1 /* $Id: man.c,v 1.137 2014/08/01 21:24:17 schwarze Exp $ */
3 * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2013, 2014 Ingo Schwarze <schwarze@openbsd.org>
5 * Copyright (c) 2011 Joerg Sonnenberger <joerg@netbsd.org>
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.
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.
23 #include <sys/types.h>
34 #include "mandoc_aux.h"
36 #include "libmandoc.h"
38 const char *const __man_macronames[MAN_MAX] = {
39 "br", "TH", "SH", "SS",
40 "TP", "LP", "PP", "P",
41 "IP", "HP", "SM", "SB",
42 "BI", "IB", "BR", "RB",
44 "RI", "na", "sp", "nf",
45 "fi", "RE", "RS", "DT",
46 "UC", "PD", "AT", "in",
47 "ft", "OP", "EX", "EE",
51 const char * const *man_macronames = __man_macronames;
53 static struct man_node *man_node_alloc(struct man *, int, int,
54 enum man_type, enum mant);
55 static int man_node_append(struct man *,
57 static void man_node_free(struct man_node *);
58 static void man_node_unlink(struct man *,
60 static int man_ptext(struct man *, int, char *, int);
61 static int man_pmacro(struct man *, int, char *, int);
62 static void man_free1(struct man *);
63 static void man_alloc1(struct man *);
64 static int man_descope(struct man *, int, int);
67 const struct man_node *
68 man_node(const struct man *man)
74 const struct man_meta *
75 man_meta(const struct man *man)
82 man_reset(struct man *man)
90 man_free(struct man *man)
98 man_alloc(struct roff *roff, struct mparse *parse, int quick)
102 p = mandoc_calloc(1, sizeof(struct man));
114 man_endparse(struct man *man)
117 return(man_macroend(man));
121 man_parseln(struct man *man, int ln, char *buf, int offs)
124 man->flags |= MAN_NEWLINE;
126 return (roff_getcontrol(man->roff, buf, &offs) ?
127 man_pmacro(man, ln, buf, offs) :
128 man_ptext(man, ln, buf, offs));
132 man_free1(struct man *man)
136 man_node_delete(man, man->first);
138 free(man->meta.title);
139 if (man->meta.source)
140 free(man->meta.source);
142 free(man->meta.date);
146 free(man->meta.msec);
150 man_alloc1(struct man *man)
153 memset(&man->meta, 0, sizeof(struct man_meta));
155 man->last = mandoc_calloc(1, sizeof(struct man_node));
156 man->first = man->last;
157 man->last->type = MAN_ROOT;
158 man->last->tok = MAN_MAX;
159 man->next = MAN_NEXT_CHILD;
164 man_node_append(struct man *man, struct man_node *p)
169 assert(MAN_ROOT != p->type);
172 case MAN_NEXT_SIBLING:
175 p->parent = man->last->parent;
178 man->last->child = p;
179 p->parent = man->last;
191 if (p->tok == MAN_SH || p->tok == MAN_SS)
192 man->flags &= ~MAN_LITERAL;
195 assert(MAN_BLOCK == p->parent->type);
199 assert(MAN_BLOCK == p->parent->type);
203 assert(MAN_BLOCK == p->parent->type);
216 if ( ! man_valid_post(man))
226 static struct man_node *
227 man_node_alloc(struct man *man, int line, int pos,
228 enum man_type type, enum mant tok)
232 p = mandoc_calloc(1, sizeof(struct man_node));
238 if (MAN_NEWLINE & man->flags)
239 p->flags |= MAN_LINE;
240 man->flags &= ~MAN_NEWLINE;
245 man_elem_alloc(struct man *man, int line, int pos, enum mant tok)
249 p = man_node_alloc(man, line, pos, MAN_ELEM, tok);
250 if ( ! man_node_append(man, p))
252 man->next = MAN_NEXT_CHILD;
257 man_tail_alloc(struct man *man, int line, int pos, enum mant tok)
261 p = man_node_alloc(man, line, pos, MAN_TAIL, tok);
262 if ( ! man_node_append(man, p))
264 man->next = MAN_NEXT_CHILD;
269 man_head_alloc(struct man *man, int line, int pos, enum mant tok)
273 p = man_node_alloc(man, line, pos, MAN_HEAD, tok);
274 if ( ! man_node_append(man, p))
276 man->next = MAN_NEXT_CHILD;
281 man_body_alloc(struct man *man, int line, int pos, enum mant tok)
285 p = man_node_alloc(man, line, pos, MAN_BODY, tok);
286 if ( ! man_node_append(man, p))
288 man->next = MAN_NEXT_CHILD;
293 man_block_alloc(struct man *man, int line, int pos, enum mant tok)
297 p = man_node_alloc(man, line, pos, MAN_BLOCK, tok);
298 if ( ! man_node_append(man, p))
300 man->next = MAN_NEXT_CHILD;
305 man_word_alloc(struct man *man, int line, int pos, const char *word)
309 n = man_node_alloc(man, line, pos, MAN_TEXT, MAN_MAX);
310 n->string = roff_strdup(man->roff, word);
312 if ( ! man_node_append(man, n))
315 man->next = MAN_NEXT_SIBLING;
320 * Free all of the resources held by a node. This does NOT unlink a
321 * node from its context; for that, see man_node_unlink().
