Annotation of mandoc/mdoc.c, Revision 1.166
1.166 ! kristaps 1: /* $Id: mdoc.c,v 1.165 2010/09/27 23:03:44 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);
1.27 kristaps 331:
332: switch (p->type) {
333: case (MDOC_HEAD):
334: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 335: p->parent->head = p;
1.27 kristaps 336: break;
337: case (MDOC_TAIL):
338: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 339: p->parent->tail = p;
1.27 kristaps 340: break;
341: case (MDOC_BODY):
1.152 schwarze 342: if (p->end)
343: break;
1.27 kristaps 344: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 345: p->parent->body = p;
1.27 kristaps 346: break;
347: default:
348: break;
349: }
350:
1.1 kristaps 351: mdoc->last = p;
1.71 kristaps 352:
353: switch (p->type) {
354: case (MDOC_TEXT):
355: if ( ! mdoc_valid_post(mdoc))
356: return(0);
357: break;
358: default:
359: break;
360: }
361:
1.23 kristaps 362: return(1);
1.1 kristaps 363: }
364:
365:
1.45 kristaps 366: static struct mdoc_node *
1.117 kristaps 367: node_alloc(struct mdoc *m, int line, int pos,
368: enum mdoct tok, enum mdoc_type type)
1.45 kristaps 369: {
1.46 kristaps 370: struct mdoc_node *p;
371:
1.113 kristaps 372: p = mandoc_calloc(1, sizeof(struct mdoc_node));
1.91 kristaps 373: p->sec = m->lastsec;
1.73 kristaps 374: p->line = line;
375: p->pos = pos;
376: p->tok = tok;
1.118 kristaps 377: p->type = type;
1.150 kristaps 378:
379: /* Flag analysis. */
380:
1.153 schwarze 381: if (MDOC_SYNOPSIS & m->flags)
382: p->flags |= MDOC_SYNPRETTY;
383: else
384: p->flags &= ~MDOC_SYNPRETTY;
1.130 kristaps 385: if (MDOC_NEWLINE & m->flags)
386: p->flags |= MDOC_LINE;
387: m->flags &= ~MDOC_NEWLINE;
1.150 kristaps 388:
1.46 kristaps 389: return(p);
1.45 kristaps 390: }
391:
392:
1.23 kristaps 393: int
1.117 kristaps 394: mdoc_tail_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
1.17 kristaps 395: {
396: struct mdoc_node *p;
397:
1.91 kristaps 398: p = node_alloc(m, line, pos, tok, MDOC_TAIL);
1.102 kristaps 399: if ( ! node_append(m, p))
400: return(0);
401: m->next = MDOC_NEXT_CHILD;
402: return(1);
1.17 kristaps 403: }
404:
405:
1.23 kristaps 406: int
1.117 kristaps 407: mdoc_head_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
1.1 kristaps 408: {
409: struct mdoc_node *p;
410:
1.91 kristaps 411: assert(m->first);
412: assert(m->last);
1.1 kristaps 413:
1.91 kristaps 414: p = node_alloc(m, line, pos, tok, MDOC_HEAD);
1.102 kristaps 415: if ( ! node_append(m, p))
416: return(0);
417: m->next = MDOC_NEXT_CHILD;
418: return(1);
1.1 kristaps 419: }
420:
421:
1.23 kristaps 422: int
1.117 kristaps 423: mdoc_body_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
