man_macro.c revision 1.4 
1/*	Id: man_macro.c,v 1.115 2017/01/10 13:47:00 schwarze Exp  */
2/*
3 * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2012, 2013, 2014, 2015 Ingo Schwarze <schwarze@openbsd.org>
5 * Copyright (c) 2013 Franco Fichtner <franco@lastsummer.de>
6 *
7 * Permission to use, copy, modify, and distribute this software for any
8 * purpose with or without fee is hereby granted, provided that the above
9 * copyright notice and this permission notice appear in all copies.
10 *
11 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
12 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
13 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
14 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
15 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
16 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
17 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
18 */
19#include "config.h"
20
21#include <sys/types.h>
22
23#include <assert.h>
24#include <ctype.h>
25#include <stdlib.h>
26#include <string.h>
27
28#include "mandoc.h"
29#include "roff.h"
30#include "man.h"
31#include "libmandoc.h"
32#include "roff_int.h"
33#include "libman.h"
34
35static	void		 blk_close(MACRO_PROT_ARGS);
36static	void		 blk_exp(MACRO_PROT_ARGS);
37static	void		 blk_imp(MACRO_PROT_ARGS);
38static	void		 in_line_eoln(MACRO_PROT_ARGS);
39static	int		 man_args(struct roff_man *, int,
40				int *, char *, char **);
41static	void		 rew_scope(struct roff_man *, int);
42
43const	struct man_macro __man_macros[MAN_MAX] = {
44	{ in_line_eoln, MAN_NSCOPED }, /* br */
45	{ in_line_eoln, MAN_BSCOPE }, /* TH */
46	{ blk_imp, MAN_BSCOPE | MAN_SCOPED }, /* SH */
47	{ blk_imp, MAN_BSCOPE | MAN_SCOPED }, /* SS */
48	{ blk_imp, MAN_BSCOPE | MAN_SCOPED }, /* TP */
49	{ blk_imp, MAN_BSCOPE }, /* LP */
50	{ blk_imp, MAN_BSCOPE }, /* PP */
51	{ blk_imp, MAN_BSCOPE }, /* P */
52	{ blk_imp, MAN_BSCOPE }, /* IP */
53	{ blk_imp, MAN_BSCOPE }, /* HP */
54	{ in_line_eoln, MAN_SCOPED | MAN_JOIN }, /* SM */
55	{ in_line_eoln, MAN_SCOPED | MAN_JOIN }, /* SB */
56	{ in_line_eoln, 0 }, /* BI */
57	{ in_line_eoln, 0 }, /* IB */
58	{ in_line_eoln, 0 }, /* BR */
59	{ in_line_eoln, 0 }, /* RB */
60	{ in_line_eoln, MAN_SCOPED | MAN_JOIN }, /* R */
61	{ in_line_eoln, MAN_SCOPED | MAN_JOIN }, /* B */
62	{ in_line_eoln, MAN_SCOPED | MAN_JOIN }, /* I */
63	{ in_line_eoln, 0 }, /* IR */
64	{ in_line_eoln, 0 }, /* RI */
65	{ in_line_eoln, MAN_NSCOPED }, /* sp */
66	{ in_line_eoln, MAN_NSCOPED }, /* nf */
67	{ in_line_eoln, MAN_NSCOPED }, /* fi */
68	{ blk_close, MAN_BSCOPE }, /* RE */
69	{ blk_exp, MAN_BSCOPE }, /* RS */
70	{ in_line_eoln, 0 }, /* DT */
71	{ in_line_eoln, 0 }, /* UC */
72	{ in_line_eoln, MAN_NSCOPED }, /* PD */
73	{ in_line_eoln, 0 }, /* AT */
74	{ in_line_eoln, 0 }, /* in */
75	{ in_line_eoln, 0 }, /* ft */
76	{ in_line_eoln, 0 }, /* OP */
77	{ in_line_eoln, MAN_BSCOPE }, /* EX */
78	{ in_line_eoln, MAN_BSCOPE }, /* EE */
79	{ blk_exp, MAN_BSCOPE }, /* UR */
80	{ blk_close, MAN_BSCOPE }, /* UE */
81	{ in_line_eoln, 0 }, /* ll */
82};
83
84const	struct man_macro * const man_macros = __man_macros;
85
86
87void
88man_unscope(struct roff_man *man, const struct roff_node *to)
89{
90	struct roff_node *n;
91
92	to = to->parent;
93	n = man->last;
94	while (n != to) {
95
96		/* Reached the end of the document? */
97
98		if (to == NULL && ! (n->flags & NODE_VALID)) {
99			if (man->flags & (MAN_BLINE | MAN_ELINE) &&
100			    man_macros[n->tok].flags & MAN_SCOPED) {
101				mandoc_vmsg(MANDOCERR_BLK_LINE,
102				    man->parse, n->line, n->pos,
103				    "EOF breaks %s",
104				    man_macronames[n->tok]);
105				if (man->flags & MAN_ELINE)
106					man->flags &= ~MAN_ELINE;
107				else {
108					assert(n->type == ROFFT_HEAD);
109					n = n->parent;
110					man->flags &= ~MAN_BLINE;
111				}
112				man->last = n;
113				n = n->parent;
114				roff_node_delete(man, man->last);
115				continue;
116			}
117			if (n->type == ROFFT_BLOCK &&
118			    man_macros[n->tok].fp == blk_exp)
119				mandoc_msg(MANDOCERR_BLK_NOEND,
120				    man->parse, n->line, n->pos,
121				    man_macronames[n->tok]);
122		}
123
124		/*
125		 * We might delete the man->last node
126		 * in the post-validation phase.
