Annotation of mandoc/mdoc.c, Revision 1.238
1.238 ! schwarze 1: /* $Id: mdoc.c,v 1.237 2015/02/12 12:24:33 schwarze Exp $ */
1.1 kristaps 2: /*
1.182 schwarze 3: * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
1.237 schwarze 4: * Copyright (c) 2010, 2012-2015 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:
1.234 schwarze 193: void
1.203 schwarze 194: mdoc_endparse(struct mdoc *mdoc)
1.20 kristaps 195: {
196:
1.232 schwarze 197: mdoc_macroend(mdoc);
1.181 kristaps 198: }
199:
1.233 schwarze 200: void
1.203 schwarze 201: mdoc_addeqn(struct mdoc *mdoc, const struct eqn *ep)
1.181 kristaps 202: {
203: struct mdoc_node *n;
204:
1.203 schwarze 205: n = node_alloc(mdoc, ep->ln, ep->pos, MDOC_MAX, MDOC_EQN);
1.181 kristaps 206: n->eqn = ep;
1.227 schwarze 207: if (ep->ln > mdoc->last->line)
208: n->flags |= MDOC_LINE;
1.230 schwarze 209: node_append(mdoc, n);
1.203 schwarze 210: mdoc->next = MDOC_NEXT_SIBLING;
1.20 kristaps 211: }
212:
1.233 schwarze 213: void
1.203 schwarze 214: mdoc_addspan(struct mdoc *mdoc, const struct tbl_span *sp)
1.175 kristaps 215: {
1.180 kristaps 216: struct mdoc_node *n;
1.175 kristaps 217:
1.203 schwarze 218: n = node_alloc(mdoc, sp->line, 0, MDOC_MAX, MDOC_TBL);
1.180 kristaps 219: n->span = sp;
1.230 schwarze 220: node_append(mdoc, n);
1.203 schwarze 221: mdoc->next = MDOC_NEXT_SIBLING;
1.175 kristaps 222: }
223:
1.50 kristaps 224: /*
1.53 kristaps 225: * Main parse routine. Parses a single line -- really just hands off to
1.123 kristaps 226: * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
1.50 kristaps 227: */
1.20 kristaps 228: int
1.203 schwarze 229: mdoc_parseln(struct mdoc *mdoc, int ln, char *buf, int offs)
1.1 kristaps 230: {
231:
1.228 schwarze 232: if (mdoc->last->type != MDOC_EQN || ln > mdoc->last->line)
233: mdoc->flags |= MDOC_NEWLINE;
1.153 schwarze 234:
235: /*
236: * Let the roff nS register switch SYNOPSIS mode early,
237: * such that the parser knows at all times
238: * whether this mode is on or off.
239: * Note that this mode is also switched by the Sh macro.
240: */
1.204 schwarze 241: if (roff_getreg(mdoc->roff, "nS"))
242: mdoc->flags |= MDOC_SYNOPSIS;
243: else
244: mdoc->flags &= ~MDOC_SYNOPSIS;
1.153 schwarze 245:
1.203 schwarze 246: return(roff_getcontrol(mdoc->roff, buf, &offs) ?
1.213 schwarze 247: mdoc_pmacro(mdoc, ln, buf, offs) :
248: mdoc_ptext(mdoc, ln, buf, offs));
1.1 kristaps 249: }
250:
1.232 schwarze 251: void
1.148 kristaps 252: mdoc_macro(MACRO_PROT_ARGS)
1.88 kristaps 253: {
1.122 kristaps 254: assert(tok < MDOC_MAX);
255:
1.223 schwarze 256: if (mdoc->flags & MDOC_PBODY) {
257: if (tok == MDOC_Dt) {
258: mandoc_vmsg(MANDOCERR_DT_LATE,
259: mdoc->parse, line, ppos,
260: "Dt %s", buf + *pos);
1.232 schwarze 261: return;
1.223 schwarze 262: }
263: } else if ( ! (mdoc_macros[tok].flags & MDOC_PROLOGUE)) {
264: if (mdoc->meta.title == NULL) {
265: mandoc_vmsg(MANDOCERR_DT_NOTITLE,
266: mdoc->parse, line, ppos, "%s %s",
267: mdoc_macronames[tok], buf + *pos);
268: mdoc->meta.title = mandoc_strdup("UNTITLED");
269: }
1.203 schwarze 270: if (NULL == mdoc->meta.vol)
271: mdoc->meta.vol = mandoc_strdup("LOCAL");
272: mdoc->flags |= MDOC_PBODY;
1.120 kristaps 273: }
1.232 schwarze 274: (*mdoc_macros[tok].fp)(mdoc, tok, line, ppos, pos, buf);
1.73 kristaps 275: }
276:
277:
1.230 schwarze 278: static void
1.73 kristaps 279: node_append(struct mdoc *mdoc, struct mdoc_node *p)
1.1 kristaps 280: {
281:
1.25 kristaps 282: assert(mdoc->last);
283: assert(mdoc->first);
284: assert(MDOC_ROOT != p->type);
1.1 kristaps 285:
1.13 kristaps 286: switch (mdoc->next) {
1.213 schwarze 287: case MDOC_NEXT_SIBLING:
1.6 kristaps 288: mdoc->last->next = p;
289: p->prev = mdoc->last;
1.13 kristaps 290: p->parent = mdoc->last->parent;
1.1 kristaps 291: break;
1.213 schwarze 292: case MDOC_NEXT_CHILD:
1.13 kristaps 293: mdoc->last->child = p;
1.1 kristaps 294: p->parent = mdoc->last;
295: break;
296: default:
1.13 kristaps 297: abort();
298: /* NOTREACHED */
1.1 kristaps 299: }
300:
1.86 kristaps 301: p->parent->nchild++;
302:
1.172 kristaps 303: /*
304: * Copy over the normalised-data pointer of our parent. Not
305: * everybody has one, but copying a null pointer is fine.