324 man_node_free(struct man_node *p)
333 man_node_delete(struct man *man, struct man_node *p)
337 man_node_delete(man, p->child);
339 man_node_unlink(man, p);
344 man_addeqn(struct man *man, const struct eqn *ep)
348 n = man_node_alloc(man, ep->ln, ep->pos, MAN_EQN, MAN_MAX);
351 if ( ! man_node_append(man, n))
354 man->next = MAN_NEXT_SIBLING;
355 return(man_descope(man, ep->ln, ep->pos));
359 man_addspan(struct man *man, const struct tbl_span *sp)
363 n = man_node_alloc(man, sp->line, 0, MAN_TBL, MAN_MAX);
366 if ( ! man_node_append(man, n))
369 man->next = MAN_NEXT_SIBLING;
370 return(man_descope(man, sp->line, 0));
374 man_descope(struct man *man, int line, int offs)
377 * Co-ordinate what happens with having a next-line scope open:
378 * first close out the element scope (if applicable), then close
379 * out the block scope (also if applicable).
382 if (MAN_ELINE & man->flags) {
383 man->flags &= ~MAN_ELINE;
384 if ( ! man_unscope(man, man->last->parent))
388 if ( ! (MAN_BLINE & man->flags))
390 man->flags &= ~MAN_BLINE;
392 if ( ! man_unscope(man, man->last->parent))
394 return(man_body_alloc(man, line, offs, man->last->tok));
398 man_ptext(struct man *man, int line, char *buf, int offs)
402 /* Literal free-form text whitespace is preserved. */
404 if (MAN_LITERAL & man->flags) {
405 if ( ! man_word_alloc(man, line, offs, buf + offs))
407 return(man_descope(man, line, offs));
410 for (i = offs; ' ' == buf[i]; i++)
411 /* Skip leading whitespace. */ ;
414 * Blank lines are ignored right after headings
415 * but add a single vertical space elsewhere.
418 if ('\0' == buf[i]) {
419 /* Allocate a blank entry. */
420 if (MAN_SH != man->last->tok &&
421 MAN_SS != man->last->tok) {
422 if ( ! man_elem_alloc(man, line, offs, MAN_sp))
424 man->next = MAN_NEXT_SIBLING;
430 * Warn if the last un-escaped character is whitespace. Then
431 * strip away the remaining spaces (tabs stay!).
434 i = (int)strlen(buf);
437 if (' ' == buf[i - 1] || '\t' == buf[i - 1]) {
438 if (i > 1 && '\\' != buf[i - 2])
439 mandoc_msg(MANDOCERR_SPACE_EOL, man->parse,
442 for (--i; i && ' ' == buf[i]; i--)
443 /* Spin back to non-space. */ ;
445 /* Jump ahead of escaped whitespace. */
446 i += '\\' == buf[i] ? 2 : 1;
451 if ( ! man_word_alloc(man, line, offs, buf + offs))
455 * End-of-sentence check. If the last character is an unescaped
456 * EOS character, then flag the node as being the end of a
457 * sentence. The front-end will know how to interpret this.
461 if (mandoc_eos(buf, (size_t)i))
462 man->last->flags |= MAN_EOS;
464 return(man_descope(man, line, offs));
468 man_pmacro(struct man *man, int ln, char *buf, int offs)
476 if ('"' == buf[offs]) {
477 mandoc_msg(MANDOCERR_COMMENT_BAD, man->parse,
480 } else if ('\0' == buf[offs])
486 * Copy the first word into a nil-terminated buffer.
487 * Stop copying when a tab, space, or eoln is encountered.
491 while (i < 4 && '\0' != buf[offs] && ' ' != buf[offs] &&
493 mac[i++] = buf[offs++];
497 tok = (i > 0 && i < 4) ? man_hash_find(mac) : MAN_MAX;
499 if (MAN_MAX == tok) {
500 mandoc_msg(MANDOCERR_MACRO, man->parse,
501 ln, ppos, buf + ppos - 1);
505 /* The macro is sane. Jump to the next word. */
507 while (buf[offs] && ' ' == buf[offs])
511 * Trailing whitespace. Note that tabs are allowed to be passed
512 * into the parser as "text", so we only warn about spaces here.
