triggering <CODE>gettext</CODE> Operations</A></H2> 63 64<P> 65<A NAME="IDX104"></A> 66The initialization of locale data should be done with more or less 67the same code in every program, as demonstrated below: 68 69</P> 70 71<PRE> 72int 73main (int argc, char *argv[]) 74{ 75 ... 76 setlocale (LC_ALL, ""); 77 bindtextdomain (PACKAGE, LOCALEDIR); 78 textdomain (PACKAGE); 79 ... 80} 81</PRE> 82 83<P> 84<VAR>PACKAGE</VAR> and <VAR>LOCALEDIR</VAR> should be provided either by 85<TT>‘config.h’</TT> or by the Makefile. For now consult the <CODE>gettext</CODE> 86or <CODE>hello</CODE> sources for more information. 87 88</P> 89<P> 90<A NAME="IDX105"></A> 91<A NAME="IDX106"></A> 92The use of <CODE>LC_ALL</CODE> might not be appropriate for you. 93<CODE>LC_ALL</CODE> includes all locale categories and especially 94<CODE>LC_CTYPE</CODE>. This later category is responsible for determining 95character classes with the <CODE>isalnum</CODE> etc. functions from 96<TT>‘ctype.h’</TT> which could especially for programs, which process some 97kind of input language, be wrong. For example this would mean that a 98source code using the ç (c-cedilla character) is runnable in 99France but not in the U.S. 100 101</P> 102<P> 103Some systems also have problems with parsing numbers using the 104<CODE>scanf</CODE> functions if an other but the <CODE>LC_ALL</CODE> locale is used. 105The standards say that additional formats but the one known in the 106<CODE>"C"</CODE> locale might be recognized. But some systems seem to reject 107numbers in the <CODE>"C"</CODE> locale format. In some situation, it might 108also be a problem with the notation itself which makes it impossible to 109recognize whether the number is in the <CODE>"C"</CODE> locale or the local 110format. This can happen if thousands separator characters are used. 111Some locales define this character according to the national 112conventions to <CODE>'.'</CODE> which is the same character used in the 113<CODE>"C"</CODE> locale to denote the decimal point. 114 115</P> 116<P> 117So it is sometimes necessary to replace the <CODE>LC_ALL</CODE> line in the 118code above by a sequence of <CODE>setlocale</CODE> lines 119 120</P> 121 122<PRE> 123{ 124 ... 125 setlocale (LC_CTYPE, ""); 126 setlocale (LC_MESSAGES, ""); 127 ... 128} 129</PRE> 130 131<P> 132<A NAME="IDX107"></A> 133<A NAME="IDX108"></A> 134<A NAME="IDX109"></A> 135<A NAME="IDX110"></A> 136<A NAME="IDX111"></A> 137<A NAME="IDX112"></A> 138<A NAME="IDX113"></A> 139On all POSIX conformant systems the locale categories <CODE>LC_CTYPE</CODE>, 140<CODE>LC_MESSAGES</CODE>, <CODE>LC_COLLATE</CODE>, <CODE>LC_MONETARY</CODE>, 141<CODE>LC_NUMERIC</CODE>, and <CODE>LC_TIME</CODE> are available. On some systems 142which are only ISO C compliant, <CODE>LC_MESSAGES</CODE> is missing, but 143a substitute for it is defined in GNU gettext's <CODE><libintl.h></CODE>. 144 145</P> 146<P> 147Note that changing the <CODE>LC_CTYPE</CODE> also affects the functions 148declared in the <CODE><ctype.h></CODE> standard header. If this is not 149desirable in your application (for example in a compiler's parser), 150you can use a set of substitute functions which hardwire the C locale, 151such as found in the <CODE><c-ctype.h></CODE> and <CODE><c-ctype.c></CODE> files 152in the gettext source distribution. 153 154</P> 155<P> 156It is also possible to switch the locale forth and back between the 157environment dependent locale and the C locale, but this approach is 158normally avoided because a <CODE>setlocale</CODE> call is expensive, 159because it is tedious to determine the places where a locale switch 160is needed in a large program's source, and because switching a locale 161is not multithread-safe. 162 163</P> 164 165 166<H2><A NAME="SEC14" HREF="gettext_toc.html#TOC14">4.3 Preparing Translatable Strings</A></H2> 167 168<P> 169<A NAME="IDX114"></A> 170Before strings can be marked for translations, they sometimes need to 171be adjusted. Usually preparing a string for translation is done right 172before marking it, during the marking phase which is described in the 173next sections. What you have to keep in mind while doing that is the 174following. 175 176</P> 177 178<UL> 179<LI> 180 181Decent English style. 182 183<LI> 184 185Entire sentences. 186 187<LI> 188 189Split at paragraphs. 190 191<LI> 192 193Use format strings instead of string concatenation. 194 195<LI> 196 197Avoid unusual markup and unusual control characters. 198</UL> 199 200<P> 201Let's look at some examples of these guidelines. 202 203</P> 204<P> 205<A NAME="IDX115"></A> 206Translatable strings should be in good English style. If slang language 207with abbreviations and shortcuts is used, often translators will not 208understand the message and will produce very inappropriate translations. 209 210</P> 211 212<PRE> 213"%s: is parameter\n" 214</PRE> 215 216<P> 217This is nearly untranslatable: Is the displayed item <EM>a</EM> parameter or 218<EM>the</EM> parameter? 219 220</P> 221 222<PRE> 223"No match" 224</PRE> 225 226<P> 227The ambiguity in this message makes it unintelligible: Is the program 228attempting to set something on fire? Does it mean "The given object does 229not match the template"? Does it mean "The template does not fit for any 230of the objects"? 