]> git.cameronkatri.com Git - mandoc.git/blob - man_validate.c
Support translating Greek characters in eqn.
[mandoc.git] / man_validate.c
1 /* $Id: man_validate.c,v 1.70 2011/07/08 09:34:06 kristaps Exp $ */
2 /*
3 * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2010 Ingo Schwarze <schwarze@openbsd.org>
5 *
6 * Permission to use, copy, modify, and distribute this software for any
7 * purpose with or without fee is hereby granted, provided that the above
8 * copyright notice and this permission notice appear in all copies.
9 *
10 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
11 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
12 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
13 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
14 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
15 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
16 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
17 */
18 #ifdef HAVE_CONFIG_H
19 #include "config.h"
20 #endif
21
22 #include <sys/types.h>
23
24 #include <assert.h>
25 #include <ctype.h>
26 #include <errno.h>
27 #include <limits.h>
28 #include <stdarg.h>
29 #include <stdlib.h>
30 #include <string.h>
31 #include <time.h>
32
33 #include "man.h"
34 #include "mandoc.h"
35 #include "libman.h"
36 #include "libmandoc.h"
37
38 #define CHKARGS struct man *m, struct man_node *n
39
40 typedef int (*v_check)(CHKARGS);
41
42 struct man_valid {
43 v_check *pres;
44 v_check *posts;
45 };
46
47 static int check_bline(CHKARGS);
48 static int check_eq0(CHKARGS);
49 static int check_ft(CHKARGS);
50 static int check_le1(CHKARGS);
51 static int check_ge2(CHKARGS);
52 static int check_le5(CHKARGS);
53 static int check_par(CHKARGS);
54 static int check_part(CHKARGS);
55 static int check_root(CHKARGS);
56 static int check_sec(CHKARGS);
57 static void check_text(CHKARGS);
58
59 static int post_AT(CHKARGS);
60 static int post_vs(CHKARGS);
61 static int post_fi(CHKARGS);
62 static int post_nf(CHKARGS);
63 static int post_TH(CHKARGS);
64 static int post_UC(CHKARGS);
65
66 static v_check posts_at[] = { post_AT, NULL };
67 static v_check posts_br[] = { post_vs, check_eq0, NULL };
68 static v_check posts_eq0[] = { check_eq0, NULL };
69 static v_check posts_fi[] = { check_eq0, post_fi, NULL };
70 static v_check posts_ft[] = { check_ft, NULL };
71 static v_check posts_nf[] = { check_eq0, post_nf, NULL };
72 static v_check posts_par[] = { check_par, NULL };
73 static v_check posts_part[] = { check_part, NULL };
74 static v_check posts_sec[] = { check_sec, NULL };
75 static v_check posts_sp[] = { post_vs, check_le1, NULL };
76 static v_check posts_th[] = { check_ge2, check_le5, post_TH, NULL };
77 static v_check posts_uc[] = { post_UC, NULL };
78 static v_check pres_bline[] = { check_bline, NULL };
79
80 static const struct man_valid man_valids[MAN_MAX] = {
81 { NULL, posts_br }, /* br */
82 { pres_bline, posts_th }, /* TH */
83 { pres_bline, posts_sec }, /* SH */
84 { pres_bline, posts_sec }, /* SS */
85 { pres_bline, NULL }, /* TP */
86 { pres_bline, posts_par }, /* LP */
87 { pres_bline, posts_par }, /* PP */
88 { pres_bline, posts_par }, /* P */
89 { pres_bline, NULL }, /* IP */
90 { pres_bline, NULL }, /* HP */
91 { NULL, NULL }, /* SM */
92 { NULL, NULL }, /* SB */
93 { NULL, NULL }, /* BI */
94 { NULL, NULL }, /* IB */
95 { NULL, NULL }, /* BR */
96 { NULL, NULL }, /* RB */
97 { NULL, NULL }, /* R */
98 { NULL, NULL }, /* B */
99 { NULL, NULL }, /* I */
100 { NULL, NULL }, /* IR */
101 { NULL, NULL }, /* RI */
