man_macro.c revision 1.2 
1/*	$Id: man_macro.c,v 1.2 2016/01/07 20:00:35 christos 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 AUTHOR DISCLAIMS ALL WARRANTIES
12 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
13 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR 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 "man.h"
29#include "mandoc.h"
30#include "libmandoc.h"
31#include "libman.h"
32
33enum	rew {
34	REW_REWIND,
35	REW_NOHALT,
36	REW_HALT
37};
38
39static	void		 blk_close(MACRO_PROT_ARGS);
40static	void		 blk_exp(MACRO_PROT_ARGS);
41static	void		 blk_imp(MACRO_PROT_ARGS);
42static	void		 in_line_eoln(MACRO_PROT_ARGS);
43static	int		 man_args(struct man *, int,
44				int *, char *, char **);
45
46static	void		 rew_scope(enum man_type,
47				struct man *, enum mant);
48static	enum rew	 rew_dohalt(enum mant, enum man_type,
49				const struct man_node *);
50static	enum rew	 rew_block(enum mant, enum man_type,
51				const struct man_node *);
52
53const	struct man_macro __man_macros[MAN_MAX] = {
54	{ in_line_eoln, MAN_NSCOPED }, /* br */
55	{ in_line_eoln, MAN_BSCOPE }, /* TH */
56	{ blk_imp, MAN_BSCOPE | MAN_SCOPED }, /* SH */
57	{ blk_imp, MAN_BSCOPE | MAN_SCOPED }, /* SS */
58	{ blk_imp, MAN_BSCOPE | MAN_SCOPED | MAN_FSCOPED }, /* TP */
59	{ blk_imp, MAN_BSCOPE }, /* LP */
60	{ blk_imp, MAN_BSCOPE }, /* PP */
61	{ blk_imp, MAN_BSCOPE }, /* P */
62	{ blk_imp, MAN_BSCOPE }, /* IP */
63	{ blk_imp, MAN_BSCOPE }, /* HP */
64	{ in_line_eoln, MAN_SCOPED | MAN_JOIN }, /* SM */
65	{ in_line_eoln, MAN_SCOPED | MAN_JOIN }, /* SB */
66	{ in_line_eoln, 0 }, /* BI */
67	{ in_line_eoln, 0 }, /* IB */
68	{ in_line_eoln, 0 }, /* BR */
69	{ in_line_eoln, 0 }, /* RB */
70	{ in_line_eoln, MAN_SCOPED | MAN_JOIN }, /* R */
71	{ in_line_eoln, MAN_SCOPED | MAN_JOIN }, /* B */
72	{ in_line_eoln, MAN_SCOPED | MAN_JOIN }, /* I */
73	{ in_line_eoln, 0 }, /* IR */
74	{ in_line_eoln, 0 }, /* RI */
75	{ in_line_eoln, MAN_NSCOPED }, /* sp */
76	{ in_line_eoln, MAN_BSCOPE }, /* nf */
77	{ in_line_eoln, MAN_BSCOPE }, /* fi */
78	{ blk_close, MAN_BSCOPE }, /* RE */
79	{ blk_exp, MAN_BSCOPE | MAN_EXPLICIT }, /* RS */
80	{ in_line_eoln, 0 }, /* DT */
81	{ in_line_eoln, 0 }, /* UC */
82	{ in_line_eoln, 0 }, /* PD */
83	{ in_line_eoln, 0 }, /* AT */
84	{ in_line_eoln, 0 }, /* in */
85	{ in_line_eoln, 0 }, /* ft */
86	{ in_line_eoln, 0 }, /* OP */
87	{ in_line_eoln, MAN_BSCOPE }, /* EX */
88	{ in_line_eoln, MAN_BSCOPE }, /* EE */
89	{ blk_exp, MAN_BSCOPE | MAN_EXPLICIT }, /* UR */
90	{ blk_close, MAN_BSCOPE }, /* UE */
91	{ in_line_eoln, 0 }, /* ll */
92};
93
94const	struct man_macro * const man_macros = __man_macros;
95
96
97void
98man_unscope(struct man *man, const struct man_node *to)
99{
100	struct man_node	*n;
101
102	to = to->parent;
103	n = man->last;
104	while (n != to) {
105
106		/* Reached the end of the document? */
107
108		if (to == NULL && ! (n->flags & MAN_VALID)) {
109			if (man->flags & (MAN_BLINE | MAN_ELINE) &&
110			    man_macros[n->tok].flags & MAN_SCOPED) {
111				mandoc_vmsg(MANDOCERR_BLK_LINE,
112				    man->parse, n->line, n->pos,
113				    "EOF breaks %s",
114				    man_macronames[n->tok]);
115				if (man->flags & MAN_ELINE)
116					man->flags &= ~MAN_ELINE;
117				else {
118					assert(n->type == MAN_HEAD);
119					n = n->parent;
120					man->flags &= ~MAN_BLINE;
121				}
122				man->last = n;
123				n = n->parent;
124				man_node_delete(man, man->last);
125				continue;
126			}
127			if (n->type == MAN_BLOCK &&
128			    man_macros[n->tok].flags & MAN_EXPLICIT)
129				mandoc_msg(MANDOCERR_BLK_NOEND,
130				    man->parse, n->line, n->pos,
131				    man_macronames[n->tok]);
132		}
133
134		/*
135		 * We might delete the man->last node
136		 * in the post-validation phase.