231 232</P> 233<P> 234<A NAME="IDX116"></A> 235In both cases, adding more words to the message will help both the 236translator and the English speaking user. 237 238</P> 239<P> 240<A NAME="IDX117"></A> 241Translatable strings should be entire sentences. It is often not possible 242to translate single verbs or adjectives in a substitutable way. 243 244</P> 245 246<PRE> 247printf ("File %s is %s protected", filename, rw ? "write" : "read"); 248</PRE> 249 250<P> 251Most translators will not look at the source and will thus only see the 252string <CODE>"File %s is %s protected"</CODE>, which is unintelligible. Change 253this to 254 255</P> 256 257<PRE> 258printf (rw ? "File %s is write protected" : "File %s is read protected", 259 filename); 260</PRE> 261 262<P> 263This way the translator will not only understand the message, she will 264also be able to find the appropriate grammatical construction. A French 265translator for example translates "write protected" like "protected 266against writing". 267 268</P> 269<P> 270Entire sentences are also important because in many languages, the 271declination of some word in a sentence depends on the gender or the 272number (singular/plural) of another part of the sentence. There are 273usually more interdependencies between words than in English. The 274consequence is that asking a translator to translate two half-sentences 275and then combining these two half-sentences through dumb string concatenation 276will not work, for many languages, even though it would work for English. 277That's why translators need to handle entire sentences. 278 279</P> 280<P> 281Often sentences don't fit into a single line. If a sentence is output 282using two subsequent <CODE>printf</CODE> statements, like this 283 284</P> 285 286<PRE> 287printf ("Locale charset \"%s\" is different from\n", lcharset); 288printf ("input file charset \"%s\".\n", fcharset); 289</PRE> 290 291<P> 292the translator would have to translate two half sentences, but nothing 293in the POT file would tell her that the two half sentences belong together. 294It is necessary to merge the two <CODE>printf</CODE> statements so that the 295translator can handle the entire sentence at once and decide at which 296place to insert a line break in the translation (if at all): 297 298</P> 299 300<PRE> 301printf ("Locale charset \"%s\" is different from\n\ 302input file charset \"%s\".\n", lcharset, fcharset); 303</PRE> 304 305<P> 306You may now ask: how about two or more adjacent sentences? Like in this case: 307 308</P> 309 310<PRE> 311puts ("Apollo 13 scenario: Stack overflow handling failed."); 312puts ("On the next stack overflow we will crash!!!"); 313</PRE> 314 315<P> 316Should these two statements merged into a single one? I would recommend to 317merge them if the two sentences are related to each other, because then it 318makes it easier for the translator to understand and translate both. On 319the other hand, if one of the two messages is a stereotypic one, occurring 320in other places as well, you will do a favour to the translator by not 321merging the two. (Identical messages occurring in several places are 322combined by xgettext, so the translator has to handle them once only.) 323 324</P> 325<P> 326<A NAME="IDX118"></A> 327Translatable strings should be limited to one paragraph; don't let a 328single message be longer than ten lines. The reason is that when the 329translatable string changes, the translator is faced with the task of 330updating the entire translated string. Maybe only a single word will 331have changed in the English string, but the translator doesn't see that 332(with the current translation tools), therefore she has to proofread 333the entire message. 334 335</P> 336<P> 337<A NAME="IDX119"></A> 338Many GNU programs have a <SAMP>‘--help’</SAMP> output that extends over several 339screen pages. It is a courtesy towards the translators to split such a 340message into several ones of five to ten lines each. While doing that, 341you can also attempt to split the documented options into groups, 342such as the input options, the output options, and the informative 343output options. This will help every user to find the option he is 344looking for. 345 346</P> 347<P> 348<A NAME="IDX120"></A> 349<A NAME="IDX121"></A> 350Hardcoded string concatenation is sometimes used to construct English 351strings: 352 353</P> 354 355<PRE> 356strcpy (s, "Replace "); 357strcat (s, object1); 358strcat (s, " with "); 359strcat (s, object2); 360strcat (s, "?"); 361</PRE> 362 363<P> 364In order to present to the translator only entire sentences, and also 365because in some languages the translator might want to swap the order 366of <CODE>object1</CODE> and <CODE>object2</CODE>, it is necessary to change this 367to use a format string: 368 369</P> 370 371<PRE> 372sprintf (s, "Replace %s with %s?", object1, object2); 373</PRE> 374 375<P> 376<A NAME="IDX122"></A> 377A similar case is compile time concatenation of strings. The ISO C 99 378include file <CODE><inttypes.h></CODE> contains a macro <CODE>PRId64</CODE> that 379can be used as a formatting directive for outputting an <SAMP>‘int64_t’</SAMP> 380integer through <CODE>printf</CODE>. It expands to a constant string, usually 381"d" or "ld" or "lld" or something like this, depending on the platform. 382Assume you have code like 383 384</P> 385 386<PRE> 387printf ("The amount is %0" PRId64 "\n", number); 388</PRE> 389 390<P> 391The <CODE>gettext</CODE> tools and library have special support for these 392<CODE><inttypes.h></CODE> macros. You can therefore simply write 393 394</P> 395 396<PRE> 397printf (gettext ("The amount is %0" PRId64 "\n"), number); 398</PRE> 399 400<P> 401The PO file will contain the string "The amount is %0<PRId64>\n". 