mdoc.c revision 1.138 
1/*	$OpenBSD: mdoc.c,v 1.138 2015/04/19 14:25:05 schwarze Exp $ */
2/*
3 * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2010, 2012-2015 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 <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 "mandoc_aux.h"
29#include "mandoc.h"
30#include "roff.h"
31#include "mdoc.h"
32#include "libmandoc.h"
33#include "roff_int.h"
34#include "libmdoc.h"
35
36const	char *const __mdoc_macronames[MDOC_MAX + 1] = {
37	"text",
38	"Ap",		"Dd",		"Dt",		"Os",
39	"Sh",		"Ss",		"Pp",		"D1",
40	"Dl",		"Bd",		"Ed",		"Bl",
41	"El",		"It",		"Ad",		"An",
42	"Ar",		"Cd",		"Cm",		"Dv",
43	"Er",		"Ev",		"Ex",		"Fa",
44	"Fd",		"Fl",		"Fn",		"Ft",
45	"Ic",		"In",		"Li",		"Nd",
46	"Nm",		"Op",		"Ot",		"Pa",
47	"Rv",		"St",		"Va",		"Vt",
48	"Xr",		"%A",		"%B",		"%D",
49	"%I",		"%J",		"%N",		"%O",
50	"%P",		"%R",		"%T",		"%V",
51	"Ac",		"Ao",		"Aq",		"At",
52	"Bc",		"Bf",		"Bo",		"Bq",
53	"Bsx",		"Bx",		"Db",		"Dc",
54	"Do",		"Dq",		"Ec",		"Ef",
55	"Em",		"Eo",		"Fx",		"Ms",
56	"No",		"Ns",		"Nx",		"Ox",
57	"Pc",		"Pf",		"Po",		"Pq",
58	"Qc",		"Ql",		"Qo",		"Qq",
59	"Re",		"Rs",		"Sc",		"So",
60	"Sq",		"Sm",		"Sx",		"Sy",
61	"Tn",		"Ux",		"Xc",		"Xo",
62	"Fo",		"Fc",		"Oo",		"Oc",
63	"Bk",		"Ek",		"Bt",		"Hf",
64	"Fr",		"Ud",		"Lb",		"Lp",
65	"Lk",		"Mt",		"Brq",		"Bro",
66	"Brc",		"%C",		"Es",		"En",
67	"Dx",		"%Q",		"br",		"sp",
68	"%U",		"Ta",		"ll",
69};
70
71const	char *const __mdoc_argnames[MDOC_ARG_MAX] = {
72	"split",		"nosplit",		"ragged",
73	"unfilled",		"literal",		"file",
74	"offset",		"bullet",		"dash",
75	"hyphen",		"item",			"enum",
76	"tag",			"diag",			"hang",
77	"ohang",		"inset",		"column",
78	"width",		"compact",		"std",
79	"filled",		"words",		"emphasis",
80	"symbolic",		"nested",		"centered"
81	};
82
83const	char * const *mdoc_macronames = __mdoc_macronames + 1;
84const	char * const *mdoc_argnames = __mdoc_argnames;
85
86static	int		  mdoc_ptext(struct roff_man *, int, char *, int);
87static	int		  mdoc_pmacro(struct roff_man *, int, char *, int);
88
89
90void
91mdoc_endparse(struct roff_man *mdoc)
92{
93
94	mdoc_macroend(mdoc);
95}
96
97/*
98 * Main parse routine.  Parses a single line -- really just hands off to
99 * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
100 */
101int
102mdoc_parseln(struct roff_man *mdoc, int ln, char *buf, int offs)
103{
104
105	if (mdoc->last->type != ROFFT_EQN || ln > mdoc->last->line)
106		mdoc->flags |= MDOC_NEWLINE;
107
108	/*
109	 * Let the roff nS register switch SYNOPSIS mode early,
110	 * such that the parser knows at all times
111	 * whether this mode is on or off.
112	 * Note that this mode is also switched by the Sh macro.
113	 */
114	if (roff_getreg(mdoc->roff, "nS"))
115		mdoc->flags |= MDOC_SYNOPSIS;
116	else
117		mdoc->flags &= ~MDOC_SYNOPSIS;
118
119	return(roff_getcontrol(mdoc->roff, buf, &offs) ?
