1 /* Copyright (C) 1998-2022 Free Software Foundation, Inc.
2    This file is part of the GNU C Library.
3 
4    This program is free software; you can redistribute it and/or modify
5    it under the terms of the GNU General Public License as published
6    by the Free Software Foundation; version 2 of the License, or
7    (at your option) any later version.
8 
9    This program is distributed in the hope that it will be useful,
10    but WITHOUT ANY WARRANTY; without even the implied warranty of
11    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
12    GNU General Public License for more details.
13 
14    You should have received a copy of the GNU General Public License
15    along with this program; if not, see <https://www.gnu.org/licenses/>.  */
16 
17 #ifdef HAVE_CONFIG_H
18 # include <config.h>
19 #endif
20 
21 #include <langinfo.h>
22 #include <stdlib.h>
23 #include <string.h>
24 #include <stdint.h>
25 #include <sys/uio.h>
26 
27 #include <assert.h>
28 
29 #include "localedef.h"
30 #include "localeinfo.h"
31 #include "locfile.h"
32 
33 
34 /* The real definition of the struct for the LC_IDENTIFICATION locale.  */
35 struct locale_identification_t
36 {
37   const char *title;
38   const char *source;
39   const char *address;
40   const char *contact;
41   const char *email;
42   const char *tel;
43   const char *fax;
44   const char *language;
45   const char *territory;
46   const char *audience;
47   const char *application;
48   const char *abbreviation;
49   const char *revision;
50   const char *date;
51   const char *category[__LC_LAST];
52 };
53 
54 
55 static const char *category_name[__LC_LAST] =
56 {
57   [LC_CTYPE] = "LC_CTYPE",
58   [LC_NUMERIC] = "LC_NUMERIC",
59   [LC_TIME] = "LC_TIME",
60   [LC_COLLATE] = "LC_COLLATE",
61   [LC_MONETARY] = "LC_MONETARY",
62   [LC_MESSAGES] = "LC_MESSAGES",
63   [LC_ALL] = "LC_ALL",
64   [LC_PAPER] = "LC_PAPER",
65   [LC_NAME] = "LC_NAME",
66   [LC_ADDRESS] = "LC_ADDRESS",
67   [LC_TELEPHONE] = "LC_TELEPHONE",
68   [LC_MEASUREMENT] = "LC_MEASUREMENT",
69   [LC_IDENTIFICATION] = "LC_IDENTIFICATION"
70 };
71 
72 
73 static void
identification_startup(struct linereader * lr,struct localedef_t * locale,int ignore_content)74 identification_startup (struct linereader *lr, struct localedef_t *locale,
75 			int ignore_content)
76 {
77   if (!ignore_content)
78     {
79       locale->categories[LC_IDENTIFICATION].identification =
80 	(struct locale_identification_t *)
81 	xcalloc (1, sizeof (struct locale_identification_t));
82 
83       locale->categories[LC_IDENTIFICATION].identification->category[LC_ALL] =
84 	"";
85     }
86 
87   if (lr != NULL)
88     {
89       lr->translate_strings = 1;
90       lr->return_widestr = 0;
91     }
92 }
93 
94 
95 void
identification_finish(struct localedef_t * locale,const struct charmap_t * charmap)96 identification_finish (struct localedef_t *locale,
97 		       const struct charmap_t *charmap)
98 {
99   struct locale_identification_t *identification
100     = locale->categories[LC_IDENTIFICATION].identification;
101   int nothing = 0;
102   size_t num;
103 
104   /* Now resolve copying and also handle completely missing definitions.  */
105   if (identification == NULL)
106     {
107       /* First see whether we were supposed to copy.  If yes, find the
108 	 actual definition.  */
109       if (locale->copy_name[LC_IDENTIFICATION] != NULL)
110 	{
111 	  /* Find the copying locale.  This has to happen transitively since
112 	     the locale we are copying from might also copying another one.  */