402The translators will provide a translation containing "%0<PRId64>" 403as well, and at runtime the <CODE>gettext</CODE> function's result will 404contain the appropriate constant string, "d" or "ld" or "lld". 405 406</P> 407<P> 408This works only for the predefined <CODE><inttypes.h></CODE> macros. If 409you have defined your own similar macros, let's say <SAMP>‘MYPRId64’</SAMP>, 410that are not known to <CODE>xgettext</CODE>, the solution for this problem 411is to change the code like this: 412 413</P> 414 415<PRE> 416char buf1[100]; 417sprintf (buf1, "%0" MYPRId64, number); 418printf (gettext ("The amount is %s\n"), buf1); 419</PRE> 420 421<P> 422This means, you put the platform dependent code in one statement, and the 423internationalization code in a different statement. Note that a buffer length 424of 100 is safe, because all available hardware integer types are limited to 425128 bits, and to print a 128 bit integer one needs at most 54 characters, 426regardless whether in decimal, octal or hexadecimal. 427 428</P> 429<P> 430<A NAME="IDX123"></A> 431<A NAME="IDX124"></A> 432All this applies to other programming languages as well. For example, in 433Java and C#, string concatenation is very frequently used, because it is a 434compiler built-in operator. Like in C, in Java, you would change 435 436</P> 437 438<PRE> 439System.out.println("Replace "+object1+" with "+object2+"?"); 440</PRE> 441 442<P> 443into a statement involving a format string: 444 445</P> 446 447<PRE> 448System.out.println( 449 MessageFormat.format("Replace {0} with {1}?", 450 new Object[] { object1, object2 })); 451</PRE> 452 453<P> 454Similarly, in C#, you would change 455 456</P> 457 458<PRE> 459Console.WriteLine("Replace "+object1+" with "+object2+"?"); 460</PRE> 461 462<P> 463into a statement involving a format string: 464 465</P> 466 467<PRE> 468Console.WriteLine( 469 String.Format("Replace {0} with {1}?", object1, object2)); 470</PRE> 471 472<P> 473<A NAME="IDX125"></A> 474<A NAME="IDX126"></A> 475Unusual markup or control characters should not be used in translatable 476strings. Translators will likely not understand the particular meaning 477of the markup or control characters. 478 479</P> 480<P> 481For example, if you have a convention that <SAMP>‘|’</SAMP> delimits the 482left-hand and right-hand part of some GUI elements, translators will 483often not understand it without specific comments. It might be 484better to have the translator translate the left-hand and right-hand 485part separately. 486 487</P> 488<P> 489Another example is the <SAMP>‘argp’</SAMP> convention to use a single <SAMP>‘\v’</SAMP> 490(vertical tab) control character to delimit two sections inside a 491string. This is flawed. Some translators may convert it to a simple 492newline, some to blank lines. With some PO file editors it may not be 493easy to even enter a vertical tab control character. So, you cannot 494be sure that the translation will contain a <SAMP>‘\v’</SAMP> character, at the 495corresponding position. The solution is, again, to let the translator 496translate two separate strings and combine at run-time the two translated 497strings with the <SAMP>‘\v’</SAMP> required by the convention. 498 499</P> 500<P> 501HTML markup, however, is common enough that it's probably ok to use in 502translatable strings. But please bear in mind that the GNU gettext tools 503don't verify that the translations are well-formed HTML. 504 505</P> 506 507 508<H2><A NAME="SEC15" HREF="gettext_toc.html#TOC15">4.4 How Marks Appear in Sources</A></H2> 509<P> 510<A NAME="IDX127"></A> 511 512</P> 513<P> 514All strings requiring translation should be marked in the C sources. Marking 515is done in such a way that each translatable string appears to be 516the sole argument of some function or preprocessor macro. There are 517only a few such possible functions or macros meant for translation, 518and their names are said to be marking keywords. The marking is 519attached to strings themselves, rather than to what we do with them. 520This approach has more uses. A blatant example is an error message 521produced by formatting. The format string needs translation, as 522well as some strings inserted through some <SAMP>‘%s’</SAMP> specification 523in the format, while the result from <CODE>sprintf</CODE> may have so many 524different instances that it is impractical to list them all in some 525<SAMP>‘error_string_out()’</SAMP> routine, say. 526 527</P> 528<P> 529This marking operation has two goals. The first goal of marking 530is for triggering the retrieval of the translation, at run time. 531The keyword is possibly resolved into a routine able to dynamically 532return the proper translation, as far as possible or wanted, for the 533argument string. Most localizable strings are found in executable 534positions, that is, attached to variables or given as parameters to 535functions. But this is not universal usage, and some translatable 536strings appear in structured initializations. See section <A HREF="gettext_4.html#SEC18">4.7 Special Cases of Translatable Strings</A>. 537 538</P> 539<P> 540The second goal of the marking operation is to help <CODE>xgettext</CODE> 541at properly extracting all translatable strings when it scans a set 542of program sources and produces PO file templates. 