4 PROPOSAL FOR HOW THIS ALL OUGHT TO WORK
5 this isn't implemented yet, but this is the plan-in-progress
8 In general, it's accepted that the best way to internationalize is for all
9 messages to be referred to by a symbolic name (or number) and come out of a
10 table or tables, which are easy to change.
12 However, with Emacs, we've got the task of internationalizing a huge body
13 of existing code, which already contains messages internally.
15 For the C code we've got two options:
17 - Use a Sun-like gettext() form, which takes an "english" string which
18 appears literally in the source, and uses that as a hash key to find
20 - Rip all of the strings out and put them in a table.
22 In this case, it's desirable to make as few changes as possible to the C
23 code, to make it easier to merge the code with the FSF version of emacs
24 which won't ever have these changes made to it. So we should go with the
27 The way it has been done (between 19.8 and 19.9) was to use gettext(), but
28 *also* to make massive changes to the source code. The goal now is to use
29 gettext() at run-time and yet not require a textual change to every line
30 in the C code which contains a string constant. A possible way to do this
33 (gettext() can be implemented in terms of catgets() for non-Sun systems, so
34 that in itself isn't a problem.)
36 For the Lisp code, we've got basically the same options: put everything in
37 a table, or translate things implicitly.
39 Another kink that lisp code introduces is that there are thousands of third-
40 party packages, so changing the source for all of those is simply not an
43 Is it a goal that if some third party package displays a message which is
44 one we know how to translate, then we translate it? I think this is a
45 worthy goal. It remains to be seen how well it will work in practice.
47 So, we should endeavor to minimize the impact on the lisp code. Certain
48 primitive lisp routines (the stuff in lisp/prim/, and especially in
49 cmdloop.el and minibuf.el) may need to be changed to know about translation,
50 but that's an ideologically clean thing to do because those are considered
51 a part of the emacs substrate.
53 However, if we find ourselves wanting to make changes to, say, RMAIL, then
54 something has gone wrong. (Except to do things like remove assumptions
55 about the order of words within a sentence, or how pluralization works.)
57 There are two parts to the task of displaying translated strings to the
58 user: the first is to extract the strings which need to be translated from
59 the sources; and the second is to make some call which will translate those
60 strings before they are presented to the user.
62 The old way was to use the same form to do both, that is, GETTEXT() was both
63 the tag that we searched for to build a catalog, and was the form which did
64 the translation. The new plan is to separate these two things more: the
65 tags that we search for to build the catalog will be stuff that was in there
66 already, and the translation will get done in some more centralized, lower
69 This program (make-msgfile.c) addresses the first part, extracting the
72 For the emacs C code, we need to recognize the following patterns:
74 message ("string" ... )
76 report_file_error ("string" ... )
77 signal_simple_error ("string" ... )
78 signal_simple_error_2 ("string" ... )
80 build_translated_string ("string")
81 #### add this and use it instead of build_string() in some places.
83 yes_or_no_p ("string" ... )
84 #### add this instead of funcalling Qyes_or_no_p directly.
86 barf_or_query_if_file_exists #### restructure this
87 check all callers of Fsignal #### restructure these
88 signal_error (Qerror ... ) #### change all of these to error()
90 And we also parse out the `interactive' prompts from DEFUN() forms.
92 #### When we've got a string which is a candidate for translation, we
93 should ignore it if it contains only format directives, that is, if
94 there are no alphabetic characters in it that are not a part of a `%'
95 directive. (Careful not to translate either "%s%s" or "%s: ".)
97 For the emacs Lisp code, we need to recognize the following patterns:
99 (message "string" ... )
100 (error "string" ... )
101 (format "string" ... )
102 (read-from-minibuffer "string" ... )
103 (read-shell-command "string" ... )
104 (y-or-n-p "string" ... )
105 (yes-or-no-p "string" ... )
106 (read-file-name "string" ... )
107 (temp-minibuffer-message "string")
108 (query-replace-read-args "string" ... )
110 I expect there will be a lot like the above; basically, any function which
111 is a commonly used wrapper around an eventual call to `message' or
112 `read-from-minibuffer' needs to be recognized by this program.
