Annotation of mandoc/man.c, Revision 1.45
1.45 ! kristaps 1: /* $Id: man.c,v 1.44 2009/10/27 08:26:12 kristaps Exp $ */
1.1 kristaps 2: /*
1.19 kristaps 3: * Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@kth.se>
1.1 kristaps 4: *
5: * Permission to use, copy, modify, and distribute this software for any
1.18 kristaps 6: * purpose with or without fee is hereby granted, provided that the above
7: * copyright notice and this permission notice appear in all copies.
1.1 kristaps 8: *
1.18 kristaps 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.
1.1 kristaps 16: */
1.41 kristaps 17: #include <sys/types.h>
18:
1.1 kristaps 19: #include <assert.h>
20: #include <ctype.h>
21: #include <stdarg.h>
22: #include <stdlib.h>
23: #include <stdio.h>
24: #include <string.h>
25:
26: #include "libman.h"
1.45 ! kristaps 27: #include "libmandoc.h"
1.1 kristaps 28:
1.27 kristaps 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 */
1.28 kristaps 39: "invalid number format", /* WNUMFMT */
1.29 kristaps 40: "expected block head arguments", /* WHEADARGS */
41: "expected block body arguments", /* WBODYARGS */
42: "expected empty block head", /* WNHEADARGS */
43: "ill-formed macro", /* WMACROFORM */
1.30 kristaps 44: "scope open on exit", /* WEXITSCOPE */
1.35 kristaps 45: "no scope context", /* WNOSCOPE */
46: "literal context already open", /* WOLITERAL */
47: "no literal context open" /* WNLITERAL */
1.27 kristaps 48: };
49:
1.1 kristaps 50: const char *const __man_macronames[MAN_MAX] = {
1.21 kristaps 51: "br", "TH", "SH", "SS",
1.1 kristaps 52: "TP", "LP", "PP", "P",
53: "IP", "HP", "SM", "SB",
54: "BI", "IB", "BR", "RB",
1.11 kristaps 55: "R", "B", "I", "IR",
1.29 kristaps 56: "RI", "na", "i", "sp",
1.30 kristaps 57: "nf", "fi", "r", "RE",
1.42 kristaps 58: "RS", "DT", "UC", "PD"
1.1 kristaps 59: };
60:
61: const char * const *man_macronames = __man_macronames;
62:
1.16 kristaps 63: static struct man_node *man_node_alloc(int, int,
64: enum man_type, int);
1.1 kristaps 65: static int man_node_append(struct man *,
66: struct man_node *);
67: static int man_ptext(struct man *, int, char *);
68: static int man_pmacro(struct man *, int, char *);
1.2 kristaps 69: static void man_free1(struct man *);
1.45 ! kristaps 70: static void man_alloc1(struct man *);
1.31 kristaps 71: static int pstring(struct man *, int, int,
72: const char *, size_t);
1.43 kristaps 73: static int macrowarn(struct man *, int, const char *);
1.1 kristaps 74:
1.32 kristaps 75: #ifdef __linux__
76: extern size_t strlcpy(char *, const char *, size_t);
77: #endif
78:
1.1 kristaps 79:
80: const struct man_node *
1.2 kristaps 81: man_node(const struct man *m)
1.1 kristaps 82: {
83:
1.2 kristaps 84: return(MAN_HALT & m->flags ? NULL : m->first);
1.1 kristaps 85: }
86:
87:
88: const struct man_meta *
1.2 kristaps 89: man_meta(const struct man *m)
1.1 kristaps 90: {
91:
1.2 kristaps 92: return(MAN_HALT & m->flags ? NULL : &m->meta);
1.1 kristaps 93: }
94:
95:
1.45 ! kristaps 96: void
1.1 kristaps 97: man_reset(struct man *man)
98: {
99:
1.2 kristaps 100: man_free1(man);