1.1 kristaps 424: {
425: struct mdoc_node *p;
426:
1.91 kristaps 427: p = node_alloc(m, line, pos, tok, MDOC_BODY);
1.102 kristaps 428: if ( ! node_append(m, p))
429: return(0);
430: m->next = MDOC_NEXT_CHILD;
1.152 schwarze 431: return(1);
432: }
433:
434:
435: int
436: mdoc_endbody_alloc(struct mdoc *m, int line, int pos, enum mdoct tok,
437: struct mdoc_node *body, enum mdoc_endbody end)
438: {
439: struct mdoc_node *p;
440:
441: p = node_alloc(m, line, pos, tok, MDOC_BODY);
442: p->pending = body;
443: p->end = end;
444: if ( ! node_append(m, p))
445: return(0);
446: m->next = MDOC_NEXT_SIBLING;
1.102 kristaps 447: return(1);
1.1 kristaps 448: }
449:
450:
1.23 kristaps 451: int
1.91 kristaps 452: mdoc_block_alloc(struct mdoc *m, int line, int pos,
1.117 kristaps 453: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 454: {
455: struct mdoc_node *p;
456:
1.91 kristaps 457: p = node_alloc(m, line, pos, tok, MDOC_BLOCK);
1.77 kristaps 458: p->args = args;
459: if (p->args)
1.53 kristaps 460: (args->refcnt)++;
1.102 kristaps 461: if ( ! node_append(m, p))
462: return(0);
463: m->next = MDOC_NEXT_CHILD;
464: return(1);
1.1 kristaps 465: }
466:
467:
1.23 kristaps 468: int
1.91 kristaps 469: mdoc_elem_alloc(struct mdoc *m, int line, int pos,
1.117 kristaps 470: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 471: {
472: struct mdoc_node *p;
473:
1.91 kristaps 474: p = node_alloc(m, line, pos, tok, MDOC_ELEM);
1.77 kristaps 475: p->args = args;
476: if (p->args)
1.53 kristaps 477: (args->refcnt)++;
1.102 kristaps 478: if ( ! node_append(m, p))
479: return(0);
480: m->next = MDOC_NEXT_CHILD;
481: return(1);
1.1 kristaps 482: }
483:
484:
1.124 kristaps 485: int
486: mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p)
1.1 kristaps 487: {
1.91 kristaps 488: struct mdoc_node *n;
1.124 kristaps 489: size_t sv, len;
490:
491: len = strlen(p);
1.1 kristaps 492:
1.125 kristaps 493: n = node_alloc(m, line, pos, MDOC_MAX, MDOC_TEXT);
1.113 kristaps 494: n->string = mandoc_malloc(len + 1);
1.91 kristaps 495: sv = strlcpy(n->string, p, len + 1);
496:
497: /* Prohibit truncation. */
498: assert(sv < len + 1);
499:
1.101 kristaps 500: if ( ! node_append(m, n))
501: return(0);
1.124 kristaps 502:
1.101 kristaps 503: m->next = MDOC_NEXT_SIBLING;
504: return(1);
1.91 kristaps 505: }
506:
507:
1.155 kristaps 508: static void
1.53 kristaps 509: mdoc_node_free(struct mdoc_node *p)
1.1 kristaps 510: {
511:
1.156 kristaps 512: /*
513: * XXX: if these end up being problematic in terms of memory
514: * management and dereferencing freed blocks, then make them
515: * into reference-counted double-pointers.
