Annotation of mandoc/man_validate.c, Revision 1.63
1.63 ! schwarze 1: /* $Id: man_validate.c,v 1.62 2011/02/09 09:18:15 kristaps Exp $ */
1.1 kristaps 2: /*
1.63 ! schwarze 3: * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
! 4: * Copyright (c) 2010 Ingo Schwarze <schwarze@openbsd.org>
1.1 kristaps 5: *
6: * Permission to use, copy, modify, and distribute this software for any
1.8 kristaps 7: * purpose with or without fee is hereby granted, provided that the above
8: * copyright notice and this permission notice appear in all copies.
1.1 kristaps 9: *
1.8 kristaps 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.
1.1 kristaps 17: */
1.28 kristaps 18: #ifdef HAVE_CONFIG_H
19: #include "config.h"
20: #endif
21:
1.1 kristaps 22: #include <sys/types.h>
23:
24: #include <assert.h>
25: #include <ctype.h>
1.16 kristaps 26: #include <errno.h>
27: #include <limits.h>
1.1 kristaps 28: #include <stdarg.h>
29: #include <stdlib.h>
1.46 kristaps 30: #include <string.h>
1.50 kristaps 31: #include <time.h>
1.1 kristaps 32:
1.41 kristaps 33: #include "mandoc.h"
1.1 kristaps 34: #include "libman.h"
1.15 kristaps 35: #include "libmandoc.h"
1.1 kristaps 36:
1.43 kristaps 37: #define CHKARGS struct man *m, struct man_node *n
1.1 kristaps 38:
1.17 kristaps 39: typedef int (*v_check)(CHKARGS);
1.1 kristaps 40:
41: struct man_valid {
1.17 kristaps 42: v_check *pres;
43: v_check *posts;
1.1 kristaps 44: };
45:
1.17 kristaps 46: static int check_bline(CHKARGS);
47: static int check_eq0(CHKARGS);
1.55 kristaps 48: static int check_ft(CHKARGS);
1.25 kristaps 49: static int check_le1(CHKARGS);
1.17 kristaps 50: static int check_ge2(CHKARGS);
51: static int check_le5(CHKARGS);
52: static int check_par(CHKARGS);
1.23 kristaps 53: static int check_part(CHKARGS);
1.17 kristaps 54: static int check_root(CHKARGS);
55: static int check_sec(CHKARGS);
56: static int check_text(CHKARGS);
57:
1.51 kristaps 58: static int post_AT(CHKARGS);
59: static int post_fi(CHKARGS);
60: static int post_nf(CHKARGS);
61: static int post_TH(CHKARGS);
62: static int post_UC(CHKARGS);
63:
64: static v_check posts_at[] = { post_AT, NULL };
1.17 kristaps 65: static v_check posts_eq0[] = { check_eq0, NULL };
1.51 kristaps 66: static v_check posts_fi[] = { check_eq0, post_fi, NULL };
67: static v_check posts_le1[] = { check_le1, NULL };
1.55 kristaps 68: static v_check posts_ft[] = { check_ft, NULL };
1.51 kristaps 69: static v_check posts_nf[] = { check_eq0, post_nf, NULL };
1.17 kristaps 70: static v_check posts_par[] = { check_par, NULL };
1.23 kristaps 71: static v_check posts_part[] = { check_part, NULL };
1.17 kristaps 72: static v_check posts_sec[] = { check_sec, NULL };
1.60 schwarze 73: static v_check posts_th[] = { check_ge2, check_le5, post_TH, NULL };
1.51 kristaps 74: static v_check posts_uc[] = { post_UC, NULL };
1.17 kristaps 75: static v_check pres_bline[] = { check_bline, NULL };
1.1 kristaps 76:
1.51 kristaps 77:
1.1 kristaps 78: static const struct man_valid man_valids[MAN_MAX] = {
1.29 kristaps 79: { NULL, posts_eq0 }, /* br */
1.32 kristaps 80: { pres_bline, posts_th }, /* TH */
1.17 kristaps 81: { pres_bline, posts_sec }, /* SH */
82: { pres_bline, posts_sec }, /* SS */
1.59 kristaps 83: { pres_bline, NULL }, /* TP */
