1 /* $Id: man.c,v 1.46 2009/11/02 08:40:31 kristaps Exp $ */
3 * Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@kth.se>
5 * Permission to use, copy, modify, and distribute this software for any
6 * purpose with or without fee is hereby granted, provided that the above
7 * copyright notice and this permission notice appear in all copies.
9 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
10 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
11 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
12 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
13 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
14 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
15 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
17 #include <sys/types.h>
27 #include "libmandoc.h"
29 const char *const __man_merrnames[WERRMAX] = {
30 "invalid character", /* WNPRINT */
31 "invalid manual section", /* WMSEC */
32 "invalid date format", /* WDATE */
33 "scope of prior line violated", /* WLNSCOPE */
34 "trailing whitespace", /* WTSPACE */
35 "unterminated quoted parameter", /* WTQUOTE */
36 "document has no body", /* WNODATA */
37 "document has no title/section", /* WNOTITLE */
38 "invalid escape sequence", /* WESCAPE */
39 "invalid number format", /* WNUMFMT */
40 "expected block head arguments", /* WHEADARGS */
41 "expected block body arguments", /* WBODYARGS */
42 "expected empty block head", /* WNHEADARGS */
43 "ill-formed macro", /* WMACROFORM */
44 "scope open on exit", /* WEXITSCOPE */
45 "no scope context", /* WNOSCOPE */
46 "literal context already open", /* WOLITERAL */
47 "no literal context open" /* WNLITERAL */
50 const char *const __man_macronames[MAN_MAX] = {
51 "br", "TH", "SH", "SS",
52 "TP", "LP", "PP", "P",
53 "IP", "HP", "SM", "SB",
54 "BI", "IB", "BR", "RB",
56 "RI", "na", "i", "sp",
57 "nf", "fi", "r", "RE",
58 "RS", "DT", "UC", "PD"
61 const char * const *man_macronames = __man_macronames;
63 static struct man_node *man_node_alloc(int, int,
65 static int man_node_append(struct man *,
67 static int man_ptext(struct man *, int, char *);
68 static int man_pmacro(struct man *, int, char *);
69 static void man_free1(struct man *);
70 static void man_alloc1(struct man *);
71 static int pstring(struct man *, int, int,
72 const char *, size_t);
73 static int macrowarn(struct man *, int, const char *);
76 extern size_t strlcpy(char *, const char *, size_t);
80 const struct man_node *
81 man_node(const struct man *m)
84 return(MAN_HALT & m->flags ? NULL : m->first);
88 const struct man_meta *
89 man_meta(const struct man *m)
92 return(MAN_HALT & m->flags ? NULL : &m->meta);
97 man_reset(struct man *man)
106 man_free(struct man *man)
115 man_alloc(void *data, int pflags, const struct man_cb *cb)
119 p = mandoc_calloc(1, sizeof(struct man));
122 memcpy(&p->cb, cb, sizeof(struct man_cb));
134 man_endparse(struct man *m)
137 if (MAN_HALT & m->flags)
139 else if (man_macroend(m))
141 m->flags |= MAN_HALT;
147 man_parseln(struct man *m, int ln, char *buf)
151 man_pmacro(m, ln, buf) :
152 man_ptext(m, ln, buf));
157 man_free1(struct man *man)
161 man_node_freelist(man->first);
163 free(man->meta.title);
164 if (man->meta.source)
165 free(man->meta.source);
172 man_alloc1(struct man *m)
175 memset(&m->meta, 0, sizeof(struct man_meta));
177 m->last = mandoc_calloc(1, sizeof(struct man_node));
179 m->last->type = MAN_ROOT;
180 m->next = MAN_NEXT_CHILD;
185 man_node_append(struct man *man, struct man_node *p)
190 assert(MAN_ROOT != p->type);
193 case (MAN_NEXT_SIBLING):
196 p->parent = man->last->parent;
198 case (MAN_NEXT_CHILD):
199 man->last->child = p;
200 p->parent = man->last;
209 if ( ! man_valid_pre(man, p))
214 assert(MAN_BLOCK == p->parent->type);
218 assert(MAN_BLOCK == p->parent->type);
229 if ( ! man_valid_post(man))
231 if ( ! man_action_post(man))
242 static struct man_node *
243 man_node_alloc(int line, int pos, enum man_type type, int tok)
247 p = mandoc_calloc(1, sizeof(struct man_node));
257 man_elem_alloc(struct man *m, int line, int pos, int tok)
261 p = man_node_alloc(line, pos, MAN_ELEM, tok);
262 if ( ! man_node_append(m, p))
264 m->next = MAN_NEXT_CHILD;
270 man_head_alloc(struct man *m, int line, int pos, int tok)
274 p = man_node_alloc(line, pos, MAN_HEAD, tok);
275 if ( ! man_node_append(m, p))
277 m->next = MAN_NEXT_CHILD;
283 man_body_alloc(struct man *m, int line, int pos, int tok)
287 p = man_node_alloc(line, pos, MAN_BODY, tok);
288 if ( ! man_node_append(m, p))
290 m->next = MAN_NEXT_CHILD;
296 man_block_alloc(struct man *m, int line, int pos, int tok)
