Annotation of mandoc/mdoc.c, Revision 1.223
1.223 ! schwarze 1: /* $Id: mdoc.c,v 1.222 2014/08/01 17:27:44 schwarze Exp $ */
1.1 kristaps 2: /*
1.182 schwarze 3: * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
1.208 schwarze 4: * Copyright (c) 2010, 2012, 2013, 2014 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>
1.211 schwarze 25: #include <ctype.h>
1.1 kristaps 26: #include <stdarg.h>
1.73 kristaps 27: #include <stdio.h>
1.1 kristaps 28: #include <stdlib.h>
29: #include <string.h>
1.120 kristaps 30: #include <time.h>
1.1 kristaps 31:
1.187 kristaps 32: #include "mdoc.h"
1.136 kristaps 33: #include "mandoc.h"
1.209 schwarze 34: #include "mandoc_aux.h"
1.70 kristaps 35: #include "libmdoc.h"
1.113 kristaps 36: #include "libmandoc.h"
1.1 kristaps 37:
1.216 schwarze 38: const char *const __mdoc_macronames[MDOC_MAX + 1] = {
1.82 kristaps 39: "Ap", "Dd", "Dt", "Os",
1.1 kristaps 40: "Sh", "Ss", "Pp", "D1",
41: "Dl", "Bd", "Ed", "Bl",
42: "El", "It", "Ad", "An",
43: "Ar", "Cd", "Cm", "Dv",
44: "Er", "Ev", "Ex", "Fa",
45: "Fd", "Fl", "Fn", "Ft",
46: "Ic", "In", "Li", "Nd",
47: "Nm", "Op", "Ot", "Pa",
48: "Rv", "St", "Va", "Vt",
1.114 kristaps 49: "Xr", "%A", "%B", "%D",
50: "%I", "%J", "%N", "%O",
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",
1.114 kristaps 67: "Brc", "%C", "Es", "En",
68: "Dx", "%Q", "br", "sp",
1.216 schwarze 69: "%U", "Ta", "ll", "text",
1.1 kristaps 70: };
71:
1.213 schwarze 72: const char *const __mdoc_argnames[MDOC_ARG_MAX] = {
1.1 kristaps 73: "split", "nosplit", "ragged",
1.213 schwarze 74: "unfilled", "literal", "file",
75: "offset", "bullet", "dash",
76: "hyphen", "item", "enum",
77: "tag", "diag", "hang",
78: "ohang", "inset", "column",
79: "width", "compact", "std",
1.52 kristaps 80: "filled", "words", "emphasis",
1.108 kristaps 81: "symbolic", "nested", "centered"
1.1 kristaps 82: };
83:
84: const char * const *mdoc_macronames = __mdoc_macronames;
85: const char * const *mdoc_argnames = __mdoc_argnames;
86:
1.121 kristaps 87: static void mdoc_node_free(struct mdoc_node *);
1.213 schwarze 88: static void mdoc_node_unlink(struct mdoc *,
1.121 kristaps 89: struct mdoc_node *);
1.73 kristaps 90: static void mdoc_free1(struct mdoc *);
1.113 kristaps 91: static void mdoc_alloc1(struct mdoc *);
1.213 schwarze 92: static struct mdoc_node *node_alloc(struct mdoc *, int, int,
1.117 kristaps 93: enum mdoct, enum mdoc_type);
1.213 schwarze 94: static int node_append(struct mdoc *,
1.71 kristaps 95: struct mdoc_node *);
1.193 kristaps 96: #if 0
1.191 kristaps 97: static int mdoc_preptext(struct mdoc *, int, char *, int);
1.193 kristaps 98: #endif
1.149 kristaps 99: static int mdoc_ptext(struct mdoc *, int, char *, int);
100: static int mdoc_pmacro(struct mdoc *, int, char *, int);
1.88 kristaps 101:
1.213 schwarze 102:
1.1 kristaps 103: const struct mdoc_node *
1.203 schwarze 104: mdoc_node(const struct mdoc *mdoc)
1.1 kristaps 105: {
106:
1.203 schwarze 107: return(mdoc->first);
1.1 kristaps 108: }
109:
1.37 kristaps 110: const struct mdoc_meta *
1.203 schwarze 111: mdoc_meta(const struct mdoc *mdoc)
1.37 kristaps 112: {
113:
1.203 schwarze 114: return(&mdoc->meta);
1.37 kristaps 115: }
116:
1.85 kristaps 117: /*
118: * Frees volatile resources (parse tree, meta-data, fields).
119: */
1.73 kristaps 120: static void
121: mdoc_free1(struct mdoc *mdoc)
1.67 kristaps 122: {
123:
124: if (mdoc->first)
1.121 kristaps 125: mdoc_node_delete(mdoc, mdoc->first);
1.223 ! schwarze 126: free(mdoc->meta.msec);
! 127: free(mdoc->meta.vol);
! 128: free(mdoc->meta.arch);
! 129: free(mdoc->meta.date);
! 130: free(mdoc->meta.title);
! 131: free(mdoc->meta.os);
! 132: free(mdoc->meta.name);
1.73 kristaps 133: }
134:
1.85 kristaps 135: /*
136: * Allocate all volatile resources (parse tree, meta-data, fields).
