1 /* Emacs Lisp format strings.
2 Copyright (C) 2001-2004, 2006-2007, 2009, 2019-2020 Free Software Foundation, Inc.
3 Written by Bruno Haible <haible@clisp.cons.org>, 2002.
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 /* Emacs Lisp format strings are implemented in emacs-21.1/src/editfns.c,
35 xemacs-21.1.14/src/editfns.c and xemacs-21.1.14/src/doprnt.c.
36 A directive
37 - starts with '%' or '%m$' where m is a positive integer,
38 - is optionally followed by any of the characters '#', '0', '-', ' ', '+',
39 each of which acts as a flag,
40 - is optionally followed by a width specification: '*' (reads an argument)
41 or a nonempty digit sequence,
42 - is optionally followed by '.' and a precision specification: '*' (reads
43 an argument) or a nonempty digit sequence,
44 - is finished by a specifier
45 - '%', that needs no argument,
46 - 'c', that need a character argument,
47 - 'd', 'i', 'x', 'X', 'o', that need an integer argument,
48 - 'e', 'E', 'f', 'g', 'G', that need a floating-point argument,
49 - 's', that need an argument and prints it using princ,
50 - 'S', that need an argument and prints it using prin1.
51 Numbered ('%m$') and unnumbered argument specifications can be used in the
52 same string. The effect of '%m$' is to set the current argument number to
53 m. The current argument number is incremented after processing a directive.
54 */
55
56 enum format_arg_type
57 {
58 FAT_NONE,
59 FAT_CHARACTER,
60 FAT_INTEGER,
61 FAT_FLOAT,
62 FAT_OBJECT_PRETTY,
63 FAT_OBJECT
64 };
65
66 struct numbered_arg
67 {
68 unsigned int number;
69 enum format_arg_type type;
70 };
71
72 struct spec
73 {
74 unsigned int directives;
75 unsigned int numbered_arg_count;
76 struct numbered_arg *numbered;
77 };
78
79 /* Locale independent test for a decimal digit.
80 Argument can be 'char' or 'unsigned char'. (Whereas the argument of
81 <ctype.h> isdigit must be an 'unsigned char'.) */
82 #undef isdigit
83 #define isdigit(c) ((unsigned int) ((c) - '0') < 10)
84
85
86 static int
87 numbered_arg_compare (const void *p1, const void *p2)
88 {
89 unsigned int n1 = ((const struct numbered_arg *) p1)->number;
90 unsigned int n2 = ((const struct numbered_arg *) p2)->number;
91
92 return (n1 > n2 ? 1 : n1 < n2 ? -1 : 0);
93 }
94
95 static void *
96 format_parse (const char *format, bool translated, char *fdi,
97 char **invalid_reason)
98 {
99 const char *const format_start = format;
100 struct spec spec;
101 unsigned int numbered_allocated;
102 struct spec *result;
103 unsigned int number;
104
105 spec.directives = 0;
106 spec.numbered_arg_count = 0;
107 spec.numbered = NULL;
108 numbered_allocated = 0;
109 number = 1;
110
111 for (; *format != '\0';)
112 if (*format++ == '%')
113 {
114 /* A directive. */
115 enum format_arg_type type;
116
117 FDI_SET (format - 1, FMTDIR_START);
118 spec.directives++;
119
120 if (isdigit (*format))
121 {
122 const char *f = format;
123 unsigned int m = 0;
124
125 do
126 {
127 m = 10 * m + (*f - '0');
128 f++;
129 }
130 while (isdigit (*f));
131
132 if (*f == '$' && m > 0)
133 {
134 number = m;
135 format = ++f;
136 }
137 }
138
139 /* Parse flags. */
140 while (*format == ' ' || *format == '+' || *format == '-'
141 || *format == '#' || *format == '0')
142 format++;
143
144 /* Parse width. */
145 if (*format == '*')
146 {
147 format++;
148
149 if (numbered_allocated == spec.numbered_arg_count)
150 {
151 numbered_allocated = 2 * numbered_allocated + 1;
152 spec.numbered = (struct numbered_arg *) xrealloc (spec.numbered, numbered_allocated * sizeof (struct numbered_arg));
153 }
154 spec.numbered[spec.numbered_arg_count].number = number;
155 spec.numbered[spec.numbered_arg_count].type = FAT_INTEGER;
156 spec.numbered_arg_count++;
157
158 number++;
159 }
160 else if (isdigit (*format))
161 {
162 do format++; while (isdigit (*format));
163 }
164
165 /* Parse precision. */
166 if (*format == '.')