1.17 kristaps 84: { pres_bline, posts_par }, /* LP */
85: { pres_bline, posts_par }, /* PP */
86: { pres_bline, posts_par }, /* P */
1.59 kristaps 87: { pres_bline, NULL }, /* IP */
88: { pres_bline, NULL }, /* HP */
1.22 kristaps 89: { NULL, NULL }, /* SM */
90: { NULL, NULL }, /* SB */
1.17 kristaps 91: { NULL, NULL }, /* BI */
92: { NULL, NULL }, /* IB */
93: { NULL, NULL }, /* BR */
94: { NULL, NULL }, /* RB */
1.22 kristaps 95: { NULL, NULL }, /* R */
96: { NULL, NULL }, /* B */
97: { NULL, NULL }, /* I */
1.17 kristaps 98: { NULL, NULL }, /* IR */
99: { NULL, NULL }, /* RI */
1.47 kristaps 100: { NULL, posts_eq0 }, /* na */ /* FIXME: should warn only. */
101: { NULL, posts_le1 }, /* sp */ /* FIXME: should warn only. */
1.51 kristaps 102: { pres_bline, posts_nf }, /* nf */
103: { pres_bline, posts_fi }, /* fi */
1.19 kristaps 104: { NULL, NULL }, /* RE */
1.23 kristaps 105: { NULL, posts_part }, /* RS */
1.21 kristaps 106: { NULL, NULL }, /* DT */
1.51 kristaps 107: { NULL, posts_uc }, /* UC */
1.26 kristaps 108: { NULL, NULL }, /* PD */
1.51 kristaps 109: { NULL, posts_at }, /* AT */
1.47 kristaps 110: { NULL, NULL }, /* in */
1.55 kristaps 111: { NULL, posts_ft }, /* ft */
1.1 kristaps 112: };
113:
114:
115: int
1.43 kristaps 116: man_valid_pre(struct man *m, struct man_node *n)
1.17 kristaps 117: {
118: v_check *cp;
119:
1.57 kristaps 120: switch (n->type) {
121: case (MAN_TEXT):
122: /* FALLTHROUGH */
123: case (MAN_ROOT):
1.61 kristaps 124: /* FALLTHROUGH */
125: case (MAN_EQN):
1.57 kristaps 126: /* FALLTHROUGH */
127: case (MAN_TBL):
1.17 kristaps 128: return(1);
1.57 kristaps 129: default:
130: break;
131: }
1.17 kristaps 132:
133: if (NULL == (cp = man_valids[n->tok].pres))
134: return(1);
135: for ( ; *cp; cp++)
136: if ( ! (*cp)(m, n))
137: return(0);
138: return(1);
139: }
140:
141:
142: int
1.1 kristaps 143: man_valid_post(struct man *m)
144: {
1.17 kristaps 145: v_check *cp;
1.1 kristaps 146:
147: if (MAN_VALID & m->last->flags)
148: return(1);
149: m->last->flags |= MAN_VALID;
150:
151: switch (m->last->type) {
152: case (MAN_TEXT):
1.11 kristaps 153: return(check_text(m, m->last));
1.1 kristaps 154: case (MAN_ROOT):
1.14 kristaps 155: return(check_root(m, m->last));
1.62 kristaps 156: case (MAN_EQN):
157: /* FALLTHROUGH */
1.57 kristaps 158: case (MAN_TBL):
159: return(1);
1.1 kristaps 160: default:
161: break;
162: }
163:
164: if (NULL == (cp = man_valids[m->last->tok].posts))
165: return(1);
166: for ( ; *cp; cp++)
167: if ( ! (*cp)(m, m->last))
168: return(0);
169:
170: return(1);
171: }
172:
173:
1.11 kristaps 174: static int
1.17 kristaps 175: check_root(CHKARGS)
1.14 kristaps 176: {
1.17 kristaps 177:
178: if (MAN_BLINE & m->flags)
1.54 kristaps 179: man_nmsg(m, n, MANDOCERR_SCOPEEXIT);
180: else if (MAN_ELINE & m->flags)
181: man_nmsg(m, n, MANDOCERR_SCOPEEXIT);
1.22 kristaps 182:
183: m->flags &= ~MAN_BLINE;
184: m->flags &= ~MAN_ELINE;
1.17 kristaps 185:
1.41 kristaps 186: if (NULL == m->first->child) {
187: man_nmsg(m, n, MANDOCERR_NODOCBODY);
188: return(0);
189: } else if (NULL == m->meta.title) {
1.54 kristaps 190: man_nmsg(m, n, MANDOCERR_NOTITLE);
191:
1.34 kristaps 192: /*
193: * If a title hasn't been set, do so now (by
194: * implication, date and section also aren't set).