137: */
1.113 kristaps 138: static void
1.73 kristaps 139: mdoc_alloc1(struct mdoc *mdoc)
140: {
1.67 kristaps 141:
1.112 kristaps 142: memset(&mdoc->meta, 0, sizeof(struct mdoc_meta));
1.67 kristaps 143: mdoc->flags = 0;
1.85 kristaps 144: mdoc->lastnamed = mdoc->lastsec = SEC_NONE;
1.113 kristaps 145: mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node));
1.70 kristaps 146: mdoc->first = mdoc->last;
1.67 kristaps 147: mdoc->last->type = MDOC_ROOT;
1.196 schwarze 148: mdoc->last->tok = MDOC_MAX;
1.67 kristaps 149: mdoc->next = MDOC_NEXT_CHILD;
1.73 kristaps 150: }
151:
152: /*
1.85 kristaps 153: * Free up volatile resources (see mdoc_free1()) then re-initialises the
154: * data with mdoc_alloc1(). After invocation, parse data has been reset
155: * and the parser is ready for re-invocation on a new tree; however,
156: * cross-parse non-volatile data is kept intact.
1.73 kristaps 157: */
1.113 kristaps 158: void
1.73 kristaps 159: mdoc_reset(struct mdoc *mdoc)
160: {
161:
162: mdoc_free1(mdoc);
1.113 kristaps 163: mdoc_alloc1(mdoc);
1.67 kristaps 164: }
165:
1.68 kristaps 166: /*
1.85 kristaps 167: * Completely free up all volatile and non-volatile parse resources.
168: * After invocation, the pointer is no longer usable.
1.68 kristaps 169: */
1.67 kristaps 170: void
1.38 kristaps 171: mdoc_free(struct mdoc *mdoc)
1.34 kristaps 172: {
173:
1.73 kristaps 174: mdoc_free1(mdoc);
1.1 kristaps 175: free(mdoc);
176: }
177:
1.85 kristaps 178: /*
1.213 schwarze 179: * Allocate volatile and non-volatile parse resources.
1.85 kristaps 180: */
1.1 kristaps 181: struct mdoc *
1.208 schwarze 182: mdoc_alloc(struct roff *roff, struct mparse *parse,
1.220 schwarze 183: const char *defos, int quick)
1.1 kristaps 184: {
185: struct mdoc *p;
186:
1.113 kristaps 187: p = mandoc_calloc(1, sizeof(struct mdoc));
188:
1.185 kristaps 189: p->parse = parse;
1.197 schwarze 190: p->defos = defos;
1.208 schwarze 191: p->quick = quick;
1.189 kristaps 192: p->roff = roff;
1.73 kristaps 193:
1.113 kristaps 194: mdoc_hash_init();
195: mdoc_alloc1(p);
196: return(p);
1.1 kristaps 197: }
198:
199: int
1.203 schwarze 200: mdoc_endparse(struct mdoc *mdoc)
1.20 kristaps 201: {
202:
1.221 schwarze 203: return(mdoc_macroend(mdoc));
1.181 kristaps 204: }
205:
206: int
1.203 schwarze 207: mdoc_addeqn(struct mdoc *mdoc, const struct eqn *ep)
1.181 kristaps 208: {
209: struct mdoc_node *n;
210:
1.203 schwarze 211: n = node_alloc(mdoc, ep->ln, ep->pos, MDOC_MAX, MDOC_EQN);
1.181 kristaps 212: n->eqn = ep;
213:
1.203 schwarze 214: if ( ! node_append(mdoc, n))
1.181 kristaps 215: return(0);
216:
1.203 schwarze 217: mdoc->next = MDOC_NEXT_SIBLING;
1.181 kristaps 218: return(1);
1.20 kristaps 219: }
220:
1.175 kristaps 221: int
1.203 schwarze 222: mdoc_addspan(struct mdoc *mdoc, const struct tbl_span *sp)
1.175 kristaps 223: {
1.180 kristaps 224: struct mdoc_node *n;
1.175 kristaps 225:
1.203 schwarze 226: n = node_alloc(mdoc, sp->line, 0, MDOC_MAX, MDOC_TBL);
1.180 kristaps 227: n->span = sp;
228:
1.203 schwarze 229: if ( ! node_append(mdoc, n))
1.180 kristaps 230: return(0);
231:
1.203 schwarze 232: mdoc->next = MDOC_NEXT_SIBLING;
1.180 kristaps 233: return(1);
1.175 kristaps 234: }
235:
1.50 kristaps 236: /*
1.53 kristaps 237: * Main parse routine. Parses a single line -- really just hands off to
1.123 kristaps 238: * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
1.50 kristaps 239: */
1.20 kristaps 240: int
1.203 schwarze 241: mdoc_parseln(struct mdoc *mdoc, int ln, char *buf, int offs)
1.1 kristaps 242: {
243:
1.203 schwarze 244: mdoc->flags |= MDOC_NEWLINE;
1.153 schwarze 245:
246: /*
247: * Let the roff nS register switch SYNOPSIS mode early,
248: * such that the parser knows at all times
249: * whether this mode is on or off.