1.45 ! kristaps 101: man_alloc1(man);
1.1 kristaps 102: }
103:
104:
105: void
106: man_free(struct man *man)
107: {
108:
1.2 kristaps 109: man_free1(man);
1.1 kristaps 110: free(man);
111: }
112:
113:
114: struct man *
1.7 kristaps 115: man_alloc(void *data, int pflags, const struct man_cb *cb)
1.1 kristaps 116: {
117: struct man *p;
118:
1.45 ! kristaps 119: p = mandoc_calloc(1, sizeof(struct man));
1.2 kristaps 120:
1.45 ! kristaps 121: if (cb)
! 122: memcpy(&p->cb, cb, sizeof(struct man_cb));
1.1 kristaps 123:
1.40 kristaps 124: man_hash_init();
1.4 kristaps 125: p->data = data;
1.7 kristaps 126: p->pflags = pflags;
1.45 ! kristaps 127:
! 128: man_alloc1(p);
1.1 kristaps 129: return(p);
130: }
131:
132:
133: int
134: man_endparse(struct man *m)
135: {
136:
1.3 kristaps 137: if (MAN_HALT & m->flags)
138: return(0);
139: else if (man_macroend(m))
140: return(1);
141: m->flags |= MAN_HALT;
142: return(0);
1.1 kristaps 143: }
144:
145:
146: int
147: man_parseln(struct man *m, int ln, char *buf)
148: {
149:
150: return('.' == *buf ?
151: man_pmacro(m, ln, buf) :
152: man_ptext(m, ln, buf));
153: }
154:
155:
1.2 kristaps 156: static void
157: man_free1(struct man *man)
158: {
159:
160: if (man->first)
161: man_node_freelist(man->first);
162: if (man->meta.title)
163: free(man->meta.title);
1.6 kristaps 164: if (man->meta.source)
165: free(man->meta.source);
1.2 kristaps 166: if (man->meta.vol)
167: free(man->meta.vol);
168: }
169:
170:
1.45 ! kristaps 171: static void
1.2 kristaps 172: man_alloc1(struct man *m)
173: {
174:
1.44 kristaps 175: memset(&m->meta, 0, sizeof(struct man_meta));
1.2 kristaps 176: m->flags = 0;
1.45 ! kristaps 177: m->last = mandoc_calloc(1, sizeof(struct man_node));
1.2 kristaps 178: m->first = m->last;
179: m->last->type = MAN_ROOT;
180: m->next = MAN_NEXT_CHILD;
181: }
182:
183:
1.1 kristaps 184: static int
185: man_node_append(struct man *man, struct man_node *p)
186: {
187:
188: assert(man->last);
189: assert(man->first);
190: assert(MAN_ROOT != p->type);
191:
192: switch (man->next) {
193: case (MAN_NEXT_SIBLING):
194: man->last->next = p;
195: p->prev = man->last;
196: p->parent = man->last->parent;
197: break;
198: case (MAN_NEXT_CHILD):
199: man->last->child = p;
200: p->parent = man->last;
201: break;
202: default:
203: abort();
204: /* NOTREACHED */
205: }
1.22 kristaps 206:
207: p->parent->nchild++;
1.1 kristaps 208:
1.29 kristaps 209: if ( ! man_valid_pre(man, p))
210: return(0);
211:
212: switch (p->type) {
213: case (MAN_HEAD):
214: assert(MAN_BLOCK == p->parent->type);
215: p->parent->head = p;
216: break;
217: case (MAN_BODY):
218: assert(MAN_BLOCK == p->parent->type);
219: p->parent->body = p;
220: break;
221: default:
222: break;
223: }
224:
1.2 kristaps 225: man->last = p;
226:
1.1 kristaps 227: switch (p->type) {
1.2 kristaps 228: case (MAN_TEXT):
229: if ( ! man_valid_post(man))
230: return(0);
231: if ( ! man_action_post(man))
232: return(0);
1.1 kristaps 233: break;
234: default:
235: break;
236: }
237:
238: return(1);
239: }
240:
241:
242: static struct man_node *
