1 /* librep format strings.
2 Copyright (C) 2001-2004, 2006-2007, 2009, 2019-2020, 2023 Free Software Foundation, Inc.
3 Written by Bruno Haible <haible@clisp.cons.org>, 2001.
4
5 This program is free software: you can redistribute it and/or modify
6 it under the terms of the GNU General Public License as published by
7 the Free Software Foundation; either version 3 of the License, or
8 (at your option) any later version.
9
10 This program is distributed in the hope that it will be useful,
11 but WITHOUT ANY WARRANTY; without even the implied warranty of
12 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 GNU General Public License for more details.
14
15 You should have received a copy of the GNU General Public License
16 along with this program. If not, see <https://www.gnu.org/licenses/>. */
17
18 #ifdef HAVE_CONFIG_H
19 # include <config.h>
20 #endif
21
22 #include <stdbool.h>
23 #include <stdlib.h>
24
25 #include "format.h"
26 #include "c-ctype.h"
27 #include "xalloc.h"
28 #include "xvasprintf.h"
29 #include "format-invalid.h"
30 #include "gettext.h"
31
32 #define _(str) gettext (str)
33
34 /* librep format strings are implemented in librep-0.14/src/streams.c.
35 A directive
36 - starts with '%' or '%m$' where m is a positive integer,
37 - is optionally followed by any of the characters '-', '^', '0', '+', ' ',
38 each of which acts as a flag,
39 - is optionally followed by a width specification: a nonempty digit
40 sequence,
41 - is optionally followed by '.' and a precision specification: a nonempty
42 digit sequence,
43 - is finished by a specifier
44 - '%', that needs no argument,
45 - 'c', that need a character argument,
46 - 'd', 'x', 'X', 'o', that need an integer argument,
47 - 's', that need an argument and prints it using princ,
48 - 'S', that need an argument and prints it using prin1.
49 Numbered ('%m$') and unnumbered argument specifications can be used in the
50 same string. The effect of '%m$' is to set the current argument number to
51 m. The current argument number is incremented after processing a directive.
52 */
53
54 enum format_arg_type
55 {
56 FAT_NONE,
57 FAT_CHARACTER,
58 FAT_INTEGER,
59 FAT_OBJECT_PRETTY,
60 FAT_OBJECT
61 };
62
63 struct numbered_arg
64 {
65 unsigned int number;
66 enum format_arg_type type;
67 };
68
69 struct spec
70 {
71 unsigned int directives;
72 unsigned int numbered_arg_count;
73 struct numbered_arg *numbered;
74 };
75
76 /* Locale independent test for a decimal digit.
77 Argument can be 'char' or 'unsigned char'. (Whereas the argument of
78 <ctype.h> isdigit must be an 'unsigned char'.) */
79 #undef isdigit
80 #define isdigit(c) ((unsigned int) ((c) - '0') < 10)
81
82
83 static int
84 numbered_arg_compare (const void *p1, const void *p2)
85 {
86 unsigned int n1 = ((const struct numbered_arg *) p1)->number;
87 unsigned int n2 = ((const struct numbered_arg *) p2)->number;
88
89 return (n1 > n2 ? 1 : n1 < n2 ? -1 : 0);
90 }
91
92 static void *
93 format_parse (const char *format, bool translated, char *fdi,
94 char **invalid_reason)
95 {
96 const char *const format_start = format;
97 struct spec spec;
98 unsigned int numbered_allocated;
99 struct spec *result;
100 unsigned int number;
101
102 spec.directives = 0;
103 spec.numbered_arg_count = 0;
104 spec.numbered = NULL;
105 numbered_allocated = 0;
106 number = 1;
107
108 for (; *format != '\0';)
109 if (*format++ == '%')
110 {
111 /* A directive. */
112 enum format_arg_type type;
113
114 FDI_SET (format - 1, FMTDIR_START);
115 spec.directives++;
116
117 if (isdigit (*format))
118 {
119 const char *f = format;
120 unsigned int m = 0;
121
122 do
123 {
124 m = 10 * m + (*f - '0');
125 f++;
126 }
127 while (isdigit (*f));
128
129 if (*f == '$' && m > 0)
130 {
131 number = m;
132 format = ++f;
133 }
134 }
135
136 /* Parse flags. */
137 while (*format == '-' || *format == '^' || *format == '0'
138 || *format == '+' || *format == ' ')
139 format++;
140
141 /* Parse width. */
142 if (isdigit (*format))
143 {
144 do format++; while (isdigit (*format));
145 }
146
147 /* Parse precision. */
148 if (*format == '.')