115 (dgettext "domain-name" "string") #### do we still need this?
117 things that should probably be restructured:
118 `princ' in cmdloop.el
121 help.el, syntax.el all messed up
124 Menu descriptors: one way to extract the strings in menu labels would be
125 to teach this program about "^(defvar .*menu\n" forms; that's probably
126 kind of hard, though, so perhaps a better approach would be to make this
127 program recognize lines of the form
129 "string" ... ;###translate
131 where the magic token ";###translate" on a line means that the string
132 constant on this line should go into the message catalog. This is analogous
133 to the magic ";###autoload" comments, and to the magic comments used in the
134 EPSF structuring conventions.
137 So this program manages to build up a catalog of strings to be translated.
138 To address the second part of the problem, of actually looking up the
139 translations, there are hooks in a small number of low level places in
142 Assume the existence of a C function gettext(str) which returns the
143 translation of `str' if there is one, otherwise returns `str'.
145 - message() takes a char* as its argument, and always filters it through
146 gettext() before displaying it.
148 - errors are printed by running the lisp function `display-error' which
149 doesn't call `message' directly (it princ's to streams), so it must be
150 carefully coded to translate its arguments. This is only a few lines
153 - Fread_minibuffer_internal() is the lowest level interface to all minibuf
154 interactions, so it is responsible for translating the value that will go
155 into Vminibuf_prompt.
157 - Fpopup_menu filters the menu titles through gettext().
159 The above take care of 99% of all messages the user ever sees.
161 - The lisp function temp-minibuffer-message translates its arg.
163 - query-replace-read-args is funny; it does
164 (setq from (read-from-minibuffer (format "%s: " string) ... ))
165 (setq to (read-from-minibuffer (format "%s %s with: " string from) ... ))
167 What should we do about this? We could hack query-replace-read-args to
168 translate its args, but might this be a more general problem? I don't
169 think we ought to translate all calls to format. We could just change
170 the calling sequence, since this is odd in that the first %s wants to be
171 translated but the second doesn't.
174 Solving the "translating too much" problem:
175 The concern has been raised that in this situation:
176 - "Help" is a string for which we know a translation;
177 - someone visits a file called Help, and someone does something
178 contrived like (error buffer-file-name)
179 then we would display the translation of Help, which would not be correct.
180 We can solve this by adding a bit to Lisp_String objects which identifies
181 them as having been read as literal constants from a .el or .elc file (as
182 opposed to having been constructed at run time as it would in the above
183 case.) To solve this:
185 - Fmessage() takes a lisp string as its first argument.
186 If that string is a constant, that is, was read from a source file
187 as a literal, then it calls message() with it, which translates.
188 Otherwise, it calls message_no_translate(), which does not translate.
190 - Ferror() (actually, Fsignal() when condition is Qerror) works similarly.
196 /* Scan specified C and Lisp files, extracting the following messages:
201 DEFUN interactive prompts
204 (dgettext "domain-name" ...)
208 The arguments given to this program are all the C and Lisp source files
209 of GNU Emacs. .el and .c files are allowed. There is no support for .elc
210 files at this time, but they may be specified; the corresponding .el file
211 will be used. Similarly, .o files can also be specified, and the corresponding
212 .c file will be used. This helps the makefile pass the correct list of files.