515 if ('\0' == buf[offs] && ' ' == buf[offs - 1])
516 mandoc_msg(MANDOCERR_SPACE_EOL, man->parse,
520 * Remove prior ELINE macro, as it's being clobbered by a new
521 * macro. Note that NSCOPED macros do not close out ELINE
522 * macros---they don't print text---so we let those slip by.
525 if ( ! (MAN_NSCOPED & man_macros[tok].flags) &&
526 man->flags & MAN_ELINE) {
528 assert(MAN_TEXT != n->type);
530 /* Remove repeated NSCOPED macros causing ELINE. */
532 if (MAN_NSCOPED & man_macros[n->tok].flags)
535 mandoc_vmsg(MANDOCERR_BLK_LINE, man->parse, n->line,
536 n->pos, "%s breaks %s", man_macronames[tok],
537 man_macronames[n->tok]);
539 man_node_delete(man, n);
540 man->flags &= ~MAN_ELINE;
544 * Remove prior BLINE macro that is being clobbered.
546 if ((man->flags & MAN_BLINE) &&
547 (MAN_BSCOPE & man_macros[tok].flags)) {
550 /* Might be a text node like 8 in
554 if (MAN_TEXT == n->type)
557 /* Remove element that didn't end BLINE, if any. */
558 if ( ! (MAN_BSCOPE & man_macros[n->tok].flags))
561 assert(MAN_HEAD == n->type);
563 assert(MAN_BLOCK == n->type);
564 assert(MAN_SCOPED & man_macros[n->tok].flags);
566 mandoc_vmsg(MANDOCERR_BLK_LINE, man->parse, n->line,
567 n->pos, "%s breaks %s", man_macronames[tok],
568 man_macronames[n->tok]);
570 man_node_delete(man, n);
571 man->flags &= ~MAN_BLINE;
574 /* Remember whether we are in next-line scope for a block head. */
576 bline = man->flags & MAN_BLINE;
578 /* Call to handler... */
580 assert(man_macros[tok].fp);
581 if ( ! (*man_macros[tok].fp)(man, tok, ln, ppos, &offs, buf))
584 /* In quick mode (for mandocdb), abort after the NAME section. */
586 if (man->quick && MAN_SH == tok) {
588 if (MAN_BODY == n->type &&
589 strcmp(n->prev->child->string, "NAME"))
594 * If we are in a next-line scope for a block head,
595 * close it out now and switch to the body,
596 * unless the next-line scope is allowed to continue.
599 if ( ! bline || man->flags & MAN_ELINE ||
600 man_macros[tok].flags & MAN_NSCOPED)
603 assert(MAN_BLINE & man->flags);
604 man->flags &= ~MAN_BLINE;
606 if ( ! man_unscope(man, man->last->parent))
608 return(man_body_alloc(man, ln, ppos, man->last->tok));
612 * Unlink a node from its context. If "man" is provided, the last parse
613 * point will also be adjusted accordingly.
616 man_node_unlink(struct man *man, struct man_node *n)
619 /* Adjust siblings. */
622 n->prev->next = n->next;
624 n->next->prev = n->prev;
630 if (n->parent->child == n)
631 n->parent->child = n->prev ? n->prev : n->next;
634 /* Adjust parse point, if applicable. */
636 if (man && man->last == n) {
637 /*XXX: this can occur when bailing from validation. */
638 /*assert(NULL == n->next);*/
641 man->next = MAN_NEXT_SIBLING;
643 man->last = n->parent;
644 man->next = MAN_NEXT_CHILD;
648 if (man && man->first == n)
652 const struct mparse *
653 man_mparse(const struct man *man)
656 assert(man && man->parse);
661 man_deroff(char **dest, const struct man_node *n)
666 if (MAN_TEXT != n->type) {
667 for (n = n->child; n; n = n->next)
672 /* Skip leading whitespace and escape sequences. */
675 while ('\0' != *cp) {
678 mandoc_escape((const char **)&cp, NULL, NULL);
679 } else if (isspace((unsigned char)*cp))
685 /* Skip trailing whitespace. */
687 for (sz = strlen(cp); sz; sz--)
688 if (0 == isspace((unsigned char)cp[sz-1]))
691 /* Skip empty strings. */
697 *dest = mandoc_strndup(cp, sz);
701 mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);