149 {
150 format++;
151
152 if (isdigit (*format))
153 {
154 do format++; while (isdigit (*format));
155 }
156 }
157
158 switch (*format)
159 {
160 case '%':
161 type = FAT_NONE;
162 break;
163 case 'c':
164 type = FAT_CHARACTER;
165 break;
166 case 'd': case 'x': case 'X': case 'o':
167 type = FAT_INTEGER;
168 break;
169 case 's':
170 type = FAT_OBJECT_PRETTY;
171 break;
172 case 'S':
173 type = FAT_OBJECT;
174 break;
175 default:
176 if (*format == '\0')
177 {
178 *invalid_reason = INVALID_UNTERMINATED_DIRECTIVE ();
179 FDI_SET (format - 1, FMTDIR_ERROR);
180 }
181 else
182 {
183 *invalid_reason =
184 INVALID_CONVERSION_SPECIFIER (spec.directives, *format);
185 FDI_SET (format, FMTDIR_ERROR);
186 }
187 goto bad_format;
188 }
189
190 if (type != FAT_NONE)
191 {
192 if (numbered_allocated == spec.numbered_arg_count)
193 {
194 numbered_allocated = 2 * numbered_allocated + 1;
195 spec.numbered = (struct numbered_arg *) xrealloc (spec.numbered, numbered_allocated * sizeof (struct numbered_arg));
196 }
197 spec.numbered[spec.numbered_arg_count].number = number;
198 spec.numbered[spec.numbered_arg_count].type = type;
199 spec.numbered_arg_count++;
200
201 number++;
202 }
203
204 FDI_SET (format, FMTDIR_END);
205
206 format++;
207 }
208
209 /* Sort the numbered argument array, and eliminate duplicates. */
210 if (spec.numbered_arg_count > 1)
211 {
212 unsigned int i, j;
213 bool err;
214
215 qsort (spec.numbered, spec.numbered_arg_count,
216 sizeof (struct numbered_arg), numbered_arg_compare);
217
218 /* Remove duplicates: Copy from i to j, keeping 0 <= j <= i. */
219 err = false;
220 for (i = j = 0; i < spec.numbered_arg_count; i++)
221 if (j > 0 && spec.numbered[i].number == spec.numbered[j-1].number)
222 {
223 enum format_arg_type type1 = spec.numbered[i].type;
224 enum format_arg_type type2 = spec.numbered[j-1].type;
225 enum format_arg_type type_both;
226
227 if (type1 == type2)
228 type_both = type1;
229 else
230 {
231 /* Incompatible types. */
232 type_both = FAT_NONE;
233 if (!err)
234 *invalid_reason =
235 INVALID_INCOMPATIBLE_ARG_TYPES (spec.numbered[i].number);
236 err = true;
237 }
238
239 spec.numbered[j-1].type = type_both;
240 }
241 else
242 {
243 if (j < i)
244 {
245 spec.numbered[j].number = spec.numbered[i].number;
246 spec.numbered[j].type = spec.numbered[i].type;
247 }
248 j++;
249 }
250 spec.numbered_arg_count = j;
251 if (err)
252 /* *invalid_reason has already been set above. */
253 goto bad_format;
254 }
255
256 result = XMALLOC (struct spec);
257 *result = spec;
258 return result;
259
260 bad_format:
261 if (spec.numbered != NULL)
262 free (spec.numbered);
263 return NULL;
264 }
265
266 static void
267 format_free (void *descr)
268 {
269 struct spec *spec = (struct spec *) descr;
270
271 if (spec->numbered != NULL)
272 free (spec->numbered);
273 free (spec);
274 }
275
276 static int
277 format_get_number_of_directives (void *descr)
278 {
279 struct spec *spec = (struct spec *) descr;
280
281 return spec->directives;
282 }
283
284 static bool
285 format_check (void *msgid_descr, void *msgstr_descr, bool equality,
286 formatstring_error_logger_t error_logger,
287 const char *pretty_msgid, const char *pretty_msgstr)
288 {
289 struct spec *spec1 = (struct spec *) msgid_descr;
290 struct spec *spec2 = (struct spec *) msgstr_descr;
291 bool err = false;
292
293 if (spec1->numbered_arg_count + spec2->numbered_arg_count > 0)
294 {
295 unsigned int i, j;
296 unsigned int n1 = spec1->numbered_arg_count;
297 unsigned int n2 = spec2->numbered_arg_count;
298
299 /* Check that the argument numbers are the same.