120	    mdoc_pmacro(mdoc, ln, buf, offs) :
121	    mdoc_ptext(mdoc, ln, buf, offs));
122}
123
124void
125mdoc_macro(MACRO_PROT_ARGS)
126{
127	assert(tok > TOKEN_NONE && tok < MDOC_MAX);
128
129	if (mdoc->flags & MDOC_PBODY) {
130		if (tok == MDOC_Dt) {
131			mandoc_vmsg(MANDOCERR_DT_LATE,
132			    mdoc->parse, line, ppos,
133			    "Dt %s", buf + *pos);
134			return;
135		}
136	} else if ( ! (mdoc_macros[tok].flags & MDOC_PROLOGUE)) {
137		if (mdoc->meta.title == NULL) {
138			mandoc_vmsg(MANDOCERR_DT_NOTITLE,
139			    mdoc->parse, line, ppos, "%s %s",
140			    mdoc_macronames[tok], buf + *pos);
141			mdoc->meta.title = mandoc_strdup("UNTITLED");
142		}
143		if (NULL == mdoc->meta.vol)
144			mdoc->meta.vol = mandoc_strdup("LOCAL");
145		mdoc->flags |= MDOC_PBODY;
146	}
147	(*mdoc_macros[tok].fp)(mdoc, tok, line, ppos, pos, buf);
148}
149
150void
151mdoc_tail_alloc(struct roff_man *mdoc, int line, int pos, int tok)
152{
153	struct roff_node *p;
154
155	p = roff_node_alloc(mdoc, line, pos, ROFFT_TAIL, tok);
156	roff_node_append(mdoc, p);
157	mdoc->next = ROFF_NEXT_CHILD;
158}
159
160struct roff_node *
161mdoc_endbody_alloc(struct roff_man *mdoc, int line, int pos, int tok,
162		struct roff_node *body, enum mdoc_endbody end)
163{
164	struct roff_node *p;
165
166	body->flags |= MDOC_ENDED;
167	body->parent->flags |= MDOC_ENDED;
168	p = roff_node_alloc(mdoc, line, pos, ROFFT_BODY, tok);
169	p->body = body;
170	p->norm = body->norm;
171	p->end = end;
172	roff_node_append(mdoc, p);
173	mdoc->next = ROFF_NEXT_SIBLING;
174	return(p);
175}
176
177struct roff_node *
178mdoc_block_alloc(struct roff_man *mdoc, int line, int pos,
179	int tok, struct mdoc_arg *args)
180{
181	struct roff_node *p;
182
183	p = roff_node_alloc(mdoc, line, pos, ROFFT_BLOCK, tok);
184	p->args = args;
185	if (p->args)
186		(args->refcnt)++;
187
188	switch (tok) {
189	case MDOC_Bd:
190		/* FALLTHROUGH */
191	case MDOC_Bf:
192		/* FALLTHROUGH */
193	case MDOC_Bl:
194		/* FALLTHROUGH */
195	case MDOC_En:
196		/* FALLTHROUGH */
197	case MDOC_Rs:
198		p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
199		break;
200	default:
201		break;
202	}
203	roff_node_append(mdoc, p);
204	mdoc->next = ROFF_NEXT_CHILD;
205	return(p);
206}
207
208void
209mdoc_elem_alloc(struct roff_man *mdoc, int line, int pos,
210	int tok, struct mdoc_arg *args)
211{
212	struct roff_node *p;
213
214	p = roff_node_alloc(mdoc, line, pos, ROFFT_ELEM, tok);
215	p->args = args;
216	if (p->args)
217		(args->refcnt)++;
218
219	switch (tok) {
220	case MDOC_An:
221		p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
222		break;
223	default:
224		break;
225	}
226	roff_node_append(mdoc, p);
227	mdoc->next = ROFF_NEXT_CHILD;
228}
229
230void
231mdoc_node_relink(struct roff_man *mdoc, struct roff_node *p)
232{
233
234	roff_node_unlink(mdoc, p);
235	roff_node_append(mdoc, p);
236}
237
238/*
239 * Parse free-form text, that is, a line that does not begin with the
240 * control character.
241 */
242static int
243mdoc_ptext(struct roff_man *mdoc, int line, char *buf, int offs)
244{
245	struct roff_node *n;
246	char		 *c, *ws, *end;
247
248	assert(mdoc->last);
249	n = mdoc->last;
250
251	/*
252	 * Divert directly to list processing if we're encountering a
253	 * columnar ROFFT_BLOCK with or without a prior ROFFT_BLOCK entry
254	 * (a ROFFT_BODY means it's already open, in which case we should
255	 * process within its context in the normal way).
