mdoc.c revision 1.120 
1/*	$OpenBSD: mdoc.c,v 1.120 2014/11/28 03:13:58 schwarze Exp $ */
2/*
3 * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2010, 2012, 2013, 2014 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#include <sys/types.h>
19
20#include <assert.h>
21#include <ctype.h>
22#include <stdarg.h>
23#include <stdio.h>
24#include <stdlib.h>
25#include <string.h>
26#include <time.h>
27
28#include "mdoc.h"
29#include "mandoc.h"
30#include "mandoc_aux.h"
31#include "libmdoc.h"
32#include "libmandoc.h"
33
34const	char *const __mdoc_macronames[MDOC_MAX + 1] = {
35	"Ap",		"Dd",		"Dt",		"Os",
36	"Sh",		"Ss",		"Pp",		"D1",
37	"Dl",		"Bd",		"Ed",		"Bl",
38	"El",		"It",		"Ad",		"An",
39	"Ar",		"Cd",		"Cm",		"Dv",
40	"Er",		"Ev",		"Ex",		"Fa",
41	"Fd",		"Fl",		"Fn",		"Ft",
42	"Ic",		"In",		"Li",		"Nd",
43	"Nm",		"Op",		"Ot",		"Pa",
44	"Rv",		"St",		"Va",		"Vt",
45	"Xr",		"%A",		"%B",		"%D",
46	"%I",		"%J",		"%N",		"%O",
47	"%P",		"%R",		"%T",		"%V",
48	"Ac",		"Ao",		"Aq",		"At",
49	"Bc",		"Bf",		"Bo",		"Bq",
50	"Bsx",		"Bx",		"Db",		"Dc",
51	"Do",		"Dq",		"Ec",		"Ef",
52	"Em",		"Eo",		"Fx",		"Ms",
53	"No",		"Ns",		"Nx",		"Ox",
54	"Pc",		"Pf",		"Po",		"Pq",
55	"Qc",		"Ql",		"Qo",		"Qq",
56	"Re",		"Rs",		"Sc",		"So",
57	"Sq",		"Sm",		"Sx",		"Sy",
58	"Tn",		"Ux",		"Xc",		"Xo",
59	"Fo",		"Fc",		"Oo",		"Oc",
60	"Bk",		"Ek",		"Bt",		"Hf",
61	"Fr",		"Ud",		"Lb",		"Lp",
62	"Lk",		"Mt",		"Brq",		"Bro",
63	"Brc",		"%C",		"Es",		"En",
64	"Dx",		"%Q",		"br",		"sp",
65	"%U",		"Ta",		"ll",		"text",
66	};
67
68const	char *const __mdoc_argnames[MDOC_ARG_MAX] = {
69	"split",		"nosplit",		"ragged",
70	"unfilled",		"literal",		"file",
71	"offset",		"bullet",		"dash",
72	"hyphen",		"item",			"enum",
73	"tag",			"diag",			"hang",
74	"ohang",		"inset",		"column",
75	"width",		"compact",		"std",
76	"filled",		"words",		"emphasis",
77	"symbolic",		"nested",		"centered"
78	};
79
80const	char * const *mdoc_macronames = __mdoc_macronames;
81const	char * const *mdoc_argnames = __mdoc_argnames;
82
83static	void		  mdoc_node_free(struct mdoc_node *);
84static	void		  mdoc_node_unlink(struct mdoc *,
85				struct mdoc_node *);
86static	void		  mdoc_free1(struct mdoc *);
87static	void		  mdoc_alloc1(struct mdoc *);
88static	struct mdoc_node *node_alloc(struct mdoc *, int, int,
89				enum mdoct, enum mdoc_type);
90static	void		  node_append(struct mdoc *, struct mdoc_node *);
91static	int		  mdoc_ptext(struct mdoc *, int, char *, int);
92static	int		  mdoc_pmacro(struct mdoc *, int, char *, int);
93
94
95const struct mdoc_node *
96mdoc_node(const struct mdoc *mdoc)
97{
98
99	return(mdoc->first);
100}
101
102const struct mdoc_meta *
103mdoc_meta(const struct mdoc *mdoc)
104{
105
106	return(&mdoc->meta);
107}
108
109/*
110 * Frees volatile resources (parse tree, meta-data, fields).
