Annotation of mandoc/mdoc.c, Revision 1.233
1.233 ! schwarze 1: /* $Id: mdoc.c,v 1.232 2014/11/28 04:47:03 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: #include "config.h"
19:
1.106 kristaps 20: #include <sys/types.h>
21:
1.1 kristaps 22: #include <assert.h>
1.211 schwarze 23: #include <ctype.h>
1.1 kristaps 24: #include <stdarg.h>
1.73 kristaps 25: #include <stdio.h>
1.1 kristaps 26: #include <stdlib.h>
27: #include <string.h>
1.120 kristaps 28: #include <time.h>
1.1 kristaps 29:
1.187 kristaps 30: #include "mdoc.h"
1.136 kristaps 31: #include "mandoc.h"
1.209 schwarze 32: #include "mandoc_aux.h"
1.70 kristaps 33: #include "libmdoc.h"
1.113 kristaps 34: #include "libmandoc.h"
1.1 kristaps 35:
1.216 schwarze 36: const char *const __mdoc_macronames[MDOC_MAX + 1] = {
1.82 kristaps 37: "Ap", "Dd", "Dt", "Os",
1.1 kristaps 38: "Sh", "Ss", "Pp", "D1",
39: "Dl", "Bd", "Ed", "Bl",
40: "El", "It", "Ad", "An",
41: "Ar", "Cd", "Cm", "Dv",
42: "Er", "Ev", "Ex", "Fa",
43: "Fd", "Fl", "Fn", "Ft",
44: "Ic", "In", "Li", "Nd",
45: "Nm", "Op", "Ot", "Pa",
46: "Rv", "St", "Va", "Vt",
1.114 kristaps 47: "Xr", "%A", "%B", "%D",
48: "%I", "%J", "%N", "%O",
49: "%P", "%R", "%T", "%V",
1.1 kristaps 50: "Ac", "Ao", "Aq", "At",
51: "Bc", "Bf", "Bo", "Bq",
52: "Bsx", "Bx", "Db", "Dc",
53: "Do", "Dq", "Ec", "Ef",
54: "Em", "Eo", "Fx", "Ms",
55: "No", "Ns", "Nx", "Ox",
56: "Pc", "Pf", "Po", "Pq",
57: "Qc", "Ql", "Qo", "Qq",
58: "Re", "Rs", "Sc", "So",
59: "Sq", "Sm", "Sx", "Sy",
60: "Tn", "Ux", "Xc", "Xo",
61: "Fo", "Fc", "Oo", "Oc",
62: "Bk", "Ek", "Bt", "Hf",
1.82 kristaps 63: "Fr", "Ud", "Lb", "Lp",
64: "Lk", "Mt", "Brq", "Bro",
1.114 kristaps 65: "Brc", "%C", "Es", "En",
66: "Dx", "%Q", "br", "sp",
1.216 schwarze 67: "%U", "Ta", "ll", "text",
1.1 kristaps 68: };
69:
1.213 schwarze 70: const char *const __mdoc_argnames[MDOC_ARG_MAX] = {
1.1 kristaps 71: "split", "nosplit", "ragged",
1.213 schwarze 72: "unfilled", "literal", "file",
73: "offset", "bullet", "dash",
74: "hyphen", "item", "enum",
75: "tag", "diag", "hang",
76: "ohang", "inset", "column",
77: "width", "compact", "std",
1.52 kristaps 78: "filled", "words", "emphasis",
1.108 kristaps 79: "symbolic", "nested", "centered"
1.1 kristaps 80: };
81:
82: const char * const *mdoc_macronames = __mdoc_macronames;
83: const char * const *mdoc_argnames = __mdoc_argnames;
84:
1.121 kristaps 85: static void mdoc_node_free(struct mdoc_node *);
1.213 schwarze 86: static void mdoc_node_unlink(struct mdoc *,
1.121 kristaps 87: struct mdoc_node *);
1.73 kristaps 88: static void mdoc_free1(struct mdoc *);
1.113 kristaps 89: static void mdoc_alloc1(struct mdoc *);
1.213 schwarze 90: static struct mdoc_node *node_alloc(struct mdoc *, int, int,
1.117 kristaps 91: enum mdoct, enum mdoc_type);
1.230 schwarze 92: static void node_append(struct mdoc *, struct mdoc_node *);
1.149 kristaps 93: static int mdoc_ptext(struct mdoc *, int, char *, int);
94: static int mdoc_pmacro(struct mdoc *, int, char *, int);
1.88 kristaps 95:
1.213 schwarze 96:
1.1 kristaps 97: const struct mdoc_node *
1.203 schwarze 98: mdoc_node(const struct mdoc *mdoc)
1.1 kristaps 99: {
100:
1.203 schwarze 101: return(mdoc->first);
1.1 kristaps 102: }
103:
1.37 kristaps 104: const struct mdoc_meta *
1.203 schwarze 105: mdoc_meta(const struct mdoc *mdoc)
1.37 kristaps 106: {
107:
1.203 schwarze 108: return(&mdoc->meta);
1.37 kristaps 109: }
110:
1.85 kristaps 111: /*
112: * Frees volatile resources (parse tree, meta-data, fields).