516: */
517:
1.154 kristaps 518: if (MDOC_Bd == p->tok && MDOC_BLOCK == p->type)
519: if (p->data.Bd)
520: free(p->data.Bd);
521: if (MDOC_Bl == p->tok && MDOC_BLOCK == p->type)
522: if (p->data.Bl)
523: free(p->data.Bl);
1.156 kristaps 524: if (MDOC_Bf == p->tok && MDOC_HEAD == p->type)
525: if (p->data.Bf)
526: free(p->data.Bf);
1.154 kristaps 527:
1.53 kristaps 528: if (p->string)
529: free(p->string);
530: if (p->args)
531: mdoc_argv_free(p->args);
1.1 kristaps 532: free(p);
533: }
534:
535:
1.121 kristaps 536: static void
537: mdoc_node_unlink(struct mdoc *m, struct mdoc_node *n)
538: {
539:
540: /* Adjust siblings. */
541:
542: if (n->prev)
543: n->prev->next = n->next;
544: if (n->next)
545: n->next->prev = n->prev;
546:
547: /* Adjust parent. */
548:
549: if (n->parent) {
550: n->parent->nchild--;
551: if (n->parent->child == n)
552: n->parent->child = n->prev ? n->prev : n->next;
553: }
554:
555: /* Adjust parse point, if applicable. */
556:
557: if (m && m->last == n) {
558: if (n->prev) {
559: m->last = n->prev;
560: m->next = MDOC_NEXT_SIBLING;
561: } else {
562: m->last = n->parent;
563: m->next = MDOC_NEXT_CHILD;
564: }
565: }
566:
567: if (m && m->first == n)
568: m->first = NULL;
569: }
570:
571:
1.53 kristaps 572: void
1.121 kristaps 573: mdoc_node_delete(struct mdoc *m, struct mdoc_node *p)
1.1 kristaps 574: {
575:
1.121 kristaps 576: while (p->child) {
577: assert(p->nchild);
578: mdoc_node_delete(m, p->child);
579: }
580: assert(0 == p->nchild);
1.1 kristaps 581:
1.121 kristaps 582: mdoc_node_unlink(m, p);
1.53 kristaps 583: mdoc_node_free(p);
1.1 kristaps 584: }
585:
586:
1.53 kristaps 587: /*
588: * Parse free-form text, that is, a line that does not begin with the
589: * control character.
590: */
591: static int
1.149 kristaps 592: mdoc_ptext(struct mdoc *m, int line, char *buf, int offs)
1.1 kristaps 593: {
1.142 kristaps 594: char *c, *ws, *end;
595: struct mdoc_node *n;
1.1 kristaps 596:
1.123 kristaps 597: /* Ignore bogus comments. */
598:
1.135 kristaps 599: if ('\\' == buf[offs] &&
600: '.' == buf[offs + 1] &&
601: '"' == buf[offs + 2])
1.136 kristaps 602: return(mdoc_pmsg(m, line, offs, MANDOCERR_BADCOMMENT));
1.123 kristaps 603:
1.124 kristaps 604: /* No text before an initial macro. */
605:
1.85 kristaps 606: if (SEC_NONE == m->lastnamed)
1.136 kristaps 607: return(mdoc_pmsg(m, line, offs, MANDOCERR_NOTEXT));
1.142 kristaps 608:
609: assert(m->last);
610: n = m->last;
611:
612: /*
1.144 kristaps 613: * Divert directly to list processing if we're encountering a
1.142 kristaps 614: * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
1.144 kristaps 615: * (a MDOC_BODY means it's already open, in which case we should
616: * process within its context in the normal way).
1.142 kristaps 617: */
618:
1.143 kristaps 619: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.154 kristaps 620: LIST_column == n->data.Bl->type) {
1.144 kristaps 621: /* `Bl' is open without any children. */
1.142 kristaps 622: m->flags |= MDOC_FREECOL;
1.149 kristaps 623: return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
1.142 kristaps 624: }
625:
626: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
627: NULL != n->parent &&
628: MDOC_Bl == n->parent->tok &&
1.154 kristaps 629: LIST_column == n->parent->data.Bl->type) {
1.144 kristaps 630: /* `Bl' has block-level `It' children. */
1.142 kristaps 631: m->flags |= MDOC_FREECOL;
1.149 kristaps 632: return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
1.142 kristaps 633: }
1.124 kristaps 634:
1.137 schwarze 635: /*
636: * Search for the beginning of unescaped trailing whitespace (ws)
637: * and for the first character not to be output (end).
638: */
1.139 kristaps 639:
640: /* FIXME: replace with strcspn(). */
1.137 schwarze 641: ws = NULL;
642: for (c = end = buf + offs; *c; c++) {
643: switch (*c) {
1.138 kristaps 644: case '-':
645: if (mandoc_hyph(buf + offs, c))
646: *c = ASCII_HYPH;
1.145 kristaps 647: ws = NULL;
1.138 kristaps 648: break;
1.137 schwarze 649: case ' ':
650: if (NULL == ws)
651: ws = c;
652: continue;
653: case '\t':
654: /*
655: * Always warn about trailing tabs,
656: * even outside literal context,
657: * where they should be put on the next line.