250: * Note that this mode is also switched by the Sh macro.
251: */
1.204 schwarze 252: if (roff_getreg(mdoc->roff, "nS"))
253: mdoc->flags |= MDOC_SYNOPSIS;
254: else
255: mdoc->flags &= ~MDOC_SYNOPSIS;
1.153 schwarze 256:
1.203 schwarze 257: return(roff_getcontrol(mdoc->roff, buf, &offs) ?
1.213 schwarze 258: mdoc_pmacro(mdoc, ln, buf, offs) :
259: mdoc_ptext(mdoc, ln, buf, offs));
1.1 kristaps 260: }
261:
1.88 kristaps 262: int
1.148 kristaps 263: mdoc_macro(MACRO_PROT_ARGS)
1.88 kristaps 264: {
1.122 kristaps 265: assert(tok < MDOC_MAX);
266:
1.223 ! schwarze 267: if (mdoc->flags & MDOC_PBODY) {
! 268: if (tok == MDOC_Dt) {
! 269: mandoc_vmsg(MANDOCERR_DT_LATE,
! 270: mdoc->parse, line, ppos,
! 271: "Dt %s", buf + *pos);
! 272: return(1);
! 273: }
! 274: } else if ( ! (mdoc_macros[tok].flags & MDOC_PROLOGUE)) {
! 275: if (mdoc->meta.title == NULL) {
! 276: mandoc_vmsg(MANDOCERR_DT_NOTITLE,
! 277: mdoc->parse, line, ppos, "%s %s",
! 278: mdoc_macronames[tok], buf + *pos);
! 279: mdoc->meta.title = mandoc_strdup("UNTITLED");
! 280: }
1.203 schwarze 281: if (NULL == mdoc->meta.vol)
282: mdoc->meta.vol = mandoc_strdup("LOCAL");
283: mdoc->flags |= MDOC_PBODY;
1.120 kristaps 284: }
1.88 kristaps 285:
1.203 schwarze 286: return((*mdoc_macros[tok].fp)(mdoc, tok, line, ppos, pos, buf));
1.73 kristaps 287: }
288:
289:
290: static int
291: node_append(struct mdoc *mdoc, struct mdoc_node *p)
1.1 kristaps 292: {
293:
1.25 kristaps 294: assert(mdoc->last);
295: assert(mdoc->first);
296: assert(MDOC_ROOT != p->type);
1.1 kristaps 297:
1.13 kristaps 298: switch (mdoc->next) {
1.213 schwarze 299: case MDOC_NEXT_SIBLING:
1.6 kristaps 300: mdoc->last->next = p;
301: p->prev = mdoc->last;
1.13 kristaps 302: p->parent = mdoc->last->parent;
1.1 kristaps 303: break;
1.213 schwarze 304: case MDOC_NEXT_CHILD:
1.13 kristaps 305: mdoc->last->child = p;
1.1 kristaps 306: p->parent = mdoc->last;
307: break;
308: default:
1.13 kristaps 309: abort();
310: /* NOTREACHED */
1.1 kristaps 311: }
312:
1.86 kristaps 313: p->parent->nchild++;
314:
1.172 kristaps 315: /*
316: * Copy over the normalised-data pointer of our parent. Not
317: * everybody has one, but copying a null pointer is fine.
318: */
319:
320: switch (p->type) {
1.213 schwarze 321: case MDOC_BODY:
1.202 schwarze 322: if (ENDBODY_NOT != p->end)
323: break;
1.172 kristaps 324: /* FALLTHROUGH */
1.213 schwarze 325: case MDOC_TAIL:
1.172 kristaps 326: /* FALLTHROUGH */
1.213 schwarze 327: case MDOC_HEAD:
1.172 kristaps 328: p->norm = p->parent->norm;
329: break;
330: default:
331: break;
332: }
333:
1.23 kristaps 334: if ( ! mdoc_valid_pre(mdoc, p))
335: return(0);
1.27 kristaps 336:
337: switch (p->type) {
1.213 schwarze 338: case MDOC_HEAD:
1.27 kristaps 339: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 340: p->parent->head = p;
1.27 kristaps 341: break;
1.213 schwarze 342: case MDOC_TAIL:
1.27 kristaps 343: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 344: p->parent->tail = p;
1.27 kristaps 345: break;
1.213 schwarze 346: case MDOC_BODY:
1.152 schwarze 347: if (p->end)
348: break;
1.27 kristaps 349: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 350: p->parent->body = p;
1.27 kristaps 351: break;
352: default:
353: break;
354: }
355:
1.1 kristaps 356: mdoc->last = p;
1.71 kristaps 357:
358: switch (p->type) {
1.213 schwarze 359: case MDOC_TBL:
1.176 kristaps 360: /* FALLTHROUGH */
1.213 schwarze 361: case MDOC_TEXT:
1.71 kristaps 362: if ( ! mdoc_valid_post(mdoc))
363: return(0);
364: break;
365: default:
366: break;
367: }
368:
1.23 kristaps 369: return(1);
1.1 kristaps 370: }
371:
1.45 kristaps 372: static struct mdoc_node *
1.213 schwarze 373: node_alloc(struct mdoc *mdoc, int line, int pos,