543 544</P> 545<P> 546The canonical keyword for marking translatable strings is 547<SAMP>‘gettext’</SAMP>, it gave its name to the whole GNU <CODE>gettext</CODE> 548package. For packages making only light use of the <SAMP>‘gettext’</SAMP> 549keyword, macro or function, it is easily used <EM>as is</EM>. However, 550for packages using the <CODE>gettext</CODE> interface more heavily, it 551is usually more convenient to give the main keyword a shorter, less 552obtrusive name. Indeed, the keyword might appear on a lot of strings 553all over the package, and programmers usually do not want nor need 554their program sources to remind them forcefully, all the time, that they 555are internationalized. Further, a long keyword has the disadvantage 556of using more horizontal space, forcing more indentation work on 557sources for those trying to keep them within 79 or 80 columns. 558 559</P> 560<P> 561<A NAME="IDX128"></A> 562Many packages use <SAMP>‘_’</SAMP> (a simple underline) as a keyword, 563and write <SAMP>‘_("Translatable string")’</SAMP> instead of <SAMP>‘gettext 564("Translatable string")’</SAMP>. Further, the coding rule, from GNU standards, 565wanting that there is a space between the keyword and the opening 566parenthesis is relaxed, in practice, for this particular usage. 567So, the textual overhead per translatable string is reduced to 568only three characters: the underline and the two parentheses. 569However, even if GNU <CODE>gettext</CODE> uses this convention internally, 570it does not offer it officially. The real, genuine keyword is truly 571<SAMP>‘gettext’</SAMP> indeed. It is fairly easy for those wanting to use 572<SAMP>‘_’</SAMP> instead of <SAMP>‘gettext’</SAMP> to declare: 573 574</P> 575 576<PRE> 577#include <libintl.h> 578#define _(String) gettext (String) 579</PRE> 580 581<P> 582instead of merely using <SAMP>‘#include <libintl.h>’</SAMP>. 583 584</P> 585<P> 586The marking keywords <SAMP>‘gettext’</SAMP> and <SAMP>‘_’</SAMP> take the translatable 587string as sole argument. It is also possible to define marking functions 588that take it at another argument position. It is even possible to make 589the marked argument position depend on the total number of arguments of 590the function call; this is useful in C++. All this is achieved using 591<CODE>xgettext</CODE>'s <SAMP>‘--keyword’</SAMP> option. 592 593</P> 594<P> 595Note also that long strings can be split across lines, into multiple 596adjacent string tokens. Automatic string concatenation is performed 597at compile time according to ISO C and ISO C++; <CODE>xgettext</CODE> also 598supports this syntax. 599 600</P> 601<P> 602Later on, the maintenance is relatively easy. If, as a programmer, 603you add or modify a string, you will have to ask yourself if the 604new or altered string requires translation, and include it within 605<SAMP>‘_()’</SAMP> if you think it should be translated. For example, <SAMP>‘"%s"’</SAMP> 606is an example of string <EM>not</EM> requiring translation. But 607<SAMP>‘"%s: %d"’</SAMP> <EM>does</EM> require translation, because in French, unlike 608in English, it's customary to put a space before a colon. 609 610</P> 611 612 613<H2><A NAME="SEC16" HREF="gettext_toc.html#TOC16">4.5 Marking Translatable Strings</A></H2> 614<P> 615<A NAME="IDX129"></A> 616 617</P> 618<P> 619In PO mode, one set of features is meant more for the programmer than 620for the translator, and allows him to interactively mark which strings, 621in a set of program sources, are translatable, and which are not. 622Even if it is a fairly easy job for a programmer to find and mark 623such strings by other means, using any editor of his choice, PO mode 624makes this work more comfortable. Further, this gives translators 625who feel a little like programmers, or programmers who feel a little 626like translators, a tool letting them work at marking translatable 627strings in the program sources, while simultaneously producing a set of 628translation in some language, for the package being internationalized. 629 630</P> 631<P> 632<A NAME="IDX130"></A> 633The set of program sources, targeted by the PO mode commands describe 634here, should have an Emacs tags table constructed for your project, 635prior to using these PO file commands. This is easy to do. In any 636shell window, change the directory to the root of your project, then 637execute a command resembling: 638 639</P> 640 641<PRE> 642etags src/*.[hc] lib/*.[hc] 643</PRE> 644 645<P> 646presuming here you want to process all <TT>‘.h’</TT> and <TT>‘.c’</TT> files 647from the <TT>‘src/’</TT> and <TT>‘lib/’</TT> directories. This command will 648explore all said files and create a <TT>‘TAGS’</TT> file in your root 649directory, somewhat summarizing the contents using a special file 650format Emacs can understand. 651 652</P> 653<P> 654<A NAME="IDX131"></A> 655For packages following the GNU coding standards, there is 656a make goal <CODE>tags</CODE> or <CODE>TAGS</CODE> which constructs the tag files in 657all directories and for all files containing source code. 658 659</P> 660<P> 661Once your <TT>‘TAGS’</TT> file is ready, the following commands assist 662the programmer at marking translatable strings in his set of sources. 663But these commands are necessarily driven from within a PO file 664window, and it is likely that you do not even have such a PO file yet. 665This is not a problem at all, as you may safely open a new, empty PO 666file, mainly for using these commands. This empty PO file will slowly 667fill in while you mark strings as translatable in your program sources. 