Annotation of mandoc/mdoc.c, Revision 1.164
1.164 ! kristaps 1: /* $Id: mdoc.c,v 1.163 2010/08/20 01:02:07 schwarze Exp $ */
1.1 kristaps 2: /*
1.157 schwarze 3: * Copyright (c) 2008, 2009, 2010 Kristaps Dzonsons <kristaps@bsd.lv>
4: * Copyright (c) 2010 Ingo Schwarze <schwarze@openbsd.org>
1.1 kristaps 5: *
6: * Permission to use, copy, modify, and distribute this software for any
1.75 kristaps 7: * purpose with or without fee is hereby granted, provided that the above
8: * copyright notice and this permission notice appear in all copies.
1.1 kristaps 9: *
1.75 kristaps 10: * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
11: * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
12: * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
13: * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
14: * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
15: * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
16: * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
1.1 kristaps 17: */
1.114 kristaps 18: #ifdef HAVE_CONFIG_H
19: #include "config.h"
20: #endif
21:
1.106 kristaps 22: #include <sys/types.h>
23:
1.1 kristaps 24: #include <assert.h>
25: #include <stdarg.h>
1.73 kristaps 26: #include <stdio.h>
1.1 kristaps 27: #include <stdlib.h>
28: #include <string.h>
1.120 kristaps 29: #include <time.h>
1.1 kristaps 30:
1.136 kristaps 31: #include "mandoc.h"
1.70 kristaps 32: #include "libmdoc.h"
1.113 kristaps 33: #include "libmandoc.h"
1.1 kristaps 34:
35: const char *const __mdoc_macronames[MDOC_MAX] = {
1.82 kristaps 36: "Ap", "Dd", "Dt", "Os",
1.1 kristaps 37: "Sh", "Ss", "Pp", "D1",
38: "Dl", "Bd", "Ed", "Bl",
39: "El", "It", "Ad", "An",
40: "Ar", "Cd", "Cm", "Dv",
41: "Er", "Ev", "Ex", "Fa",
42: "Fd", "Fl", "Fn", "Ft",
43: "Ic", "In", "Li", "Nd",
44: "Nm", "Op", "Ot", "Pa",
45: "Rv", "St", "Va", "Vt",
46: /* LINTED */
1.114 kristaps 47: "Xr", "%A", "%B", "%D",
1.1 kristaps 48: /* LINTED */
1.114 kristaps 49: "%I", "%J", "%N", "%O",
1.1 kristaps 50: /* LINTED */
1.114 kristaps 51: "%P", "%R", "%T", "%V",
1.1 kristaps 52: "Ac", "Ao", "Aq", "At",
53: "Bc", "Bf", "Bo", "Bq",
54: "Bsx", "Bx", "Db", "Dc",
55: "Do", "Dq", "Ec", "Ef",
56: "Em", "Eo", "Fx", "Ms",
57: "No", "Ns", "Nx", "Ox",
58: "Pc", "Pf", "Po", "Pq",
59: "Qc", "Ql", "Qo", "Qq",
60: "Re", "Rs", "Sc", "So",
61: "Sq", "Sm", "Sx", "Sy",
62: "Tn", "Ux", "Xc", "Xo",
63: "Fo", "Fc", "Oo", "Oc",
64: "Bk", "Ek", "Bt", "Hf",
1.82 kristaps 65: "Fr", "Ud", "Lb", "Lp",
66: "Lk", "Mt", "Brq", "Bro",
1.64 kristaps 67: /* LINTED */
1.114 kristaps 68: "Brc", "%C", "Es", "En",
1.69 kristaps 69: /* LINTED */
1.114 kristaps 70: "Dx", "%Q", "br", "sp",
1.110 kristaps 71: /* LINTED */
1.141 kristaps 72: "%U", "Ta"
1.1 kristaps 73: };
74:
75: const char *const __mdoc_argnames[MDOC_ARG_MAX] = {
76: "split", "nosplit", "ragged",
77: "unfilled", "literal", "file",
78: "offset", "bullet", "dash",
79: "hyphen", "item", "enum",
80: "tag", "diag", "hang",
81: "ohang", "inset", "column",
82: "width", "compact", "std",
1.52 kristaps 83: "filled", "words", "emphasis",
1.108 kristaps 84: "symbolic", "nested", "centered"
1.1 kristaps 85: };
86:
87: const char * const *mdoc_macronames = __mdoc_macronames;
88: const char * const *mdoc_argnames = __mdoc_argnames;
89:
1.121 kristaps 90: static void mdoc_node_free(struct mdoc_node *);
91: static void mdoc_node_unlink(struct mdoc *,
92: struct mdoc_node *);
1.73 kristaps 93: static void mdoc_free1(struct mdoc *);
1.113 kristaps 94: static void mdoc_alloc1(struct mdoc *);
1.73 kristaps 95: static struct mdoc_node *node_alloc(struct mdoc *, int, int,
1.117 kristaps 96: enum mdoct, enum mdoc_type);
1.73 kristaps 97: static int node_append(struct mdoc *,
1.71 kristaps 98: struct mdoc_node *);
1.149 kristaps 99: static int mdoc_ptext(struct mdoc *, int, char *, int);
100: static int mdoc_pmacro(struct mdoc *, int, char *, int);
1.124 kristaps 101:
1.88 kristaps 102:
1.1 kristaps 103: const struct mdoc_node *
1.71 kristaps 104: mdoc_node(const struct mdoc *m)
1.1 kristaps 105: {
106:
1.71 kristaps 107: return(MDOC_HALT & m->flags ? NULL : m->first);
1.1 kristaps 108: }
109:
110:
1.37 kristaps 111: const struct mdoc_meta *
1.71 kristaps 112: mdoc_meta(const struct mdoc *m)
1.37 kristaps 113: {
114:
1.71 kristaps 115: return(MDOC_HALT & m->flags ? NULL : &m->meta);
1.37 kristaps 116: }
117:
118:
1.85 kristaps 119: /*
120: * Frees volatile resources (parse tree, meta-data, fields).
121: */
1.73 kristaps 122: static void
123: mdoc_free1(struct mdoc *mdoc)
1.67 kristaps 124: {
125:
126: if (mdoc->first)
1.121 kristaps 127: mdoc_node_delete(mdoc, mdoc->first);
1.67 kristaps 128: if (mdoc->meta.title)
129: free(mdoc->meta.title);
130: if (mdoc->meta.os)
131: free(mdoc->meta.os);
132: if (mdoc->meta.name)
133: free(mdoc->meta.name);
134: if (mdoc->meta.arch)
135: free(mdoc->meta.arch);
136: if (mdoc->meta.vol)
137: free(mdoc->meta.vol);
1.133 kristaps 138: if (mdoc->meta.msec)
139: free(mdoc->meta.msec);
1.73 kristaps 140: }
141:
142:
1.85 kristaps 143: /*
144: * Allocate all volatile resources (parse tree, meta-data, fields).
145: */
1.113 kristaps 146: static void
1.73 kristaps 147: mdoc_alloc1(struct mdoc *mdoc)
148: {
1.67 kristaps 149:
1.112 kristaps 150: memset(&mdoc->meta, 0, sizeof(struct mdoc_meta));
1.67 kristaps 151: mdoc->flags = 0;
1.85 kristaps 152: mdoc->lastnamed = mdoc->lastsec = SEC_NONE;
1.113 kristaps 153: mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node));
1.70 kristaps 154: mdoc->first = mdoc->last;
1.67 kristaps 155: mdoc->last->type = MDOC_ROOT;
156: mdoc->next = MDOC_NEXT_CHILD;
1.73 kristaps 157: }
158:
159:
160: /*
1.85 kristaps 161: * Free up volatile resources (see mdoc_free1()) then re-initialises the
162: * data with mdoc_alloc1(). After invocation, parse data has been reset
163: * and the parser is ready for re-invocation on a new tree; however,
164: * cross-parse non-volatile data is kept intact.
1.73 kristaps 165: */
1.113 kristaps 166: void
1.73 kristaps 167: mdoc_reset(struct mdoc *mdoc)
168: {
169:
170: mdoc_free1(mdoc);
1.113 kristaps 171: mdoc_alloc1(mdoc);
1.67 kristaps 172: }
173:
174:
1.68 kristaps 175: /*
1.85 kristaps 176: * Completely free up all volatile and non-volatile parse resources.
177: * After invocation, the pointer is no longer usable.