214 The results, which go to standard output or to a file specified with -a or -o
215 (-a to append, -o to start from nothing), are quoted strings wrapped in
216 gettext(...). The results can be passed to xgettext to produce a .po message
224 #define GET_LINE fgets (line, LINESIZE, infile)
225 #define CHECK_EOL(p) if (*(p) == '\0') (p) = GET_LINE
226 #define SKIP_BLANKS(p) while ((*p) == ' ' || (*p) == '\t') (p)++
228 enum filetype { C_FILE, LISP_FILE, INVALID_FILE };
229 /* some brain-dead headers define this ... */
232 enum boolean { FALSE, TRUE };
239 void scan_file (char *filename);
240 void process_C_file (void);
241 void process_Lisp_file (void);
242 char *copy_up_to_paren (register char *p);
243 char *copy_quoted_string (register char *p);
244 enum boolean no_interactive_prompt (register char *q);
245 char *skip_blanks (register char *p);
248 main (int argc, char *argv[])
254 /* If first two args are -o FILE, output to FILE. */
256 if (argc > i + 1 && strcmp (argv[i], "-o") == 0) {
257 outfile = fopen (argv[++i], "w");
260 /* ...Or if args are -a FILE, append to FILE. */
261 if (argc > i + 1 && strcmp (argv[i], "-a") == 0) {
262 outfile = fopen (argv[++i], "a");
266 fprintf (stderr, "Unable to open output file %s\n", argv[--i]);
270 for (; i < argc; i++)
277 void scan_file (char *filename)
279 enum filetype type = INVALID_FILE;
280 register char *p = filename + strlen (filename);
282 if (strcmp (p - 4, ".elc") == 0) {
283 *--p = '\0'; /* Use .el file instead */
285 } else if (strcmp (p - 3, ".el") == 0)
287 else if (strcmp (p - 2, ".o") == 0) {
288 *--p = 'c'; /* Use .c file instead */
290 } else if (strcmp (p - 2, ".c") == 0)
293 if (type == INVALID_FILE) {
294 fprintf (stderr, "File %s being ignored\n", filename);
297 infile = fopen (filename, "r");
299 fprintf (stderr, "Unable to open input file %s\n", filename);
303 fprintf (outfile, "/* %s */\n", filename);
307 process_Lisp_file ();
308 fputc ('\n', outfile);
314 void process_C_file (void)
317 char *gettext, *defun;
319 while (p = GET_LINE) {
320 gettext = strstr (p, "GETTEXT");
321 defun = strstr (p, "DEFUN");
322 if (gettext || defun) {
325 p += 7; /* Skip over "GETTEXT" */
329 p += 5; /* Skip over "DEFUN" */
339 for (i = 0; i < 5; i++) /* Skip over commas to doc string */
347 if (*p != '\"') /* Make sure there is a quoted string */
350 if (defun && no_interactive_prompt (p))
353 fprintf (outfile, "gettext(");
355 p = copy_up_to_paren (p);
357 p = copy_quoted_string (p);
358 fprintf (outfile, ")\n");
364 void process_Lisp_file (void)
367 char *gettext, *interactive;
368 enum boolean dgettext = FALSE;
370 while (p = GET_LINE) {
371 gettext = strstr (p, "gettext");
372 interactive = strstr (p, "(interactive");
373 if (gettext || interactive) {
378 else if (gettext < interactive) {
387 if (p > line && *(p-1) == 'd')
389 p += 7; /* Skip over "gettext" */
391 p += 12; /* Skip over "(interactive" */
394 if (*p != '\"') /* Make sure there is a quoted string */
397 if (dgettext) { /* Skip first quoted string (domain name) */
399 ; /* null statement */
402 if (*p != '\"') /* Check for second quoted string (message) */
406 if (interactive && no_interactive_prompt (p))
409 fprintf (outfile, "gettext(");
410 p = copy_up_to_paren (p);
411 fprintf (outfile, ")\n");
417 /* Assuming p points to some character beyond an opening parenthesis, copy
418 everything to outfile up to but not including the closing parenthesis.
420 char *copy_up_to_paren (register char *p)
423 SKIP_BLANKS (p); /* We don't call skip_blanks() in order to */
424 CHECK_EOL (p); /* preserve blanks at the beginning of the line */
429 p = copy_quoted_string (p);
431 fputc (*p++, outfile);
437 /* Assuming p points to a quote character, copy the quoted string to outfile.
439 char *copy_quoted_string (register char *p)
443 fputc (*p++, outfile);
444 fputc (*p++, outfile);
446 } while (*p != '\"');
448 fputc (*p++, outfile);
453 /* Return TRUE if the interactive specification consists only
454 of code letters and no prompt.
456 enum boolean no_interactive_prompt (register char *q)
458 while (++q, *q == '*' || *q == '@')
459 ; /* null statement */
465 if (*q == '\\' && *++q == 'n') {
467 goto skip_code_letter;
473 char *skip_blanks (register char *p)
475 while (*p == ' ' || *p == '\t' || *p == '\n') {