1.16 kristaps 243: man_node_alloc(int line, int pos, enum man_type type, int tok)
1.1 kristaps 244: {
245: struct man_node *p;
246:
1.45 ! kristaps 247: p = mandoc_calloc(1, sizeof(struct man_node));
1.1 kristaps 248: p->line = line;
249: p->pos = pos;
250: p->type = type;
1.16 kristaps 251: p->tok = tok;
1.1 kristaps 252: return(p);
253: }
254:
255:
256: int
1.30 kristaps 257: man_elem_alloc(struct man *m, int line, int pos, int tok)
1.1 kristaps 258: {
259: struct man_node *p;
260:
1.16 kristaps 261: p = man_node_alloc(line, pos, MAN_ELEM, tok);
1.30 kristaps 262: if ( ! man_node_append(m, p))
263: return(0);
264: m->next = MAN_NEXT_CHILD;
265: return(1);
1.1 kristaps 266: }
267:
268:
269: int
1.29 kristaps 270: man_head_alloc(struct man *m, int line, int pos, int tok)
271: {
272: struct man_node *p;
273:
274: p = man_node_alloc(line, pos, MAN_HEAD, tok);
275: if ( ! man_node_append(m, p))
276: return(0);
277: m->next = MAN_NEXT_CHILD;
278: return(1);
279: }
280:
281:
282: int
283: man_body_alloc(struct man *m, int line, int pos, int tok)
284: {
285: struct man_node *p;
286:
287: p = man_node_alloc(line, pos, MAN_BODY, tok);
288: if ( ! man_node_append(m, p))
289: return(0);
290: m->next = MAN_NEXT_CHILD;
291: return(1);
292: }
293:
294:
295: int
296: man_block_alloc(struct man *m, int line, int pos, int tok)
297: {
298: struct man_node *p;
299:
300: p = man_node_alloc(line, pos, MAN_BLOCK, tok);
301: if ( ! man_node_append(m, p))
302: return(0);
303: m->next = MAN_NEXT_CHILD;
304: return(1);
305: }
306:
307:
1.31 kristaps 308: static int
309: pstring(struct man *m, int line, int pos,
310: const char *p, size_t len)
1.1 kristaps 311: {
1.31 kristaps 312: struct man_node *n;
313: size_t sv;
1.1 kristaps 314:
1.31 kristaps 315: n = man_node_alloc(line, pos, MAN_TEXT, -1);
1.45 ! kristaps 316: n->string = mandoc_malloc(len + 1);
1.31 kristaps 317: sv = strlcpy(n->string, p, len + 1);
318:
319: /* Prohibit truncation. */
320: assert(sv < len + 1);
321:
322: if ( ! man_node_append(m, n))
1.30 kristaps 323: return(0);
324: m->next = MAN_NEXT_SIBLING;
325: return(1);
1.1 kristaps 326: }
327:
328:
1.31 kristaps 329: int
330: man_word_alloc(struct man *m, int line, int pos, const char *word)
331: {
332:
333: return(pstring(m, line, pos, word, strlen(word)));
334: }
335:
336:
1.1 kristaps 337: void
338: man_node_free(struct man_node *p)
339: {
340:
341: if (p->string)
342: free(p->string);
1.24 kristaps 343: if (p->parent)
344: p->parent->nchild--;
1.1 kristaps 345: free(p);
346: }
347:
348:
349: void
350: man_node_freelist(struct man_node *p)
351: {
1.37 kristaps 352: struct man_node *n;
1.1 kristaps 353:
354: if (p->child)
355: man_node_freelist(p->child);
1.24 kristaps 356: assert(0 == p->nchild);
1.37 kristaps 357: n = p->next;
1.1 kristaps 358: man_node_free(p);
1.37 kristaps 359: if (n)
360: man_node_freelist(n);
1.1 kristaps 361: }
362:
363:
364: static int
365: man_ptext(struct man *m, int line, char *buf)
366: {
1.31 kristaps 367: int i, j;
368:
1.35 kristaps 369: /* Literal free-form text whitespace is preserved. */
370:
371: if (MAN_LITERAL & m->flags) {