137		 * Save a pointer to the parent such that
138		 * we know where to continue the iteration.
139		 */
140
141		man->last = n;
142		n = n->parent;
143		man_valid_post(man);
144	}
145
146	/*
147	 * If we ended up at the parent of the node we were
148	 * supposed to rewind to, that means the target node
149	 * got deleted, so add the next node we parse as a child
150	 * of the parent instead of as a sibling of the target.
151	 */
152
153	man->next = (man->last == to) ?
154	    MAN_NEXT_CHILD : MAN_NEXT_SIBLING;
155}
156
157static enum rew
158rew_block(enum mant ntok, enum man_type type, const struct man_node *n)
159{
160
161	if (type == MAN_BLOCK && ntok == n->parent->tok &&
162	    n->parent->type == MAN_BODY)
163		return(REW_REWIND);
164	return(ntok == n->tok ? REW_HALT : REW_NOHALT);
165}
166
167/*
168 * There are three scope levels: scoped to the root (all), scoped to the
169 * section (all less sections), and scoped to subsections (all less
170 * sections and subsections).
171 */
172static enum rew
173rew_dohalt(enum mant tok, enum man_type type, const struct man_node *n)
174{
175	enum rew	 c;
176
177	/* We cannot progress beyond the root ever. */
178	if (MAN_ROOT == n->type)
179		return(REW_HALT);
180
181	assert(n->parent);
182
183	/* Normal nodes shouldn't go to the level of the root. */
184	if (MAN_ROOT == n->parent->type)
185		return(REW_REWIND);
186
187	/* Already-validated nodes should be closed out. */
188	if (MAN_VALID & n->flags)
189		return(REW_NOHALT);
190
191	/* First: rewind to ourselves. */
192	if (type == n->type && tok == n->tok) {
193		if (MAN_EXPLICIT & man_macros[n->tok].flags)
194			return(REW_HALT);
195		else
196			return(REW_REWIND);
197	}
198
199	/*
200	 * Next follow the implicit scope-smashings as defined by man.7:
201	 * section, sub-section, etc.
202	 */
203
204	switch (tok) {
205	case MAN_SH:
206		break;
207	case MAN_SS:
208		/* Rewind to a section, if a block. */
209		if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
210			return(c);
211		break;
212	case MAN_RS:
213		/* Preserve empty paragraphs before RS. */
214		if (0 == n->nchild && (MAN_P == n->tok ||
215		    MAN_PP == n->tok || MAN_LP == n->tok))
216			return(REW_HALT);
217		/* Rewind to a subsection, if a block. */
218		if (REW_NOHALT != (c = rew_block(MAN_SS, type, n)))
219			return(c);
220		/* Rewind to a section, if a block. */
221		if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
222			return(c);
223		break;
224	default:
225		/* Rewind to an offsetter, if a block. */
226		if (REW_NOHALT != (c = rew_block(MAN_RS, type, n)))
227			return(c);
228		/* Rewind to a subsection, if a block. */
229		if (REW_NOHALT != (c = rew_block(MAN_SS, type, n)))
230			return(c);
231		/* Rewind to a section, if a block. */
232		if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
233			return(c);
234		break;
235	}
236
237	return(REW_NOHALT);
238}
239
240/*
241 * Rewinding entails ascending the parse tree until a coherent point,
242 * for example, the `SH' macro will close out any intervening `SS'
243 * scopes.  When a scope is closed, it must be validated and actioned.