113 	  struct localedef_t *from = locale;
114 
115 	  do
116 	    from = find_locale (LC_IDENTIFICATION,
117 				from->copy_name[LC_IDENTIFICATION],
118 				from->repertoire_name, charmap);
119 	  while (from->categories[LC_IDENTIFICATION].identification == NULL
120 		 && from->copy_name[LC_IDENTIFICATION] != NULL);
121 
122 	  identification = locale->categories[LC_IDENTIFICATION].identification
123 	    = from->categories[LC_IDENTIFICATION].identification;
124 	}
125 
126       /* If there is still no definition issue an warning and create an
127 	 empty one.  */
128       if (identification == NULL)
129 	{
130 	  record_warning (_("\
131 No definition for %s category found"), "LC_IDENTIFICATION");
132 	  identification_startup (NULL, locale, 0);
133 	  identification
134 	    = locale->categories[LC_IDENTIFICATION].identification;
135 	  nothing = 1;
136 	}
137     }
138 
139 #define TEST_ELEM(cat) \
140   if (identification->cat == NULL)					      \
141     {									      \
142       if (verbose && ! nothing)						      \
143 	record_warning (_("%s: field `%s' not defined"), "LC_IDENTIFICATION", \
144 			#cat);						      \
145       identification->cat = "";						      \
146     }
147 
148   TEST_ELEM (title);
149   TEST_ELEM (source);
150   TEST_ELEM (address);
151   TEST_ELEM (contact);
152   TEST_ELEM (email);
153   TEST_ELEM (tel);
154   TEST_ELEM (fax);
155   TEST_ELEM (language);
156   TEST_ELEM (territory);
157   TEST_ELEM (audience);
158   TEST_ELEM (application);
159   TEST_ELEM (abbreviation);
160   TEST_ELEM (revision);
161   TEST_ELEM (date);
162 
163   for (num = 0; num < __LC_LAST; ++num)
164     {
165       /* We don't accept/parse this category, so skip it early.  */
166       if (num == LC_ALL)
167 	continue;
168 
169       if (identification->category[num] == NULL)
170 	{
171 	  if (verbose && ! nothing)
172 	    record_warning (_("\
173 %s: no identification for category `%s'"), "LC_IDENTIFICATION",
174 			    category_name[num]);
175 	  identification->category[num] = "";
176 	}
177       else
178 	{
179 	  /* Only list the standards we care about.  This is based on the
180 	     ISO 30112 WD10 [2014] standard which supersedes all previous
181 	     revisions of the ISO 14652 standard.  */
182 	  static const char * const standards[] =
183 	    {
184 	      "posix:1993",
185 	      "i18n:2004",
186 	      "i18n:2012",
187 	    };
188 	  size_t i;
189 	  bool matched = false;
190 
191 	  for (i = 0; i < sizeof (standards) / sizeof (standards[0]); ++i)
192 	    if (strcmp (identification->category[num], standards[i]) == 0)
193 	      matched = true;
194 
195 	  if (matched != true)
196 	    record_error (0, 0, _("\
197 %s: unknown standard `%s' for category `%s'"),
198 			  "LC_IDENTIFICATION",
199 			  identification->category[num],
200 			  category_name[num]);
201 	}
202     }
203 }
204 
205 
206 void
identification_output(struct localedef_t * locale,const struct charmap_t * charmap,const char * output_path)207 identification_output (struct localedef_t *locale,
208 		       const struct charmap_t *charmap,
209 		       const char *output_path)
210 {
211   struct locale_identification_t *identification
212     = locale->categories[LC_IDENTIFICATION].identification;
213   struct locale_file file;
214   size_t num;
215 
216   init_locale_data (&file, _NL_ITEM_INDEX (_NL_NUM_LC_IDENTIFICATION));
217   add_locale_string (&file, identification->title);
218   add_locale_string (&file, identification->source);