1.68 kristaps 178: */
1.67 kristaps 179: void
1.38 kristaps 180: mdoc_free(struct mdoc *mdoc)
1.34 kristaps 181: {
182:
1.73 kristaps 183: mdoc_free1(mdoc);
1.1 kristaps 184: free(mdoc);
185: }
186:
187:
1.85 kristaps 188: /*
189: * Allocate volatile and non-volatile parse resources.
190: */
1.1 kristaps 191: struct mdoc *
1.163 schwarze 192: mdoc_alloc(struct regset *regs, void *data, mandocmsg msg)
1.1 kristaps 193: {
194: struct mdoc *p;
195:
1.113 kristaps 196: p = mandoc_calloc(1, sizeof(struct mdoc));
197:
1.136 kristaps 198: p->msg = msg;
1.1 kristaps 199: p->data = data;
1.149 kristaps 200: p->regs = regs;
1.73 kristaps 201:
1.113 kristaps 202: mdoc_hash_init();
203: mdoc_alloc1(p);
204: return(p);
1.1 kristaps 205: }
206:
207:
1.68 kristaps 208: /*
209: * Climb back up the parse tree, validating open scopes. Mostly calls
1.85 kristaps 210: * through to macro_end() in macro.c.
1.68 kristaps 211: */
1.1 kristaps 212: int
1.72 kristaps 213: mdoc_endparse(struct mdoc *m)
1.20 kristaps 214: {
215:
1.72 kristaps 216: if (MDOC_HALT & m->flags)
1.20 kristaps 217: return(0);
1.72 kristaps 218: else if (mdoc_macroend(m))
1.20 kristaps 219: return(1);
1.72 kristaps 220: m->flags |= MDOC_HALT;
221: return(0);
1.20 kristaps 222: }
223:
224:
1.50 kristaps 225: /*
1.53 kristaps 226: * Main parse routine. Parses a single line -- really just hands off to
1.123 kristaps 227: * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
1.50 kristaps 228: */
1.20 kristaps 229: int
1.149 kristaps 230: mdoc_parseln(struct mdoc *m, int ln, char *buf, int offs)
1.1 kristaps 231: {
232:
1.53 kristaps 233: if (MDOC_HALT & m->flags)
1.20 kristaps 234: return(0);
1.50 kristaps 235:
1.130 kristaps 236: m->flags |= MDOC_NEWLINE;
1.153 schwarze 237:
238: /*
239: * Let the roff nS register switch SYNOPSIS mode early,
240: * such that the parser knows at all times
241: * whether this mode is on or off.
242: * Note that this mode is also switched by the Sh macro.
243: */
244: if (m->regs->regs[(int)REG_nS].set) {
245: if (m->regs->regs[(int)REG_nS].v.u)
246: m->flags |= MDOC_SYNOPSIS;
247: else
248: m->flags &= ~MDOC_SYNOPSIS;
249: }
250:
1.135 kristaps 251: return(('.' == buf[offs] || '\'' == buf[offs]) ?
1.149 kristaps 252: mdoc_pmacro(m, ln, buf, offs) :
253: mdoc_ptext(m, ln, buf, offs));
1.1 kristaps 254: }
255:
256:
257: int
1.136 kristaps 258: mdoc_vmsg(struct mdoc *mdoc, enum mandocerr t,
259: int ln, int pos, const char *fmt, ...)