111 */
112static void
113mdoc_free1(struct mdoc *mdoc)
114{
115
116	if (mdoc->first)
117		mdoc_node_delete(mdoc, mdoc->first);
118	free(mdoc->meta.msec);
119	free(mdoc->meta.vol);
120	free(mdoc->meta.arch);
121	free(mdoc->meta.date);
122	free(mdoc->meta.title);
123	free(mdoc->meta.os);
124	free(mdoc->meta.name);
125}
126
127/*
128 * Allocate all volatile resources (parse tree, meta-data, fields).
129 */
130static void
131mdoc_alloc1(struct mdoc *mdoc)
132{
133
134	memset(&mdoc->meta, 0, sizeof(struct mdoc_meta));
135	mdoc->flags = 0;
136	mdoc->lastnamed = mdoc->lastsec = SEC_NONE;
137	mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node));
138	mdoc->first = mdoc->last;
139	mdoc->last->type = MDOC_ROOT;
140	mdoc->last->tok = MDOC_MAX;
141	mdoc->next = MDOC_NEXT_CHILD;
142}
143
144/*
145 * Free up volatile resources (see mdoc_free1()) then re-initialises the
146 * data with mdoc_alloc1().  After invocation, parse data has been reset
147 * and the parser is ready for re-invocation on a new tree; however,
148 * cross-parse non-volatile data is kept intact.
149 */
150void
151mdoc_reset(struct mdoc *mdoc)
152{
153
154	mdoc_free1(mdoc);
155	mdoc_alloc1(mdoc);
156}
157
158/*
159 * Completely free up all volatile and non-volatile parse resources.
160 * After invocation, the pointer is no longer usable.
161 */
162void
163mdoc_free(struct mdoc *mdoc)
164{
165
166	mdoc_free1(mdoc);
167	free(mdoc);
168}
169
170/*
171 * Allocate volatile and non-volatile parse resources.
172 */
173struct mdoc *
174mdoc_alloc(struct roff *roff, struct mparse *parse,
175	const char *defos, int quick)
176{
177	struct mdoc	*p;
178
179	p = mandoc_calloc(1, sizeof(struct mdoc));
180
181	p->parse = parse;
182	p->defos = defos;
183	p->quick = quick;
184	p->roff = roff;
185
186	mdoc_hash_init();
187	mdoc_alloc1(p);
188	return(p);
189}
190
191int
192mdoc_endparse(struct mdoc *mdoc)
193{
194
195	return(mdoc_macroend(mdoc));
196}
197
198int
199mdoc_addeqn(struct mdoc *mdoc, const struct eqn *ep)
200{
201	struct mdoc_node *n;
202
203	n = node_alloc(mdoc, ep->ln, ep->pos, MDOC_MAX, MDOC_EQN);
204	n->eqn = ep;
205	if (ep->ln > mdoc->last->line)
206		n->flags |= MDOC_LINE;
207	node_append(mdoc, n);
208	mdoc->next = MDOC_NEXT_SIBLING;
209	return(1);
210}
211
212int
213mdoc_addspan(struct mdoc *mdoc, const struct tbl_span *sp)
214{
215	struct mdoc_node *n;
216
217	n = node_alloc(mdoc, sp->line, 0, MDOC_MAX, MDOC_TBL);
218	n->span = sp;
219	node_append(mdoc, n);
220	mdoc->next = MDOC_NEXT_SIBLING;
221	return(1);
222}
223
224/*
225 * Main parse routine.  Parses a single line -- really just hands off to
226 * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
227 */
228int
229mdoc_parseln(struct mdoc *mdoc, int ln, char *buf, int offs)
230{
231
232	if (mdoc->last->type != MDOC_EQN || ln > mdoc->last->line)
233		mdoc->flags |= MDOC_NEWLINE;
234
235	/*
236	 * Let the roff nS register switch SYNOPSIS mode early,
237	 * such that the parser knows at all times
238	 * whether this mode is on or off.