102 { NULL, posts_eq0 }, /* na */ /* FIXME: should warn only. */
103 { NULL, posts_sp }, /* sp */ /* FIXME: should warn only. */
104 { pres_bline, posts_nf }, /* nf */
105 { pres_bline, posts_fi }, /* fi */
106 { NULL, NULL }, /* RE */
107 { NULL, posts_part }, /* RS */
108 { NULL, NULL }, /* DT */
109 { NULL, posts_uc }, /* UC */
110 { NULL, NULL }, /* PD */
111 { NULL, posts_at }, /* AT */
112 { NULL, NULL }, /* in */
113 { NULL, posts_ft }, /* ft */
114 };
115
116
117 int
118 man_valid_pre(struct man *m, struct man_node *n)
119 {
120 v_check *cp;
121
122 switch (n->type) {
123 case (MAN_TEXT):
124 /* FALLTHROUGH */
125 case (MAN_ROOT):
126 /* FALLTHROUGH */
127 case (MAN_EQN):
128 /* FALLTHROUGH */
129 case (MAN_TBL):
130 return(1);
131 default:
132 break;
133 }
134
135 if (NULL == (cp = man_valids[n->tok].pres))
136 return(1);
137 for ( ; *cp; cp++)
138 if ( ! (*cp)(m, n))
139 return(0);
140 return(1);
141 }
142
143
144 int
145 man_valid_post(struct man *m)
146 {
147 v_check *cp;
148
149 if (MAN_VALID & m->last->flags)
150 return(1);
151 m->last->flags |= MAN_VALID;
152
153 switch (m->last->type) {
154 case (MAN_TEXT):
155 check_text(m, m->last);
156 return(1);
157 case (MAN_ROOT):
158 return(check_root(m, m->last));
159 case (MAN_EQN):
160 /* FALLTHROUGH */
161 case (MAN_TBL):
162 return(1);
163 default:
164 break;
165 }
166
167 if (NULL == (cp = man_valids[m->last->tok].posts))
168 return(1);
169 for ( ; *cp; cp++)
170 if ( ! (*cp)(m, m->last))
171 return(0);
172
173 return(1);
174 }
175
176
177 static int
178 check_root(CHKARGS)
179 {
180
181 if (MAN_BLINE & m->flags)
182 man_nmsg(m, n, MANDOCERR_SCOPEEXIT);
183 else if (MAN_ELINE & m->flags)
184 man_nmsg(m, n, MANDOCERR_SCOPEEXIT);
185
186 m->flags &= ~MAN_BLINE;
187 m->flags &= ~MAN_ELINE;
188
189 if (NULL == m->first->child) {
190 man_nmsg(m, n, MANDOCERR_NODOCBODY);
191 return(0);
192 } else if (NULL == m->meta.title) {
193 man_nmsg(m, n, MANDOCERR_NOTITLE);
194
195 /*
196 * If a title hasn't been set, do so now (by
197 * implication, date and section also aren't set).
198 */
199
200 m->meta.title = mandoc_strdup("unknown");
201 m->meta.msec = mandoc_strdup("1");
202 m->meta.date = mandoc_normdate
203 (m->parse, NULL, n->line, n->pos);
204 }
205
206 return(1);
207 }
208
209 static void
210 check_text(CHKARGS)
211 {
212 char *p, *pp, *cpp;
213 int pos;
214 size_t sz;
215
216 p = n->string;
217 pos = n->pos + 1;
218
219 while ('\0' != *p) {
220 sz = strcspn(p, "\t\\");
221
222 p += (int)sz;
223 pos += (int)sz;
224
225 if ('\t' == *p) {
226 if ( ! (MAN_LITERAL & m->flags))
227 man_pmsg(m, n->line, pos, MANDOCERR_BADTAB);
228 p++;
229 pos++;
230 continue;
231 } else if ('\0' == *p)
232 break;
233
234 pos++;
235 pp = ++p;
236
237 if (ESCAPE_ERROR == mandoc_escape
238 ((const char **)&pp, NULL, NULL)) {
239 man_pmsg(m, n->line, pos, MANDOCERR_BADESCAPE);
240 break;
241 }
242
243 cpp = p;
244 while (NULL != (cpp = memchr(cpp, ASCII_HYPH, pp - cpp)))
245 *cpp = '-';
246
247 pos += pp - p;
248 p = pp;
249 }
250 }
251
252 #define INEQ_DEFINE(x, ineq, name) \
253 static int \
254 check_##name(CHKARGS) \
255 { \
256 if (n->nchild ineq (x)) \
257 return(1); \
258 mandoc_vmsg(MANDOCERR_ARGCOUNT, m->parse, n->line, n->pos, \
259 "line arguments %s %d (have %d)", \
260 #ineq, (x), n->nchild); \
261 return(1); \
262 }
263
264 INEQ_DEFINE(0, ==, eq0)
265 INEQ_DEFINE(1, <=, le1)
266 INEQ_DEFINE(2, >=, ge2)
267 INEQ_DEFINE(5, <=, le5)
268
269 static int
270 check_ft(CHKARGS)
271 {
272 char *cp;