372: if ( ! man_word_alloc(m, line, 0, buf))
373: return(0);
374: goto descope;
375: }
376:
1.31 kristaps 377: /* First de-chunk and allocate words. */
378:
379: for (i = 0; ' ' == buf[i]; i++)
380: /* Skip leading whitespace. */ ;
381: if (0 == buf[i]) {
382: if ( ! pstring(m, line, 0, &buf[i], 0))
383: return(0);
384: goto descope;
385: }
386:
387: for (j = i; buf[i]; i++) {
388: if (' ' != buf[i])
389: continue;
390:
391: /* Escaped whitespace. */
392: if (i && ' ' == buf[i] && '\\' == buf[i - 1])
393: continue;
1.1 kristaps 394:
1.31 kristaps 395: buf[i++] = 0;
396: if ( ! pstring(m, line, j, &buf[j], (size_t)(i - j)))
397: return(0);
1.29 kristaps 398:
1.31 kristaps 399: for ( ; ' ' == buf[i]; i++)
400: /* Skip trailing whitespace. */ ;
1.30 kristaps 401:
1.31 kristaps 402: j = i;
403: if (0 == buf[i])
404: break;
405: }
406:
407: if (j != i && ! pstring(m, line, j, &buf[j], (size_t)(i - j)))
1.1 kristaps 408: return(0);
1.31 kristaps 409:
410: descope:
1.11 kristaps 411:
412: /*
1.29 kristaps 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).
1.11 kristaps 416: */
417:
1.29 kristaps 418: if (MAN_ELINE & m->flags) {
419: m->flags &= ~MAN_ELINE;
420: if ( ! man_unscope(m, m->last->parent))
421: return(0);
422: }
423:
424: if ( ! (MAN_BLINE & m->flags))
1.11 kristaps 425: return(1);
1.29 kristaps 426: m->flags &= ~MAN_BLINE;
1.11 kristaps 427:
1.29 kristaps 428: if ( ! man_unscope(m, m->last->parent))
1.11 kristaps 429: return(0);
1.29 kristaps 430: return(man_body_alloc(m, line, 0, m->last->tok));
1.1 kristaps 431: }
432:
433:
1.43 kristaps 434: static int
435: macrowarn(struct man *m, int ln, const char *buf)
436: {
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 ? "..." : ""));
443: }
444:
445:
1.1 kristaps 446: int
447: man_pmacro(struct man *m, int ln, char *buf)
448: {
1.34 kristaps 449: int i, j, c, ppos, fl;
450: char mac[5];
451: struct man_node *n;
1.1 kristaps 452:
453: /* Comments and empties are quickly ignored. */
454:
1.29 kristaps 455: fl = m->flags;
1.11 kristaps 456:
1.1 kristaps 457: if (0 == buf[1])
1.11 kristaps 458: goto out;
1.1 kristaps 459:
1.9 kristaps 460: i = 1;
461:
462: if (' ' == buf[i]) {
463: i++;
1.1 kristaps 464: while (buf[i] && ' ' == buf[i])
465: i++;
466: if (0 == buf[i])
1.11 kristaps 467: goto out;
1.1 kristaps 468: }
469:
1.10 kristaps 470: ppos = i;
471:
1.1 kristaps 472: /* Copy the first word into a nil-terminated buffer. */
473:
1.10 kristaps 474: for (j = 0; j < 4; j++, i++) {
475: if (0 == (mac[j] = buf[i]))
1.1 kristaps 476: break;
1.10 kristaps 477: else if (' ' == buf[i])
1.1 kristaps 478: break;
1.38 kristaps 479:
480: /* Check for invalid characters. */
481:
482: if (isgraph((u_char)buf[i]))
483: continue;
484: return(man_perr(m, ln, i, WNPRINT));
1.1 kristaps 485: }
486:
1.9 kristaps 487: mac[j] = 0;
1.1 kristaps 488:
1.9 kristaps 489: if (j == 4 || j < 1) {
1.7 kristaps 490: if ( ! (MAN_IGN_MACRO & m->pflags)) {
1.29 kristaps 491: (void)man_perr(m, ln, ppos, WMACROFORM);
1.7 kristaps 492: goto err;
493: }
1.29 kristaps 494: if ( ! man_pwarn(m, ln, ppos, WMACROFORM))