668 669</P> 670<DL COMPACT> 671 672<DT><KBD>,</KBD> 673<DD> 674<A NAME="IDX132"></A> 675Search through program sources for a string which looks like a 676candidate for translation (<CODE>po-tags-search</CODE>). 677 678<DT><KBD>M-,</KBD> 679<DD> 680<A NAME="IDX133"></A> 681Mark the last string found with <SAMP>‘_()’</SAMP> (<CODE>po-mark-translatable</CODE>). 682 683<DT><KBD>M-.</KBD> 684<DD> 685<A NAME="IDX134"></A> 686Mark the last string found with a keyword taken from a set of possible 687keywords. This command with a prefix allows some management of these 688keywords (<CODE>po-select-mark-and-mark</CODE>). 689 690</DL> 691 692<P> 693<A NAME="IDX135"></A> 694The <KBD>,</KBD> (<CODE>po-tags-search</CODE>) command searches for the next 695occurrence of a string which looks like a possible candidate for 696translation, and displays the program source in another Emacs window, 697positioned in such a way that the string is near the top of this other 698window. If the string is too big to fit whole in this window, it is 699positioned so only its end is shown. In any case, the cursor 700is left in the PO file window. If the shown string would be better 701presented differently in different native languages, you may mark it 702using <KBD>M-,</KBD> or <KBD>M-.</KBD>. Otherwise, you might rather ignore it 703and skip to the next string by merely repeating the <KBD>,</KBD> command. 704 705</P> 706<P> 707A string is a good candidate for translation if it contains a sequence 708of three or more letters. A string containing at most two letters in 709a row will be considered as a candidate if it has more letters than 710non-letters. The command disregards strings containing no letters, 711or isolated letters only. It also disregards strings within comments, 712or strings already marked with some keyword PO mode knows (see below). 713 714</P> 715<P> 716If you have never told Emacs about some <TT>‘TAGS’</TT> file to use, the 717command will request that you specify one from the minibuffer, the 718first time you use the command. You may later change your <TT>‘TAGS’</TT> 719file by using the regular Emacs command <KBD>M-x visit-tags-table</KBD>, 720which will ask you to name the precise <TT>‘TAGS’</TT> file you want 721to use. See section ���Tag Tables��� in <CITE>The Emacs Editor</CITE>. 722 723</P> 724<P> 725Each time you use the <KBD>,</KBD> command, the search resumes from where it was 726left by the previous search, and goes through all program sources, 727obeying the <TT>‘TAGS’</TT> file, until all sources have been processed. 728However, by giving a prefix argument to the command (<KBD>C-u 729,)</KBD>, you may request that the search be restarted all over again 730from the first program source; but in this case, strings that you 731recently marked as translatable will be automatically skipped. 732 733</P> 734<P> 735Using this <KBD>,</KBD> command does not prevent using of other regular 736Emacs tags commands. For example, regular <CODE>tags-search</CODE> or 737<CODE>tags-query-replace</CODE> commands may be used without disrupting the 738independent <KBD>,</KBD> search sequence. However, as implemented, the 739<EM>initial</EM> <KBD>,</KBD> command (or the <KBD>,</KBD> command is used with a 740prefix) might also reinitialize the regular Emacs tags searching to the 741first tags file, this reinitialization might be considered spurious. 742 743</P> 744<P> 745<A NAME="IDX136"></A> 746<A NAME="IDX137"></A> 747The <KBD>M-,</KBD> (<CODE>po-mark-translatable</CODE>) command will mark the 748recently found string with the <SAMP>‘_’</SAMP> keyword. The <KBD>M-.</KBD> 749(<CODE>po-select-mark-and-mark</CODE>) command will request that you type 750one keyword from the minibuffer and use that keyword for marking 751the string. Both commands will automatically create a new PO file 752untranslated entry for the string being marked, and make it the 753current entry (making it easy for you to immediately proceed to its 754translation, if you feel like doing it right away). It is possible 755that the modifications made to the program source by <KBD>M-,</KBD> or 756<KBD>M-.</KBD> render some source line longer than 80 columns, forcing you 757to break and re-indent this line differently. You may use the <KBD>O</KBD> 758command from PO mode, or any other window changing command from 759Emacs, to break out into the program source window, and do any 760needed adjustments. You will have to use some regular Emacs command 761to return the cursor to the PO file window, if you want command 762<KBD>,</KBD> for the next string, say. 763 764</P> 765<P> 766The <KBD>M-.</KBD> command has a few built-in speedups, so you do not 767have to explicitly type all keywords all the time. The first such 768speedup is that you are presented with a <EM>preferred</EM> keyword, 769which you may accept by merely typing <KBD><KBD>RET</KBD></KBD> at the prompt. 770The second speedup is that you may type any non-ambiguous prefix of the 771keyword you really mean, and the command will complete it automatically 772for you. This also means that PO mode has to <EM>know</EM> all 773your possible keywords, and that it will not accept mistyped keywords. 774 775</P> 776<P> 777If you reply <KBD>?</KBD> to the keyword request, the command gives a 778list of all known keywords, from which you may choose. When the 779command is prefixed by an argument (<KBD>C-u M-.