256	 */
257
258	if (n->tok == MDOC_Bl && n->type == ROFFT_BODY &&
259	    n->end == ENDBODY_NOT && n->norm->Bl.type == LIST_column) {
260		/* `Bl' is open without any children. */
261		mdoc->flags |= MDOC_FREECOL;
262		mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf);
263		return(1);
264	}
265
266	if (n->tok == MDOC_It && n->type == ROFFT_BLOCK &&
267	    NULL != n->parent &&
268	    MDOC_Bl == n->parent->tok &&
269	    LIST_column == n->parent->norm->Bl.type) {
270		/* `Bl' has block-level `It' children. */
271		mdoc->flags |= MDOC_FREECOL;
272		mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf);
273		return(1);
274	}
275
276	/*
277	 * Search for the beginning of unescaped trailing whitespace (ws)
278	 * and for the first character not to be output (end).
279	 */
280
281	/* FIXME: replace with strcspn(). */
282	ws = NULL;
283	for (c = end = buf + offs; *c; c++) {
284		switch (*c) {
285		case ' ':
286			if (NULL == ws)
287				ws = c;
288			continue;
289		case '\t':
290			/*
291			 * Always warn about trailing tabs,
292			 * even outside literal context,
293			 * where they should be put on the next line.
294			 */
295			if (NULL == ws)
296				ws = c;
297			/*
298			 * Strip trailing tabs in literal context only;
299			 * outside, they affect the next line.
300			 */
301			if (MDOC_LITERAL & mdoc->flags)
302				continue;
303			break;
304		case '\\':
305			/* Skip the escaped character, too, if any. */
306			if (c[1])
307				c++;
308			/* FALLTHROUGH */
309		default:
310			ws = NULL;
311			break;
312		}
313		end = c + 1;
314	}
315	*end = '\0';
316
317	if (ws)
318		mandoc_msg(MANDOCERR_SPACE_EOL, mdoc->parse,
319		    line, (int)(ws-buf), NULL);
320
321	if (buf[offs] == '\0' && ! (mdoc->flags & MDOC_LITERAL)) {
322		mandoc_msg(MANDOCERR_FI_BLANK, mdoc->parse,
323		    line, (int)(c - buf), NULL);
324
325		/*
326		 * Insert a `sp' in the case of a blank line.  Technically,
327		 * blank lines aren't allowed, but enough manuals assume this
328		 * behaviour that we want to work around it.
329		 */
330		mdoc_elem_alloc(mdoc, line, offs, MDOC_sp, NULL);
331		mdoc->next = ROFF_NEXT_SIBLING;
332		mdoc_valid_post(mdoc);
333		return(1);
334	}
335
336	roff_word_alloc(mdoc, line, offs, buf+offs);
337
338	if (mdoc->flags & MDOC_LITERAL)
339		return(1);
340
341	/*
342	 * End-of-sentence check.  If the last character is an unescaped
343	 * EOS character, then flag the node as being the end of a
344	 * sentence.  The front-end will know how to interpret this.
345	 */
346
347	assert(buf < end);
348
349	if (mandoc_eos(buf+offs, (size_t)(end-buf-offs)))
350		mdoc->last->flags |= MDOC_EOS;
351	return(1);
352}
353
354/*
355 * Parse a macro line, that is, a line beginning with the control
356 * character.
357 */
358static int
359mdoc_pmacro(struct roff_man *mdoc, int ln, char *buf, int offs)
360{
361	struct roff_node *n;
362	const char	 *cp;
363	int		  tok;
364	int		  i, sv;
365	char		  mac[5];
366
367	sv = offs;
368
369	/*
370	 * Copy the first word into a nil-terminated buffer.
371	 * Stop when a space, tab, escape, or eoln is encountered.
372	 */
373
374	i = 0;
375	while (i < 4 && strchr(" \t\\", buf[offs]) == NULL)
376		mac[i++] = buf[offs++];
377
378	mac[i] = '\0';
379
380	tok = (i > 1 && i < 4) ? mdoc_hash_find(mac) : TOKEN_NONE;
381
382	if (tok == TOKEN_NONE) {
383		mandoc_msg(MANDOCERR_MACRO, mdoc->parse,
384		    ln, sv, buf + sv - 1);
385		return(1);
386	}
387
388	/* Skip a leading escape sequence or tab. */
389
390	switch (buf[offs]) {
391	case '\\':
392		cp = buf + offs + 1;
393		mandoc_escape(&cp, NULL, NULL);
394		offs = cp - buf;
395		break;
396	case '\t':
397		offs++;
398		break;
399	default:
400		break;
401	}
402
403	/* Jump to the next non-whitespace word. */
404
405	while (buf[offs] && ' ' == buf[offs])
406		offs++;
407
408	/*
409	 * Trailing whitespace.  Note that tabs are allowed to be passed
410	 * into the parser as "text", so we only warn about spaces here.