1.7 kristaps 495: goto err;
1.12 kristaps 496: return(1);
1.7 kristaps 497: }
1.1 kristaps 498:
1.40 kristaps 499: if (MAN_MAX == (c = man_hash_find(mac))) {
1.43 kristaps 500: if ( ! macrowarn(m, ln, mac))
1.7 kristaps 501: goto err;
1.12 kristaps 502: return(1);
1.1 kristaps 503: }
504:
505: /* The macro is sane. Jump to the next word. */
506:
507: while (buf[i] && ' ' == buf[i])
508: i++;
509:
1.34 kristaps 510: /* Remove prior ELINE macro, if applicable. */
511:
512: if (m->flags & MAN_ELINE) {
513: n = m->last;
514: assert(NULL == n->child);
1.37 kristaps 515: assert(0 == n->nchild);
1.34 kristaps 516: if ( ! man_nwarn(m, n, WLNSCOPE))
517: return(0);
518:
519: if (n->prev) {
520: assert(n != n->parent->child);
521: assert(n == n->prev->next);
522: n->prev->next = NULL;
523: m->last = n->prev;
1.37 kristaps 524: m->next = MAN_NEXT_SIBLING;
1.34 kristaps 525: } else {
526: assert(n == n->parent->child);
527: n->parent->child = NULL;
528: m->last = n->parent;
1.37 kristaps 529: m->next = MAN_NEXT_CHILD;
1.34 kristaps 530: }
531:
532: man_node_free(n);
533: m->flags &= ~MAN_ELINE;
534: }
535:
1.1 kristaps 536: /* Begin recursive parse sequence. */
537:
1.29 kristaps 538: assert(man_macros[c].fp);
539:
540: if ( ! (*man_macros[c].fp)(m, c, ln, ppos, &i, buf))
1.1 kristaps 541: goto err;
542:
1.11 kristaps 543: out:
1.29 kristaps 544: if ( ! (MAN_BLINE & fl))
545: return(1);
546:
547: /*
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.
550: */
551:
552: if (MAN_ELINE & m->flags)
553: return(1);
554:
555: /* Close out the block scope opened in the prior line. */
1.11 kristaps 556:
1.29 kristaps 557: assert(MAN_BLINE & m->flags);
558: m->flags &= ~MAN_BLINE;
1.11 kristaps 559:
1.29 kristaps 560: if ( ! man_unscope(m, m->last->parent))
561: return(0);
562: return(man_body_alloc(m, ln, 0, m->last->tok));
1.1 kristaps 563:
564: err: /* Error out. */
565:
1.2 kristaps 566: m->flags |= MAN_HALT;
1.1 kristaps 567: return(0);
568: }
1.3 kristaps 569:
1.4 kristaps 570:
571: int
572: man_verr(struct man *man, int ln, int pos, const char *fmt, ...)
573: {
574: char buf[256];
575: va_list ap;
576:
577: if (NULL == man->cb.man_err)
578: return(0);
579:
580: va_start(ap, fmt);
581: (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
582: va_end(ap);
583: return((*man->cb.man_err)(man->data, ln, pos, buf));
584: }
585:
586:
587: int
588: man_vwarn(struct man *man, int ln, int pos, const char *fmt, ...)
589: {
590: char buf[256];
591: va_list ap;
592:
593: if (NULL == man->cb.man_warn)
594: return(0);
595:
596: va_start(ap, fmt);
597: (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
598: va_end(ap);
599: return((*man->cb.man_warn)(man->data, ln, pos, buf));
600: }
601:
602:
1.23 kristaps 603: int
1.27 kristaps 604: man_err(struct man *m, int line, int pos, int iserr, enum merr type)
1.23 kristaps 605: {
606: const char *p;
607:
1.27 kristaps 608: p = __man_merrnames[(int)type];
1.23 kristaps 609: assert(p);
610:
611: if (iserr)
612: return(man_verr(m, line, pos, p));
613:
614: return(man_vwarn(m, line, pos, p));
615: }
CVSweb