1/*	Id: man_validate.c,v 1.146 2018/12/31 10:04:39 schwarze Exp  */
2/*
3 * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2010, 2012-2018 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 AUTHORS DISCLAIM ALL WARRANTIES
11 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
12 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS 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#include "config.h"
19
20#include <sys/types.h>
21
22#include <assert.h>
23#include <ctype.h>
24#include <errno.h>
25#include <limits.h>
26#include <stdarg.h>
27#include <stdio.h>
28#include <stdlib.h>
29#include <string.h>
30#include <time.h>
31
32#include "mandoc_aux.h"
33#include "mandoc.h"
34#include "roff.h"
35#include "man.h"
36#include "libmandoc.h"
37#include "roff_int.h"
38#include "libman.h"
39
40#define	CHKARGS	  struct roff_man *man, struct roff_node *n
41
42typedef	void	(*v_check)(CHKARGS);
43
44static	void	  check_abort(CHKARGS) __dead;
45static	void	  check_par(CHKARGS);
46static	void	  check_part(CHKARGS);
47static	void	  check_root(CHKARGS);
48static	void	  check_text(CHKARGS);
49
50static	void	  post_AT(CHKARGS);
51static	void	  post_EE(CHKARGS);
52static	void	  post_EX(CHKARGS);
53static	void	  post_IP(CHKARGS);
54static	void	  post_OP(CHKARGS);
55static	void	  post_SH(CHKARGS);
56static	void	  post_TH(CHKARGS);
57static	void	  post_UC(CHKARGS);
58static	void	  post_UR(CHKARGS);
59static	void	  post_in(CHKARGS);
60
61static	const v_check man_valids[MAN_MAX - MAN_TH] = {
62	post_TH,    /* TH */
63	post_SH,    /* SH */
64	post_SH,    /* SS */
65	NULL,       /* TP */
66	NULL,       /* TQ */
67	check_abort,/* LP */
68	check_par,  /* PP */
69	check_abort,/* P */
70	post_IP,    /* IP */
71	NULL,       /* HP */
72	NULL,       /* SM */
73	NULL,       /* SB */
74	NULL,       /* BI */
75	NULL,       /* IB */
76	NULL,       /* BR */
77	NULL,       /* RB */
78	NULL,       /* R */
79	NULL,       /* B */
80	NULL,       /* I */
81	NULL,       /* IR */
82	NULL,       /* RI */
83	NULL,       /* RE */
84	check_part, /* RS */
85	NULL,       /* DT */
86	post_UC,    /* UC */
87	NULL,       /* PD */
88	post_AT,    /* AT */
89	post_in,    /* in */
90	NULL,       /* SY */
91	NULL,       /* YS */
92	post_OP,    /* OP */
93	post_EX,    /* EX */
94	post_EE,    /* EE */
95	post_UR,    /* UR */
96	NULL,       /* UE */
97	post_UR,    /* MT */
98	NULL,       /* ME */
99};
100
101
102/* Validate the subtree rooted at man->last. */
103void
104man_validate(struct roff_man *man)
105{
106	struct roff_node *n;
107	const v_check	 *cp;
108
109	/*
110	 * Translate obsolete macros such that later code
111	 * does not need to look for them.
112	 */
113
114	n = man->last;
115	switch (n->tok) {
116	case MAN_LP:
117	case MAN_P:
118		n->tok = MAN_PP;
119		break;
120	default:
121		break;
122	}
123
124	/*
125	 * Iterate over all children, recursing into each one
126	 * in turn, depth-first.