300 Both arrays are sorted. We search for the first difference. */
301 for (i = 0, j = 0; i < n1 || j < n2; )
302 {
303 int cmp = (i >= n1 ? 1 :
304 j >= n2 ? -1 :
305 spec1->numbered[i].number > spec2->numbered[j].number ? 1 :
306 spec1->numbered[i].number < spec2->numbered[j].number ? -1 :
307 0);
308
309 if (cmp > 0)
310 {
311 if (error_logger)
312 error_logger (_("a format specification for argument %u, as in '%s', doesn't exist in '%s'"),
313 spec2->numbered[j].number, pretty_msgstr,
314 pretty_msgid);
315 err = true;
316 break;
317 }
318 else if (cmp < 0)
319 {
320 if (equality)
321 {
322 if (error_logger)
323 error_logger (_("a format specification for argument %u doesn't exist in '%s'"),
324 spec1->numbered[i].number, pretty_msgstr);
325 err = true;
326 break;
327 }
328 else
329 i++;
330 }
331 else
332 j++, i++;
333 }
334 /* Check the argument types are the same. */
335 if (!err)
336 for (i = 0, j = 0; j < n2; )
337 {
338 if (spec1->numbered[i].number == spec2->numbered[j].number)
339 {
340 if (spec1->numbered[i].type != spec2->numbered[j].type)
341 {
342 if (error_logger)
343 error_logger (_("format specifications in '%s' and '%s' for argument %u are not the same"),
344 pretty_msgid, pretty_msgstr,
345 spec2->numbered[j].number);
346 err = true;
347 break;
348 }
349 j++, i++;
350 }
351 else
352 i++;
353 }
354 }
355
356 return err;
357 }
358
359
360 struct formatstring_parser formatstring_librep =
361 {
362 format_parse,
363 format_free,
364 format_get_number_of_directives,
365 NULL,
366 format_check
367 };
368
369
370 #ifdef TEST
371
372 /* Test program: Print the argument list specification returned by
373 format_parse for strings read from standard input. */
374
375 #include <stdio.h>
376
377 static void
378 format_print (void *descr)
379 {
380 struct spec *spec = (struct spec *) descr;
381 unsigned int last;
382 unsigned int i;
383
384 if (spec == NULL)
385 {
386 printf ("INVALID");
387 return;
388 }
389
390 printf ("(");
391 last = 1;
392 for (i = 0; i < spec->numbered_arg_count; i++)
393 {
394 unsigned int number = spec->numbered[i].number;
395
396 if (i > 0)
397 printf (" ");
398 if (number < last)
399 abort ();
400 for (; last < number; last++)
401 printf ("_ ");
402 switch (spec->numbered[i].type)
403 {
404 case FAT_CHARACTER:
405 printf ("c");
406 break;
407 case FAT_INTEGER:
408 printf ("i");
409 break;
410 case FAT_OBJECT_PRETTY:
411 printf ("s");
412 break;
413 case FAT_OBJECT:
414 printf ("*");
415 break;
416 default:
417 abort ();
418 }
419 last = number + 1;
420 }
421 printf (")");
422 }
423
424 int
425 main ()
426 {
427 for (;;)
428 {
429 char *line = NULL;
430 size_t line_size = 0;
431 int line_len;
432 char *invalid_reason;
433 void *descr;
434
435 line_len = getline (&line, &line_size, stdin);
436 if (line_len < 0)
437 break;
438 if (line_len > 0 && line[line_len - 1] == '\n')
439 line[--line_len] = '\0';
440
441 invalid_reason = NULL;
442 descr = format_parse (line, false, NULL, &invalid_reason);
443
444 format_print (descr);
445 printf ("\n");
446 if (descr == NULL)
447 printf ("%s\n", invalid_reason);
448
449 free (invalid_reason);
450 free (line);
451 }
452
453 return 0;
454 }
455
456 /*
457 * For Emacs M-x compile
458 * Local Variables:
459 * compile-command: "/bin/sh ../libtool --tag=CC --mode=link gcc -o a.out -static -O -g -Wall -I.. -I../gnulib-lib -I../../gettext-runtime/intl -DHAVE_CONFIG_H -DTEST format-librep.c ../gnulib-lib/libgettextlib.la"
460 * End:
461 */
462
463 #endif /* TEST */