</KBD>), it inhibits 780updating any program source or PO file buffer, and does some simple 781keyword management instead. In this case, the command asks for a 782keyword, written in full, which becomes a new allowed keyword for 783later <KBD>M-.</KBD> commands. Moreover, this new keyword automatically 784becomes the <EM>preferred</EM> keyword for later commands. By typing 785an already known keyword in response to <KBD>C-u M-.</KBD>, one merely 786changes the <EM>preferred</EM> keyword and does nothing more. 787 788</P> 789<P> 790All keywords known for <KBD>M-.</KBD> are recognized by the <KBD>,</KBD> command 791when scanning for strings, and strings already marked by any of those 792known keywords are automatically skipped. If many PO files are opened 793simultaneously, each one has its own independent set of known keywords. 794There is no provision in PO mode, currently, for deleting a known 795keyword, you have to quit the file (maybe using <KBD>q</KBD>) and reopen 796it afresh. When a PO file is newly brought up in an Emacs window, only 797<SAMP>‘gettext’</SAMP> and <SAMP>‘_’</SAMP> are known as keywords, and <SAMP>‘gettext’</SAMP> 798is preferred for the <KBD>M-.</KBD> command. In fact, this is not useful to 799prefer <SAMP>‘_’</SAMP>, as this one is already built in the <KBD>M-,</KBD> command. 800 801</P> 802 803 804<H2><A NAME="SEC17" HREF="gettext_toc.html#TOC17">4.6 Special Comments preceding Keywords</A></H2> 805 806<P> 807<A NAME="IDX138"></A> 808In C programs strings are often used within calls of functions from the 809<CODE>printf</CODE> family. The special thing about these format strings is 810that they can contain format specifiers introduced with <KBD>%</KBD>. Assume 811we have the code 812 813</P> 814 815<PRE> 816printf (gettext ("String `%s' has %d characters\n"), s, strlen (s)); 817</PRE> 818 819<P> 820A possible German translation for the above string might be: 821 822</P> 823 824<PRE> 825"%d Zeichen lang ist die Zeichenkette `%s'" 826</PRE> 827 828<P> 829A C programmer, even if he cannot speak German, will recognize that 830there is something wrong here. The order of the two format specifiers 831is changed but of course the arguments in the <CODE>printf</CODE> don't have. 832This will most probably lead to problems because now the length of the 833string is regarded as the address. 834 835</P> 836<P> 837To prevent errors at runtime caused by translations the <CODE>msgfmt</CODE> 838tool can check statically whether the arguments in the original and the 839translation string match in type and number. If this is not the case 840and the <SAMP>‘-c’</SAMP> option has been passed to <CODE>msgfmt</CODE>, <CODE>msgfmt</CODE> 841will give an error and refuse to produce a MO file. Thus consequent 842use of <SAMP>‘msgfmt -c’</SAMP> will catch the error, so that it cannot cause 843cause problems at runtime. 844 845</P> 846<P> 847If the word order in the above German translation would be correct one 848would have to write 849 850</P> 851 852<PRE> 853"%2$d Zeichen lang ist die Zeichenkette `%1$s'" 854</PRE> 855 856<P> 857The routines in <CODE>msgfmt</CODE> know about this special notation. 858 859</P> 860<P> 861Because not all strings in a program must be format strings it is not 862useful for <CODE>msgfmt</CODE> to test all the strings in the <TT>‘.po’</TT> file. 863This might cause problems because the string might contain what looks 864like a format specifier, but the string is not used in <CODE>printf</CODE>. 865 866</P> 867<P> 868Therefore the <CODE>xgettext</CODE> adds a special tag to those messages it 869thinks might be a format string. There is no absolute rule for this, 870only a heuristic. In the <TT>‘.po’</TT> file the entry is marked using the 871<CODE>c-format</CODE> flag in the <CODE>#,</CODE> comment line (see section <A HREF="gettext_3.html#SEC10">3 The Format of PO Files</A>). 872 873</P> 874<P> 875<A NAME="IDX139"></A> 876<A NAME="IDX140"></A> 877The careful reader now might say that this again can cause problems. 878The heuristic might guess it wrong. This is true and therefore 879<CODE>xgettext</CODE> knows about a special kind of comment which lets 880the programmer take over the decision. If in the same line as or 881the immediately preceding line to the <CODE>gettext</CODE> keyword 882the <CODE>xgettext</CODE> program finds a comment containing the words 883<CODE>xgettext:c-format</CODE>, it will mark the string in any case with 884the <CODE>c-format</CODE> flag. This kind of comment should be used when 885<CODE>xgettext</CODE> does not recognize the string as a format string but 886it really is one and it should be tested. Please note that when the 887comment is in the same line as the <CODE>gettext</CODE> keyword, it must be 888before the string to be translated. 889 890</P> 891<P> 892This situation happens quite often. The <CODE>printf</CODE> function is often 893called with strings which do not contain a format specifier. Of course 894one would normally use <CODE>fputs</CODE> but it does happen. In this case 895<CODE>xgettext</CODE> does not recognize this as a format string but what 896happens if the translation introduces a valid format specifier? The 897<CODE>printf</CODE> function will try to access one of the parameters but none 898exists because the original code does not pass any parameters. 899 900</P> 901<P> 902<CODE>xgettext</CODE> of course could make a wrong decision the other way 903round, i.e. a string marked as a format string actually is not a format 904string. In this case the <CODE>msgfmt</CODE> might give too many warnings and 905would prevent translating the <TT>‘.po’</TT> file. The method to prevent 906this wrong decision is similar to the one used above, only the comment 907to use must contain the string <CODE>xgettext:no-c-format</CODE>. 