195: */
1.54 kristaps 196:
1.34 kristaps 197: m->meta.title = mandoc_strdup("unknown");
1.37 kristaps 198: m->meta.msec = mandoc_strdup("1");
1.63 ! schwarze 199: m->meta.date = mandoc_normdate(NULL,
! 200: m->msg, m->data, n->line, n->pos);
1.34 kristaps 201: }
1.32 kristaps 202:
203: return(1);
204: }
205:
206:
207: static int
1.17 kristaps 208: check_text(CHKARGS)
1.11 kristaps 209: {
1.43 kristaps 210: char *p;
1.15 kristaps 211: int pos, c;
1.46 kristaps 212: size_t sz;
1.11 kristaps 213:
1.46 kristaps 214: for (p = n->string, pos = n->pos + 1; *p; p++, pos++) {
215: sz = strcspn(p, "\t\\");
216: p += (int)sz;
217:
218: if ('\0' == *p)
219: break;
220:
221: pos += (int)sz;
1.11 kristaps 222:
1.46 kristaps 223: if ('\t' == *p) {
224: if (MAN_LITERAL & m->flags)
225: continue;
226: if (man_pmsg(m, n->line, pos, MANDOCERR_BADTAB))
1.15 kristaps 227: continue;
1.46 kristaps 228: return(0);
1.15 kristaps 229: }
1.45 kristaps 230:
1.46 kristaps 231: /* Check the special character. */
1.15 kristaps 232:
1.46 kristaps 233: c = mandoc_special(p);
234: if (c) {
235: p += c - 1;
236: pos += c - 1;
1.49 schwarze 237: } else
238: man_pmsg(m, n->line, pos, MANDOCERR_BADESCAPE);
1.11 kristaps 239: }
240:
241: return(1);
1.1 kristaps 242: }
243:
244:
245: #define INEQ_DEFINE(x, ineq, name) \
246: static int \
1.17 kristaps 247: check_##name(CHKARGS) \
1.1 kristaps 248: { \
1.11 kristaps 249: if (n->nchild ineq (x)) \
1.1 kristaps 250: return(1); \
1.60 schwarze 251: man_vmsg(m, MANDOCERR_ARGCOUNT, n->line, n->pos, \
1.41 kristaps 252: "line arguments %s %d (have %d)", \
253: #ineq, (x), n->nchild); \
1.60 schwarze 254: return(1); \
1.1 kristaps 255: }
256:
257: INEQ_DEFINE(0, ==, eq0)
1.25 kristaps 258: INEQ_DEFINE(1, <=, le1)
1.1 kristaps 259: INEQ_DEFINE(2, >=, ge2)
260: INEQ_DEFINE(5, <=, le5)
261:
1.55 kristaps 262: static int
263: check_ft(CHKARGS)
264: {
265: char *cp;
266: int ok;
267:
268: if (0 == n->nchild)
269: return(1);
270:
271: ok = 0;
272: cp = n->child->string;
273: switch (*cp) {
274: case ('1'):
275: /* FALLTHROUGH */
276: case ('2'):
277: /* FALLTHROUGH */
278: case ('3'):
279: /* FALLTHROUGH */
280: case ('4'):
281: /* FALLTHROUGH */
282: case ('I'):
283: /* FALLTHROUGH */
284: case ('P'):
285: /* FALLTHROUGH */
286: case ('R'):
287: if ('\0' == cp[1])
288: ok = 1;
289: break;
290: case ('B'):
291: if ('\0' == cp[1] || ('I' == cp[1] && '\0' == cp[2]))
292: ok = 1;
293: break;
294: case ('C'):
295: if ('W' == cp[1] && '\0' == cp[2])
296: ok = 1;
297: break;
298: default:
299: break;
300: }
301:
302: if (0 == ok) {
1.63 ! schwarze 303: man_vmsg(m, MANDOCERR_BADFONT,
1.55 kristaps 304: n->line, n->pos, "%s", cp);
305: *cp = '\0';
306: }
307:
308: if (1 < n->nchild)
309: man_vmsg(m, MANDOCERR_ARGCOUNT, n->line, n->pos,
310: "want one child (have %d)", n->nchild);
311:
312: return(1);
313: }
1.16 kristaps 314:
315: static int
1.17 kristaps 316: check_sec(CHKARGS)
317: {
1.16 kristaps 318:
1.41 kristaps 319: if (MAN_HEAD == n->type && 0 == n->nchild) {
320: man_nmsg(m, n, MANDOCERR_SYNTARGCOUNT);