113: */
1.73 kristaps 114: static void
115: mdoc_free1(struct mdoc *mdoc)
1.67 kristaps 116: {
117:
118: if (mdoc->first)
1.121 kristaps 119: mdoc_node_delete(mdoc, mdoc->first);
1.223 schwarze 120: free(mdoc->meta.msec);
121: free(mdoc->meta.vol);
122: free(mdoc->meta.arch);
123: free(mdoc->meta.date);
124: free(mdoc->meta.title);
125: free(mdoc->meta.os);
126: free(mdoc->meta.name);
1.73 kristaps 127: }
128:
1.85 kristaps 129: /*
130: * Allocate all volatile resources (parse tree, meta-data, fields).
131: */
1.113 kristaps 132: static void
1.73 kristaps 133: mdoc_alloc1(struct mdoc *mdoc)
134: {
1.67 kristaps 135:
1.112 kristaps 136: memset(&mdoc->meta, 0, sizeof(struct mdoc_meta));
1.67 kristaps 137: mdoc->flags = 0;
1.85 kristaps 138: mdoc->lastnamed = mdoc->lastsec = SEC_NONE;
1.113 kristaps 139: mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node));
1.70 kristaps 140: mdoc->first = mdoc->last;
1.67 kristaps 141: mdoc->last->type = MDOC_ROOT;
1.196 schwarze 142: mdoc->last->tok = MDOC_MAX;
1.67 kristaps 143: mdoc->next = MDOC_NEXT_CHILD;
1.73 kristaps 144: }
145:
146: /*
1.85 kristaps 147: * Free up volatile resources (see mdoc_free1()) then re-initialises the
148: * data with mdoc_alloc1(). After invocation, parse data has been reset
149: * and the parser is ready for re-invocation on a new tree; however,
150: * cross-parse non-volatile data is kept intact.
1.73 kristaps 151: */
1.113 kristaps 152: void
1.73 kristaps 153: mdoc_reset(struct mdoc *mdoc)
154: {
155:
156: mdoc_free1(mdoc);
1.113 kristaps 157: mdoc_alloc1(mdoc);
1.67 kristaps 158: }
159:
1.68 kristaps 160: /*
1.85 kristaps 161: * Completely free up all volatile and non-volatile parse resources.
162: * After invocation, the pointer is no longer usable.
1.68 kristaps 163: */
1.67 kristaps 164: void
1.38 kristaps 165: mdoc_free(struct mdoc *mdoc)
1.34 kristaps 166: {
167:
1.73 kristaps 168: mdoc_free1(mdoc);
1.1 kristaps 169: free(mdoc);
170: }
171:
1.85 kristaps 172: /*
1.213 schwarze 173: * Allocate volatile and non-volatile parse resources.
1.85 kristaps 174: */
1.1 kristaps 175: struct mdoc *
1.208 schwarze 176: mdoc_alloc(struct roff *roff, struct mparse *parse,
1.220 schwarze 177: const char *defos, int quick)
1.1 kristaps 178: {
179: struct mdoc *p;
180:
1.113 kristaps 181: p = mandoc_calloc(1, sizeof(struct mdoc));
182:
1.185 kristaps 183: p->parse = parse;
1.197 schwarze 184: p->defos = defos;
1.208 schwarze 185: p->quick = quick;
1.189 kristaps 186: p->roff = roff;
1.73 kristaps 187:
1.113 kristaps 188: mdoc_hash_init();
189: mdoc_alloc1(p);
190: return(p);
1.1 kristaps 191: }
192:
193: int
1.203 schwarze 194: mdoc_endparse(struct mdoc *mdoc)
1.20 kristaps 195: {
196:
1.232 schwarze 197: mdoc_macroend(mdoc);
198: return(1);
1.181 kristaps 199: }
200:
1.233 ! schwarze 201: void
1.203 schwarze 202: mdoc_addeqn(struct mdoc *mdoc, const struct eqn *ep)
1.181 kristaps 203: {
204: struct mdoc_node *n;
205:
1.203 schwarze 206: n = node_alloc(mdoc, ep->ln, ep->pos, MDOC_MAX, MDOC_EQN);
1.181 kristaps 207: n->eqn = ep;
1.227 schwarze 208: if (ep->ln > mdoc->last->line)
209: n->flags |= MDOC_LINE;
1.230 schwarze 210: node_append(mdoc, n);
1.203 schwarze 211: mdoc->next = MDOC_NEXT_SIBLING;
1.20 kristaps 212: }
213:
1.233 ! schwarze 214: void
1.203 schwarze 215: mdoc_addspan(struct mdoc *mdoc, const struct tbl_span *sp)
1.175 kristaps 216: {
1.180 kristaps 217: struct mdoc_node *n;
1.175 kristaps 218:
1.203 schwarze 219: n = node_alloc(mdoc, sp->line, 0, MDOC_MAX, MDOC_TBL);
1.180 kristaps 220: n->span = sp;
1.230 schwarze 221: node_append(mdoc, n);
1.203 schwarze 222: mdoc->next = MDOC_NEXT_SIBLING;
1.175 kristaps 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
1.203 schwarze 230: mdoc_parseln(struct mdoc *mdoc, int ln, char *buf, int offs)
1.1 kristaps 231: {
232:
1.228 schwarze 233: if (mdoc->last->type != MDOC_EQN || ln > mdoc->last->line)
234: mdoc->flags |= MDOC_NEWLINE;
1.153 schwarze 235:
236: /*
237: * Let the roff nS register switch SYNOPSIS mode early,
238: * such that the parser knows at all times
239: * whether this mode is on or off.