300 p = man_node_alloc(line, pos, MAN_BLOCK, tok);
301 if ( ! man_node_append(m, p))
303 m->next = MAN_NEXT_CHILD;
309 pstring(struct man *m, int line, int pos,
310 const char *p, size_t len)
315 n = man_node_alloc(line, pos, MAN_TEXT, -1);
316 n->string = mandoc_malloc(len + 1);
317 sv = strlcpy(n->string, p, len + 1);
319 /* Prohibit truncation. */
320 assert(sv < len + 1);
322 if ( ! man_node_append(m, n))
324 m->next = MAN_NEXT_SIBLING;
330 man_word_alloc(struct man *m, int line, int pos, const char *word)
333 return(pstring(m, line, pos, word, strlen(word)));
338 man_node_free(struct man_node *p)
350 man_node_freelist(struct man_node *p)
355 man_node_freelist(p->child);
356 assert(0 == p->nchild);
360 man_node_freelist(n);
365 man_ptext(struct man *m, int line, char *buf)
369 /* Literal free-form text whitespace is preserved. */
371 if (MAN_LITERAL & m->flags) {
372 if ( ! man_word_alloc(m, line, 0, buf))
377 /* First de-chunk and allocate words. */
379 for (i = 0; ' ' == buf[i]; i++)
380 /* Skip leading whitespace. */ ;
382 if ( ! pstring(m, line, 0, &buf[i], 0))
387 for (j = i; buf[i]; i++) {
391 /* Escaped whitespace. */
392 if (i && ' ' == buf[i] && '\\' == buf[i - 1])
396 if ( ! pstring(m, line, j, &buf[j], (size_t)(i - j)))
399 for ( ; ' ' == buf[i]; i++)
400 /* Skip trailing whitespace. */ ;
407 if (j != i && ! pstring(m, line, j, &buf[j], (size_t)(i - j)))
413 * Co-ordinate what happens with having a next-line scope open:
414 * first close out the element scope (if applicable), then close
415 * out the block scope (also if applicable).
418 if (MAN_ELINE & m->flags) {
419 m->flags &= ~MAN_ELINE;
420 if ( ! man_unscope(m, m->last->parent))
424 if ( ! (MAN_BLINE & m->flags))
426 m->flags &= ~MAN_BLINE;
428 if ( ! man_unscope(m, m->last->parent))
430 return(man_body_alloc(m, line, 0, m->last->tok));
435 macrowarn(struct man *m, int ln, const char *buf)
437 if ( ! (MAN_IGN_MACRO & m->pflags))
438 return(man_verr(m, ln, 0,
439 "unknown macro: %s%s",
440 buf, strlen(buf) > 3 ? "..." : ""));
441 return(man_vwarn(m, ln, 0, "unknown macro: %s%s",
442 buf, strlen(buf) > 3 ? "..." : ""));
447 man_pmacro(struct man *m, int ln, char *buf)
449 int i, j, c, ppos, fl;
453 /* Comments and empties are quickly ignored. */
464 while (buf[i] && ' ' == buf[i])
472 /* Copy the first word into a nil-terminated buffer. */
474 for (j = 0; j < 4; j++, i++) {
475 if (0 == (mac[j] = buf[i]))
477 else if (' ' == buf[i])
480 /* Check for invalid characters. */
482 if (isgraph((u_char)buf[i]))
484 return(man_perr(m, ln, i, WNPRINT));
489 if (j == 4 || j < 1) {
490 if ( ! (MAN_IGN_MACRO & m->pflags)) {
491 (void)man_perr(m, ln, ppos, WMACROFORM);
494 if ( ! man_pwarn(m, ln, ppos, WMACROFORM))
499 if (MAN_MAX == (c = man_hash_find(mac))) {
500 if ( ! macrowarn(m, ln, mac))
505 /* The macro is sane. Jump to the next word. */
507 while (buf[i] && ' ' == buf[i])
510 /* Remove prior ELINE macro, if applicable. */
512 if (m->flags & MAN_ELINE) {
514 assert(NULL == n->child);
515 assert(0 == n->nchild);
516 if ( ! man_nwarn(m, n, WLNSCOPE))
520 assert(n != n->parent->child);
521 assert(n == n->prev->next);
522 n->prev->next = NULL;
524 m->next = MAN_NEXT_SIBLING;
526 assert(n == n->parent->child);
527 n->parent->child = NULL;
529 m->next = MAN_NEXT_CHILD;
533 m->flags &= ~MAN_ELINE;
536 /* Begin recursive parse sequence. */
538 assert(man_macros[c].fp);
540 if ( ! (*man_macros[c].fp)(m, c, ln, ppos, &i, buf))
544 if ( ! (MAN_BLINE & fl))
548 * If we've opened a new next-line element scope, then return
549 * now, as the next line will close out the block scope.
552 if (MAN_ELINE & m->flags)
555 /* Close out the block scope opened in the prior line. */
557 assert(MAN_BLINE & m->flags);
558 m->flags &= ~MAN_BLINE;
560 if ( ! man_unscope(m, m->last->parent))
562 return(man_body_alloc(m, ln, 0, m->last->tok));
564 err: /* Error out. */
566 m->flags |= MAN_HALT;
572 man_verr(struct man *man, int ln, int pos, const char *fmt, ...)
577 if (NULL == man->cb.man_err)
581 (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
583 return((*man->cb.man_err)(man->data, ln, pos, buf));
588 man_vwarn(struct man *man, int ln, int pos, const char *fmt, ...)
593 if (NULL == man->cb.man_warn)
597 (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
599 return((*man->cb.man_warn)(man->data, ln, pos, buf));
604 man_err(struct man *m, int line, int pos, int iserr, enum merr type)
608 p = __man_merrnames[(int)type];
612 return(man_verr(m, line, pos, p));
614 return(man_vwarn(m, line, pos, p));