273 int ok;
274
275 if (0 == n->nchild)
276 return(1);
277
278 ok = 0;
279 cp = n->child->string;
280 switch (*cp) {
281 case ('1'):
282 /* FALLTHROUGH */
283 case ('2'):
284 /* FALLTHROUGH */
285 case ('3'):
286 /* FALLTHROUGH */
287 case ('4'):
288 /* FALLTHROUGH */
289 case ('I'):
290 /* FALLTHROUGH */
291 case ('P'):
292 /* FALLTHROUGH */
293 case ('R'):
294 if ('\0' == cp[1])
295 ok = 1;
296 break;
297 case ('B'):
298 if ('\0' == cp[1] || ('I' == cp[1] && '\0' == cp[2]))
299 ok = 1;
300 break;
301 case ('C'):
302 if ('W' == cp[1] && '\0' == cp[2])
303 ok = 1;
304 break;
305 default:
306 break;
307 }
308
309 if (0 == ok) {
310 mandoc_vmsg
311 (MANDOCERR_BADFONT, m->parse,
312 n->line, n->pos, "%s", cp);
313 *cp = '\0';
314 }
315
316 if (1 < n->nchild)
317 mandoc_vmsg
318 (MANDOCERR_ARGCOUNT, m->parse, n->line,
319 n->pos, "want one child (have %d)",
320 n->nchild);
321
322 return(1);
323 }
324
325 static int
326 check_sec(CHKARGS)
327 {
328
329 if ( ! (MAN_HEAD == n->type && 0 == n->nchild))
330 return(1);
331
332 man_nmsg(m, n, MANDOCERR_SYNTARGCOUNT);
333 return(0);
334 }
335
336
337 static int
338 check_part(CHKARGS)
339 {
340
341 if (MAN_BODY == n->type && 0 == n->nchild)
342 mandoc_msg(MANDOCERR_ARGCWARN, m->parse, n->line,
343 n->pos, "want children (have none)");
344
345 return(1);
346 }
347
348
349 static int
350 check_par(CHKARGS)
351 {
352
353 switch (n->type) {
354 case (MAN_BLOCK):
355 if (0 == n->body->nchild)
356 man_node_delete(m, n);
357 break;
358 case (MAN_BODY):
359 if (0 == n->nchild)
360 man_nmsg(m, n, MANDOCERR_IGNPAR);
361 break;
362 case (MAN_HEAD):
363 if (n->nchild)
364 man_nmsg(m, n, MANDOCERR_ARGSLOST);
365 break;
366 default:
367 break;
368 }
369
370 return(1);
371 }
372
373
374 static int
375 check_bline(CHKARGS)
376 {
377
378 assert( ! (MAN_ELINE & m->flags));
379 if (MAN_BLINE & m->flags) {
380 man_nmsg(m, n, MANDOCERR_SYNTLINESCOPE);
381 return(0);
382 }
383
384 return(1);
385 }
386
387 static int
388 post_TH(CHKARGS)
389 {
390 const char *p;
391 int line, pos;
392
393 if (m->meta.title)
394 free(m->meta.title);
395 if (m->meta.vol)
396 free(m->meta.vol);
397 if (m->meta.source)
398 free(m->meta.source);
399 if (m->meta.msec)
400 free(m->meta.msec);
401 if (m->meta.date)
402 free(m->meta.date);
403
404 line = n->line;
405 pos = n->pos;
406 m->meta.title = m->meta.vol = m->meta.date =
407 m->meta.msec = m->meta.source = NULL;
408
409 /* ->TITLE<- MSEC DATE SOURCE VOL */
410
411 n = n->child;
412 if (n && n->string) {
413 for (p = n->string; '\0' != *p; p++) {
414 /* Only warn about this once... */
415 if (isalpha((u_char)*p) && ! isupper((u_char)*p)) {
416 man_nmsg(m, n, MANDOCERR_UPPERCASE);
417 break;
418 }
419 }
420 m->meta.title = mandoc_strdup(n->string);
421 } else
422 m->meta.title = mandoc_strdup("");
423
424 /* TITLE ->MSEC<- DATE SOURCE VOL */
425
426 if (n)
427 n = n->next;
428 if (n && n->string)
429 m->meta.msec = mandoc_strdup(n->string);
430 else
431 m->meta.msec = mandoc_strdup("");
432
433 /* TITLE MSEC ->DATE<- SOURCE VOL */
434
435 if (n)
436 n = n->next;
437 if (n)
438 pos = n->pos;
439 m->meta.date = mandoc_normdate
440 (m->parse, n ? n->string : NULL, line, pos);
441
442 /* TITLE MSEC DATE ->SOURCE<- VOL */
443
444 if (n && (n = n->next))
445 m->meta.source = mandoc_strdup(n->string);
446
447 /* TITLE MSEC DATE SOURCE ->VOL<- */
448
449 if (n && (n = n->next))
450 m->meta.vol = mandoc_strdup(n->string);
451
452 /*
453 * Remove the `TH' node after we've processed it for our
454 * meta-data.