1.1 kristaps 260: {
1.31 kristaps 261: char buf[256];
262: va_list ap;
1.1 kristaps 263:
1.31 kristaps 264: va_start(ap, fmt);
1.136 kristaps 265: vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
1.31 kristaps 266: va_end(ap);
1.88 kristaps 267:
1.136 kristaps 268: return((*mdoc->msg)(t, mdoc->data, ln, pos, buf));
1.88 kristaps 269: }
270:
271:
272: int
1.148 kristaps 273: mdoc_macro(MACRO_PROT_ARGS)
1.88 kristaps 274: {
1.122 kristaps 275: assert(tok < MDOC_MAX);
276:
277: /* If we're in the body, deny prologue calls. */
1.117 kristaps 278:
1.88 kristaps 279: if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
1.122 kristaps 280: MDOC_PBODY & m->flags)
1.148 kristaps 281: return(mdoc_pmsg(m, line, ppos, MANDOCERR_BADBODY));
1.122 kristaps 282:
283: /* If we're in the prologue, deny "body" macros. */
284:
285: if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
286: ! (MDOC_PBODY & m->flags)) {
1.148 kristaps 287: if ( ! mdoc_pmsg(m, line, ppos, MANDOCERR_BADPROLOG))
1.120 kristaps 288: return(0);
289: if (NULL == m->meta.title)
1.140 kristaps 290: m->meta.title = mandoc_strdup("UNKNOWN");
1.120 kristaps 291: if (NULL == m->meta.vol)
1.140 kristaps 292: m->meta.vol = mandoc_strdup("LOCAL");
1.120 kristaps 293: if (NULL == m->meta.os)
1.140 kristaps 294: m->meta.os = mandoc_strdup("LOCAL");
1.120 kristaps 295: if (0 == m->meta.date)
296: m->meta.date = time(NULL);
297: m->flags |= MDOC_PBODY;
298: }
1.88 kristaps 299:
1.149 kristaps 300: return((*mdoc_macros[tok].fp)(m, tok, line, ppos, pos, buf));
1.73 kristaps 301: }
302:
303:
304: static int
305: node_append(struct mdoc *mdoc, struct mdoc_node *p)
1.1 kristaps 306: {
307:
1.25 kristaps 308: assert(mdoc->last);
309: assert(mdoc->first);
310: assert(MDOC_ROOT != p->type);
1.1 kristaps 311:
1.13 kristaps 312: switch (mdoc->next) {
313: case (MDOC_NEXT_SIBLING):
1.6 kristaps 314: mdoc->last->next = p;
315: p->prev = mdoc->last;
1.13 kristaps 316: p->parent = mdoc->last->parent;
1.1 kristaps 317: break;
1.13 kristaps 318: case (MDOC_NEXT_CHILD):
319: mdoc->last->child = p;
1.1 kristaps 320: p->parent = mdoc->last;
321: break;
322: default:
1.13 kristaps 323: abort();
324: /* NOTREACHED */
1.1 kristaps 325: }
326:
1.86 kristaps 327: p->parent->nchild++;
328:
1.23 kristaps 329: if ( ! mdoc_valid_pre(mdoc, p))
330: return(0);
1.68 kristaps 331: if ( ! mdoc_action_pre(mdoc, p))
332: return(0);
1.27 kristaps 333:
334: switch (p->type) {
335: case (MDOC_HEAD):
336: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 337: p->parent->head = p;
1.27 kristaps 338: break;
339: case (MDOC_TAIL):
340: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 341: p->parent->tail = p;
1.27 kristaps 342: break;
343: case (MDOC_BODY):
1.152 schwarze 344: if (p->end)
345: break;
1.27 kristaps 346: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 347: p->parent->body = p;
1.27 kristaps 348: break;
349: default:
350: break;
351: }
352:
1.1 kristaps 353: mdoc->last = p;
1.71 kristaps 354:
355: switch (p->type) {
356: case (MDOC_TEXT):
357: if ( ! mdoc_valid_post(mdoc))
358: return(0);
359: if ( ! mdoc_action_post(mdoc))
360: return(0);
361: break;
362: default:
363: break;
364: }
365:
1.23 kristaps 366: return(1);
1.1 kristaps 367: }
368:
369:
1.45 kristaps 370: static struct mdoc_node *
1.117 kristaps 371: node_alloc(struct mdoc *m, int line, int pos,
372: enum mdoct tok, enum mdoc_type type)
1.45 kristaps 373: {
1.46 kristaps 374: struct mdoc_node *p;
375:
1.113 kristaps 376: p = mandoc_calloc(1, sizeof(struct mdoc_node));
1.91 kristaps 377: p->sec = m->lastsec;
1.73 kristaps 378: p->line = line;
379: p->pos = pos;
380: p->tok = tok;
1.118 kristaps 381: p->type = type;
1.150 kristaps 382:
383: /* Flag analysis. */
384:
1.153 schwarze 385: if (MDOC_SYNOPSIS & m->flags)
386: p->flags |= MDOC_SYNPRETTY;