127		 * Save a pointer to the parent such that
128		 * we know where to continue the iteration.
129		 */
130
131		man->last = n;
132		n = n->parent;
133		man->last->flags |= NODE_VALID;
134	}
135
136	/*
137	 * If we ended up at the parent of the node we were
138	 * supposed to rewind to, that means the target node
139	 * got deleted, so add the next node we parse as a child
140	 * of the parent instead of as a sibling of the target.
141	 */
142
143	man->next = (man->last == to) ?
144	    ROFF_NEXT_CHILD : ROFF_NEXT_SIBLING;
145}
146
147/*
148 * Rewinding entails ascending the parse tree until a coherent point,
149 * for example, the `SH' macro will close out any intervening `SS'
150 * scopes.  When a scope is closed, it must be validated and actioned.
151 */
152static void
153rew_scope(struct roff_man *man, int tok)
154{
155	struct roff_node *n;
156
157	/* Preserve empty paragraphs before RS. */
158
159	n = man->last;
160	if (tok == MAN_RS && n->child == NULL &&
161	    (n->tok == MAN_P || n->tok == MAN_PP || n->tok == MAN_LP))
162		return;
163
164	for (;;) {
165		if (n->type == ROFFT_ROOT)
166			return;
167		if (n->flags & NODE_VALID) {
168			n = n->parent;
169			continue;
170		}
171		if (n->type != ROFFT_BLOCK) {
172			if (n->parent->type == ROFFT_ROOT) {
173				man_unscope(man, n);
174				return;
175			} else {
176				n = n->parent;
177				continue;
178			}
179		}
180		if (tok != MAN_SH && (n->tok == MAN_SH ||
181		    (tok != MAN_SS && (n->tok == MAN_SS ||
182		     man_macros[n->tok].fp == blk_exp))))
183			return;
184		man_unscope(man, n);
185		n = man->last;
186	}
187}
188
189
190/*
191 * Close out a generic explicit macro.
192 */
193void
194blk_close(MACRO_PROT_ARGS)
195{
196	int			 ntok;
197	const struct roff_node	*nn;
198	char			*p;
199	int			 nrew, target;
200
201	nrew = 1;
202	switch (tok) {
203	case MAN_RE:
204		ntok = MAN_RS;
205		if ( ! man_args(man, line, pos, buf, &p))
206			break;
207		for (nn = man->last->parent; nn; nn = nn->parent)
208			if (nn->tok == ntok && nn->type == ROFFT_BLOCK)
209				nrew++;
210		target = strtol(p, &p, 10);
211		if (*p != '\0')
212			mandoc_vmsg(MANDOCERR_ARG_EXCESS, man->parse,
213			    line, p - buf, "RE ... %s", p);
214		if (target == 0)
215			target = 1;
216		nrew -= target;
217		if (nrew < 1) {
218			mandoc_vmsg(MANDOCERR_RE_NOTOPEN, man->parse,
219			    line, ppos, "RE %d", target);
220			return;
221		}
222		break;
223	case MAN_UE:
224		ntok = MAN_UR;
225		break;
226	default:
227		abort();
228	}
229
230	for (nn = man->last->parent; nn; nn = nn->parent)
231		if (nn->tok == ntok && nn->type == ROFFT_BLOCK && ! --nrew)
232			break;
233
234	if (nn == NULL) {
235		mandoc_msg(MANDOCERR_BLK_NOTOPEN, man->parse,
236		    line, ppos, man_macronames[tok]);
237		rew_scope(man, MAN_PP);
238	} else {
239		line = man->last->line;
240		ppos = man->last->pos;
241		ntok = man->last->tok;
242		man_unscope(man, nn);
243
244		/* Move a trailing paragraph behind the block. */
245
246		if (ntok == MAN_LP || ntok == MAN_PP || ntok == MAN_P) {
247			*pos = strlen(buf);
248			blk_imp(man, ntok, line, ppos, pos, buf);
249		}
250	}
251}
252
253void
254blk_exp(MACRO_PROT_ARGS)
255{
256	struct roff_node *head;
257	char		*p;
258	int		 la;
259
260	rew_scope(man, tok);
261	roff_block_alloc(man, line, ppos, tok);
262	head = roff_head_alloc(man, line, ppos, tok);
263
264	la = *pos;
265	if (man_args(man, line, pos, buf, &p))
266		roff_word_alloc(man, line, la, p);
267
268	if (buf[*pos] != '\0')
269		mandoc_vmsg(MANDOCERR_ARG_EXCESS,
270		    man->parse, line, *pos, "%s ... %s",
271		    man_macronames[tok], buf + *pos);
272
273	man_unscope(man, head);
274	roff_body_alloc(man, line, ppos, tok);
275}
276
277/*
278 * Parse an implicit-block macro.  These contain a ROFFT_HEAD and a
279 * ROFFT_BODY contained within a ROFFT_BLOCK.  Rules for closing out other
280 * scopes, such as `SH' closing out an `SS', are defined in the rew
281 * routines.