1.117 kristaps 374: enum mdoct tok, enum mdoc_type type)
1.45 kristaps 375: {
1.46 kristaps 376: struct mdoc_node *p;
377:
1.113 kristaps 378: p = mandoc_calloc(1, sizeof(struct mdoc_node));
1.203 schwarze 379: p->sec = mdoc->lastsec;
1.73 kristaps 380: p->line = line;
381: p->pos = pos;
1.206 schwarze 382: p->lastline = line;
1.73 kristaps 383: p->tok = tok;
1.118 kristaps 384: p->type = type;
1.150 kristaps 385:
386: /* Flag analysis. */
387:
1.203 schwarze 388: if (MDOC_SYNOPSIS & mdoc->flags)
1.153 schwarze 389: p->flags |= MDOC_SYNPRETTY;
390: else
391: p->flags &= ~MDOC_SYNPRETTY;
1.203 schwarze 392: if (MDOC_NEWLINE & mdoc->flags)
1.130 kristaps 393: p->flags |= MDOC_LINE;
1.203 schwarze 394: mdoc->flags &= ~MDOC_NEWLINE;
1.150 kristaps 395:
1.46 kristaps 396: return(p);
1.45 kristaps 397: }
398:
1.23 kristaps 399: int
1.203 schwarze 400: mdoc_tail_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.17 kristaps 401: {
402: struct mdoc_node *p;
403:
1.203 schwarze 404: p = node_alloc(mdoc, line, pos, tok, MDOC_TAIL);
405: if ( ! node_append(mdoc, p))
1.102 kristaps 406: return(0);
1.203 schwarze 407: mdoc->next = MDOC_NEXT_CHILD;
1.102 kristaps 408: return(1);
1.17 kristaps 409: }
410:
1.23 kristaps 411: int
1.203 schwarze 412: mdoc_head_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.1 kristaps 413: {
414: struct mdoc_node *p;
415:
1.203 schwarze 416: assert(mdoc->first);
417: assert(mdoc->last);
1.1 kristaps 418:
1.203 schwarze 419: p = node_alloc(mdoc, line, pos, tok, MDOC_HEAD);
420: if ( ! node_append(mdoc, p))
1.102 kristaps 421: return(0);
1.203 schwarze 422: mdoc->next = MDOC_NEXT_CHILD;
1.102 kristaps 423: return(1);
1.1 kristaps 424: }
425:
1.23 kristaps 426: int
1.203 schwarze 427: mdoc_body_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.1 kristaps 428: {
429: struct mdoc_node *p;
430:
1.203 schwarze 431: p = node_alloc(mdoc, line, pos, tok, MDOC_BODY);
432: if ( ! node_append(mdoc, p))
1.102 kristaps 433: return(0);
1.203 schwarze 434: mdoc->next = MDOC_NEXT_CHILD;
1.152 schwarze 435: return(1);
436: }
437:
438: int
1.203 schwarze 439: mdoc_endbody_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok,
1.152 schwarze 440: struct mdoc_node *body, enum mdoc_endbody end)
441: {
442: struct mdoc_node *p;
443:
1.203 schwarze 444: p = node_alloc(mdoc, line, pos, tok, MDOC_BODY);
1.152 schwarze 445: p->pending = body;
1.202 schwarze 446: p->norm = body->norm;
1.152 schwarze 447: p->end = end;
1.203 schwarze 448: if ( ! node_append(mdoc, p))
1.152 schwarze 449: return(0);
1.203 schwarze 450: mdoc->next = MDOC_NEXT_SIBLING;
1.102 kristaps 451: return(1);
1.1 kristaps 452: }
453:
1.23 kristaps 454: int
1.213 schwarze 455: mdoc_block_alloc(struct mdoc *mdoc, int line, int pos,
1.117 kristaps 456: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 457: {
458: struct mdoc_node *p;
459:
1.203 schwarze 460: p = node_alloc(mdoc, line, pos, tok, MDOC_BLOCK);
1.77 kristaps 461: p->args = args;
462: if (p->args)
1.53 kristaps 463: (args->refcnt)++;
1.172 kristaps 464:
465: switch (tok) {
1.213 schwarze 466: case MDOC_Bd:
1.172 kristaps 467: /* FALLTHROUGH */
1.213 schwarze 468: case MDOC_Bf:
1.172 kristaps 469: /* FALLTHROUGH */
1.213 schwarze 470: case MDOC_Bl:
1.217 schwarze 471: /* FALLTHROUGH */
472: case MDOC_En:
1.173 kristaps 473: /* FALLTHROUGH */
1.213 schwarze 474: case MDOC_Rs:
1.172 kristaps 475: p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
476: break;
477: default:
478: break;
479: }
480:
1.203 schwarze 481: if ( ! node_append(mdoc, p))
1.102 kristaps 482: return(0);
1.203 schwarze 483: mdoc->next = MDOC_NEXT_CHILD;
1.102 kristaps 484: return(1);
1.1 kristaps 485: }
486:
1.23 kristaps 487: int
1.213 schwarze 488: mdoc_elem_alloc(struct mdoc *mdoc, int line, int pos,