658: */
659: if (NULL == ws)
660: ws = c;
661: /*
662: * Strip trailing tabs in literal context only;
663: * outside, they affect the next line.
664: */
665: if (MDOC_LITERAL & m->flags)
666: continue;
667: break;
668: case '\\':
669: /* Skip the escaped character, too, if any. */
670: if (c[1])
671: c++;
672: /* FALLTHROUGH */
673: default:
674: ws = NULL;
675: break;
676: }
677: end = c + 1;
678: }
679: *end = '\0';
1.91 kristaps 680:
1.137 schwarze 681: if (ws)
682: if ( ! mdoc_pmsg(m, line, (int)(ws-buf), MANDOCERR_EOLNSPACE))
683: return(0);
1.115 kristaps 684:
1.137 schwarze 685: if ('\0' == buf[offs] && ! (MDOC_LITERAL & m->flags)) {
686: if ( ! mdoc_pmsg(m, line, (int)(c-buf), MANDOCERR_NOBLANKLN))
1.119 kristaps 687: return(0);
1.124 kristaps 688:
1.119 kristaps 689: /*
1.165 schwarze 690: * Insert a `sp' in the case of a blank line. Technically,
1.124 kristaps 691: * blank lines aren't allowed, but enough manuals assume this
692: * behaviour that we want to work around it.
1.119 kristaps 693: */
1.165 schwarze 694: if ( ! mdoc_elem_alloc(m, line, offs, MDOC_sp, NULL))
1.119 kristaps 695: return(0);
1.124 kristaps 696:
697: m->next = MDOC_NEXT_SIBLING;
698: return(1);
1.119 kristaps 699: }
1.68 kristaps 700:
1.137 schwarze 701: if ( ! mdoc_word_alloc(m, line, offs, buf+offs))
702: return(0);
1.91 kristaps 703:
1.137 schwarze 704: if (MDOC_LITERAL & m->flags)
705: return(1);
1.128 kristaps 706:
707: /*
708: * End-of-sentence check. If the last character is an unescaped
709: * EOS character, then flag the node as being the end of a
710: * sentence. The front-end will know how to interpret this.
711: */
1.132 kristaps 712:
1.137 schwarze 713: assert(buf < end);
714:
1.159 schwarze 715: if (mandoc_eos(buf+offs, (size_t)(end-buf-offs), 0))
1.128 kristaps 716: m->last->flags |= MDOC_EOS;
717:
718: return(1);
1.1 kristaps 719: }
720:
721:
1.53 kristaps 722: /*
723: * Parse a macro line, that is, a line beginning with the control
724: * character.
725: */
1.155 kristaps 726: static int
1.149 kristaps 727: mdoc_pmacro(struct mdoc *m, int ln, char *buf, int offs)
1.1 kristaps 728: {
1.144 kristaps 729: enum mdoct tok;
730: int i, j, sv;
731: char mac[5];
732: struct mdoc_node *n;
1.1 kristaps 733:
1.81 kristaps 734: /* Empty lines are ignored. */
1.63 kristaps 735:
1.135 kristaps 736: offs++;
737:
738: if ('\0' == buf[offs])
1.63 kristaps 739: return(1);
740:
1.135 kristaps 741: i = offs;
1.100 kristaps 742:
1.160 kristaps 743: /* Accept tabs/whitespace after the initial control char. */
1.100 kristaps 744:
1.160 kristaps 745: if (' ' == buf[i] || '\t' == buf[i]) {
1.100 kristaps 746: i++;
1.160 kristaps 747: while (buf[i] && (' ' == buf[i] || '\t' == buf[i]))
1.63 kristaps 748: i++;
1.115 kristaps 749: if ('\0' == buf[i])
1.63 kristaps 750: return(1);
751: }
1.1 kristaps 752:
1.130 kristaps 753: sv = i;
754:
1.160 kristaps 755: /*
1.162 schwarze 756: * Copy the first word into a nil-terminated buffer.