387: else
388: p->flags &= ~MDOC_SYNPRETTY;
1.130 kristaps 389: if (MDOC_NEWLINE & m->flags)
390: p->flags |= MDOC_LINE;
391: m->flags &= ~MDOC_NEWLINE;
1.150 kristaps 392:
1.46 kristaps 393: return(p);
1.45 kristaps 394: }
395:
396:
1.23 kristaps 397: int
1.117 kristaps 398: mdoc_tail_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
1.17 kristaps 399: {
400: struct mdoc_node *p;
401:
1.91 kristaps 402: p = node_alloc(m, line, pos, tok, MDOC_TAIL);
1.102 kristaps 403: if ( ! node_append(m, p))
404: return(0);
405: m->next = MDOC_NEXT_CHILD;
406: return(1);
1.17 kristaps 407: }
408:
409:
1.23 kristaps 410: int
1.117 kristaps 411: mdoc_head_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
1.1 kristaps 412: {
413: struct mdoc_node *p;
414:
1.91 kristaps 415: assert(m->first);
416: assert(m->last);
1.1 kristaps 417:
1.91 kristaps 418: p = node_alloc(m, line, pos, tok, MDOC_HEAD);
1.102 kristaps 419: if ( ! node_append(m, p))
420: return(0);
421: m->next = MDOC_NEXT_CHILD;
422: return(1);
1.1 kristaps 423: }
424:
425:
1.23 kristaps 426: int
1.117 kristaps 427: mdoc_body_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
1.1 kristaps 428: {
429: struct mdoc_node *p;
430:
1.91 kristaps 431: p = node_alloc(m, line, pos, tok, MDOC_BODY);
1.102 kristaps 432: if ( ! node_append(m, p))
433: return(0);
434: m->next = MDOC_NEXT_CHILD;
1.152 schwarze 435: return(1);
436: }
437:
438:
439: int
440: mdoc_endbody_alloc(struct mdoc *m, int line, int pos, enum mdoct tok,
441: struct mdoc_node *body, enum mdoc_endbody end)
442: {
443: struct mdoc_node *p;
444:
445: p = node_alloc(m, line, pos, tok, MDOC_BODY);
446: p->pending = body;
447: p->end = end;
448: if ( ! node_append(m, p))
449: return(0);
450: m->next = MDOC_NEXT_SIBLING;
1.102 kristaps 451: return(1);
1.1 kristaps 452: }
453:
454:
1.23 kristaps 455: int
1.91 kristaps 456: mdoc_block_alloc(struct mdoc *m, int line, int pos,
1.117 kristaps 457: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 458: {
459: struct mdoc_node *p;
460:
1.91 kristaps 461: p = node_alloc(m, line, pos, tok, MDOC_BLOCK);
1.77 kristaps 462: p->args = args;
463: if (p->args)
1.53 kristaps 464: (args->refcnt)++;
1.102 kristaps 465: if ( ! node_append(m, p))
466: return(0);
467: m->next = MDOC_NEXT_CHILD;
468: return(1);
1.1 kristaps 469: }
470:
471:
1.23 kristaps 472: int
1.91 kristaps 473: mdoc_elem_alloc(struct mdoc *m, int line, int pos,
1.117 kristaps 474: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 475: {
476: struct mdoc_node *p;
477:
1.91 kristaps 478: p = node_alloc(m, line, pos, tok, MDOC_ELEM);
1.77 kristaps 479: p->args = args;
480: if (p->args)
1.53 kristaps 481: (args->refcnt)++;
1.102 kristaps 482: if ( ! node_append(m, p))
483: return(0);
484: m->next = MDOC_NEXT_CHILD;
485: return(1);
1.1 kristaps 486: }
487:
488:
1.124 kristaps 489: int
490: mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p)
1.1 kristaps 491: {
1.91 kristaps 492: struct mdoc_node *n;
1.124 kristaps 493: size_t sv, len;
494:
495: len = strlen(p);
1.1 kristaps 496:
1.125 kristaps 497: n = node_alloc(m, line, pos, MDOC_MAX, MDOC_TEXT);
1.113 kristaps 498: n->string = mandoc_malloc(len + 1);
1.91 kristaps 499: sv = strlcpy(n->string, p, len + 1);
500:
501: /* Prohibit truncation. */
502: assert(sv < len + 1);
503:
1.101 kristaps 504: if ( ! node_append(m, n))
505: return(0);
1.124 kristaps 506:
1.101 kristaps 507: m->next = MDOC_NEXT_SIBLING;
508: return(1);
1.91 kristaps 509: }
510:
511:
1.155 kristaps 512: static void
1.53 kristaps 513: mdoc_node_free(struct mdoc_node *p)
1.1 kristaps 514: {
515:
1.156 kristaps 516: /*
517: * XXX: if these end up being problematic in terms of memory
518: * management and dereferencing freed blocks, then make them
519: * into reference-counted double-pointers.