908 909</P> 910<P> 911If a string is marked with <CODE>c-format</CODE> and this is not correct the 912user can find out who is responsible for the decision. See 913section <A HREF="gettext_5.html#SEC22">5.1 Invoking the <CODE>xgettext</CODE> Program</A> to see how the <CODE>--debug</CODE> option can be 914used for solving this problem. 915 916</P> 917 918 919<H2><A NAME="SEC18" HREF="gettext_toc.html#TOC18">4.7 Special Cases of Translatable Strings</A></H2> 920 921<P> 922<A NAME="IDX141"></A> 923The attentive reader might now point out that it is not always possible 924to mark translatable string with <CODE>gettext</CODE> or something like this. 925Consider the following case: 926 927</P> 928 929<PRE> 930{ 931 static const char *messages[] = { 932 "some very meaningful message", 933 "and another one" 934 }; 935 const char *string; 936 ... 937 string 938 = index > 1 ? "a default message" : messages[index]; 939 940 fputs (string); 941 ... 942} 943</PRE> 944 945<P> 946While it is no problem to mark the string <CODE>"a default message"</CODE> it 947is not possible to mark the string initializers for <CODE>messages</CODE>. 948What is to be done? We have to fulfill two tasks. First we have to mark the 949strings so that the <CODE>xgettext</CODE> program (see section <A HREF="gettext_5.html#SEC22">5.1 Invoking the <CODE>xgettext</CODE> Program</A>) 950can find them, and second we have to translate the string at runtime 951before printing them. 952 953</P> 954<P> 955The first task can be fulfilled by creating a new keyword, which names a 956no-op. For the second we have to mark all access points to a string 957from the array. So one solution can look like this: 958 959</P> 960 961<PRE> 962#define gettext_noop(String) String 963 964{ 965 static const char *messages[] = { 966 gettext_noop ("some very meaningful message"), 967 gettext_noop ("and another one") 968 }; 969 const char *string; 970 ... 971 string 972 = index > 1 ? gettext ("a default message") : gettext (messages[index]); 973 974 fputs (string); 975 ... 976} 977</PRE> 978 979<P> 980Please convince yourself that the string which is written by 981<CODE>fputs</CODE> is translated in any case. How to get <CODE>xgettext</CODE> know 982the additional keyword <CODE>gettext_noop</CODE> is explained in section <A HREF="gettext_5.html#SEC22">5.1 Invoking the <CODE>xgettext</CODE> Program</A>. 983 984</P> 985<P> 986The above is of course not the only solution. You could also come along 987with the following one: 988 989</P> 990 991<PRE> 992#define gettext_noop(String) String 993 994{ 995 static const char *messages[] = { 996 gettext_noop ("some very meaningful message", 997 gettext_noop ("and another one") 998 }; 999 const char *string; 1000 ... 1001 string 1002 = index > 1 ? gettext_noop ("a default message") : messages[index]; 1003 1004 fputs (gettext (string)); 1005 ... 1006} 1007</PRE> 1008 1009<P> 1010But this has a drawback. The programmer has to take care that 1011he uses <CODE>gettext_noop</CODE> for the string <CODE>"a default message"</CODE>. 1012A use of <CODE>gettext</CODE> could have in rare cases unpredictable results. 1013 1014</P> 1015<P> 1016One advantage is that you need not make control flow analysis to make 1017sure the output is really translated in any case. But this analysis is 1018generally not very difficult. If it should be in any situation you can 1019use this second method in this situation. 1020 1021</P> 1022 1023 1024<H2><A NAME="SEC19" HREF="gettext_toc.html#TOC19">4.8 Marking Proper Names for Translation</A></H2> 1025 1026<P> 1027Should names of persons, cities, locations etc. be marked for translation 1028or not? People who only know languages that can be written with Latin 1029letters (English, Spanish, French, German, etc.) are tempted to say ���no���, 1030because names usually do not change when transported between these languages. 1031However, in general when translating from one script to another, names 1032are translated too, usually phonetically or by transliteration. For 1033example, Russian or Greek names are converted to the Latin alphabet when 1034being translated to English, and English or French names are converted 1035to the Katakana script when being translated to Japanese. This is 1036necessary because the speakers of the target language in general cannot 1037read the script the name is originally written in. 1038 1039</P> 1040<P> 1041As a programmer, you should therefore make sure that names are marked 1042for translation, with a special comment telling the translators that it 1043is a proper name and how to pronounce it. Like this: 1044 1045</P> 1046 1047<PRE> 1048printf (_("Written by %s.\n"), 1049 /* TRANSLATORS: This is a proper name. See the gettext 1050 manual, section Names. Note this is actually a non-ASCII 1051 name: The first name is (with Unicode escapes) 1052 "Fran\u00e7ois" or (with HTML entities) "Fran&ccedil;ois". 