411	 */
412
413	if ('\0' == buf[offs] && ' ' == buf[offs - 1])
414		mandoc_msg(MANDOCERR_SPACE_EOL, mdoc->parse,
415		    ln, offs - 1, NULL);
416
417	/*
418	 * If an initial macro or a list invocation, divert directly
419	 * into macro processing.
420	 */
421
422	if (NULL == mdoc->last || MDOC_It == tok || MDOC_El == tok) {
423		mdoc_macro(mdoc, tok, ln, sv, &offs, buf);
424		return(1);
425	}
426
427	n = mdoc->last;
428	assert(mdoc->last);
429
430	/*
431	 * If the first macro of a `Bl -column', open an `It' block
432	 * context around the parsed macro.
433	 */
434
435	if (n->tok == MDOC_Bl && n->type == ROFFT_BODY &&
436	    n->end == ENDBODY_NOT && n->norm->Bl.type == LIST_column) {
437		mdoc->flags |= MDOC_FREECOL;
438		mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf);
439		return(1);
440	}
441
442	/*
443	 * If we're following a block-level `It' within a `Bl -column'
444	 * context (perhaps opened in the above block or in ptext()),
445	 * then open an `It' block context around the parsed macro.
446	 */
447
448	if (n->tok == MDOC_It && n->type == ROFFT_BLOCK &&
449	    NULL != n->parent &&
450	    MDOC_Bl == n->parent->tok &&
451	    LIST_column == n->parent->norm->Bl.type) {
452		mdoc->flags |= MDOC_FREECOL;
453		mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf);
454		return(1);
455	}
456
457	/* Normal processing of a macro. */
458
459	mdoc_macro(mdoc, tok, ln, sv, &offs, buf);
460
461	/* In quick mode (for mandocdb), abort after the NAME section. */
462
463	if (mdoc->quick && MDOC_Sh == tok &&
464	    SEC_NAME != mdoc->last->sec)
465		return(2);
466
467	return(1);
468}
469
470enum mdelim
471mdoc_isdelim(const char *p)
472{
473
474	if ('\0' == p[0])
475		return(DELIM_NONE);
476
477	if ('\0' == p[1])
478		switch (p[0]) {
479		case '(':
480			/* FALLTHROUGH */
481		case '[':
482			return(DELIM_OPEN);
483		case '|':
484			return(DELIM_MIDDLE);
485		case '.':
486			/* FALLTHROUGH */
487		case ',':
488			/* FALLTHROUGH */
489		case ';':
490			/* FALLTHROUGH */
491		case ':':
492			/* FALLTHROUGH */
493		case '?':
494			/* FALLTHROUGH */
495		case '!':
496			/* FALLTHROUGH */
497		case ')':
498			/* FALLTHROUGH */
499		case ']':
500			return(DELIM_CLOSE);
501		default:
502			return(DELIM_NONE);
503		}
504
505	if ('\\' != p[0])
506		return(DELIM_NONE);
507
508	if (0 == strcmp(p + 1, "."))
509		return(DELIM_CLOSE);
510	if (0 == strcmp(p + 1, "fR|\\fP"))
511		return(DELIM_MIDDLE);
512
513	return(DELIM_NONE);
514}
515
516void
517mdoc_deroff(char **dest, const struct roff_node *n)
518{
519	char	*cp;
520	size_t	 sz;
521
522	if (n->type != ROFFT_TEXT) {
523		for (n = n->child; n; n = n->next)
524			mdoc_deroff(dest, n);
525		return;
526	}
527
528	/* Skip leading whitespace. */
529
530	for (cp = n->string; '\0' != *cp; cp++)
531		if (0 == isspace((unsigned char)*cp))
532			break;
533
534	/* Skip trailing whitespace. */
535
536	for (sz = strlen(cp); sz; sz--)
537		if (0 == isspace((unsigned char)cp[sz-1]))
538			break;
539
540	/* Skip empty strings. */
541
542	if (0 == sz)
543		return;
544
545	if (NULL == *dest) {
546		*dest = mandoc_strndup(cp, sz);
547		return;
548	}
549
550	mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
551	free(*dest);
552	*dest = cp;
553}
554