127	 */
128
129	man->last = man->last->child;
130	while (man->last != NULL) {
131		man_validate(man);
132		if (man->last == n)
133			man->last = man->last->child;
134		else
135			man->last = man->last->next;
136	}
137
138	/* Finally validate the macro itself. */
139
140	man->last = n;
141	man->next = ROFF_NEXT_SIBLING;
142	switch (n->type) {
143	case ROFFT_TEXT:
144		check_text(man, n);
145		break;
146	case ROFFT_ROOT:
147		check_root(man, n);
148		break;
149	case ROFFT_COMMENT:
150	case ROFFT_EQN:
151	case ROFFT_TBL:
152		break;
153	default:
154		if (n->tok < ROFF_MAX) {
155			roff_validate(man);
156			break;
157		}
158		assert(n->tok >= MAN_TH && n->tok < MAN_MAX);
159		cp = man_valids + (n->tok - MAN_TH);
160		if (*cp)
161			(*cp)(man, n);
162		if (man->last == n)
163			n->flags |= NODE_VALID;
164		break;
165	}
166}
167
168static void
169check_root(CHKARGS)
170{
171	assert((man->flags & (MAN_BLINE | MAN_ELINE)) == 0);
172
173	if (n->last == NULL || n->last->type == ROFFT_COMMENT)
174		mandoc_msg(MANDOCERR_DOC_EMPTY, n->line, n->pos, NULL);
175	else
176		man->meta.hasbody = 1;
177
178	if (NULL == man->meta.title) {
179		mandoc_msg(MANDOCERR_TH_NOTITLE, n->line, n->pos, NULL);
180
181		/*
182		 * If a title hasn't been set, do so now (by
183		 * implication, date and section also aren't set).
184		 */
185
186		man->meta.title = mandoc_strdup("");
187		man->meta.msec = mandoc_strdup("");
188		man->meta.date = man->quick ? mandoc_strdup("") :
189		    mandoc_normdate(man, NULL, n->line, n->pos);
190	}
191
192	if (man->meta.os_e &&
193	    (man->meta.rcsids & (1 << man->meta.os_e)) == 0)
194		mandoc_msg(MANDOCERR_RCS_MISSING, 0, 0,
195		    man->meta.os_e == MANDOC_OS_OPENBSD ?
196		    "(OpenBSD)" : "(NetBSD)");
197}
198
199static void
200check_abort(CHKARGS)
201{
202	abort();
203}
204
205static void
206check_text(CHKARGS)
207{
208	char		*cp, *p;
209
210	if (n->flags & NODE_NOFILL)
211		return;
212
213	cp = n->string;
214	for (p = cp; NULL != (p = strchr(p, '\t')); p++)
215		mandoc_msg(MANDOCERR_FI_TAB,
216		    n->line, n->pos + (int)(p - cp), NULL);
217}
218
219static void
220post_EE(CHKARGS)
221{
222	if ((n->flags & NODE_NOFILL) == 0)
223		mandoc_msg(MANDOCERR_FI_SKIP, n->line, n->pos, "EE");
224}
225
226static void
227post_EX(CHKARGS)
228{
229	if (n->flags & NODE_NOFILL)
230		mandoc_msg(MANDOCERR_NF_SKIP, n->line, n->pos, "EX");
231}
232
233static void
234post_OP(CHKARGS)
235{
236
237	if (n->child == NULL)
238		mandoc_msg(MANDOCERR_OP_EMPTY, n->line, n->pos, "OP");
239	else if (n->child->next != NULL && n->child->next->next != NULL) {
240		n = n->child->next->next;
241		mandoc_msg(MANDOCERR_ARG_EXCESS,
242		    n->line, n->pos, "OP ... %s", n->string);
243	}
244}
245
246static void
247post_SH(CHKARGS)
248{
249	struct roff_node	*nc;
250
251	if (n->type != ROFFT_BODY || (nc = n->child) == NULL)
252		return;
253
254	if (nc->tok == MAN_PP && nc->body->child != NULL) {
255		while (nc->body->last != NULL) {
256			man->next = ROFF_NEXT_CHILD;
257			roff_node_relink(man, nc->body->last);
258			man->last = n;
259		}
260	}
261
262	if (nc->tok == MAN_PP || nc->tok == ROFF_sp || nc->tok == ROFF_br) {
263		mandoc_msg(MANDOCERR_PAR_SKIP, nc->line, nc->pos,
264		    "%s after %s", roff_name[nc->tok], roff_name[n->tok]);
265		roff_node_delete(man, nc);
266	}
267
268	/*
269	 * Trailing PP is empty, so it is deleted by check_par().
270	 * Trailing sp is significant.