240: * Note that this mode is also switched by the Sh macro.
241: */
1.204 schwarze 242: if (roff_getreg(mdoc->roff, "nS"))
243: mdoc->flags |= MDOC_SYNOPSIS;
244: else
245: mdoc->flags &= ~MDOC_SYNOPSIS;
1.153 schwarze 246:
1.203 schwarze 247: return(roff_getcontrol(mdoc->roff, buf, &offs) ?
1.213 schwarze 248: mdoc_pmacro(mdoc, ln, buf, offs) :
249: mdoc_ptext(mdoc, ln, buf, offs));
1.1 kristaps 250: }
251:
1.232 schwarze 252: void
1.148 kristaps 253: mdoc_macro(MACRO_PROT_ARGS)
1.88 kristaps 254: {
1.122 kristaps 255: assert(tok < MDOC_MAX);
256:
1.223 schwarze 257: if (mdoc->flags & MDOC_PBODY) {
258: if (tok == MDOC_Dt) {
259: mandoc_vmsg(MANDOCERR_DT_LATE,
260: mdoc->parse, line, ppos,
261: "Dt %s", buf + *pos);
1.232 schwarze 262: return;
1.223 schwarze 263: }
264: } else if ( ! (mdoc_macros[tok].flags & MDOC_PROLOGUE)) {
265: if (mdoc->meta.title == NULL) {
266: mandoc_vmsg(MANDOCERR_DT_NOTITLE,
267: mdoc->parse, line, ppos, "%s %s",
268: mdoc_macronames[tok], buf + *pos);
269: mdoc->meta.title = mandoc_strdup("UNTITLED");
270: }
1.203 schwarze 271: if (NULL == mdoc->meta.vol)
272: mdoc->meta.vol = mandoc_strdup("LOCAL");
273: mdoc->flags |= MDOC_PBODY;
1.120 kristaps 274: }
1.232 schwarze 275: (*mdoc_macros[tok].fp)(mdoc, tok, line, ppos, pos, buf);
1.73 kristaps 276: }
277:
278:
1.230 schwarze 279: static void
1.73 kristaps 280: node_append(struct mdoc *mdoc, struct mdoc_node *p)
1.1 kristaps 281: {
282:
1.25 kristaps 283: assert(mdoc->last);
284: assert(mdoc->first);
285: assert(MDOC_ROOT != p->type);
1.1 kristaps 286:
1.13 kristaps 287: switch (mdoc->next) {
1.213 schwarze 288: case MDOC_NEXT_SIBLING:
1.6 kristaps 289: mdoc->last->next = p;
290: p->prev = mdoc->last;
1.13 kristaps 291: p->parent = mdoc->last->parent;
1.1 kristaps 292: break;
1.213 schwarze 293: case MDOC_NEXT_CHILD:
1.13 kristaps 294: mdoc->last->child = p;
1.1 kristaps 295: p->parent = mdoc->last;
296: break;
297: default:
1.13 kristaps 298: abort();
299: /* NOTREACHED */
1.1 kristaps 300: }
301:
1.86 kristaps 302: p->parent->nchild++;
303:
1.172 kristaps 304: /*
305: * Copy over the normalised-data pointer of our parent. Not
306: * everybody has one, but copying a null pointer is fine.