321: return(0);
322: } else if (MAN_BODY == n->type && 0 == n->nchild)
1.54 kristaps 323: man_nmsg(m, n, MANDOCERR_NOBODY);
1.41 kristaps 324:
1.16 kristaps 325: return(1);
326: }
1.17 kristaps 327:
328:
329: static int
1.23 kristaps 330: check_part(CHKARGS)
331: {
332:
333: if (MAN_BODY == n->type && 0 == n->nchild)
1.54 kristaps 334: man_nmsg(m, n, MANDOCERR_NOBODY);
335:
1.23 kristaps 336: return(1);
337: }
338:
339:
340: static int
1.17 kristaps 341: check_par(CHKARGS)
342: {
343:
1.59 kristaps 344: switch (n->type) {
345: case (MAN_BLOCK):
346: if (0 == n->body->nchild)
347: man_node_delete(m, n);
348: break;
349: case (MAN_BODY):
350: if (0 == n->nchild)
351: man_nmsg(m, n, MANDOCERR_IGNPAR);
352: break;
353: case (MAN_HEAD):
354: if (n->nchild)
355: man_nmsg(m, n, MANDOCERR_ARGSLOST);
356: break;
357: default:
358: break;
359: }
1.17 kristaps 360:
361: return(1);
362: }
363:
364:
365: static int
366: check_bline(CHKARGS)
367: {
368:
1.22 kristaps 369: assert( ! (MAN_ELINE & m->flags));
1.41 kristaps 370: if (MAN_BLINE & m->flags) {
371: man_nmsg(m, n, MANDOCERR_SYNTLINESCOPE);
372: return(0);
373: }
1.31 kristaps 374:
1.18 kristaps 375: return(1);
1.17 kristaps 376: }
377:
1.51 kristaps 378: static int
379: post_TH(CHKARGS)
380: {
1.60 schwarze 381: const char *p;
1.63 ! schwarze 382: int line, pos;
1.51 kristaps 383:
384: if (m->meta.title)
385: free(m->meta.title);
386: if (m->meta.vol)
387: free(m->meta.vol);
388: if (m->meta.source)
389: free(m->meta.source);
390: if (m->meta.msec)
391: free(m->meta.msec);
1.63 ! schwarze 392: if (m->meta.date)
! 393: free(m->meta.date);
1.51 kristaps 394:
1.63 ! schwarze 395: line = n->line;
! 396: pos = n->pos;
! 397: m->meta.title = m->meta.vol = m->meta.date =
1.51 kristaps 398: m->meta.msec = m->meta.source = NULL;
399:
400: /* ->TITLE<- MSEC DATE SOURCE VOL */
401:
402: n = n->child;
1.60 schwarze 403: if (n && n->string) {
404: for (p = n->string; '\0' != *p; p++) {
405: /* Only warn about this once... */
406: if (isalpha((u_char)*p) && ! isupper((u_char)*p)) {
407: man_nmsg(m, n, MANDOCERR_UPPERCASE);
408: break;
409: }
410: }
411: m->meta.title = mandoc_strdup(n->string);
412: } else
413: m->meta.title = mandoc_strdup("");
1.51 kristaps 414:
415: /* TITLE ->MSEC<- DATE SOURCE VOL */
416:
1.60 schwarze 417: if (n)
418: n = n->next;
419: if (n && n->string)
420: m->meta.msec = mandoc_strdup(n->string);
421: else
422: m->meta.msec = mandoc_strdup("");
1.51 kristaps 423:
424: /* TITLE MSEC ->DATE<- SOURCE VOL */
425:
1.60 schwarze 426: if (n)
427: n = n->next;
1.63 ! schwarze 428: if (n)
! 429: pos = n->pos;
! 430: m->meta.date = mandoc_normdate(n ? n->string : NULL,
! 431: m->msg, m->data, line, pos);
1.51 kristaps 432:
433: /* TITLE MSEC DATE ->SOURCE<- VOL */
434:
435: if (n && (n = n->next))
436: m->meta.source = mandoc_strdup(n->string);
437:
438: /* TITLE MSEC DATE SOURCE ->VOL<- */
439:
440: if (n && (n = n->next))
441: m->meta.vol = mandoc_strdup(n->string);
442:
443: /*
444: * Remove the `TH' node after we've processed it for our
445: * meta-data.