239	 * Note that this mode is also switched by the Sh macro.
240	 */
241	if (roff_getreg(mdoc->roff, "nS"))
242		mdoc->flags |= MDOC_SYNOPSIS;
243	else
244		mdoc->flags &= ~MDOC_SYNOPSIS;
245
246	return(roff_getcontrol(mdoc->roff, buf, &offs) ?
247	    mdoc_pmacro(mdoc, ln, buf, offs) :
248	    mdoc_ptext(mdoc, ln, buf, offs));
249}
250
251int
252mdoc_macro(MACRO_PROT_ARGS)
253{
254	assert(tok < MDOC_MAX);
255
256	if (mdoc->flags & MDOC_PBODY) {
257		if (tok == MDOC_Dt) {
258			mandoc_vmsg(MANDOCERR_DT_LATE,
259			    mdoc->parse, line, ppos,
260			    "Dt %s", buf + *pos);
261			return(1);
262		}
263	} else if ( ! (mdoc_macros[tok].flags & MDOC_PROLOGUE)) {
264		if (mdoc->meta.title == NULL) {
265			mandoc_vmsg(MANDOCERR_DT_NOTITLE,
266			    mdoc->parse, line, ppos, "%s %s",
267			    mdoc_macronames[tok], buf + *pos);
268			mdoc->meta.title = mandoc_strdup("UNTITLED");
269		}
270		if (NULL == mdoc->meta.vol)
271			mdoc->meta.vol = mandoc_strdup("LOCAL");
272		mdoc->flags |= MDOC_PBODY;
273	}
274
275	return((*mdoc_macros[tok].fp)(mdoc, tok, line, ppos, pos, buf));
276}
277
278
279static void
280node_append(struct mdoc *mdoc, struct mdoc_node *p)
281{
282
283	assert(mdoc->last);
284	assert(mdoc->first);
285	assert(MDOC_ROOT != p->type);
286
287	switch (mdoc->next) {
288	case MDOC_NEXT_SIBLING:
289		mdoc->last->next = p;
290		p->prev = mdoc->last;
291		p->parent = mdoc->last->parent;
292		break;
293	case MDOC_NEXT_CHILD:
294		mdoc->last->child = p;
295		p->parent = mdoc->last;
296		break;
297	default:
298		abort();
299		/* NOTREACHED */
300	}
301
302	p->parent->nchild++;
303
304	/*
305	 * Copy over the normalised-data pointer of our parent.  Not
306	 * everybody has one, but copying a null pointer is fine.