1.117 kristaps 489: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 490: {
491: struct mdoc_node *p;
492:
1.203 schwarze 493: p = node_alloc(mdoc, line, pos, tok, MDOC_ELEM);
1.77 kristaps 494: p->args = args;
495: if (p->args)
1.53 kristaps 496: (args->refcnt)++;
1.172 kristaps 497:
498: switch (tok) {
1.213 schwarze 499: case MDOC_An:
1.172 kristaps 500: p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
501: break;
502: default:
503: break;
504: }
505:
1.203 schwarze 506: if ( ! node_append(mdoc, p))
1.102 kristaps 507: return(0);
1.203 schwarze 508: mdoc->next = MDOC_NEXT_CHILD;
1.175 kristaps 509: return(1);
510: }
1.1 kristaps 511:
1.124 kristaps 512: int
1.203 schwarze 513: mdoc_word_alloc(struct mdoc *mdoc, int line, int pos, const char *p)
1.1 kristaps 514: {
1.91 kristaps 515: struct mdoc_node *n;
1.1 kristaps 516:
1.203 schwarze 517: n = node_alloc(mdoc, line, pos, MDOC_MAX, MDOC_TEXT);
518: n->string = roff_strdup(mdoc->roff, p);
1.91 kristaps 519:
1.203 schwarze 520: if ( ! node_append(mdoc, n))
1.101 kristaps 521: return(0);
1.124 kristaps 522:
1.203 schwarze 523: mdoc->next = MDOC_NEXT_SIBLING;
1.101 kristaps 524: return(1);
1.91 kristaps 525: }
526:
1.205 schwarze 527: void
528: mdoc_word_append(struct mdoc *mdoc, const char *p)
529: {
530: struct mdoc_node *n;
531: char *addstr, *newstr;
532:
533: n = mdoc->last;
534: addstr = roff_strdup(mdoc->roff, p);
1.210 schwarze 535: mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
1.205 schwarze 536: free(addstr);
537: free(n->string);
538: n->string = newstr;
539: mdoc->next = MDOC_NEXT_SIBLING;
540: }
1.91 kristaps 541:
1.155 kristaps 542: static void
1.53 kristaps 543: mdoc_node_free(struct mdoc_node *p)
1.1 kristaps 544: {
545:
1.172 kristaps 546: if (MDOC_BLOCK == p->type || MDOC_ELEM == p->type)
1.171 kristaps 547: free(p->norm);
1.53 kristaps 548: if (p->string)
549: free(p->string);
550: if (p->args)
551: mdoc_argv_free(p->args);
1.1 kristaps 552: free(p);
553: }
554:
1.121 kristaps 555: static void
1.203 schwarze 556: mdoc_node_unlink(struct mdoc *mdoc, struct mdoc_node *n)
1.121 kristaps 557: {
558:
559: /* Adjust siblings. */
560:
561: if (n->prev)
562: n->prev->next = n->next;
563: if (n->next)
564: n->next->prev = n->prev;
565:
566: /* Adjust parent. */
567:
568: if (n->parent) {
569: n->parent->nchild--;
570: if (n->parent->child == n)
571: n->parent->child = n->prev ? n->prev : n->next;
1.169 kristaps 572: if (n->parent->last == n)
573: n->parent->last = n->prev ? n->prev : NULL;
1.121 kristaps 574: }
575:
576: /* Adjust parse point, if applicable. */
577:
1.203 schwarze 578: if (mdoc && mdoc->last == n) {
1.121 kristaps 579: if (n->prev) {
1.203 schwarze 580: mdoc->last = n->prev;
581: mdoc->next = MDOC_NEXT_SIBLING;
1.121 kristaps 582: } else {
1.203 schwarze 583: mdoc->last = n->parent;
584: mdoc->next = MDOC_NEXT_CHILD;
1.121 kristaps 585: }
586: }
587:
1.203 schwarze 588: if (mdoc && mdoc->first == n)
589: mdoc->first = NULL;
1.121 kristaps 590: }
591:
1.53 kristaps 592: void
1.203 schwarze 593: mdoc_node_delete(struct mdoc *mdoc, struct mdoc_node *p)
1.1 kristaps 594: {
595:
1.121 kristaps 596: while (p->child) {
597: assert(p->nchild);
1.203 schwarze 598: mdoc_node_delete(mdoc, p->child);
1.121 kristaps 599: }
600: assert(0 == p->nchild);
1.1 kristaps 601:
1.203 schwarze 602: mdoc_node_unlink(mdoc, p);
1.53 kristaps 603: mdoc_node_free(p);
1.201 schwarze 604: }
605:
606: int
1.203 schwarze 607: mdoc_node_relink(struct mdoc *mdoc, struct mdoc_node *p)
1.201 schwarze 608: {
609:
1.203 schwarze 610: mdoc_node_unlink(mdoc, p);
611: return(node_append(mdoc, p));
1.1 kristaps 612: }
613:
1.193 kristaps 614: #if 0
1.191 kristaps 615: /*
616: * Pre-treat a text line.