307: */
308:
309: switch (p->type) {
1.213 schwarze 310: case MDOC_BODY:
1.202 schwarze 311: if (ENDBODY_NOT != p->end)
312: break;
1.172 kristaps 313: /* FALLTHROUGH */
1.213 schwarze 314: case MDOC_TAIL:
1.172 kristaps 315: /* FALLTHROUGH */
1.213 schwarze 316: case MDOC_HEAD:
1.172 kristaps 317: p->norm = p->parent->norm;
318: break;
319: default:
320: break;
321: }
322:
1.230 schwarze 323: mdoc_valid_pre(mdoc, p);
1.27 kristaps 324:
325: switch (p->type) {
1.213 schwarze 326: case MDOC_HEAD:
1.27 kristaps 327: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 328: p->parent->head = p;
1.27 kristaps 329: break;
1.213 schwarze 330: case MDOC_TAIL:
1.27 kristaps 331: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 332: p->parent->tail = p;
1.27 kristaps 333: break;
1.213 schwarze 334: case MDOC_BODY:
1.152 schwarze 335: if (p->end)
336: break;
1.27 kristaps 337: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 338: p->parent->body = p;
1.27 kristaps 339: break;
340: default:
341: break;
342: }
343:
1.1 kristaps 344: mdoc->last = p;
1.71 kristaps 345:
346: switch (p->type) {
1.213 schwarze 347: case MDOC_TBL:
1.176 kristaps 348: /* FALLTHROUGH */
1.213 schwarze 349: case MDOC_TEXT:
1.230 schwarze 350: mdoc_valid_post(mdoc);
1.71 kristaps 351: break;
352: default:
353: break;
354: }
1.1 kristaps 355: }
356:
1.45 kristaps 357: static struct mdoc_node *
1.213 schwarze 358: node_alloc(struct mdoc *mdoc, int line, int pos,
1.117 kristaps 359: enum mdoct tok, enum mdoc_type type)
1.45 kristaps 360: {
1.46 kristaps 361: struct mdoc_node *p;
362:
1.113 kristaps 363: p = mandoc_calloc(1, sizeof(struct mdoc_node));
1.203 schwarze 364: p->sec = mdoc->lastsec;
1.73 kristaps 365: p->line = line;
366: p->pos = pos;
1.206 schwarze 367: p->lastline = line;
1.73 kristaps 368: p->tok = tok;
1.118 kristaps 369: p->type = type;
1.150 kristaps 370:
371: /* Flag analysis. */
372:
1.203 schwarze 373: if (MDOC_SYNOPSIS & mdoc->flags)
1.153 schwarze 374: p->flags |= MDOC_SYNPRETTY;
375: else
376: p->flags &= ~MDOC_SYNPRETTY;
1.203 schwarze 377: if (MDOC_NEWLINE & mdoc->flags)
1.130 kristaps 378: p->flags |= MDOC_LINE;
1.203 schwarze 379: mdoc->flags &= ~MDOC_NEWLINE;
1.150 kristaps 380:
1.46 kristaps 381: return(p);
1.45 kristaps 382: }
383:
1.231 schwarze 384: void
1.203 schwarze 385: mdoc_tail_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.17 kristaps 386: {
387: struct mdoc_node *p;
388:
1.203 schwarze 389: p = node_alloc(mdoc, line, pos, tok, MDOC_TAIL);
1.230 schwarze 390: node_append(mdoc, p);
1.203 schwarze 391: mdoc->next = MDOC_NEXT_CHILD;
1.17 kristaps 392: }
393:
1.231 schwarze 394: struct mdoc_node *
1.203 schwarze 395: mdoc_head_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.1 kristaps 396: {
397: struct mdoc_node *p;
398:
1.203 schwarze 399: assert(mdoc->first);
400: assert(mdoc->last);
401: p = node_alloc(mdoc, line, pos, tok, MDOC_HEAD);
1.230 schwarze 402: node_append(mdoc, p);
1.203 schwarze 403: mdoc->next = MDOC_NEXT_CHILD;
1.231 schwarze 404: return(p);
1.1 kristaps 405: }
406:
1.231 schwarze 407: struct mdoc_node *
1.203 schwarze 408: mdoc_body_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.1 kristaps 409: {
410: struct mdoc_node *p;
411:
1.203 schwarze 412: p = node_alloc(mdoc, line, pos, tok, MDOC_BODY);
1.230 schwarze 413: node_append(mdoc, p);
1.203 schwarze 414: mdoc->next = MDOC_NEXT_CHILD;
1.231 schwarze 415: return(p);
1.152 schwarze 416: }
417:
1.231 schwarze 418: void
1.203 schwarze 419: mdoc_endbody_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok,
1.152 schwarze 420: struct mdoc_node *body, enum mdoc_endbody end)
421: {
422: struct mdoc_node *p;
423:
1.203 schwarze 424: p = node_alloc(mdoc, line, pos, tok, MDOC_BODY);
1.152 schwarze 425: p->pending = body;
1.202 schwarze 426: p->norm = body->norm;
1.152 schwarze 427: p->end = end;
1.230 schwarze 428: node_append(mdoc, p);