1053 Pronunciation is like "fraa-swa pee-nar". */ 1054 _("Francois Pinard")); 1055</PRE> 1056 1057<P> 1058As a translator, you should use some care when translating names, because 1059it is frustrating if people see their names mutilated or distorted. If 1060your language uses the Latin script, all you need to do is to reproduce 1061the name as perfectly as you can within the usual character set of your 1062language. In this particular case, this means to provide a translation 1063containing the c-cedilla character. If your language uses a different 1064script and the people speaking it don't usually read Latin words, it means 1065transliteration; but you should still give, in parentheses, the original 1066writing of the name -- for the sake of the people that do read the Latin 1067script. Here is an example, using Greek as the target script: 1068 1069</P> 1070 1071<PRE> 1072#. This is a proper name. See the gettext 1073#. manual, section Names. Note this is actually a non-ASCII 1074#. name: The first name is (with Unicode escapes) 1075#. "Fran\u00e7ois" or (with HTML entities) "Fran&ccedil;ois". 1076#. Pronunciation is like "fraa-swa pee-nar". 1077msgid "Francois Pinard" 1078msgstr "\phi\rho\alpha\sigma\omicron\alpha \pi\iota\nu\alpha\rho" 1079 " (Francois Pinard)" 1080</PRE> 1081 1082<P> 1083Because translation of names is such a sensitive domain, it is a good 1084idea to test your translation before submitting it. 1085 1086</P> 1087<P> 1088The translation project <A HREF="http://sourceforge.net/projects/translation">http://sourceforge.net/projects/translation</A> 1089has set up a POT file and translation domain consisting of program author 1090names, with better facilities for the translator than those presented here. 1091Namely, there the original name is written directly in Unicode (rather 1092than with Unicode escapes or HTML entities), and the pronunciation is 1093denoted using the International Phonetic Alphabet (see 1094<A HREF="http://www.wikipedia.org/wiki/International_Phonetic_Alphabet">http://www.wikipedia.org/wiki/International_Phonetic_Alphabet</A>). 1095 1096</P> 1097<P> 1098However, we don't recommend this approach for all POT files in all packages, 1099because this would force translators to use PO files in UTF-8 encoding, 1100which is - in the current state of software (as of 2003) - a major hassle 1101for translators using GNU Emacs or XEmacs with po-mode. 1102 1103</P> 1104 1105 1106<H2><A NAME="SEC20" HREF="gettext_toc.html#TOC20">4.9 Preparing Library Sources</A></H2> 1107 1108<P> 1109When you are preparing a library, not a program, for the use of 1110<CODE>gettext</CODE>, only a few details are different. Here we assume that 1111the library has a translation domain and a POT file of its own. (If 1112it uses the translation domain and POT file of the main program, then 1113the previous sections apply without changes.) 1114 1115</P> 1116 1117<OL> 1118<LI> 1119 1120The library code doesn't call <CODE>setlocale (LC_ALL, "")</CODE>. It's the 1121responsibility of the main program to set the locale. The library's 1122documentation should mention this fact, so that developers of programs 1123using the library are aware of it. 1124 1125<LI> 1126 1127The library code doesn't call <CODE>textdomain (PACKAGE)</CODE>, because it 1128would interfere with the text domain set by the main program. 1129 1130<LI> 1131 1132The initialization code for a program was 1133 1134 1135<PRE> 1136 setlocale (LC_ALL, ""); 1137 bindtextdomain (PACKAGE, LOCALEDIR); 1138 textdomain (PACKAGE); 1139</PRE> 1140 1141For a library it is reduced to 1142 1143 1144<PRE> 1145 bindtextdomain (PACKAGE, LOCALEDIR); 1146</PRE> 1147 1148If your library's API doesn't already have an initialization function, 1149you need to create one, containing at least the <CODE>bindtextdomain</CODE> 1150invocation. However, you usually don't need to export and document this 1151initialization function: It is sufficient that all entry points of the 1152library call the initialization function if it hasn't been called before. 1153The typical idiom used to achieve this is a static boolean variable that 1154indicates whether the initialization function has been called. Like this: 1155 1156 1157<PRE> 1158static bool libfoo_initialized; 1159 1160static void 1161libfoo_initialize (void) 1162{ 1163 bindtextdomain (PACKAGE, LOCALEDIR); 1164 libfoo_initialized = true; 1165} 1166 1167/* This function is part of the exported API. */ 1168struct foo * 1169create_foo (...) 1170{ 1171 /* Must ensure the initialization is performed. */ 1172 if (!libfoo_initialized) 1173 libfoo_initialize (); 1174 ... 1175} 1176 1177/* This function is part of the exported API. The argument must be 1178 non-NULL and have been created through create_foo(). */ 1179int 1180foo_refcount (struct foo *argument) 1181{ 1182 /* No need to invoke the initialization function here, because 1183 create_foo() must already have been called before. */ 1184 ... 1185} 1186</PRE> 1187 1188<LI> 1189 1190The usual declaration of the <SAMP>‘_’</SAMP> macro in each source file was 1191 1192 1193<PRE> 1194#include <libintl.h> 1195#define _(String) gettext (String) 1196</PRE> 1197 1198for a program. For a library, which has its own translation domain, 1199it reads like this: 1200 1201 1202<PRE> 1203#include <libintl.h> 1204#define _(String) dgettext (PACKAGE, String) 1205</PRE> 1206 1207In other words, <CODE>dgettext</CODE> is used instead of <CODE>gettext</CODE>. 1208Similarly, the <CODE>dngettext</CODE> function should be used in place of the 1209<CODE>ngettext</CODE> function. 1210</OL> 1211 1212<P><HR><P> 1213Go to the <A HREF="gettext_1.html">first</A>, <A HREF="gettext_3.html">previous</A>, <A HREF="gettext_5.html">next</A>, <A HREF="gettext_25.html">last</A> section, <A HREF="gettext_toc.html">table of contents</A>. 1214</BODY> 1215</HTML> 1216