307	 */
308
309	switch (p->type) {
310	case MDOC_BODY:
311		if (ENDBODY_NOT != p->end)
312			break;
313		/* FALLTHROUGH */
314	case MDOC_TAIL:
315		/* FALLTHROUGH */
316	case MDOC_HEAD:
317		p->norm = p->parent->norm;
318		break;
319	default:
320		break;
321	}
322
323	mdoc_valid_pre(mdoc, p);
324
325	switch (p->type) {
326	case MDOC_HEAD:
327		assert(MDOC_BLOCK == p->parent->type);
328		p->parent->head = p;
329		break;
330	case MDOC_TAIL:
331		assert(MDOC_BLOCK == p->parent->type);
332		p->parent->tail = p;
333		break;
334	case MDOC_BODY:
335		if (p->end)
336			break;
337		assert(MDOC_BLOCK == p->parent->type);
338		p->parent->body = p;
339		break;
340	default:
341		break;
342	}
343
344	mdoc->last = p;
345
346	switch (p->type) {
347	case MDOC_TBL:
348		/* FALLTHROUGH */
349	case MDOC_TEXT:
350		mdoc_valid_post(mdoc);
351		break;
352	default:
353		break;
354	}
355}
356
357static struct mdoc_node *
358node_alloc(struct mdoc *mdoc, int line, int pos,
359		enum mdoct tok, enum mdoc_type type)
360{
361	struct mdoc_node *p;
362
363	p = mandoc_calloc(1, sizeof(struct mdoc_node));
364	p->sec = mdoc->lastsec;
365	p->line = line;
366	p->pos = pos;
367	p->lastline = line;
368	p->tok = tok;
369	p->type = type;
370
371	/* Flag analysis. */
372
373	if (MDOC_SYNOPSIS & mdoc->flags)
374		p->flags |= MDOC_SYNPRETTY;
375	else
376		p->flags &= ~MDOC_SYNPRETTY;
377	if (MDOC_NEWLINE & mdoc->flags)
378		p->flags |= MDOC_LINE;
379	mdoc->flags &= ~MDOC_NEWLINE;
380
381	return(p);
382}
383
384void
385mdoc_tail_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
386{
387	struct mdoc_node *p;
388
389	p = node_alloc(mdoc, line, pos, tok, MDOC_TAIL);
390	node_append(mdoc, p);
391	mdoc->next = MDOC_NEXT_CHILD;
392}
393
394struct mdoc_node *
395mdoc_head_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
396{
397	struct mdoc_node *p;
398
399	assert(mdoc->first);
400	assert(mdoc->last);
401	p = node_alloc(mdoc, line, pos, tok, MDOC_HEAD);
402	node_append(mdoc, p);
403	mdoc->next = MDOC_NEXT_CHILD;
404	return(p);
405}
406
407struct mdoc_node *
408mdoc_body_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
409{
410	struct mdoc_node *p;
411
412	p = node_alloc(mdoc, line, pos, tok, MDOC_BODY);
413	node_append(mdoc, p);
414	mdoc->next = MDOC_NEXT_CHILD;
415	return(p);
416}
417
418void
419mdoc_endbody_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok,
420		struct mdoc_node *body, enum mdoc_endbody end)
421{
422	struct mdoc_node *p;
423
424	p = node_alloc(mdoc, line, pos, tok, MDOC_BODY);
425	p->pending = body;
426	p->norm = body->norm;
427	p->end = end;
428	node_append(mdoc, p);
429	mdoc->next = MDOC_NEXT_SIBLING;
430}
431
432struct mdoc_node *
433mdoc_block_alloc(struct mdoc *mdoc, int line, int pos,
434		enum mdoct tok, struct mdoc_arg *args)
435{
436	struct mdoc_node *p;
437
438	p = node_alloc(mdoc, line, pos, tok, MDOC_BLOCK);
439	p->args = args;
440	if (p->args)
441		(args->refcnt)++;
442
443	switch (tok) {
444	case MDOC_Bd:
445		/* FALLTHROUGH */
446	case MDOC_Bf:
447		/* FALLTHROUGH */
448	case MDOC_Bl:
449		/* FALLTHROUGH */
450	case MDOC_En:
451		/* FALLTHROUGH */
452	case MDOC_Rs:
453		p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
454		break;
455	default:
456		break;
457	}
458	node_append(mdoc, p);
459	mdoc->next = MDOC_NEXT_CHILD;
460	return(p);
461}
462
463void
464mdoc_elem_alloc(struct mdoc *mdoc, int line, int pos,
465		enum mdoct tok, struct mdoc_arg *args)
466{
467	struct mdoc_node *p;
468
469	p = node_alloc(mdoc, line, pos, tok, MDOC_ELEM);
470	p->args = args;
471	if (p->args)
472		(args->refcnt)++;
473
474	switch (tok) {