520: */
521:
1.154 kristaps 522: if (MDOC_Bd == p->tok && MDOC_BLOCK == p->type)
523: if (p->data.Bd)
524: free(p->data.Bd);
525: if (MDOC_Bl == p->tok && MDOC_BLOCK == p->type)
526: if (p->data.Bl)
527: free(p->data.Bl);
1.156 kristaps 528: if (MDOC_Bf == p->tok && MDOC_HEAD == p->type)
529: if (p->data.Bf)
530: free(p->data.Bf);
1.154 kristaps 531:
1.53 kristaps 532: if (p->string)
533: free(p->string);
534: if (p->args)
535: mdoc_argv_free(p->args);
1.1 kristaps 536: free(p);
537: }
538:
539:
1.121 kristaps 540: static void
541: mdoc_node_unlink(struct mdoc *m, struct mdoc_node *n)
542: {
543:
544: /* Adjust siblings. */
545:
546: if (n->prev)
547: n->prev->next = n->next;
548: if (n->next)
549: n->next->prev = n->prev;
550:
551: /* Adjust parent. */
552:
553: if (n->parent) {
554: n->parent->nchild--;
555: if (n->parent->child == n)
556: n->parent->child = n->prev ? n->prev : n->next;
557: }
558:
559: /* Adjust parse point, if applicable. */
560:
561: if (m && m->last == n) {
562: if (n->prev) {
563: m->last = n->prev;
564: m->next = MDOC_NEXT_SIBLING;
565: } else {
566: m->last = n->parent;
567: m->next = MDOC_NEXT_CHILD;
568: }
569: }
570:
571: if (m && m->first == n)
572: m->first = NULL;
573: }
574:
575:
1.53 kristaps 576: void
1.121 kristaps 577: mdoc_node_delete(struct mdoc *m, struct mdoc_node *p)
1.1 kristaps 578: {
579:
1.121 kristaps 580: while (p->child) {
581: assert(p->nchild);
582: mdoc_node_delete(m, p->child);
583: }
584: assert(0 == p->nchild);
1.1 kristaps 585:
1.121 kristaps 586: mdoc_node_unlink(m, p);
1.53 kristaps 587: mdoc_node_free(p);
1.1 kristaps 588: }
589:
590:
1.53 kristaps 591: /*
592: * Parse free-form text, that is, a line that does not begin with the
593: * control character.
594: */
595: static int
1.149 kristaps 596: mdoc_ptext(struct mdoc *m, int line, char *buf, int offs)
1.1 kristaps 597: {
1.142 kristaps 598: char *c, *ws, *end;
599: struct mdoc_node *n;
1.1 kristaps 600:
1.123 kristaps 601: /* Ignore bogus comments. */
602:
1.135 kristaps 603: if ('\\' == buf[offs] &&
604: '.' == buf[offs + 1] &&
605: '"' == buf[offs + 2])
1.136 kristaps 606: return(mdoc_pmsg(m, line, offs, MANDOCERR_BADCOMMENT));
1.123 kristaps 607:
1.124 kristaps 608: /* No text before an initial macro. */
609:
1.85 kristaps 610: if (SEC_NONE == m->lastnamed)
1.136 kristaps 611: return(mdoc_pmsg(m, line, offs, MANDOCERR_NOTEXT));
1.142 kristaps 612:
613: assert(m->last);
614: n = m->last;
615:
616: /*
1.144 kristaps 617: * Divert directly to list processing if we're encountering a
1.142 kristaps 618: * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
1.144 kristaps 619: * (a MDOC_BODY means it's already open, in which case we should
620: * process within its context in the normal way).
1.142 kristaps 621: */
622:
1.143 kristaps 623: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.154 kristaps 624: LIST_column == n->data.Bl->type) {
1.144 kristaps 625: /* `Bl' is open without any children. */
1.142 kristaps 626: m->flags |= MDOC_FREECOL;
1.149 kristaps 627: return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
1.142 kristaps 628: }
629:
630: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
631: NULL != n->parent &&
632: MDOC_Bl == n->parent->tok &&
1.154 kristaps 633: LIST_column == n->parent->data.Bl->type) {
1.144 kristaps 634: /* `Bl' has block-level `It' children. */
1.142 kristaps 635: m->flags |= MDOC_FREECOL;
1.149 kristaps 636: return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
1.142 kristaps 637: }
1.124 kristaps 638:
1.137 schwarze 639: /*
640: * Search for the beginning of unescaped trailing whitespace (ws)
641: * and for the first character not to be output (end).