1.203 schwarze 429: mdoc->next = MDOC_NEXT_SIBLING;
1.1 kristaps 430: }
431:
1.231 schwarze 432: struct mdoc_node *
1.213 schwarze 433: mdoc_block_alloc(struct mdoc *mdoc, int line, int pos,
1.117 kristaps 434: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 435: {
436: struct mdoc_node *p;
437:
1.203 schwarze 438: p = node_alloc(mdoc, line, pos, tok, MDOC_BLOCK);
1.77 kristaps 439: p->args = args;
440: if (p->args)
1.53 kristaps 441: (args->refcnt)++;
1.172 kristaps 442:
443: switch (tok) {
1.213 schwarze 444: case MDOC_Bd:
1.172 kristaps 445: /* FALLTHROUGH */
1.213 schwarze 446: case MDOC_Bf:
1.172 kristaps 447: /* FALLTHROUGH */
1.213 schwarze 448: case MDOC_Bl:
1.217 schwarze 449: /* FALLTHROUGH */
450: case MDOC_En:
1.173 kristaps 451: /* FALLTHROUGH */
1.213 schwarze 452: case MDOC_Rs:
1.172 kristaps 453: p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
454: break;
455: default:
456: break;
457: }
1.230 schwarze 458: node_append(mdoc, p);
1.203 schwarze 459: mdoc->next = MDOC_NEXT_CHILD;
1.231 schwarze 460: return(p);
1.1 kristaps 461: }
462:
1.231 schwarze 463: void
1.213 schwarze 464: mdoc_elem_alloc(struct mdoc *mdoc, int line, int pos,
1.117 kristaps 465: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 466: {
467: struct mdoc_node *p;
468:
1.203 schwarze 469: p = node_alloc(mdoc, line, pos, tok, MDOC_ELEM);
1.77 kristaps 470: p->args = args;
471: if (p->args)
1.53 kristaps 472: (args->refcnt)++;
1.172 kristaps 473:
474: switch (tok) {
1.213 schwarze 475: case MDOC_An:
1.172 kristaps 476: p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
477: break;
478: default:
479: break;
480: }
1.230 schwarze 481: node_append(mdoc, p);
1.203 schwarze 482: mdoc->next = MDOC_NEXT_CHILD;
1.175 kristaps 483: }
1.1 kristaps 484:
1.231 schwarze 485: void
1.203 schwarze 486: mdoc_word_alloc(struct mdoc *mdoc, int line, int pos, const char *p)
1.1 kristaps 487: {
1.91 kristaps 488: struct mdoc_node *n;
1.1 kristaps 489:
1.203 schwarze 490: n = node_alloc(mdoc, line, pos, MDOC_MAX, MDOC_TEXT);
491: n->string = roff_strdup(mdoc->roff, p);
1.230 schwarze 492: node_append(mdoc, n);
1.203 schwarze 493: mdoc->next = MDOC_NEXT_SIBLING;
1.91 kristaps 494: }
495:
1.205 schwarze 496: void
497: mdoc_word_append(struct mdoc *mdoc, const char *p)
498: {
499: struct mdoc_node *n;
500: char *addstr, *newstr;
501:
502: n = mdoc->last;
503: addstr = roff_strdup(mdoc->roff, p);
1.210 schwarze 504: mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
1.205 schwarze 505: free(addstr);
506: free(n->string);
507: n->string = newstr;
508: mdoc->next = MDOC_NEXT_SIBLING;
509: }
1.91 kristaps 510:
1.155 kristaps 511: static void
1.53 kristaps 512: mdoc_node_free(struct mdoc_node *p)
1.1 kristaps 513: {
514:
1.172 kristaps 515: if (MDOC_BLOCK == p->type || MDOC_ELEM == p->type)
1.171 kristaps 516: free(p->norm);
1.53 kristaps 517: if (p->string)
518: free(p->string);
519: if (p->args)
520: mdoc_argv_free(p->args);
1.1 kristaps 521: free(p);
522: }
523:
1.121 kristaps 524: static void
1.203 schwarze 525: mdoc_node_unlink(struct mdoc *mdoc, struct mdoc_node *n)
1.121 kristaps 526: {
527:
528: /* Adjust siblings. */
529:
530: if (n->prev)
531: n->prev->next = n->next;
532: if (n->next)
533: n->next->prev = n->prev;
534:
535: /* Adjust parent. */
536:
537: if (n->parent) {
538: n->parent->nchild--;
539: if (n->parent->child == n)
540: n->parent->child = n->prev ? n->prev : n->next;
1.169 kristaps 541: if (n->parent->last == n)
542: n->parent->last = n->prev ? n->prev : NULL;
1.121 kristaps 543: }
544:
545: /* Adjust parse point, if applicable. */
546:
1.203 schwarze 547: if (mdoc && mdoc->last == n) {
1.121 kristaps 548: if (n->prev) {
1.203 schwarze 549: mdoc->last = n->prev;
550: mdoc->next = MDOC_NEXT_SIBLING;
1.121 kristaps 551: } else {
1.203 schwarze 552: mdoc->last = n->parent;
553: mdoc->next = MDOC_NEXT_CHILD;