306: */
307:
308: switch (p->type) {
1.213 schwarze 309: case MDOC_BODY:
1.202 schwarze 310: if (ENDBODY_NOT != p->end)
311: break;
1.172 kristaps 312: /* FALLTHROUGH */
1.213 schwarze 313: case MDOC_TAIL:
1.172 kristaps 314: /* FALLTHROUGH */
1.213 schwarze 315: case MDOC_HEAD:
1.172 kristaps 316: p->norm = p->parent->norm;
317: break;
318: default:
319: break;
320: }
321:
1.230 schwarze 322: mdoc_valid_pre(mdoc, p);
1.27 kristaps 323:
324: switch (p->type) {
1.213 schwarze 325: case MDOC_HEAD:
1.27 kristaps 326: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 327: p->parent->head = p;
1.27 kristaps 328: break;
1.213 schwarze 329: case MDOC_TAIL:
1.27 kristaps 330: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 331: p->parent->tail = p;
1.27 kristaps 332: break;
1.213 schwarze 333: case MDOC_BODY:
1.152 schwarze 334: if (p->end)
335: break;
1.27 kristaps 336: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 337: p->parent->body = p;
1.27 kristaps 338: break;
339: default:
340: break;
341: }
342:
1.1 kristaps 343: mdoc->last = p;
1.71 kristaps 344:
345: switch (p->type) {
1.213 schwarze 346: case MDOC_TBL:
1.176 kristaps 347: /* FALLTHROUGH */
1.213 schwarze 348: case MDOC_TEXT:
1.230 schwarze 349: mdoc_valid_post(mdoc);
1.71 kristaps 350: break;
351: default:
352: break;
353: }
1.1 kristaps 354: }
355:
1.45 kristaps 356: static struct mdoc_node *
1.213 schwarze 357: node_alloc(struct mdoc *mdoc, int line, int pos,
1.117 kristaps 358: enum mdoct tok, enum mdoc_type type)
1.45 kristaps 359: {
1.46 kristaps 360: struct mdoc_node *p;
361:
1.113 kristaps 362: p = mandoc_calloc(1, sizeof(struct mdoc_node));
1.203 schwarze 363: p->sec = mdoc->lastsec;
1.73 kristaps 364: p->line = line;
365: p->pos = pos;
366: p->tok = tok;
1.118 kristaps 367: p->type = type;
1.150 kristaps 368:
369: /* Flag analysis. */
370:
1.203 schwarze 371: if (MDOC_SYNOPSIS & mdoc->flags)
1.153 schwarze 372: p->flags |= MDOC_SYNPRETTY;
373: else
374: p->flags &= ~MDOC_SYNPRETTY;
1.203 schwarze 375: if (MDOC_NEWLINE & mdoc->flags)
1.130 kristaps 376: p->flags |= MDOC_LINE;
1.203 schwarze 377: mdoc->flags &= ~MDOC_NEWLINE;
1.150 kristaps 378:
1.46 kristaps 379: return(p);
1.45 kristaps 380: }
381:
1.231 schwarze 382: void
1.203 schwarze 383: mdoc_tail_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.17 kristaps 384: {
385: struct mdoc_node *p;
386:
1.203 schwarze 387: p = node_alloc(mdoc, line, pos, tok, MDOC_TAIL);
1.230 schwarze 388: node_append(mdoc, p);
1.203 schwarze 389: mdoc->next = MDOC_NEXT_CHILD;
1.17 kristaps 390: }
391:
1.231 schwarze 392: struct mdoc_node *
1.203 schwarze 393: mdoc_head_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.1 kristaps 394: {
395: struct mdoc_node *p;
396:
1.203 schwarze 397: assert(mdoc->first);
398: assert(mdoc->last);
399: p = node_alloc(mdoc, line, pos, tok, MDOC_HEAD);
1.230 schwarze 400: node_append(mdoc, p);
1.203 schwarze 401: mdoc->next = MDOC_NEXT_CHILD;
1.231 schwarze 402: return(p);
1.1 kristaps 403: }
404:
1.231 schwarze 405: struct mdoc_node *
1.203 schwarze 406: mdoc_body_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.1 kristaps 407: {
408: struct mdoc_node *p;
409:
1.203 schwarze 410: p = node_alloc(mdoc, line, pos, tok, MDOC_BODY);
1.230 schwarze 411: node_append(mdoc, p);
1.203 schwarze 412: mdoc->next = MDOC_NEXT_CHILD;
1.231 schwarze 413: return(p);
1.152 schwarze 414: }
415:
1.235 schwarze 416: struct mdoc_node *