475	case MDOC_An:
476		p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
477		break;
478	default:
479		break;
480	}
481	node_append(mdoc, p);
482	mdoc->next = MDOC_NEXT_CHILD;
483}
484
485void
486mdoc_word_alloc(struct mdoc *mdoc, int line, int pos, const char *p)
487{
488	struct mdoc_node *n;
489
490	n = node_alloc(mdoc, line, pos, MDOC_MAX, MDOC_TEXT);
491	n->string = roff_strdup(mdoc->roff, p);
492	node_append(mdoc, n);
493	mdoc->next = MDOC_NEXT_SIBLING;
494}
495
496void
497mdoc_word_append(struct mdoc *mdoc, const char *p)
498{
499	struct mdoc_node	*n;
500	char			*addstr, *newstr;
501
502	n = mdoc->last;
503	addstr = roff_strdup(mdoc->roff, p);
504	mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
505	free(addstr);
506	free(n->string);
507	n->string = newstr;
508	mdoc->next = MDOC_NEXT_SIBLING;
509}
510
511static void
512mdoc_node_free(struct mdoc_node *p)
513{
514
515	if (MDOC_BLOCK == p->type || MDOC_ELEM == p->type)
516		free(p->norm);
517	if (p->string)
518		free(p->string);
519	if (p->args)
520		mdoc_argv_free(p->args);
521	free(p);
522}
523
524static void
525mdoc_node_unlink(struct mdoc *mdoc, struct mdoc_node *n)
526{
527
528	/* Adjust siblings. */
529
530	if (n->prev)
531		n->prev->next = n->next;
532	if (n->next)
533		n->next->prev = n->prev;
534
535	/* Adjust parent. */
536
537	if (n->parent) {
538		n->parent->nchild--;
539		if (n->parent->child == n)
540			n->parent->child = n->prev ? n->prev : n->next;
541		if (n->parent->last == n)
542			n->parent->last = n->prev ? n->prev : NULL;
543	}
544
545	/* Adjust parse point, if applicable. */
546
547	if (mdoc && mdoc->last == n) {
548		if (n->prev) {
549			mdoc->last = n->prev;
550			mdoc->next = MDOC_NEXT_SIBLING;
551		} else {
552			mdoc->last = n->parent;
553			mdoc->next = MDOC_NEXT_CHILD;
554		}
555	}
556
557	if (mdoc && mdoc->first == n)
558		mdoc->first = NULL;
559}
560
561void
562mdoc_node_delete(struct mdoc *mdoc, struct mdoc_node *p)
563{
564
565	while (p->child) {
566		assert(p->nchild);
567		mdoc_node_delete(mdoc, p->child);
568	}
569	assert(0 == p->nchild);
570
571	mdoc_node_unlink(mdoc, p);
572	mdoc_node_free(p);
573}
574
575void
576mdoc_node_relink(struct mdoc *mdoc, struct mdoc_node *p)
577{
578
579	mdoc_node_unlink(mdoc, p);
580	node_append(mdoc, p);
581}
582
583/*
584 * Parse free-form text, that is, a line that does not begin with the
585 * control character.
586 */
587static int
588mdoc_ptext(struct mdoc *mdoc, int line, char *buf, int offs)
589{
590	char		 *c, *ws, *end;
591	struct mdoc_node *n;
592
593	assert(mdoc->last);
594	n = mdoc->last;
595
596	/*
597	 * Divert directly to list processing if we're encountering a
598	 * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
599	 * (a MDOC_BODY means it's already open, in which case we should
600	 * process within its context in the normal way).
601	 */
602
603	if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
604	    LIST_column == n->norm->Bl.type) {
605		/* `Bl' is open without any children. */
606		mdoc->flags |= MDOC_FREECOL;
607		return(mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf));
608	}
609
610	if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
611	    NULL != n->parent &&
612	    MDOC_Bl == n->parent->tok &&
613	    LIST_column == n->parent->norm->Bl.type) {
614		/* `Bl' has block-level `It' children. */
615		mdoc->flags |= MDOC_FREECOL;
616		return(mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf));
617	}
618
619	/*
620	 * Search for the beginning of unescaped trailing whitespace (ws)
621	 * and for the first character not to be output (end).