271	 */
272
273	if ((nc = n->last) != NULL && nc->tok == ROFF_br) {
274		mandoc_msg(MANDOCERR_PAR_SKIP,
275		    nc->line, nc->pos, "%s at the end of %s",
276		    roff_name[nc->tok], roff_name[n->tok]);
277		roff_node_delete(man, nc);
278	}
279}
280
281static void
282post_UR(CHKARGS)
283{
284	if (n->type == ROFFT_HEAD && n->child == NULL)
285		mandoc_msg(MANDOCERR_UR_NOHEAD, n->line, n->pos,
286		    "%s", roff_name[n->tok]);
287	check_part(man, n);
288}
289
290static void
291check_part(CHKARGS)
292{
293
294	if (n->type == ROFFT_BODY && n->child == NULL)
295		mandoc_msg(MANDOCERR_BLK_EMPTY, n->line, n->pos,
296		    "%s", roff_name[n->tok]);
297}
298
299static void
300check_par(CHKARGS)
301{
302
303	switch (n->type) {
304	case ROFFT_BLOCK:
305		if (n->body->child == NULL)
306			roff_node_delete(man, n);
307		break;
308	case ROFFT_BODY:
309		if (n->child != NULL &&
310		    (n->child->tok == ROFF_sp || n->child->tok == ROFF_br)) {
311			mandoc_msg(MANDOCERR_PAR_SKIP,
312			    n->child->line, n->child->pos,
313			    "%s after %s", roff_name[n->child->tok],
314			    roff_name[n->tok]);
315			roff_node_delete(man, n->child);
316		}
317		if (n->child == NULL)
318			mandoc_msg(MANDOCERR_PAR_SKIP, n->line, n->pos,
319			    "%s empty", roff_name[n->tok]);
320		break;
321	case ROFFT_HEAD:
322		if (n->child != NULL)
323			mandoc_msg(MANDOCERR_ARG_SKIP,
324			    n->line, n->pos, "%s %s%s",
325			    roff_name[n->tok], n->child->string,
326			    n->child->next != NULL ? " ..." : "");
327		break;
328	default:
329		break;
330	}
331}
332
333static void
334post_IP(CHKARGS)
335{
336
337	switch (n->type) {
338	case ROFFT_BLOCK:
339		if (n->head->child == NULL && n->body->child == NULL)
340			roff_node_delete(man, n);
341		break;
342	case ROFFT_BODY:
343		if (n->parent->head->child == NULL && n->child == NULL)
344			mandoc_msg(MANDOCERR_PAR_SKIP, n->line, n->pos,
345			    "%s empty", roff_name[n->tok]);
346		break;
347	default:
348		break;
349	}
350}
351
352static void
353post_TH(CHKARGS)
354{
355	struct roff_node *nb;
356	const char	*p;
357
358	free(man->meta.title);
359	free(man->meta.vol);
360	free(man->meta.os);
361	free(man->meta.msec);
362	free(man->meta.date);
363
364	man->meta.title = man->meta.vol = man->meta.date =
365	    man->meta.msec = man->meta.os = NULL;
366
367	nb = n;
368
369	/* ->TITLE<- MSEC DATE OS VOL */
370
371	n = n->child;
372	if (n && n->string) {
373		for (p = n->string; '\0' != *p; p++) {
374			/* Only warn about this once... */
375			if (isalpha((unsigned char)*p) &&
376			    ! isupper((unsigned char)*p)) {
377				mandoc_msg(MANDOCERR_TITLE_CASE, n->line,
378				    n->pos + (int)(p - n->string),
379				    "TH %s", n->string);
380				break;
381			}
382		}
383		man->meta.title = mandoc_strdup(n->string);
384	} else {
385		man->meta.title = mandoc_strdup("");
386		mandoc_msg(MANDOCERR_TH_NOTITLE, nb->line, nb->pos, "TH");
387	}
388
389	/* TITLE ->MSEC<- DATE OS VOL */
390
391	if (n)
392		n = n->next;
393	if (n && n->string)
394		man->meta.msec = mandoc_strdup(n->string);
395	else {
396		man->meta.msec = mandoc_strdup("");
397		mandoc_msg(MANDOCERR_MSEC_MISSING,
398		    nb->line, nb->pos, "TH %s", man->meta.title);
399	}
400
401	/* TITLE MSEC ->DATE<- OS VOL */
402
403	if (n)
404		n = n->next;
405	if (n && n->string && '\0' != n->string[0]) {
406		man->meta.date = man->quick ?
