Annotation of src/usr.bin/mandoc/man.c, Revision 1.12
1.12 ! schwarze 1: /* $Id: man.c,v 1.11 2009/09/18 22:37:05 schwarze Exp $ */
1.1 kristaps 2: /*
1.2 schwarze 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
1.2 schwarze 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: *
1.2 schwarze 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: */
17: #include <assert.h>
18: #include <ctype.h>
19: #include <stdarg.h>
20: #include <stdlib.h>
21: #include <stdio.h>
22: #include <string.h>
23:
24: #include "libman.h"
25:
1.7 schwarze 26: const char *const __man_merrnames[WERRMAX] = {
27: "invalid character", /* WNPRINT */
28: "system: malloc error", /* WNMEM */
29: "invalid manual section", /* WMSEC */
30: "invalid date format", /* WDATE */
31: "scope of prior line violated", /* WLNSCOPE */
32: "trailing whitespace", /* WTSPACE */
33: "unterminated quoted parameter", /* WTQUOTE */
34: "document has no body", /* WNODATA */
35: "document has no title/section", /* WNOTITLE */
36: "invalid escape sequence", /* WESCAPE */
1.8 schwarze 37: "invalid number format", /* WNUMFMT */
1.9 schwarze 38: "expected block head arguments", /* WHEADARGS */
39: "expected block body arguments", /* WBODYARGS */
40: "expected empty block head", /* WNHEADARGS */
41: "unknown macro", /* WMACRO */
42: "ill-formed macro", /* WMACROFORM */
1.10 schwarze 43: "scope open on exit", /* WEXITSCOPE */
44: "no scope context", /* WNOSCOPE */
45: "literal context already open", /* WOLITERAL */
46: "no literal context open" /* WNLITERAL */
1.7 schwarze 47: };
48:
1.1 kristaps 49: const char *const __man_macronames[MAN_MAX] = {
1.3 schwarze 50: "br", "TH", "SH", "SS",
1.1 kristaps 51: "TP", "LP", "PP", "P",
52: "IP", "HP", "SM", "SB",
53: "BI", "IB", "BR", "RB",
54: "R", "B", "I", "IR",
1.9 schwarze 55: "RI", "na", "i", "sp",
1.10 schwarze 56: "nf", "fi", "r", "RE",
1.12 ! schwarze 57: "RS", "DT", "UC"
1.1 kristaps 58: };
59:
60: const char * const *man_macronames = __man_macronames;
61:
62: static struct man_node *man_node_alloc(int, int,
63: enum man_type, int);
64: static int man_node_append(struct man *,
65: struct man_node *);
66: static int man_ptext(struct man *, int, char *);
67: static int man_pmacro(struct man *, int, char *);
68: static void man_free1(struct man *);
69: static int man_alloc1(struct man *);
1.10 schwarze 70: static int pstring(struct man *, int, int,
71: const char *, size_t);
1.1 kristaps 72:
73:
74: const struct man_node *
75: man_node(const struct man *m)
76: {
77:
78: return(MAN_HALT & m->flags ? NULL : m->first);
79: }
80:
81:
82: const struct man_meta *
83: man_meta(const struct man *m)
84: {
85:
86: return(MAN_HALT & m->flags ? NULL : &m->meta);
87: }
88:
89:
90: int
91: man_reset(struct man *man)
92: {
93:
94: man_free1(man);
95: return(man_alloc1(man));
96: }
97:
98:
99: void
100: man_free(struct man *man)
101: {
102:
103: man_free1(man);
104:
105: if (man->htab)
106: man_hash_free(man->htab);
107: free(man);
108: }
109:
110:
111: struct man *
112: man_alloc(void *data, int pflags, const struct man_cb *cb)
113: {
114: struct man *p;
115:
116: if (NULL == (p = calloc(1, sizeof(struct man))))
117: return(NULL);
118:
119: if ( ! man_alloc1(p)) {
120: free(p);
121: return(NULL);
122: }
123:
124: p->data = data;
125: p->pflags = pflags;
126: (void)memcpy(&p->cb, cb, sizeof(struct man_cb));
127:
128: if (NULL == (p->htab = man_hash_alloc())) {
129: free(p);
130: return(NULL);
131: }
132: return(p);
133: }
134:
135:
136: int
137: man_endparse(struct man *m)
138: {
139:
140: if (MAN_HALT & m->flags)
141: return(0);
142: else if (man_macroend(m))
143: return(1);
144: m->flags |= MAN_HALT;
145: return(0);
146: }
147:
148:
149: int
150: man_parseln(struct man *m, int ln, char *buf)
151: {
152:
153: return('.' == *buf ?