622	 */
623
624	/* FIXME: replace with strcspn(). */
625	ws = NULL;
626	for (c = end = buf + offs; *c; c++) {
627		switch (*c) {
628		case ' ':
629			if (NULL == ws)
630				ws = c;
631			continue;
632		case '\t':
633			/*
634			 * Always warn about trailing tabs,
635			 * even outside literal context,
636			 * where they should be put on the next line.
637			 */
638			if (NULL == ws)
639				ws = c;
640			/*
641			 * Strip trailing tabs in literal context only;
642			 * outside, they affect the next line.
643			 */
644			if (MDOC_LITERAL & mdoc->flags)
645				continue;
646			break;
647		case '\\':
648			/* Skip the escaped character, too, if any. */
649			if (c[1])
650				c++;
651			/* FALLTHROUGH */
652		default:
653			ws = NULL;
654			break;
655		}
656		end = c + 1;
657	}
658	*end = '\0';
659
660	if (ws)
661		mandoc_msg(MANDOCERR_SPACE_EOL, mdoc->parse,
662		    line, (int)(ws-buf), NULL);
663
664	if (buf[offs] == '\0' && ! (mdoc->flags & MDOC_LITERAL)) {
665		mandoc_msg(MANDOCERR_FI_BLANK, mdoc->parse,
666		    line, (int)(c - buf), NULL);
667
668		/*
669		 * Insert a `sp' in the case of a blank line.  Technically,
670		 * blank lines aren't allowed, but enough manuals assume this
671		 * behaviour that we want to work around it.
672		 */
673		mdoc_elem_alloc(mdoc, line, offs, MDOC_sp, NULL);
674		mdoc->next = MDOC_NEXT_SIBLING;
675		mdoc_valid_post(mdoc);
676		return(1);
677	}
678
679	mdoc_word_alloc(mdoc, line, offs, buf+offs);
680
681	if (mdoc->flags & MDOC_LITERAL)
682		return(1);
683
684	/*
685	 * End-of-sentence check.  If the last character is an unescaped
686	 * EOS character, then flag the node as being the end of a
687	 * sentence.  The front-end will know how to interpret this.
688	 */
689
690	assert(buf < end);
691
692	if (mandoc_eos(buf+offs, (size_t)(end-buf-offs)))
693		mdoc->last->flags |= MDOC_EOS;
694
695	return(1);
696}
697
698/*
699 * Parse a macro line, that is, a line beginning with the control
700 * character.
701 */
702static int
703mdoc_pmacro(struct mdoc *mdoc, int ln, char *buf, int offs)
704{
705	struct mdoc_node *n;
706	const char	 *cp;
707	enum mdoct	  tok;
708	int		  i, sv;
709	char		  mac[5];
710
711	sv = offs;
712
713	/*
714	 * Copy the first word into a nil-terminated buffer.
715	 * Stop when a space, tab, escape, or eoln is encountered.
716	 */
717
718	i = 0;
719	while (i < 4 && strchr(" \t\\", buf[offs]) == NULL)
720		mac[i++] = buf[offs++];
721
722	mac[i] = '\0';
723
724	tok = (i > 1 && i < 4) ? mdoc_hash_find(mac) : MDOC_MAX;
725
726	if (tok == MDOC_MAX) {
727		mandoc_msg(MANDOCERR_MACRO, mdoc->parse,
728		    ln, sv, buf + sv - 1);
729		return(1);
730	}
731
732	/* Skip a leading escape sequence or tab. */
733
734	switch (buf[offs]) {
735	case '\\':
736		cp = buf + offs + 1;
737		mandoc_escape(&cp, NULL, NULL);
738		offs = cp - buf;
739		break;
740	case '\t':
741		offs++;
742		break;
743	default:
744		break;
745	}
746
747	/* Jump to the next non-whitespace word. */
748
749	while (buf[offs] && ' ' == buf[offs])
750		offs++;
751
752	/*
753	 * Trailing whitespace.  Note that tabs are allowed to be passed
754	 * into the parser as "text", so we only warn about spaces here.