446: */
447: man_node_delete(m, m->last);
448: return(1);
449: }
450:
451: static int
452: post_nf(CHKARGS)
453: {
454:
455: if (MAN_LITERAL & m->flags)
456: man_nmsg(m, n, MANDOCERR_SCOPEREP);
457:
458: m->flags |= MAN_LITERAL;
459: return(1);
460: }
461:
462: static int
463: post_fi(CHKARGS)
464: {
465:
466: if ( ! (MAN_LITERAL & m->flags))
1.58 kristaps 467: man_nmsg(m, n, MANDOCERR_WNOSCOPE);
1.51 kristaps 468:
469: m->flags &= ~MAN_LITERAL;
470: return(1);
471: }
472:
473: static int
474: post_UC(CHKARGS)
475: {
476: static const char * const bsd_versions[] = {
477: "3rd Berkeley Distribution",
478: "4th Berkeley Distribution",
479: "4.2 Berkeley Distribution",
480: "4.3 Berkeley Distribution",
481: "4.4 Berkeley Distribution",
482: };
483:
484: const char *p, *s;
485:
486: n = n->child;
487: n = m->last->child;
488:
489: if (NULL == n || MAN_TEXT != n->type)
490: p = bsd_versions[0];
491: else {
492: s = n->string;
493: if (0 == strcmp(s, "3"))
494: p = bsd_versions[0];
495: else if (0 == strcmp(s, "4"))
496: p = bsd_versions[1];
497: else if (0 == strcmp(s, "5"))
498: p = bsd_versions[2];
499: else if (0 == strcmp(s, "6"))
500: p = bsd_versions[3];
501: else if (0 == strcmp(s, "7"))
502: p = bsd_versions[4];
503: else
504: p = bsd_versions[0];
505: }
506:
507: if (m->meta.source)
508: free(m->meta.source);
509:
510: m->meta.source = mandoc_strdup(p);
511: return(1);
512: }
513:
514: static int
515: post_AT(CHKARGS)
516: {
517: static const char * const unix_versions[] = {
518: "7th Edition",
519: "System III",
520: "System V",
521: "System V Release 2",
522: };
523:
524: const char *p, *s;
525: struct man_node *nn;
526:
527: n = n->child;
528:
529: if (NULL == n || MAN_TEXT != n->type)
530: p = unix_versions[0];
531: else {
532: s = n->string;
533: if (0 == strcmp(s, "3"))
534: p = unix_versions[0];
535: else if (0 == strcmp(s, "4"))
536: p = unix_versions[1];
537: else if (0 == strcmp(s, "5")) {
538: nn = n->next;
539: if (nn && MAN_TEXT == nn->type && nn->string[0])
540: p = unix_versions[3];
541: else
542: p = unix_versions[2];
543: } else
544: p = unix_versions[0];
545: }
546:
547: if (m->meta.source)
548: free(m->meta.source);
549:
550: m->meta.source = mandoc_strdup(p);
551: return(1);
552: }
CVSweb