1.121 kristaps 554: }
555: }
556:
1.203 schwarze 557: if (mdoc && mdoc->first == n)
558: mdoc->first = NULL;
1.121 kristaps 559: }
560:
1.53 kristaps 561: void
1.203 schwarze 562: mdoc_node_delete(struct mdoc *mdoc, struct mdoc_node *p)
1.1 kristaps 563: {
564:
1.121 kristaps 565: while (p->child) {
566: assert(p->nchild);
1.203 schwarze 567: mdoc_node_delete(mdoc, p->child);
1.121 kristaps 568: }
569: assert(0 == p->nchild);
1.1 kristaps 570:
1.203 schwarze 571: mdoc_node_unlink(mdoc, p);
1.53 kristaps 572: mdoc_node_free(p);
1.201 schwarze 573: }
574:
1.231 schwarze 575: void
1.203 schwarze 576: mdoc_node_relink(struct mdoc *mdoc, struct mdoc_node *p)
1.201 schwarze 577: {
578:
1.203 schwarze 579: mdoc_node_unlink(mdoc, p);
1.230 schwarze 580: node_append(mdoc, p);
1.1 kristaps 581: }
582:
1.53 kristaps 583: /*
584: * Parse free-form text, that is, a line that does not begin with the
585: * control character.
586: */
587: static int
1.203 schwarze 588: mdoc_ptext(struct mdoc *mdoc, int line, char *buf, int offs)
1.1 kristaps 589: {
1.142 kristaps 590: char *c, *ws, *end;
591: struct mdoc_node *n;
592:
1.203 schwarze 593: assert(mdoc->last);
594: n = mdoc->last;
1.142 kristaps 595:
596: /*
1.144 kristaps 597: * Divert directly to list processing if we're encountering a
1.142 kristaps 598: * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
1.144 kristaps 599: * (a MDOC_BODY means it's already open, in which case we should
600: * process within its context in the normal way).
1.142 kristaps 601: */
602:
1.143 kristaps 603: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.213 schwarze 604: LIST_column == n->norm->Bl.type) {
1.144 kristaps 605: /* `Bl' is open without any children. */
1.203 schwarze 606: mdoc->flags |= MDOC_FREECOL;
1.232 schwarze 607: mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf);
608: return(1);
1.142 kristaps 609: }
610:
611: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
1.213 schwarze 612: NULL != n->parent &&
613: MDOC_Bl == n->parent->tok &&
614: LIST_column == n->parent->norm->Bl.type) {
1.144 kristaps 615: /* `Bl' has block-level `It' children. */
1.203 schwarze 616: mdoc->flags |= MDOC_FREECOL;
1.232 schwarze 617: mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf);
618: return(1);
1.142 kristaps 619: }
1.124 kristaps 620:
1.137 schwarze 621: /*
622: * Search for the beginning of unescaped trailing whitespace (ws)
623: * and for the first character not to be output (end).
624: */
1.139 kristaps 625:
626: /* FIXME: replace with strcspn(). */
1.137 schwarze 627: ws = NULL;
628: for (c = end = buf + offs; *c; c++) {
629: switch (*c) {
630: case ' ':
631: if (NULL == ws)
632: ws = c;
633: continue;
634: case '\t':
635: /*
636: * Always warn about trailing tabs,
637: * even outside literal context,
638: * where they should be put on the next line.
639: */
640: if (NULL == ws)
641: ws = c;
642: /*
643: * Strip trailing tabs in literal context only;
644: * outside, they affect the next line.
645: */
1.203 schwarze 646: if (MDOC_LITERAL & mdoc->flags)
1.137 schwarze 647: continue;
648: break;
649: case '\\':
650: /* Skip the escaped character, too, if any. */
651: if (c[1])
652: c++;
653: /* FALLTHROUGH */
654: default:
655: ws = NULL;
656: break;
657: }
658: end = c + 1;
659: }
660: *end = '\0';
1.91 kristaps 661:
1.137 schwarze 662: if (ws)
1.218 schwarze 663: mandoc_msg(MANDOCERR_SPACE_EOL, mdoc->parse,
664: line, (int)(ws-buf), NULL);
1.115 kristaps 665:
1.231 schwarze 666: if (buf[offs] == '\0' && ! (mdoc->flags & MDOC_LITERAL)) {
1.218 schwarze 667: mandoc_msg(MANDOCERR_FI_BLANK, mdoc->parse,
668: line, (int)(c - buf), NULL);
1.124 kristaps 669:
1.119 kristaps 670: /*
1.165 schwarze 671: * Insert a `sp' in the case of a blank line. Technically,
1.124 kristaps 672: * blank lines aren't allowed, but enough manuals assume this
673: * behaviour that we want to work around it.