755	 */
756
757	if ('\0' == buf[offs] && ' ' == buf[offs - 1])
758		mandoc_msg(MANDOCERR_SPACE_EOL, mdoc->parse,
759		    ln, offs - 1, NULL);
760
761	/*
762	 * If an initial macro or a list invocation, divert directly
763	 * into macro processing.
764	 */
765
766	if (NULL == mdoc->last || MDOC_It == tok || MDOC_El == tok)
767		return(mdoc_macro(mdoc, tok, ln, sv, &offs, buf));
768
769	n = mdoc->last;
770	assert(mdoc->last);
771
772	/*
773	 * If the first macro of a `Bl -column', open an `It' block
774	 * context around the parsed macro.
775	 */
776
777	if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
778	    LIST_column == n->norm->Bl.type) {
779		mdoc->flags |= MDOC_FREECOL;
780		return(mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf));
781	}
782
783	/*
784	 * If we're following a block-level `It' within a `Bl -column'
785	 * context (perhaps opened in the above block or in ptext()),
786	 * then open an `It' block context around the parsed macro.
787	 */
788
789	if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
790	    NULL != n->parent &&
791	    MDOC_Bl == n->parent->tok &&
792	    LIST_column == n->parent->norm->Bl.type) {
793		mdoc->flags |= MDOC_FREECOL;
794		return(mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf));
795	}
796
797	/* Normal processing of a macro. */
798
799	if ( ! mdoc_macro(mdoc, tok, ln, sv, &offs, buf))
800		return(0);
801
802	/* In quick mode (for mandocdb), abort after the NAME section. */
803
804	if (mdoc->quick && MDOC_Sh == tok &&
805	    SEC_NAME != mdoc->last->sec)
806		return(2);
807
808	return(1);
809}
810
811enum mdelim
812mdoc_isdelim(const char *p)
813{
814
815	if ('\0' == p[0])
816		return(DELIM_NONE);
817
818	if ('\0' == p[1])
819		switch (p[0]) {
820		case '(':
821			/* FALLTHROUGH */
822		case '[':
823			return(DELIM_OPEN);
824		case '|':
825			return(DELIM_MIDDLE);
826		case '.':
827			/* FALLTHROUGH */
828		case ',':
829			/* FALLTHROUGH */
830		case ';':
831			/* FALLTHROUGH */
832		case ':':
833			/* FALLTHROUGH */
834		case '?':
835			/* FALLTHROUGH */
836		case '!':
837			/* FALLTHROUGH */
838		case ')':
839			/* FALLTHROUGH */
840		case ']':
841			return(DELIM_CLOSE);
842		default:
843			return(DELIM_NONE);
844		}
845
846	if ('\\' != p[0])
847		return(DELIM_NONE);
848
849	if (0 == strcmp(p + 1, "."))
850		return(DELIM_CLOSE);
851	if (0 == strcmp(p + 1, "fR|\\fP"))
852		return(DELIM_MIDDLE);
853
854	return(DELIM_NONE);
855}
856
857void
858mdoc_deroff(char **dest, const struct mdoc_node *n)
859{
860	char	*cp;
861	size_t	 sz;
862
863	if (MDOC_TEXT != n->type) {
864		for (n = n->child; n; n = n->next)
865			mdoc_deroff(dest, n);
866		return;
867	}
868
869	/* Skip leading whitespace. */
870
871	for (cp = n->string; '\0' != *cp; cp++)
872		if (0 == isspace((unsigned char)*cp))
873			break;
874
875	/* Skip trailing whitespace. */
876
877	for (sz = strlen(cp); sz; sz--)
878		if (0 == isspace((unsigned char)cp[sz-1]))
879			break;
880
881	/* Skip empty strings. */
882
883	if (0 == sz)
884		return;
885
886	if (NULL == *dest) {
887		*dest = mandoc_strndup(cp, sz);
888		return;
889	}
890
891	mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
892	free(*dest);
893	*dest = cp;
894}
895