642: */
1.139 kristaps 643:
644: /* FIXME: replace with strcspn(). */
1.137 schwarze 645: ws = NULL;
646: for (c = end = buf + offs; *c; c++) {
647: switch (*c) {
1.138 kristaps 648: case '-':
649: if (mandoc_hyph(buf + offs, c))
650: *c = ASCII_HYPH;
1.145 kristaps 651: ws = NULL;
1.138 kristaps 652: break;
1.137 schwarze 653: case ' ':
654: if (NULL == ws)
655: ws = c;
656: continue;
657: case '\t':
658: /*
659: * Always warn about trailing tabs,
660: * even outside literal context,
661: * where they should be put on the next line.
662: */
663: if (NULL == ws)
664: ws = c;
665: /*
666: * Strip trailing tabs in literal context only;
667: * outside, they affect the next line.
668: */
669: if (MDOC_LITERAL & m->flags)
670: continue;
671: break;
672: case '\\':
673: /* Skip the escaped character, too, if any. */
674: if (c[1])
675: c++;
676: /* FALLTHROUGH */
677: default:
678: ws = NULL;
679: break;
680: }
681: end = c + 1;
682: }
683: *end = '\0';
1.91 kristaps 684:
1.137 schwarze 685: if (ws)
686: if ( ! mdoc_pmsg(m, line, (int)(ws-buf), MANDOCERR_EOLNSPACE))
687: return(0);
1.115 kristaps 688:
1.137 schwarze 689: if ('\0' == buf[offs] && ! (MDOC_LITERAL & m->flags)) {
690: if ( ! mdoc_pmsg(m, line, (int)(c-buf), MANDOCERR_NOBLANKLN))
1.119 kristaps 691: return(0);
1.124 kristaps 692:
1.119 kristaps 693: /*
1.124 kristaps 694: * Insert a `Pp' in the case of a blank line. Technically,
695: * blank lines aren't allowed, but enough manuals assume this
696: * behaviour that we want to work around it.
1.119 kristaps 697: */
1.135 kristaps 698: if ( ! mdoc_elem_alloc(m, line, offs, MDOC_Pp, NULL))
1.119 kristaps 699: return(0);
1.124 kristaps 700:
701: m->next = MDOC_NEXT_SIBLING;
702: return(1);
1.119 kristaps 703: }
1.68 kristaps 704:
1.137 schwarze 705: if ( ! mdoc_word_alloc(m, line, offs, buf+offs))
706: return(0);
1.91 kristaps 707:
1.137 schwarze 708: if (MDOC_LITERAL & m->flags)
709: return(1);
1.128 kristaps 710:
711: /*
712: * End-of-sentence check. If the last character is an unescaped
713: * EOS character, then flag the node as being the end of a
714: * sentence. The front-end will know how to interpret this.
715: */
1.132 kristaps 716:
1.137 schwarze 717: assert(buf < end);
718:
1.159 schwarze 719: if (mandoc_eos(buf+offs, (size_t)(end-buf-offs), 0))
1.128 kristaps 720: m->last->flags |= MDOC_EOS;
721:
722: return(1);
1.1 kristaps 723: }
724:
725:
1.53 kristaps 726: /*
727: * Parse a macro line, that is, a line beginning with the control
728: * character.
729: */
1.155 kristaps 730: static int
1.149 kristaps 731: mdoc_pmacro(struct mdoc *m, int ln, char *buf, int offs)
1.1 kristaps 732: {
1.144 kristaps 733: enum mdoct tok;
734: int i, j, sv;
735: char mac[5];
736: struct mdoc_node *n;
1.1 kristaps 737:
1.81 kristaps 738: /* Empty lines are ignored. */
1.63 kristaps 739:
1.135 kristaps 740: offs++;
741:
742: if ('\0' == buf[offs])
1.63 kristaps 743: return(1);
744:
1.135 kristaps 745: i = offs;
1.100 kristaps 746:
1.160 kristaps 747: /* Accept tabs/whitespace after the initial control char. */
1.100 kristaps 748:
1.160 kristaps 749: if (' ' == buf[i] || '\t' == buf[i]) {
1.100 kristaps 750: i++;
1.160 kristaps 751: while (buf[i] && (' ' == buf[i] || '\t' == buf[i]))
1.63 kristaps 752: i++;
1.115 kristaps 753: if ('\0' == buf[i])
1.63 kristaps 754: return(1);
755: }
1.1 kristaps 756:
1.130 kristaps 757: sv = i;
758:
1.160 kristaps 759: /*
1.162 schwarze 760: * Copy the first word into a nil-terminated buffer.