1.119 kristaps 674: */
1.231 schwarze 675: mdoc_elem_alloc(mdoc, line, offs, MDOC_sp, NULL);
1.203 schwarze 676: mdoc->next = MDOC_NEXT_SIBLING;
1.230 schwarze 677: mdoc_valid_post(mdoc);
678: return(1);
1.119 kristaps 679: }
1.68 kristaps 680:
1.231 schwarze 681: mdoc_word_alloc(mdoc, line, offs, buf+offs);
1.91 kristaps 682:
1.231 schwarze 683: if (mdoc->flags & MDOC_LITERAL)
1.137 schwarze 684: return(1);
1.128 kristaps 685:
686: /*
687: * End-of-sentence check. If the last character is an unescaped
688: * EOS character, then flag the node as being the end of a
689: * sentence. The front-end will know how to interpret this.
690: */
1.132 kristaps 691:
1.137 schwarze 692: assert(buf < end);
693:
1.207 schwarze 694: if (mandoc_eos(buf+offs, (size_t)(end-buf-offs)))
1.203 schwarze 695: mdoc->last->flags |= MDOC_EOS;
1.128 kristaps 696: return(1);
1.1 kristaps 697: }
698:
1.53 kristaps 699: /*
700: * Parse a macro line, that is, a line beginning with the control
701: * character.
702: */
1.155 kristaps 703: static int
1.203 schwarze 704: mdoc_pmacro(struct mdoc *mdoc, int ln, char *buf, int offs)
1.1 kristaps 705: {
1.229 schwarze 706: struct mdoc_node *n;
707: const char *cp;
1.144 kristaps 708: enum mdoct tok;
1.188 kristaps 709: int i, sv;
1.144 kristaps 710: char mac[5];
1.63 kristaps 711:
1.188 kristaps 712: sv = offs;
1.130 kristaps 713:
1.213 schwarze 714: /*
1.162 schwarze 715: * Copy the first word into a nil-terminated buffer.
1.229 schwarze 716: * Stop when a space, tab, escape, or eoln is encountered.
1.160 kristaps 717: */
1.1 kristaps 718:
1.188 kristaps 719: i = 0;
1.229 schwarze 720: while (i < 4 && strchr(" \t\\", buf[offs]) == NULL)
1.188 kristaps 721: mac[i++] = buf[offs++];
722:
723: mac[i] = '\0';
724:
1.214 schwarze 725: tok = (i > 1 && i < 4) ? mdoc_hash_find(mac) : MDOC_MAX;
1.1 kristaps 726:
1.229 schwarze 727: if (tok == MDOC_MAX) {
1.222 schwarze 728: mandoc_msg(MANDOCERR_MACRO, mdoc->parse,
729: ln, sv, buf + sv - 1);
1.58 kristaps 730: return(1);
1.53 kristaps 731: }
1.1 kristaps 732:
1.229 schwarze 733: /* Skip a leading escape sequence or tab. */
1.160 kristaps 734:
1.229 schwarze 735: switch (buf[offs]) {
736: case '\\':
737: cp = buf + offs + 1;
738: mandoc_escape(&cp, NULL, NULL);
739: offs = cp - buf;
740: break;
741: case '\t':
1.188 kristaps 742: offs++;
1.229 schwarze 743: break;
744: default:
745: break;
746: }
1.160 kristaps 747:
748: /* Jump to the next non-whitespace word. */
1.1 kristaps 749:
1.188 kristaps 750: while (buf[offs] && ' ' == buf[offs])
751: offs++;
1.1 kristaps 752:
1.213 schwarze 753: /*
1.125 kristaps 754: * Trailing whitespace. Note that tabs are allowed to be passed
755: * into the parser as "text", so we only warn about spaces here.
756: */
1.115 kristaps 757:
1.188 kristaps 758: if ('\0' == buf[offs] && ' ' == buf[offs - 1])
1.218 schwarze 759: mandoc_msg(MANDOCERR_SPACE_EOL, mdoc->parse,
760: ln, offs - 1, NULL);
1.115 kristaps 761:
1.144 kristaps 762: /*
763: * If an initial macro or a list invocation, divert directly
764: * into macro processing.