219   add_locale_string (&file, identification->address);
220   add_locale_string (&file, identification->contact);
221   add_locale_string (&file, identification->email);
222   add_locale_string (&file, identification->tel);
223   add_locale_string (&file, identification->fax);
224   add_locale_string (&file, identification->language);
225   add_locale_string (&file, identification->territory);
226   add_locale_string (&file, identification->audience);
227   add_locale_string (&file, identification->application);
228   add_locale_string (&file, identification->abbreviation);
229   add_locale_string (&file, identification->revision);
230   add_locale_string (&file, identification->date);
231   start_locale_structure (&file);
232   for (num = 0; num < __LC_LAST; ++num)
233     if (num != LC_ALL)
234       add_locale_string (&file, identification->category[num]);
235   end_locale_structure (&file);
236   add_locale_string (&file, charmap->code_set_name);
237   write_locale_data (output_path, LC_IDENTIFICATION, "LC_IDENTIFICATION",
238 		     &file);
239 }
240 
241 
242 /* The parser for the LC_IDENTIFICATION section of the locale definition.  */
243 void
identification_read(struct linereader * ldfile,struct localedef_t * result,const struct charmap_t * charmap,const char * repertoire_name,int ignore_content)244 identification_read (struct linereader *ldfile, struct localedef_t *result,
245 	       const struct charmap_t *charmap, const char *repertoire_name,
246 	       int ignore_content)
247 {
248   struct locale_identification_t *identification;
249   struct token *now;
250   struct token *arg;
251   struct token *cattok;
252   int category;
253   enum token_t nowtok;
254 
255   /* The rest of the line containing `LC_IDENTIFICATION' must be free.  */
256   lr_ignore_rest (ldfile, 1);
257 
258   do
259     {
260       now = lr_token (ldfile, charmap, result, NULL, verbose);
261       nowtok = now->tok;
262     }
263   while (nowtok == tok_eol);
264 
265   /* If we see `copy' now we are almost done.  */
266   if (nowtok == tok_copy)
267     {
268       handle_copy (ldfile, charmap, repertoire_name, result,
269 		   tok_lc_identification, LC_IDENTIFICATION,
270 		   "LC_IDENTIFICATION", ignore_content);
271       return;
272     }
273 
274   /* Prepare the data structures.  */
275   identification_startup (ldfile, result, ignore_content);
276   identification = result->categories[LC_IDENTIFICATION].identification;
277 
278   while (1)
279     {
280       /* Of course we don't proceed beyond the end of file.  */
281       if (nowtok == tok_eof)
282 	break;
283 
284       /* Ignore empty lines.  */
285       if (nowtok == tok_eol)
286 	{
287 	  now = lr_token (ldfile, charmap, result, NULL, verbose);
288 	  nowtok = now->tok;
289 	  continue;
290 	}
291 
292       switch (nowtok)
293 	{
294 #define STR_ELEM(cat) \
295 	case tok_##cat:							      \
296 	  /* Ignore the rest of the line if we don't need the input of	      \
297 	     this line.  */						      \
298 	  if (ignore_content)						      \
299 	    {								      \
300 	      lr_ignore_rest (ldfile, 0);				      \
301 	      break;							      \
302 	    }								      \
303 									      \
304 	  arg = lr_token (ldfile, charmap, result, NULL, verbose);	      \
305 	  if (arg->tok != tok_string)					      \
306 	    goto err_label;						      \
307 	  if (identification->cat != NULL)				      \
308 	    lr_error (ldfile, _("\
309 %s: field `%s' declared more than once"), "LC_IDENTIFICATION", #cat);	      \