282 */
283void
284blk_imp(MACRO_PROT_ARGS)
285{
286	int		 la;
287	char		*p;
288	struct roff_node *n;
289
290	rew_scope(man, tok);
291	n = roff_block_alloc(man, line, ppos, tok);
292	if (n->tok == MAN_SH || n->tok == MAN_SS)
293		man->flags &= ~MAN_LITERAL;
294	n = roff_head_alloc(man, line, ppos, tok);
295
296	/* Add line arguments. */
297
298	for (;;) {
299		la = *pos;
300		if ( ! man_args(man, line, pos, buf, &p))
301			break;
302		roff_word_alloc(man, line, la, p);
303	}
304
305	/*
306	 * For macros having optional next-line scope,
307	 * keep the head open if there were no arguments.
308	 * For `TP', always keep the head open.
309	 */
310
311	if (man_macros[tok].flags & MAN_SCOPED &&
312	    (tok == MAN_TP || n == man->last)) {
313		man->flags |= MAN_BLINE;
314		return;
315	}
316
317	/* Close out the head and open the body. */
318
319	man_unscope(man, n);
320	roff_body_alloc(man, line, ppos, tok);
321}
322
323void
324in_line_eoln(MACRO_PROT_ARGS)
325{
326	int		 la;
327	char		*p;
328	struct roff_node *n;
329
330	roff_elem_alloc(man, line, ppos, tok);
331	n = man->last;
332
333	for (;;) {
334		if (buf[*pos] != '\0' && (tok == MAN_br ||
335		    tok == MAN_fi || tok == MAN_nf)) {
336			mandoc_vmsg(MANDOCERR_ARG_SKIP,
337			    man->parse, line, *pos, "%s %s",
338			    man_macronames[tok], buf + *pos);
339			break;
340		}
341		if (buf[*pos] != '\0' && man->last != n &&
342		    (tok == MAN_PD || tok == MAN_ft || tok == MAN_sp)) {
343			mandoc_vmsg(MANDOCERR_ARG_EXCESS,
344			    man->parse, line, *pos, "%s ... %s",
345			    man_macronames[tok], buf + *pos);
346			break;
347		}
348		la = *pos;
349		if ( ! man_args(man, line, pos, buf, &p))
350			break;
351		if (man_macros[tok].flags & MAN_JOIN &&
352		    man->last->type == ROFFT_TEXT)
353			roff_word_append(man, p);
354		else
355			roff_word_alloc(man, line, la, p);
356	}
357
358	/*
359	 * Append NODE_EOS in case the last snipped argument
360	 * ends with a dot, e.g. `.IR syslog (3).'
361	 */
362
363	if (n != man->last &&
364	    mandoc_eos(man->last->string, strlen(man->last->string)))
365		man->last->flags |= NODE_EOS;
366
367	/*
368	 * If no arguments are specified and this is MAN_SCOPED (i.e.,
369	 * next-line scoped), then set our mode to indicate that we're
370	 * waiting for terms to load into our context.
371	 */
372
373	if (n == man->last && man_macros[tok].flags & MAN_SCOPED) {
374		assert( ! (man_macros[tok].flags & MAN_NSCOPED));
375		man->flags |= MAN_ELINE;
376		return;
377	}
378
379	assert(man->last->type != ROFFT_ROOT);
380	man->next = ROFF_NEXT_SIBLING;
381
382	/* Rewind our element scope. */
383
384	for ( ; man->last; man->last = man->last->parent) {
385		man_state(man, man->last);
386		if (man->last == n)
387			break;
388	}
389}
390
391void
392man_endparse(struct roff_man *man)
393{
394
395	man_unscope(man, man->first);
396	man->flags &= ~MAN_LITERAL;
397}
398
399static int
400man_args(struct roff_man *man, int line, int *pos, char *buf, char **v)
401{
402	char	 *start;
403
404	assert(*pos);
405	*v = start = buf + *pos;
406	assert(' ' != *start);
407
408	if ('\0' == *start)
409		return 0;
410
411	*v = mandoc_getarg(man->parse, v, line, pos);
412	return 1;
413}
414