455 */
456 man_node_delete(m, m->last);
457 return(1);
458 }
459
460 static int
461 post_nf(CHKARGS)
462 {
463
464 if (MAN_LITERAL & m->flags)
465 man_nmsg(m, n, MANDOCERR_SCOPEREP);
466
467 m->flags |= MAN_LITERAL;
468 return(1);
469 }
470
471 static int
472 post_fi(CHKARGS)
473 {
474
475 if ( ! (MAN_LITERAL & m->flags))
476 man_nmsg(m, n, MANDOCERR_WNOSCOPE);
477
478 m->flags &= ~MAN_LITERAL;
479 return(1);
480 }
481
482 static int
483 post_UC(CHKARGS)
484 {
485 static const char * const bsd_versions[] = {
486 "3rd Berkeley Distribution",
487 "4th Berkeley Distribution",
488 "4.2 Berkeley Distribution",
489 "4.3 Berkeley Distribution",
490 "4.4 Berkeley Distribution",
491 };
492
493 const char *p, *s;
494
495 n = n->child;
496 n = m->last->child;
497
498 if (NULL == n || MAN_TEXT != n->type)
499 p = bsd_versions[0];
500 else {
501 s = n->string;
502 if (0 == strcmp(s, "3"))
503 p = bsd_versions[0];
504 else if (0 == strcmp(s, "4"))
505 p = bsd_versions[1];
506 else if (0 == strcmp(s, "5"))
507 p = bsd_versions[2];
508 else if (0 == strcmp(s, "6"))
509 p = bsd_versions[3];
510 else if (0 == strcmp(s, "7"))
511 p = bsd_versions[4];
512 else
513 p = bsd_versions[0];
514 }
515
516 if (m->meta.source)
517 free(m->meta.source);
518
519 m->meta.source = mandoc_strdup(p);
520 return(1);
521 }
522
523 static int
524 post_AT(CHKARGS)
525 {
526 static const char * const unix_versions[] = {
527 "7th Edition",
528 "System III",
529 "System V",
530 "System V Release 2",
531 };
532
533 const char *p, *s;
534 struct man_node *nn;
535
536 n = n->child;
537
538 if (NULL == n || MAN_TEXT != n->type)
539 p = unix_versions[0];
540 else {
541 s = n->string;
542 if (0 == strcmp(s, "3"))
543 p = unix_versions[0];
544 else if (0 == strcmp(s, "4"))
545 p = unix_versions[1];
546 else if (0 == strcmp(s, "5")) {
547 nn = n->next;
548 if (nn && MAN_TEXT == nn->type && nn->string[0])
549 p = unix_versions[3];
550 else
551 p = unix_versions[2];
552 } else
553 p = unix_versions[0];
554 }
555
556 if (m->meta.source)
557 free(m->meta.source);
558
559 m->meta.source = mandoc_strdup(p);
560 return(1);
561 }
562
563 static int
564 post_vs(CHKARGS)
565 {
566
567 /*
568 * Don't warn about this because it occurs in pod2man and would
569 * cause considerable (unfixable) warnage.
570 */
571 if (NULL == n->prev && MAN_ROOT == n->parent->type)
572 man_node_delete(m, n);
573
574 return(1);
575 }