765: */
766:
1.232 schwarze 767: if (NULL == mdoc->last || MDOC_It == tok || MDOC_El == tok) {
768: mdoc_macro(mdoc, tok, ln, sv, &offs, buf);
769: return(1);
770: }
1.144 kristaps 771:
1.203 schwarze 772: n = mdoc->last;
773: assert(mdoc->last);
1.144 kristaps 774:
775: /*
776: * If the first macro of a `Bl -column', open an `It' block
777: * context around the parsed macro.
778: */
779:
780: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.213 schwarze 781: LIST_column == n->norm->Bl.type) {
1.203 schwarze 782: mdoc->flags |= MDOC_FREECOL;
1.232 schwarze 783: mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf);
784: return(1);
1.144 kristaps 785: }
786:
787: /*
788: * If we're following a block-level `It' within a `Bl -column'
789: * context (perhaps opened in the above block or in ptext()),
790: * then open an `It' block context around the parsed macro.
1.98 kristaps 791: */
1.144 kristaps 792:
793: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
1.213 schwarze 794: NULL != n->parent &&
795: MDOC_Bl == n->parent->tok &&
796: LIST_column == n->parent->norm->Bl.type) {
1.203 schwarze 797: mdoc->flags |= MDOC_FREECOL;
1.232 schwarze 798: mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf);
799: return(1);
1.144 kristaps 800: }
801:
802: /* Normal processing of a macro. */
803:
1.232 schwarze 804: mdoc_macro(mdoc, tok, ln, sv, &offs, buf);
1.208 schwarze 805:
806: /* In quick mode (for mandocdb), abort after the NAME section. */
807:
808: if (mdoc->quick && MDOC_Sh == tok &&
809: SEC_NAME != mdoc->last->sec)
810: return(2);
1.1 kristaps 811:
1.53 kristaps 812: return(1);
1.1 kristaps 813: }
1.100 kristaps 814:
1.186 kristaps 815: enum mdelim
816: mdoc_isdelim(const char *p)
817: {
818:
819: if ('\0' == p[0])
820: return(DELIM_NONE);
821:
822: if ('\0' == p[1])
823: switch (p[0]) {
1.213 schwarze 824: case '(':
1.186 kristaps 825: /* FALLTHROUGH */
1.213 schwarze 826: case '[':
1.186 kristaps 827: return(DELIM_OPEN);
1.213 schwarze 828: case '|':
1.186 kristaps 829: return(DELIM_MIDDLE);
1.213 schwarze 830: case '.':
1.186 kristaps 831: /* FALLTHROUGH */
1.213 schwarze 832: case ',':
1.186 kristaps 833: /* FALLTHROUGH */
1.213 schwarze 834: case ';':
1.186 kristaps 835: /* FALLTHROUGH */
1.213 schwarze 836: case ':':
1.186 kristaps 837: /* FALLTHROUGH */
1.213 schwarze 838: case '?':
1.186 kristaps 839: /* FALLTHROUGH */
1.213 schwarze 840: case '!':
1.186 kristaps 841: /* FALLTHROUGH */
1.213 schwarze 842: case ')':
1.186 kristaps 843: /* FALLTHROUGH */
1.213 schwarze 844: case ']':
1.186 kristaps 845: return(DELIM_CLOSE);
846: default:
847: return(DELIM_NONE);
848: }
849:
850: if ('\\' != p[0])
851: return(DELIM_NONE);
1.100 kristaps 852:
1.186 kristaps 853: if (0 == strcmp(p + 1, "."))
854: return(DELIM_CLOSE);
1.200 schwarze 855: if (0 == strcmp(p + 1, "fR|\\fP"))
1.186 kristaps 856: return(DELIM_MIDDLE);
857:
858: return(DELIM_NONE);
1.211 schwarze 859: }
860:
861: void
862: mdoc_deroff(char **dest, const struct mdoc_node *n)
863: {
864: char *cp;
865: size_t sz;
866:
867: if (MDOC_TEXT != n->type) {
868: for (n = n->child; n; n = n->next)
869: mdoc_deroff(dest, n);
870: return;
871: }
872:
873: /* Skip leading whitespace. */
874:
875: for (cp = n->string; '\0' != *cp; cp++)
876: if (0 == isspace((unsigned char)*cp))
877: break;
878:
879: /* Skip trailing whitespace. */
880:
881: for (sz = strlen(cp); sz; sz--)
882: if (0 == isspace((unsigned char)cp[sz-1]))
883: break;
884:
885: /* Skip empty strings. */
886:
887: if (0 == sz)
888: return;
889:
890: if (NULL == *dest) {
891: *dest = mandoc_strndup(cp, sz);
892: return;
893: }
894:
895: mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
896: free(*dest);
897: *dest = cp;
1.186 kristaps 898: }
CVSweb