617: * Text lines can consist of equations, which must be handled apart from
618: * the regular text.
619: * Thus, use this function to step through a line checking if it has any
620: * equations embedded in it.
621: * This must handle multiple equations AND equations that do not end at
622: * the end-of-line, i.e., will re-enter in the next roff parse.
623: */
624: static int
1.203 schwarze 625: mdoc_preptext(struct mdoc *mdoc, int line, char *buf, int offs)
1.191 kristaps 626: {
627: char *start, *end;
628: char delim;
629:
630: while ('\0' != buf[offs]) {
631: /* Mark starting position if eqn is set. */
632: start = NULL;
1.203 schwarze 633: if ('\0' != (delim = roff_eqndelim(mdoc->roff)))
1.191 kristaps 634: if (NULL != (start = strchr(buf + offs, delim)))
635: *start++ = '\0';
636:
637: /* Parse text as normal. */
1.203 schwarze 638: if ( ! mdoc_ptext(mdoc, line, buf, offs))
1.191 kristaps 639: return(0);
640:
641: /* Continue only if an equation exists. */
642: if (NULL == start)
643: break;
644:
645: /* Read past the end of the equation. */
646: offs += start - (buf + offs);
647: assert(start == &buf[offs]);
648: if (NULL != (end = strchr(buf + offs, delim))) {
649: *end++ = '\0';
650: while (' ' == *end)
651: end++;
652: }
653:
654: /* Parse the equation itself. */
1.203 schwarze 655: roff_openeqn(mdoc->roff, NULL, line, offs, buf);
1.191 kristaps 656:
657: /* Process a finished equation? */
1.203 schwarze 658: if (roff_closeeqn(mdoc->roff))
659: if ( ! mdoc_addeqn(mdoc, roff_eqn(mdoc->roff)))
1.191 kristaps 660: return(0);
661: offs += (end - (buf + offs));
1.213 schwarze 662: }
1.191 kristaps 663:
664: return(1);
665: }
1.193 kristaps 666: #endif
1.1 kristaps 667:
1.53 kristaps 668: /*
669: * Parse free-form text, that is, a line that does not begin with the
670: * control character.
671: */
672: static int
1.203 schwarze 673: mdoc_ptext(struct mdoc *mdoc, int line, char *buf, int offs)
1.1 kristaps 674: {
1.142 kristaps 675: char *c, *ws, *end;
676: struct mdoc_node *n;
677:
1.203 schwarze 678: assert(mdoc->last);
679: n = mdoc->last;
1.142 kristaps 680:
681: /*
1.144 kristaps 682: * Divert directly to list processing if we're encountering a
1.142 kristaps 683: * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
1.144 kristaps 684: * (a MDOC_BODY means it's already open, in which case we should
685: * process within its context in the normal way).
1.142 kristaps 686: */
687:
1.143 kristaps 688: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.213 schwarze 689: LIST_column == n->norm->Bl.type) {
1.144 kristaps 690: /* `Bl' is open without any children. */
1.203 schwarze 691: mdoc->flags |= MDOC_FREECOL;
692: return(mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf));
1.142 kristaps 693: }
694:
695: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
1.213 schwarze 696: NULL != n->parent &&
697: MDOC_Bl == n->parent->tok &&
698: LIST_column == n->parent->norm->Bl.type) {
1.144 kristaps 699: /* `Bl' has block-level `It' children. */
1.203 schwarze 700: mdoc->flags |= MDOC_FREECOL;
701: return(mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf));
1.142 kristaps 702: }
1.124 kristaps 703:
1.137 schwarze 704: /*
705: * Search for the beginning of unescaped trailing whitespace (ws)
706: * and for the first character not to be output (end).
707: */
1.139 kristaps 708:
709: /* FIXME: replace with strcspn(). */
1.137 schwarze 710: ws = NULL;
711: for (c = end = buf + offs; *c; c++) {
712: switch (*c) {
713: case ' ':
714: if (NULL == ws)
715: ws = c;
716: continue;
717: case '\t':
718: /*
719: * Always warn about trailing tabs,
720: * even outside literal context,
721: * where they should be put on the next line.
722: */
723: if (NULL == ws)
724: ws = c;
725: /*
726: * Strip trailing tabs in literal context only;
727: * outside, they affect the next line.