1.165 schwarze 757: * Stop copying when a tab, space, or eoln is encountered.
1.160 kristaps 758: */
1.1 kristaps 759:
1.162 schwarze 760: j = 0;
1.165 schwarze 761: while (j < 4 && '\0' != buf[i] && ' ' != buf[i] && '\t' != buf[i])
1.162 schwarze 762: mac[j++] = buf[i++];
1.135 kristaps 763: mac[j] = '\0';
1.1 kristaps 764:
1.163 schwarze 765: tok = (j > 1 || j < 4) ? mdoc_hash_find(mac) : MDOC_MAX;
766: if (MDOC_MAX == tok) {
767: mdoc_vmsg(m, MANDOCERR_MACRO, ln, sv,
768: "unknown macro: %s%s",
769: buf, strlen(buf) > 3 ? "..." : "");
1.58 kristaps 770: return(1);
1.53 kristaps 771: }
1.1 kristaps 772:
1.160 kristaps 773: /* Disregard the first trailing tab, if applicable. */
774:
775: if ('\t' == buf[i])
776: i++;
777:
778: /* Jump to the next non-whitespace word. */
1.1 kristaps 779:
1.65 kristaps 780: while (buf[i] && ' ' == buf[i])
1.53 kristaps 781: i++;
1.1 kristaps 782:
1.125 kristaps 783: /*
784: * Trailing whitespace. Note that tabs are allowed to be passed
785: * into the parser as "text", so we only warn about spaces here.
786: */
1.115 kristaps 787:
788: if ('\0' == buf[i] && ' ' == buf[i - 1])
1.136 kristaps 789: if ( ! mdoc_pmsg(m, ln, i - 1, MANDOCERR_EOLNSPACE))
1.115 kristaps 790: goto err;
791:
1.144 kristaps 792: /*
793: * If an initial macro or a list invocation, divert directly
794: * into macro processing.
795: */
796:
797: if (NULL == m->last || MDOC_It == tok || MDOC_El == tok) {
1.149 kristaps 798: if ( ! mdoc_macro(m, tok, ln, sv, &i, buf))
1.144 kristaps 799: goto err;
800: return(1);
801: }
802:
803: n = m->last;
804: assert(m->last);
805:
806: /*
807: * If the first macro of a `Bl -column', open an `It' block
808: * context around the parsed macro.
809: */
810:
811: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.154 kristaps 812: LIST_column == n->data.Bl->type) {
1.144 kristaps 813: m->flags |= MDOC_FREECOL;
1.149 kristaps 814: if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
1.144 kristaps 815: goto err;
816: return(1);
817: }
818:
819: /*
820: * If we're following a block-level `It' within a `Bl -column'
821: * context (perhaps opened in the above block or in ptext()),
822: * then open an `It' block context around the parsed macro.
1.98 kristaps 823: */
1.144 kristaps 824:
825: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
826: NULL != n->parent &&
827: MDOC_Bl == n->parent->tok &&
1.154 kristaps 828: LIST_column == n->parent->data.Bl->type) {
1.144 kristaps 829: m->flags |= MDOC_FREECOL;
1.149 kristaps 830: if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
1.144 kristaps 831: goto err;
832: return(1);
833: }
834:
835: /* Normal processing of a macro. */
836:
1.149 kristaps 837: if ( ! mdoc_macro(m, tok, ln, sv, &i, buf))
1.53 kristaps 838: goto err;
1.1 kristaps 839:
1.53 kristaps 840: return(1);
1.1 kristaps 841:
1.53 kristaps 842: err: /* Error out. */
1.1 kristaps 843:
1.53 kristaps 844: m->flags |= MDOC_HALT;
845: return(0);
1.1 kristaps 846: }
1.100 kristaps 847:
848:
CVSweb