310 	  else if (!ignore_content && arg->val.str.startmb == NULL)	      \
311 	    {								      \
312 	      lr_error (ldfile, _("\
313 %s: unknown character in field `%s'"), "LC_IDENTIFICATION", #cat);	      \
314 	      identification->cat = "";					      \
315 	    }								      \
316 	  else if (!ignore_content)					      \
317 	    identification->cat = arg->val.str.startmb;			      \
318 	  break
319 
320 	  STR_ELEM (title);
321 	  STR_ELEM (source);
322 	  STR_ELEM (address);
323 	  STR_ELEM (contact);
324 	  STR_ELEM (email);
325 	  STR_ELEM (tel);
326 	  STR_ELEM (fax);
327 	  STR_ELEM (language);
328 	  STR_ELEM (territory);
329 	  STR_ELEM (audience);
330 	  STR_ELEM (application);
331 	  STR_ELEM (abbreviation);
332 	  STR_ELEM (revision);
333 	  STR_ELEM (date);
334 
335 	case tok_category:
336 	  /* Ignore the rest of the line if we don't need the input of
337 	     this line.  */
338 	  if (ignore_content)
339 	    {
340 	      lr_ignore_rest (ldfile, 0);
341 	      break;
342 	    }
343 
344 	  /* We expect two operands.  */
345 	  arg = lr_token (ldfile, charmap, result, NULL, verbose);
346 	  if (arg->tok != tok_string && arg->tok != tok_ident)
347 	    goto err_label;
348 	  /* Next is a semicolon.  */
349 	  cattok = lr_token (ldfile, charmap, result, NULL, verbose);
350 	  if (cattok->tok != tok_semicolon)
351 	    goto err_label;
352 	  /* Now a LC_xxx identifier.  */
353 	  cattok = lr_token (ldfile, charmap, result, NULL, verbose);
354 	  switch (cattok->tok)
355 	    {
356 #define CATEGORY(lname, uname) \
357 	    case tok_lc_##lname:					      \
358 	      category = LC_##uname;					      \
359 	      break
360 
361 	      CATEGORY (identification, IDENTIFICATION);
362 	      CATEGORY (ctype, CTYPE);
363 	      CATEGORY (collate, COLLATE);
364 	      CATEGORY (time, TIME);
365 	      CATEGORY (numeric, NUMERIC);
366 	      CATEGORY (monetary, MONETARY);
367 	      CATEGORY (messages, MESSAGES);
368 	      CATEGORY (paper, PAPER);
369 	      CATEGORY (name, NAME);
370 	      CATEGORY (address, ADDRESS);
371 	      CATEGORY (telephone, TELEPHONE);
372 	      CATEGORY (measurement, MEASUREMENT);
373 
374 	    default:
375 	      goto err_label;
376 	    }
377 	  if (identification->category[category] != NULL)
378 	    {
379 	      lr_error (ldfile, _("\
380 %s: duplicate category version definition"), "LC_IDENTIFICATION");
381 	      free (arg->val.str.startmb);
382 	    }
383 	  else
384 	    identification->category[category] = arg->val.str.startmb;
385 	  break;
386 
387 	case tok_end:
388 	  /* Next we assume `LC_IDENTIFICATION'.  */
389 	  arg = lr_token (ldfile, charmap, result, NULL, verbose);
390 	  if (arg->tok == tok_eof)
391 	    break;
392 	  if (arg->tok == tok_eol)
393 	    lr_error (ldfile, _("%s: incomplete `END' line"),
394 		      "LC_IDENTIFICATION");
395 	  else if (arg->tok != tok_lc_identification)
396 	    lr_error (ldfile, _("\
397 %1$s: definition does not end with `END %1$s'"), "LC_IDENTIFICATION");
398 	  lr_ignore_rest (ldfile, arg->tok == tok_lc_identification);
399 	  return;
400 
401 	default:
402 	err_label:
403 	  SYNTAX_ERROR (_("%s: syntax error"), "LC_IDENTIFICATION");
404 	}
405 
406       /* Prepare for the next round.  */
407       now = lr_token (ldfile, charmap, result, NULL, verbose);
408       nowtok = now->tok;
409     }
410 
411   /* When we come here we reached the end of the file.  */
412   lr_error (ldfile, _("%s: premature end of file"), "LC_IDENTIFICATION");
413 }
414