167 {
168 format++;
169
170 if (*format == '*')
171 {
172 format++;
173
174 if (numbered_allocated == spec.numbered_arg_count)
175 {
176 numbered_allocated = 2 * numbered_allocated + 1;
177 spec.numbered = (struct numbered_arg *) xrealloc (spec.numbered, numbered_allocated * sizeof (struct numbered_arg));
178 }
179 spec.numbered[spec.numbered_arg_count].number = number;
180 spec.numbered[spec.numbered_arg_count].type = FAT_INTEGER;
181 spec.numbered_arg_count++;
182
183 number++;
184 }
185 else if (isdigit (*format))
186 {
187 do format++; while (isdigit (*format));
188 }
189 }
190
191 switch (*format)
192 {
193 case '%':
194 type = FAT_NONE;
195 break;
196 case 'c':
197 type = FAT_CHARACTER;
198 break;
199 case 'd': case 'i': case 'x': case 'X': case 'o':
200 type = FAT_INTEGER;
201 break;
202 case 'e': case 'E': case 'f': case 'g': case 'G':
203 type = FAT_FLOAT;
204 break;
205 case 's':
206 type = FAT_OBJECT_PRETTY;
207 break;
208 case 'S':
209 type = FAT_OBJECT;
210 break;
211 default:
212 if (*format == '\0')
213 {
214 *invalid_reason = INVALID_UNTERMINATED_DIRECTIVE ();
215 FDI_SET (format - 1, FMTDIR_ERROR);
216 }
217 else
218 {
219 *invalid_reason =
220 INVALID_CONVERSION_SPECIFIER (spec.directives, *format);
221 FDI_SET (format, FMTDIR_ERROR);
222 }
223 goto bad_format;
224 }
225
226 if (type != FAT_NONE)
227 {
228 if (numbered_allocated == spec.numbered_arg_count)
229 {
230 numbered_allocated = 2 * numbered_allocated + 1;
231 spec.numbered = (struct numbered_arg *) xrealloc (spec.numbered, numbered_allocated * sizeof (struct numbered_arg));
232 }
233 spec.numbered[spec.numbered_arg_count].number = number;
234 spec.numbered[spec.numbered_arg_count].type = type;
235 spec.numbered_arg_count++;
236
237 number++;
238 }
239
240 FDI_SET (format, FMTDIR_END);
241
242 format++;
243 }
244
245 /* Sort the numbered argument array, and eliminate duplicates. */
246 if (spec.numbered_arg_count > 1)
247 {
248 unsigned int i, j;
249 bool err;
250
251 qsort (spec.numbered, spec.numbered_arg_count,
252 sizeof (struct numbered_arg), numbered_arg_compare);
253
254 /* Remove duplicates: Copy from i to j, keeping 0 <= j <= i. */
255 err = false;
256 for (i = j = 0; i < spec.numbered_arg_count; i++)
257 if (j > 0 && spec.numbered[i].number == spec.numbered[j-1].number)
258 {
259 enum format_arg_type type1 = spec.numbered[i].type;
260 enum format_arg_type type2 = spec.numbered[j-1].type;
261 enum format_arg_type type_both;
262
263 if (type1 == type2)
264 type_both = type1;
265 else
266 {
267 /* Incompatible types. */
268 type_both = FAT_NONE;
269 if (!err)
270 *invalid_reason =
271 INVALID_INCOMPATIBLE_ARG_TYPES (spec.numbered[i].number);
272 err = true;
273 }
274
275 spec.numbered[j-1].type = type_both;
276 }
277 else
278 {
279 if (j < i)
280 {
281 spec.numbered[j].number = spec.numbered[i].number;
282 spec.numbered[j].type = spec.numbered[i].type;
283 }
284 j++;
285 }
286 spec.numbered_arg_count = j;
287 if (err)
288 /* *invalid_reason has already been set above. */
289 goto bad_format;
290 }
291
292 result = XMALLOC (struct spec);
293 *result = spec;
294 return result;
295
296 bad_format:
297 if (spec.numbered != NULL)
298 free (spec.numbered);
299 return NULL;
300 }
301
302 static void
303 format_free (void *descr)
304 {
305 struct spec *spec = (struct spec *) descr;
306
307 if (spec->numbered != NULL)
308 free (spec->numbered);
309 free (spec);
310 }
311
312 static int
313 format_get_number_of_directives (void *descr)
314 {
315 struct spec *spec = (struct spec *) descr;
316
317 return spec->directives;
318 }
319
320 static bool
321 format_check (void *msgid_descr, void *msgstr_descr, bool equality,
322 formatstring_error_logger_t error_logger,
323 const char *pretty_msgid, const char *pretty_msgstr)
324 {
325 struct spec *spec1 = (struct spec *) msgid_descr;
326 struct spec *spec2 = (struct spec *) msgstr_descr;
327 bool err = false;
328
329 if (spec1->numbered_arg_count + spec2->numbered_arg_count > 0)
330 {
331 unsigned int i, j;
332 unsigned int n1 = spec1->numbered_arg_count;
333 unsigned int n2 = spec2->numbered_arg_count;
334
335 /* Check that the argument numbers are the same.