728: */
1.203 schwarze 729: if (MDOC_LITERAL & mdoc->flags)
1.137 schwarze 730: continue;
731: break;
732: case '\\':
733: /* Skip the escaped character, too, if any. */
734: if (c[1])
735: c++;
736: /* FALLTHROUGH */
737: default:
738: ws = NULL;
739: break;
740: }
741: end = c + 1;
742: }
743: *end = '\0';
1.91 kristaps 744:
1.137 schwarze 745: if (ws)
1.218 schwarze 746: mandoc_msg(MANDOCERR_SPACE_EOL, mdoc->parse,
747: line, (int)(ws-buf), NULL);
1.115 kristaps 748:
1.203 schwarze 749: if ('\0' == buf[offs] && ! (MDOC_LITERAL & mdoc->flags)) {
1.218 schwarze 750: mandoc_msg(MANDOCERR_FI_BLANK, mdoc->parse,
751: line, (int)(c - buf), NULL);
1.124 kristaps 752:
1.119 kristaps 753: /*
1.165 schwarze 754: * Insert a `sp' in the case of a blank line. Technically,
1.124 kristaps 755: * blank lines aren't allowed, but enough manuals assume this
756: * behaviour that we want to work around it.
1.119 kristaps 757: */
1.203 schwarze 758: if ( ! mdoc_elem_alloc(mdoc, line, offs, MDOC_sp, NULL))
1.119 kristaps 759: return(0);
1.124 kristaps 760:
1.203 schwarze 761: mdoc->next = MDOC_NEXT_SIBLING;
1.199 schwarze 762:
1.203 schwarze 763: return(mdoc_valid_post(mdoc));
1.119 kristaps 764: }
1.68 kristaps 765:
1.203 schwarze 766: if ( ! mdoc_word_alloc(mdoc, line, offs, buf+offs))
1.137 schwarze 767: return(0);
1.91 kristaps 768:
1.203 schwarze 769: if (MDOC_LITERAL & mdoc->flags)
1.137 schwarze 770: return(1);
1.128 kristaps 771:
772: /*
773: * End-of-sentence check. If the last character is an unescaped
774: * EOS character, then flag the node as being the end of a
775: * sentence. The front-end will know how to interpret this.
776: */
1.132 kristaps 777:
1.137 schwarze 778: assert(buf < end);
779:
1.207 schwarze 780: if (mandoc_eos(buf+offs, (size_t)(end-buf-offs)))
1.203 schwarze 781: mdoc->last->flags |= MDOC_EOS;
1.128 kristaps 782:
783: return(1);
1.1 kristaps 784: }
785:
1.53 kristaps 786: /*
787: * Parse a macro line, that is, a line beginning with the control
788: * character.
789: */
1.155 kristaps 790: static int
1.203 schwarze 791: mdoc_pmacro(struct mdoc *mdoc, int ln, char *buf, int offs)
1.1 kristaps 792: {
1.144 kristaps 793: enum mdoct tok;
1.188 kristaps 794: int i, sv;
1.144 kristaps 795: char mac[5];
796: struct mdoc_node *n;
1.1 kristaps 797:
1.188 kristaps 798: /* Empty post-control lines are ignored. */
1.63 kristaps 799:
1.188 kristaps 800: if ('"' == buf[offs]) {
1.218 schwarze 801: mandoc_msg(MANDOCERR_COMMENT_BAD, mdoc->parse,
802: ln, offs, NULL);
1.188 kristaps 803: return(1);
804: } else if ('\0' == buf[offs])
1.63 kristaps 805: return(1);
806:
1.188 kristaps 807: sv = offs;
1.130 kristaps 808:
1.213 schwarze 809: /*
1.162 schwarze 810: * Copy the first word into a nil-terminated buffer.
1.165 schwarze 811: * Stop copying when a tab, space, or eoln is encountered.
1.160 kristaps 812: */
1.1 kristaps 813:
1.188 kristaps 814: i = 0;
1.213 schwarze 815: while (i < 4 && '\0' != buf[offs] && ' ' != buf[offs] &&
816: '\t' != buf[offs])
1.188 kristaps 817: mac[i++] = buf[offs++];
818:
819: mac[i] = '\0';
820:
1.214 schwarze 821: tok = (i > 1 && i < 4) ? mdoc_hash_find(mac) : MDOC_MAX;
1.1 kristaps 822:
1.163 schwarze 823: if (MDOC_MAX == tok) {
1.222 schwarze 824: mandoc_msg(MANDOCERR_MACRO, mdoc->parse,
825: ln, sv, buf + sv - 1);
1.58 kristaps 826: return(1);
1.53 kristaps 827: }
1.1 kristaps 828:
1.160 kristaps 829: /* Disregard the first trailing tab, if applicable. */
830:
1.188 kristaps 831: if ('\t' == buf[offs])
832: offs++;
1.160 kristaps 833:
834: /* Jump to the next non-whitespace word. */
1.1 kristaps 835:
1.188 kristaps 836: while (buf[offs] && ' ' == buf[offs])
837: offs++;
1.1 kristaps 838:
1.213 schwarze 839: /*
1.125 kristaps 840: * Trailing whitespace. Note that tabs are allowed to be passed
841: * into the parser as "text", so we only warn about spaces here.