407		    mandoc_strdup(n->string) :
408		    mandoc_normdate(man, n->string, n->line, n->pos);
409	} else {
410		man->meta.date = mandoc_strdup("");
411		mandoc_msg(MANDOCERR_DATE_MISSING,
412		    n ? n->line : nb->line,
413		    n ? n->pos : nb->pos, "TH");
414	}
415
416	/* TITLE MSEC DATE ->OS<- VOL */
417
418	if (n && (n = n->next))
419		man->meta.os = mandoc_strdup(n->string);
420	else if (man->os_s != NULL)
421		man->meta.os = mandoc_strdup(man->os_s);
422	if (man->meta.os_e == MANDOC_OS_OTHER && man->meta.os != NULL) {
423		if (strstr(man->meta.os, "OpenBSD") != NULL)
424			man->meta.os_e = MANDOC_OS_OPENBSD;
425		else if (strstr(man->meta.os, "NetBSD") != NULL)
426			man->meta.os_e = MANDOC_OS_NETBSD;
427	}
428
429	/* TITLE MSEC DATE OS ->VOL<- */
430	/* If missing, use the default VOL name for MSEC. */
431
432	if (n && (n = n->next))
433		man->meta.vol = mandoc_strdup(n->string);
434	else if ('\0' != man->meta.msec[0] &&
435	    (NULL != (p = mandoc_a2msec(man->meta.msec))))
436		man->meta.vol = mandoc_strdup(p);
437
438	if (n != NULL && (n = n->next) != NULL)
439		mandoc_msg(MANDOCERR_ARG_EXCESS,
440		    n->line, n->pos, "TH ... %s", n->string);
441
442	/*
443	 * Remove the `TH' node after we've processed it for our
444	 * meta-data.
445	 */
446	roff_node_delete(man, man->last);
447}
448
449static void
450post_UC(CHKARGS)
451{
452	static const char * const bsd_versions[] = {
453	    "3rd Berkeley Distribution",
454	    "4th Berkeley Distribution",
455	    "4.2 Berkeley Distribution",
456	    "4.3 Berkeley Distribution",
457	    "4.4 Berkeley Distribution",
458	};
459
460	const char	*p, *s;
461
462	n = n->child;
463
464	if (n == NULL || n->type != ROFFT_TEXT)
465		p = bsd_versions[0];
466	else {
467		s = n->string;
468		if (0 == strcmp(s, "3"))
469			p = bsd_versions[0];
470		else if (0 == strcmp(s, "4"))
471			p = bsd_versions[1];
472		else if (0 == strcmp(s, "5"))
473			p = bsd_versions[2];
474		else if (0 == strcmp(s, "6"))
475			p = bsd_versions[3];
476		else if (0 == strcmp(s, "7"))
477			p = bsd_versions[4];
478		else
479			p = bsd_versions[0];
480	}
481
482	free(man->meta.os);
483	man->meta.os = mandoc_strdup(p);
484}
485
486static void
487post_AT(CHKARGS)
488{
489	static const char * const unix_versions[] = {
490	    "7th Edition",
491	    "System III",
492	    "System V",
493	    "System V Release 2",
494	};
495
496	struct roff_node *nn;
497	const char	*p, *s;
498
499	n = n->child;
500
501	if (n == NULL || n->type != ROFFT_TEXT)
502		p = unix_versions[0];
503	else {
504		s = n->string;
505		if (0 == strcmp(s, "3"))
506			p = unix_versions[0];
507		else if (0 == strcmp(s, "4"))
508			p = unix_versions[1];
509		else if (0 == strcmp(s, "5")) {
510			nn = n->next;
511			if (nn != NULL &&
512			    nn->type == ROFFT_TEXT &&
513			    nn->string[0] != '\0')
514				p = unix_versions[3];
515			else
516				p = unix_versions[2];
517		} else
518			p = unix_versions[0];
519	}
520
521	free(man->meta.os);
522	man->meta.os = mandoc_strdup(p);
523}
524
525static void
526post_in(CHKARGS)
527{
528	char	*s;
529
530	if (n->parent->tok != MAN_TP ||
531	    n->parent->type != ROFFT_HEAD ||
532	    n->child == NULL ||
533	    *n->child->string == '+' ||
534	    *n->child->string == '-')
535		return;
536	mandoc_asprintf(&s, "+%s", n->child->string);
537	free(n->child->string);
538	n->child->string = s;
539}
540