244 */
245static void
246rew_scope(enum man_type type, struct man *man, enum mant tok)
247{
248	struct man_node	*n;
249	enum rew	 c;
250
251	for (n = man->last; n; n = n->parent) {
252		/*
253		 * Whether we should stop immediately (REW_HALT), stop
254		 * and rewind until this point (REW_REWIND), or keep
255		 * rewinding (REW_NOHALT).
256		 */
257		c = rew_dohalt(tok, type, n);
258		if (REW_HALT == c)
259			return;
260		if (REW_REWIND == c)
261			break;
262	}
263
264	/*
265	 * Rewind until the current point.  Warn if we're a roff
266	 * instruction that's mowing over explicit scopes.
267	 */
268	if (n != NULL)
269		man_unscope(man, n);
270}
271
272
273/*
274 * Close out a generic explicit macro.
275 */
276void
277blk_close(MACRO_PROT_ARGS)
278{
279	enum mant		 ntok;
280	const struct man_node	*nn;
281	char			*p;
282	int			 nrew, target;
283
284	nrew = 1;
285	switch (tok) {
286	case MAN_RE:
287		ntok = MAN_RS;
288		if ( ! man_args(man, line, pos, buf, &p))
289			break;
290		for (nn = man->last->parent; nn; nn = nn->parent)
291			if (nn->tok == ntok && nn->type == MAN_BLOCK)
292				nrew++;
293		target = strtol(p, &p, 10);
294		if (*p != '\0')
295			mandoc_vmsg(MANDOCERR_ARG_EXCESS, man->parse,
296			    line, p - buf, "RE ... %s", p);
297		if (target == 0)
298			target = 1;
299		nrew -= target;
300		if (nrew < 1) {
301			mandoc_vmsg(MANDOCERR_RE_NOTOPEN, man->parse,
302			    line, ppos, "RE %d", target);
303			return;
304		}
305		break;
306	case MAN_UE:
307		ntok = MAN_UR;
308		break;
309	default:
310		abort();
311		/* NOTREACHED */
312	}
313
314	for (nn = man->last->parent; nn; nn = nn->parent)
315		if (nn->tok == ntok && nn->type == MAN_BLOCK && ! --nrew)
316			break;
317
318	if (nn == NULL) {
319		mandoc_msg(MANDOCERR_BLK_NOTOPEN, man->parse,
320		    line, ppos, man_macronames[tok]);
321		rew_scope(MAN_BLOCK, man, MAN_PP);
322	} else {
323		line = man->last->line;
324		ppos = man->last->pos;
325		ntok = man->last->tok;
326		man_unscope(man, nn);
327
328		/* Move a trailing paragraph behind the block. */
329
330		if (ntok == MAN_LP || ntok == MAN_PP || ntok == MAN_P) {
331			*pos = strlen(buf);
332			blk_imp(man, ntok, line, ppos, pos, buf);
333		}
334	}
335}
336
337void
338blk_exp(MACRO_PROT_ARGS)
339{
340	struct man_node	*head;
341	char		*p;
342	int		 la;
343
344	rew_scope(MAN_BLOCK, man, tok);
345	man_block_alloc(man, line, ppos, tok);
346	man_head_alloc(man, line, ppos, tok);
347	head = man->last;
348
349	la = *pos;
350	if (man_args(man, line, pos, buf, &p))
351		man_word_alloc(man, line, la, p);
352
353	if (buf[*pos] != '\0')
354		mandoc_vmsg(MANDOCERR_ARG_EXCESS,
355		    man->parse, line, *pos, "%s ... %s",
356		    man_macronames[tok], buf + *pos);
357
358	man_unscope(man, head);
359	man_body_alloc(man, line, ppos, tok);
360}
361
362/*
363 * Parse an implicit-block macro.  These contain a MAN_HEAD and a
364 * MAN_BODY contained within a MAN_BLOCK.  Rules for closing out other
365 * scopes, such as `SH' closing out an `SS', are defined in the rew
366 * routines.