842: */
1.115 kristaps 843:
1.188 kristaps 844: if ('\0' == buf[offs] && ' ' == buf[offs - 1])
1.218 schwarze 845: mandoc_msg(MANDOCERR_SPACE_EOL, mdoc->parse,
846: ln, offs - 1, NULL);
1.115 kristaps 847:
1.144 kristaps 848: /*
849: * If an initial macro or a list invocation, divert directly
850: * into macro processing.
851: */
852:
1.221 schwarze 853: if (NULL == mdoc->last || MDOC_It == tok || MDOC_El == tok)
854: return(mdoc_macro(mdoc, tok, ln, sv, &offs, buf));
1.144 kristaps 855:
1.203 schwarze 856: n = mdoc->last;
857: assert(mdoc->last);
1.144 kristaps 858:
859: /*
860: * If the first macro of a `Bl -column', open an `It' block
861: * context around the parsed macro.
862: */
863:
864: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.213 schwarze 865: LIST_column == n->norm->Bl.type) {
1.203 schwarze 866: mdoc->flags |= MDOC_FREECOL;
1.221 schwarze 867: return(mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf));
1.144 kristaps 868: }
869:
870: /*
871: * If we're following a block-level `It' within a `Bl -column'
872: * context (perhaps opened in the above block or in ptext()),
873: * then open an `It' block context around the parsed macro.
1.98 kristaps 874: */
1.144 kristaps 875:
876: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
1.213 schwarze 877: NULL != n->parent &&
878: MDOC_Bl == n->parent->tok &&
879: LIST_column == n->parent->norm->Bl.type) {
1.203 schwarze 880: mdoc->flags |= MDOC_FREECOL;
1.221 schwarze 881: return(mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf));
1.144 kristaps 882: }
883:
884: /* Normal processing of a macro. */
885:
1.213 schwarze 886: if ( ! mdoc_macro(mdoc, tok, ln, sv, &offs, buf))
1.221 schwarze 887: return(0);
1.208 schwarze 888:
889: /* In quick mode (for mandocdb), abort after the NAME section. */
890:
891: if (mdoc->quick && MDOC_Sh == tok &&
892: SEC_NAME != mdoc->last->sec)
893: return(2);
1.1 kristaps 894:
1.53 kristaps 895: return(1);
1.1 kristaps 896: }
1.100 kristaps 897:
1.186 kristaps 898: enum mdelim
899: mdoc_isdelim(const char *p)
900: {
901:
902: if ('\0' == p[0])
903: return(DELIM_NONE);
904:
905: if ('\0' == p[1])
906: switch (p[0]) {
1.213 schwarze 907: case '(':
1.186 kristaps 908: /* FALLTHROUGH */
1.213 schwarze 909: case '[':
1.186 kristaps 910: return(DELIM_OPEN);
1.213 schwarze 911: case '|':
1.186 kristaps 912: return(DELIM_MIDDLE);
1.213 schwarze 913: case '.':
1.186 kristaps 914: /* FALLTHROUGH */
1.213 schwarze 915: case ',':
1.186 kristaps 916: /* FALLTHROUGH */
1.213 schwarze 917: case ';':
1.186 kristaps 918: /* FALLTHROUGH */
1.213 schwarze 919: case ':':
1.186 kristaps 920: /* FALLTHROUGH */
1.213 schwarze 921: case '?':
1.186 kristaps 922: /* FALLTHROUGH */
1.213 schwarze 923: case '!':
1.186 kristaps 924: /* FALLTHROUGH */
1.213 schwarze 925: case ')':
1.186 kristaps 926: /* FALLTHROUGH */
1.213 schwarze 927: case ']':
1.186 kristaps 928: return(DELIM_CLOSE);
929: default:
930: return(DELIM_NONE);
931: }
932:
933: if ('\\' != p[0])
934: return(DELIM_NONE);
1.100 kristaps 935:
1.186 kristaps 936: if (0 == strcmp(p + 1, "."))
937: return(DELIM_CLOSE);
1.200 schwarze 938: if (0 == strcmp(p + 1, "fR|\\fP"))
1.186 kristaps 939: return(DELIM_MIDDLE);
940:
941: return(DELIM_NONE);
1.211 schwarze 942: }
943:
944: void
945: mdoc_deroff(char **dest, const struct mdoc_node *n)
946: {
947: char *cp;
948: size_t sz;
949:
950: if (MDOC_TEXT != n->type) {
951: for (n = n->child; n; n = n->next)
952: mdoc_deroff(dest, n);
953: return;
954: }
955:
956: /* Skip leading whitespace. */
957:
958: for (cp = n->string; '\0' != *cp; cp++)
959: if (0 == isspace((unsigned char)*cp))
960: break;
961:
962: /* Skip trailing whitespace. */
963:
964: for (sz = strlen(cp); sz; sz--)
965: if (0 == isspace((unsigned char)cp[sz-1]))
966: break;
967:
968: /* Skip empty strings. */
969:
970: if (0 == sz)
971: return;
972:
973: if (NULL == *dest) {
974: *dest = mandoc_strndup(cp, sz);
975: return;
976: }
977:
978: mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
979: free(*dest);
980: *dest = cp;
1.186 kristaps 981: }
CVSweb