154: man_pmacro(m, ln, buf) :
155: man_ptext(m, ln, buf));
156: }
157:
158:
159: static void
160: man_free1(struct man *man)
161: {
162:
163: if (man->first)
164: man_node_freelist(man->first);
165: if (man->meta.title)
166: free(man->meta.title);
167: if (man->meta.source)
168: free(man->meta.source);
169: if (man->meta.vol)
170: free(man->meta.vol);
171: }
172:
173:
174: static int
175: man_alloc1(struct man *m)
176: {
177:
178: bzero(&m->meta, sizeof(struct man_meta));
179: m->flags = 0;
180: m->last = calloc(1, sizeof(struct man_node));
181: if (NULL == m->last)
182: return(0);
183: m->first = m->last;
184: m->last->type = MAN_ROOT;
185: m->next = MAN_NEXT_CHILD;
186: return(1);
187: }
188:
189:
190: static int
191: man_node_append(struct man *man, struct man_node *p)
192: {
193:
194: assert(man->last);
195: assert(man->first);
196: assert(MAN_ROOT != p->type);
197:
198: switch (man->next) {
199: case (MAN_NEXT_SIBLING):
200: man->last->next = p;
201: p->prev = man->last;
202: p->parent = man->last->parent;
203: break;
204: case (MAN_NEXT_CHILD):
205: man->last->child = p;
206: p->parent = man->last;
207: break;
208: default:
209: abort();
210: /* NOTREACHED */
211: }
1.5 schwarze 212:
213: p->parent->nchild++;
1.1 kristaps 214:
1.9 schwarze 215: if ( ! man_valid_pre(man, p))
216: return(0);
217:
218: switch (p->type) {
219: case (MAN_HEAD):
220: assert(MAN_BLOCK == p->parent->type);
221: p->parent->head = p;
222: break;
223: case (MAN_BODY):
224: assert(MAN_BLOCK == p->parent->type);
225: p->parent->body = p;
226: break;
227: default:
228: break;
229: }
230:
1.1 kristaps 231: man->last = p;
232:
233: switch (p->type) {
234: case (MAN_TEXT):
235: if ( ! man_valid_post(man))
236: return(0);
237: if ( ! man_action_post(man))
238: return(0);
239: break;
240: default:
241: break;
242: }
243:
244: return(1);
245: }
246:
247:
248: static struct man_node *
249: man_node_alloc(int line, int pos, enum man_type type, int tok)
250: {
251: struct man_node *p;
252:
253: p = calloc(1, sizeof(struct man_node));
254: if (NULL == p)
255: return(NULL);
256:
257: p->line = line;
258: p->pos = pos;
259: p->type = type;
260: p->tok = tok;
261: return(p);
262: }
263:
264:
265: int
1.10 schwarze 266: man_elem_alloc(struct man *m, int line, int pos, int tok)
1.1 kristaps 267: {
268: struct man_node *p;
269:
270: p = man_node_alloc(line, pos, MAN_ELEM, tok);
271: if (NULL == p)
272: return(0);
1.10 schwarze 273: if ( ! man_node_append(m, p))
274: return(0);
275: m->next = MAN_NEXT_CHILD;
276: return(1);
1.1 kristaps 277: }
278:
279:
280: int
1.9 schwarze 281: man_head_alloc(struct man *m, int line, int pos, int tok)
282: {
283: struct man_node *p;
284:
285: p = man_node_alloc(line, pos, MAN_HEAD, tok);
286: if (NULL == p)
287: return(0);
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_body_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_BODY, tok);
301: if (NULL == p)
302: return(0);
303: if ( ! man_node_append(m, p))
304: return(0);
305: m->next = MAN_NEXT_CHILD;
306: return(1);
307: }
308:
309:
310: int
311: man_block_alloc(struct man *m, int line, int pos, int tok)
312: {
313: struct man_node *p;
314:
315: p = man_node_alloc(line, pos, MAN_BLOCK, tok);
316: if (NULL == p)
317: return(0);
318: if ( ! man_node_append(m, p))
319: return(0);
320: m->next = MAN_NEXT_CHILD;
321: return(1);
322: }
323:
324:
1.10 schwarze 325: static int
326: pstring(struct man *m, int line, int pos,
327: const char *p, size_t len)
1.1 kristaps 328: {
1.10 schwarze 329: struct man_node *n;
330: size_t sv;
1.1 kristaps 331:
1.10 schwarze 332: n = man_node_alloc(line, pos, MAN_TEXT, -1);