336 Both arrays are sorted. We search for the first difference. */
337 for (i = 0, j = 0; i < n1 || j < n2; )
338 {
339 int cmp = (i >= n1 ? 1 :
340 j >= n2 ? -1 :
341 spec1->numbered[i].number > spec2->numbered[j].number ? 1 :
342 spec1->numbered[i].number < spec2->numbered[j].number ? -1 :
343 0);
344
345 if (cmp > 0)
346 {
347 if (error_logger)
348 error_logger (_("a format specification for argument %u, as in '%s', doesn't exist in '%s'"),
349 spec2->numbered[j].number, pretty_msgstr,
350 pretty_msgid);
351 err = true;
352 break;
353 }
354 else if (cmp < 0)
355 {
356 if (equality)
357 {
358 if (error_logger)
359 error_logger (_("a format specification for argument %u doesn't exist in '%s'"),
360 spec1->numbered[i].number, pretty_msgstr);
361 err = true;
362 break;
363 }
364 else
365 i++;
366 }
367 else
368 j++, i++;
369 }
370 /* Check the argument types are the same. */
371 if (!err)
372 for (i = 0, j = 0; j < n2; )
373 {
374 if (spec1->numbered[i].number == spec2->numbered[j].number)
375 {
376 if (spec1->numbered[i].type != spec2->numbered[j].type)
377 {
378 if (error_logger)
379 error_logger (_("format specifications in '%s' and '%s' for argument %u are not the same"),
380 pretty_msgid, pretty_msgstr,
381 spec2->numbered[j].number);
382 err = true;
383 break;
384 }
385 j++, i++;
386 }
387 else
388 i++;
389 }
390 }
391
392 return err;
393 }
394
395
396 struct formatstring_parser formatstring_elisp =
397 {
398 format_parse,
399 format_free,
400 format_get_number_of_directives,
401 NULL,
402 format_check
403 };
404
405
406 #ifdef TEST
407
408 /* Test program: Print the argument list specification returned by
409 format_parse for strings read from standard input. */
410
411 #include <stdio.h>
412
413 static void
414 format_print (void *descr)
415 {
416 struct spec *spec = (struct spec *) descr;
417 unsigned int last;
418 unsigned int i;
419
420 if (spec == NULL)
421 {
422 printf ("INVALID");
423 return;
424 }
425
426 printf ("(");
427 last = 1;
428 for (i = 0; i < spec->numbered_arg_count; i++)
429 {
430 unsigned int number = spec->numbered[i].number;
431
432 if (i > 0)
433 printf (" ");
434 if (number < last)
435 abort ();
436 for (; last < number; last++)
437 printf ("_ ");
438 switch (spec->numbered[i].type)
439 {
440 case FAT_CHARACTER:
441 printf ("c");
442 break;
443 case FAT_INTEGER:
444 printf ("i");
445 break;
446 case FAT_FLOAT:
447 printf ("f");
448 break;
449 case FAT_OBJECT_PRETTY:
450 printf ("s");
451 break;
452 case FAT_OBJECT:
453 printf ("*");
454 break;
455 default:
456 abort ();
457 }
458 last = number + 1;
459 }
460 printf (")");
461 }
462
463 int
464 main ()
465 {
466 for (;;)
467 {
468 char *line = NULL;
469 size_t line_size = 0;
470 int line_len;
471 char *invalid_reason;
472 void *descr;
473
474 line_len = getline (&line, &line_size, stdin);
475 if (line_len < 0)
476 break;
477 if (line_len > 0 && line[line_len - 1] == '\n')
478 line[--line_len] = '\0';
479
480 invalid_reason = NULL;
481 descr = format_parse (line, false, NULL, &invalid_reason);
482
483 format_print (descr);
484 printf ("\n");
485 if (descr == NULL)
486 printf ("%s\n", invalid_reason);
487
488 free (invalid_reason);
489 free (line);
490 }
491
492 return 0;
493 }
494
495 /*
496 * For Emacs M-x compile
497 * Local Variables:
498 * 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-elisp.c ../gnulib-lib/libgettextlib.la"
499 * End:
500 */
501
502 #endif /* TEST */