1.164 ! kristaps 761: * Stop copying when a tab, space, backslash, or eoln is encountered.
1.160 kristaps 762: */
1.1 kristaps 763:
1.162 schwarze 764: j = 0;
1.164 ! kristaps 765: while (j < 4 && '\0' != buf[i] &&
! 766: ' ' != buf[i] &&
! 767: '\t' != buf[i] &&
! 768: '\\' != buf[i])
1.162 schwarze 769: mac[j++] = buf[i++];
1.135 kristaps 770: mac[j] = '\0';
1.1 kristaps 771:
1.163 schwarze 772: tok = (j > 1 || j < 4) ? mdoc_hash_find(mac) : MDOC_MAX;
773: if (MDOC_MAX == tok) {
774: mdoc_vmsg(m, MANDOCERR_MACRO, ln, sv,
775: "unknown macro: %s%s",
776: buf, strlen(buf) > 3 ? "..." : "");
1.58 kristaps 777: return(1);
1.53 kristaps 778: }
1.1 kristaps 779:
1.160 kristaps 780: /* Disregard the first trailing tab, if applicable. */
781:
782: if ('\t' == buf[i])
783: i++;
784:
785: /* Jump to the next non-whitespace word. */
1.1 kristaps 786:
1.65 kristaps 787: while (buf[i] && ' ' == buf[i])
1.53 kristaps 788: i++;
1.1 kristaps 789:
1.125 kristaps 790: /*
791: * Trailing whitespace. Note that tabs are allowed to be passed
792: * into the parser as "text", so we only warn about spaces here.
793: */
1.115 kristaps 794:
795: if ('\0' == buf[i] && ' ' == buf[i - 1])
1.136 kristaps 796: if ( ! mdoc_pmsg(m, ln, i - 1, MANDOCERR_EOLNSPACE))
1.115 kristaps 797: goto err;
798:
1.144 kristaps 799: /*
800: * If an initial macro or a list invocation, divert directly
801: * into macro processing.
802: */
803:
804: if (NULL == m->last || MDOC_It == tok || MDOC_El == tok) {
1.149 kristaps 805: if ( ! mdoc_macro(m, tok, ln, sv, &i, buf))
1.144 kristaps 806: goto err;
807: return(1);
808: }
809:
810: n = m->last;
811: assert(m->last);
812:
813: /*
814: * If the first macro of a `Bl -column', open an `It' block
815: * context around the parsed macro.
816: */
817:
818: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.154 kristaps 819: LIST_column == n->data.Bl->type) {
1.144 kristaps 820: m->flags |= MDOC_FREECOL;
1.149 kristaps 821: if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
1.144 kristaps 822: goto err;
823: return(1);
824: }
825:
826: /*
827: * If we're following a block-level `It' within a `Bl -column'
828: * context (perhaps opened in the above block or in ptext()),
829: * then open an `It' block context around the parsed macro.
1.98 kristaps 830: */
1.144 kristaps 831:
832: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
833: NULL != n->parent &&
834: MDOC_Bl == n->parent->tok &&
1.154 kristaps 835: LIST_column == n->parent->data.Bl->type) {
1.144 kristaps 836: m->flags |= MDOC_FREECOL;
1.149 kristaps 837: if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
1.144 kristaps 838: goto err;
839: return(1);
840: }
841:
842: /* Normal processing of a macro. */
843:
1.149 kristaps 844: if ( ! mdoc_macro(m, tok, ln, sv, &i, buf))
1.53 kristaps 845: goto err;
1.1 kristaps 846:
1.53 kristaps 847: return(1);
1.1 kristaps 848:
1.53 kristaps 849: err: /* Error out. */
1.1 kristaps 850:
1.53 kristaps 851: m->flags |= MDOC_HALT;
852: return(0);
1.1 kristaps 853: }
1.100 kristaps 854:
855:
CVSweb