333: if (NULL == n)
334: return(0);
335:
336: n->string = malloc(len + 1);
337: if (NULL == n->string) {
338: free(n);
1.1 kristaps 339: return(0);
1.10 schwarze 340: }
341:
342: sv = strlcpy(n->string, p, len + 1);
343:
344: /* Prohibit truncation. */
345: assert(sv < len + 1);
346:
347: if ( ! man_node_append(m, n))
1.1 kristaps 348: return(0);
1.10 schwarze 349: m->next = MAN_NEXT_SIBLING;
350: return(1);
351: }
352:
353:
354: int
355: man_word_alloc(struct man *m, int line, int pos, const char *word)
356: {
357:
358: return(pstring(m, line, pos, word, strlen(word)));
1.1 kristaps 359: }
360:
361:
362: void
363: man_node_free(struct man_node *p)
364: {
365:
366: if (p->string)
367: free(p->string);
1.5 schwarze 368: if (p->parent)
369: p->parent->nchild--;
1.1 kristaps 370: free(p);
371: }
372:
373:
374: void
375: man_node_freelist(struct man_node *p)
376: {
1.10 schwarze 377: struct man_node *n;
1.1 kristaps 378:
379: if (p->child)
380: man_node_freelist(p->child);
1.5 schwarze 381: assert(0 == p->nchild);
1.10 schwarze 382: n = p->next;
1.1 kristaps 383: man_node_free(p);
1.10 schwarze 384: if (n)
385: man_node_freelist(n);
1.1 kristaps 386: }
387:
388:
389: static int
390: man_ptext(struct man *m, int line, char *buf)
391: {
1.10 schwarze 392: int i, j;
393:
394: /* Literal free-form text whitespace is preserved. */
395:
396: if (MAN_LITERAL & m->flags) {
397: if ( ! man_word_alloc(m, line, 0, buf))
398: return(0);
399: goto descope;
400: }
401:
402: /* First de-chunk and allocate words. */
403:
404: for (i = 0; ' ' == buf[i]; i++)
405: /* Skip leading whitespace. */ ;
406: if (0 == buf[i]) {
407: if ( ! pstring(m, line, 0, &buf[i], 0))
408: return(0);
409: goto descope;
410: }
1.1 kristaps 411:
1.10 schwarze 412: for (j = i; buf[i]; i++) {
413: if (' ' != buf[i])
414: continue;
415:
416: /* Escaped whitespace. */
417: if (i && ' ' == buf[i] && '\\' == buf[i - 1])
418: continue;
419:
420: buf[i++] = 0;
421: if ( ! pstring(m, line, j, &buf[j], (size_t)(i - j)))
422: return(0);
423:
424: for ( ; ' ' == buf[i]; i++)
425: /* Skip trailing whitespace. */ ;
426:
427: j = i;
428: if (0 == buf[i])
429: break;
430: }
1.9 schwarze 431:
1.10 schwarze 432: if (j != i && ! pstring(m, line, j, &buf[j], (size_t)(i - j)))
1.1 kristaps 433: return(0);
1.10 schwarze 434:
435: descope:
1.1 kristaps 436:
437: /*
1.9 schwarze 438: * Co-ordinate what happens with having a next-line scope open:
439: * first close out the element scope (if applicable), then close
440: * out the block scope (also if applicable).
1.1 kristaps 441: */
442:
1.9 schwarze 443: if (MAN_ELINE & m->flags) {
444: m->flags &= ~MAN_ELINE;
445: if ( ! man_unscope(m, m->last->parent))
446: return(0);
447: }
448:
449: if ( ! (MAN_BLINE & m->flags))
1.1 kristaps 450: return(1);
1.9 schwarze 451: m->flags &= ~MAN_BLINE;
1.1 kristaps 452:
1.9 schwarze 453: if ( ! man_unscope(m, m->last->parent))
1.1 kristaps 454: return(0);
1.9 schwarze 455: return(man_body_alloc(m, line, 0, m->last->tok));
1.1 kristaps 456: }
457:
458:
459: int
460: man_pmacro(struct man *m, int ln, char *buf)
461: {
1.10 schwarze 462: int i, j, c, ppos, fl;
463: char mac[5];
464: struct man_node *n;
1.1 kristaps 465:
466: /* Comments and empties are quickly ignored. */
467:
1.9 schwarze 468: fl = m->flags;
1.1 kristaps 469:
470: if (0 == buf[1])
471: goto out;
472:
473: i = 1;
474:
475: if (' ' == buf[i]) {
476: i++;
477: while (buf[i] && ' ' == buf[i])
478: i++;
479: if (0 == buf[i])
480: goto out;
481: }
482:
483: ppos = i;
484:
485: /* Copy the first word into a nil-terminated buffer. */
486:
487: for (j = 0; j < 4; j++, i++) {
488: if (0 == (mac[j] = buf[i]))
489: break;
490: else if (' ' == buf[i])
491: break;
1.11 schwarze 492:
493: /* Check for invalid characters. */
494:
495: if (isgraph((u_char)buf[i]))
496: continue;
497: return(man_perr(m, ln, i, WNPRINT));
1.1 kristaps 498: }
499:
500: mac[j] = 0;
501:
502: if (j == 4 || j < 1) {
503: if ( ! (MAN_IGN_MACRO & m->pflags)) {
1.9 schwarze 504: (void)man_perr(m, ln, ppos, WMACROFORM);
1.1 kristaps 505: goto err;
506: }
1.9 schwarze 507: if ( ! man_pwarn(m, ln, ppos, WMACROFORM))
1.1 kristaps 508: goto err;
509: return(1);
510: }
511:
512: if (MAN_MAX == (c = man_hash_find(m->htab, mac))) {
513: if ( ! (MAN_IGN_MACRO & m->pflags)) {
1.9 schwarze 514: (void)man_perr(m, ln, ppos, WMACRO);
1.1 kristaps 515: goto err;
516: }
1.9 schwarze 517: if ( ! man_pwarn(m, ln, ppos, WMACRO))
1.1 kristaps 518: goto err;
519: return(1);
520: }
521:
522: /* The macro is sane. Jump to the next word. */
523:
524: while (buf[i] && ' ' == buf[i])
525: i++;
526:
1.10 schwarze 527: /* Remove prior ELINE macro, if applicable. */
528:
529: if (m->flags & MAN_ELINE) {
530: n = m->last;
531: assert(NULL == n->child);
532: assert(0 == n->nchild);
533: if ( ! man_nwarn(m, n, WLNSCOPE))
534: return(0);
535:
536: if (n->prev) {
537: assert(n != n->parent->child);
538: assert(n == n->prev->next);
539: n->prev->next = NULL;
540: m->last = n->prev;
541: m->next = MAN_NEXT_SIBLING;
542: } else {
543: assert(n == n->parent->child);
544: n->parent->child = NULL;
545: m->last = n->parent;
546: m->next = MAN_NEXT_CHILD;
547: }
548:
549: man_node_free(n);
550: m->flags &= ~MAN_ELINE;
551: }
552:
1.1 kristaps 553: /* Begin recursive parse sequence. */
554:
1.9 schwarze 555: assert(man_macros[c].fp);
556:
557: if ( ! (*man_macros[c].fp)(m, c, ln, ppos, &i, buf))
1.1 kristaps 558: goto err;
559:
560: out:
1.9 schwarze 561: if ( ! (MAN_BLINE & fl))
562: return(1);
563:
564: /*
565: * If we've opened a new next-line element scope, then return
566: * now, as the next line will close out the block scope.
567: */
568:
569: if (MAN_ELINE & m->flags)
570: return(1);
571:
572: /* Close out the block scope opened in the prior line. */
1.1 kristaps 573:
1.9 schwarze 574: assert(MAN_BLINE & m->flags);
575: m->flags &= ~MAN_BLINE;
1.1 kristaps 576:
1.9 schwarze 577: if ( ! man_unscope(m, m->last->parent))
578: return(0);
579: return(man_body_alloc(m, ln, 0, m->last->tok));
1.1 kristaps 580:
581: err: /* Error out. */
582:
583: m->flags |= MAN_HALT;
584: return(0);
585: }
586:
587:
588: int
589: man_verr(struct man *man, int ln, int pos, const char *fmt, ...)
590: {
591: char buf[256];
592: va_list ap;
593:
594: if (NULL == man->cb.man_err)
595: return(0);
596:
597: va_start(ap, fmt);
598: (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
599: va_end(ap);
600: return((*man->cb.man_err)(man->data, ln, pos, buf));
601: }
602:
603:
604: int
605: man_vwarn(struct man *man, int ln, int pos, const char *fmt, ...)
606: {
607: char buf[256];
608: va_list ap;
609:
610: if (NULL == man->cb.man_warn)
611: return(0);
612:
613: va_start(ap, fmt);
614: (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
615: va_end(ap);
616: return((*man->cb.man_warn)(man->data, ln, pos, buf));
617: }
618:
619:
1.4 schwarze 620: int
1.7 schwarze 621: man_err(struct man *m, int line, int pos, int iserr, enum merr type)
1.4 schwarze 622: {
623: const char *p;
624:
1.7 schwarze 625: p = __man_merrnames[(int)type];
1.4 schwarze 626: assert(p);
627:
628: if (iserr)
629: return(man_verr(m, line, pos, p));
630:
631: return(man_vwarn(m, line, pos, p));
632: }