367 */
368void
369blk_imp(MACRO_PROT_ARGS)
370{
371	int		 la;
372	char		*p;
373	struct man_node	*n;
374
375	rew_scope(MAN_BODY, man, tok);
376	rew_scope(MAN_BLOCK, man, tok);
377	man_block_alloc(man, line, ppos, tok);
378	man_head_alloc(man, line, ppos, tok);
379	n = man->last;
380
381	/* Add line arguments. */
382
383	for (;;) {
384		la = *pos;
385		if ( ! man_args(man, line, pos, buf, &p))
386			break;
387		man_word_alloc(man, line, la, p);
388	}
389
390	/* Close out head and open body (unless MAN_SCOPE). */
391
392	if (man_macros[tok].flags & MAN_SCOPED) {
393		/* If we're forcing scope (`TP'), keep it open. */
394		if (man_macros[tok].flags & MAN_FSCOPED) {
395			man->flags |= MAN_BLINE;
396			return;
397		} else if (n == man->last) {
398			man->flags |= MAN_BLINE;
399			return;
400		}
401	}
402	rew_scope(MAN_HEAD, man, tok);
403	man_body_alloc(man, line, ppos, tok);
404}
405
406void
407in_line_eoln(MACRO_PROT_ARGS)
408{
409	int		 la;
410	char		*p;
411	struct man_node	*n;
412
413	man_elem_alloc(man, line, ppos, tok);
414	n = man->last;
415
416	for (;;) {
417		if (buf[*pos] != '\0' && (tok == MAN_br ||
418		    tok == MAN_fi || tok == MAN_nf)) {
419			mandoc_vmsg(MANDOCERR_ARG_SKIP,
420			    man->parse, line, *pos, "%s %s",
421			    man_macronames[tok], buf + *pos);
422			break;
423		}
424		if (buf[*pos] != '\0' && man->last != n &&
425		    (tok == MAN_PD || tok == MAN_ft || tok == MAN_sp)) {
426			mandoc_vmsg(MANDOCERR_ARG_EXCESS,
427			    man->parse, line, *pos, "%s ... %s",
428			    man_macronames[tok], buf + *pos);
429			break;
430		}
431		la = *pos;
432		if ( ! man_args(man, line, pos, buf, &p))
433			break;
434		if (man_macros[tok].flags & MAN_JOIN &&
435		    man->last->type == MAN_TEXT)
436			man_word_append(man, p);
437		else
438			man_word_alloc(man, line, la, p);
439	}
440
441	/*
442	 * Append MAN_EOS in case the last snipped argument
443	 * ends with a dot, e.g. `.IR syslog (3).'
444	 */
445
446	if (n != man->last &&
447	    mandoc_eos(man->last->string, strlen(man->last->string)))
448		man->last->flags |= MAN_EOS;
449
450	/*
451	 * If no arguments are specified and this is MAN_SCOPED (i.e.,
452	 * next-line scoped), then set our mode to indicate that we're
453	 * waiting for terms to load into our context.
454	 */
455
456	if (n == man->last && man_macros[tok].flags & MAN_SCOPED) {
457		assert( ! (man_macros[tok].flags & MAN_NSCOPED));
458		man->flags |= MAN_ELINE;
459		return;
460	}
461
462	assert(man->last->type != MAN_ROOT);
463	man->next = MAN_NEXT_SIBLING;
464
465	/*
466	 * Rewind our element scope.  Note that when TH is pruned, we'll
467	 * be back at the root, so make sure that we don't clobber as
468	 * its sibling.
469	 */
470
471	for ( ; man->last; man->last = man->last->parent) {
472		if (man->last == n)
473			break;
474		if (man->last->type == MAN_ROOT)
475			break;
476		man_valid_post(man);
477	}
478
479	assert(man->last);
480
481	/*
482	 * Same here regarding whether we're back at the root.
483	 */
484
485	if (man->last->type != MAN_ROOT)
486		man_valid_post(man);
487}
488
489
490void
491man_macroend(struct man *man)
492{
493
494	man_unscope(man, man->first);
495}
496
497static int
498man_args(struct man *man, int line, int *pos, char *buf, char **v)
499{
500	char	 *start;
501
502	assert(*pos);
503	*v = start = buf + *pos;
504	assert(' ' != *start);
505
506	if ('\0' == *start)
507		return(0);
508
509	*v = mandoc_getarg(man->parse, v, line, pos);
510	return(1);
511}
512