1.203 schwarze 417: mdoc_endbody_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok,
1.152 schwarze 418: struct mdoc_node *body, enum mdoc_endbody end)
419: {
420: struct mdoc_node *p;
421:
1.237 schwarze 422: body->flags |= MDOC_ENDED;
423: body->parent->flags |= MDOC_ENDED;
1.203 schwarze 424: p = node_alloc(mdoc, line, pos, tok, MDOC_BODY);
1.237 schwarze 425: p->body = 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;
1.235 schwarze 430: return(p);
1.1 kristaps 431: }
432:
1.231 schwarze 433: struct mdoc_node *
1.213 schwarze 434: mdoc_block_alloc(struct mdoc *mdoc, int line, int pos,
1.117 kristaps 435: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 436: {
437: struct mdoc_node *p;
438:
1.203 schwarze 439: p = node_alloc(mdoc, line, pos, tok, MDOC_BLOCK);
1.77 kristaps 440: p->args = args;
441: if (p->args)
1.53 kristaps 442: (args->refcnt)++;
1.172 kristaps 443:
444: switch (tok) {
1.213 schwarze 445: case MDOC_Bd:
1.172 kristaps 446: /* FALLTHROUGH */
1.213 schwarze 447: case MDOC_Bf:
1.172 kristaps 448: /* FALLTHROUGH */
1.213 schwarze 449: case MDOC_Bl:
1.217 schwarze 450: /* FALLTHROUGH */
451: case MDOC_En:
1.173 kristaps 452: /* FALLTHROUGH */
1.213 schwarze 453: case MDOC_Rs:
1.172 kristaps 454: p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
455: break;
456: default:
457: break;
458: }
1.230 schwarze 459: node_append(mdoc, p);
1.203 schwarze 460: mdoc->next = MDOC_NEXT_CHILD;
1.231 schwarze 461: return(p);
1.1 kristaps 462: }
463:
1.231 schwarze 464: void
1.213 schwarze 465: mdoc_elem_alloc(struct mdoc *mdoc, int line, int pos,
1.117 kristaps 466: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 467: {
468: struct mdoc_node *p;
469:
1.203 schwarze 470: p = node_alloc(mdoc, line, pos, tok, MDOC_ELEM);
1.77 kristaps 471: p->args = args;
472: if (p->args)
1.53 kristaps 473: (args->refcnt)++;
1.172 kristaps 474:
475: switch (tok) {
1.213 schwarze 476: case MDOC_An:
1.172 kristaps 477: p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
478: break;
479: default:
480: break;
481: }
1.230 schwarze 482: node_append(mdoc, p);
1.203 schwarze 483: mdoc->next = MDOC_NEXT_CHILD;
1.175 kristaps 484: }
1.1 kristaps 485:
1.231 schwarze 486: void
1.203 schwarze 487: mdoc_word_alloc(struct mdoc *mdoc, int line, int pos, const char *p)
1.1 kristaps 488: {
1.91 kristaps 489: struct mdoc_node *n;
1.1 kristaps 490:
1.203 schwarze 491: n = node_alloc(mdoc, line, pos, MDOC_MAX, MDOC_TEXT);
492: n->string = roff_strdup(mdoc->roff, p);
1.230 schwarze 493: node_append(mdoc, n);
1.203 schwarze 494: mdoc->next = MDOC_NEXT_SIBLING;
1.91 kristaps 495: }
496:
1.205 schwarze 497: void
498: mdoc_word_append(struct mdoc *mdoc, const char *p)
499: {
500: struct mdoc_node *n;
501: char *addstr, *newstr;
502:
503: n = mdoc->last;
504: addstr = roff_strdup(mdoc->roff, p);
1.210 schwarze 505: mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
1.205 schwarze 506: free(addstr);
507: free(n->string);
508: n->string = newstr;
509: mdoc->next = MDOC_NEXT_SIBLING;
510: }
1.91 kristaps 511:
1.155 kristaps 512: static void
1.53 kristaps 513: mdoc_node_free(struct mdoc_node *p)
1.1 kristaps 514: {
515:
1.172 kristaps 516: if (MDOC_BLOCK == p->type || MDOC_ELEM == p->type)
1.171 kristaps 517: free(p->norm);
1.53 kristaps 518: if (p->string)
519: free(p->string);
520: if (p->args)
521: mdoc_argv_free(p->args);
1.1 kristaps 522: free(p);
523: }
524:
1.121 kristaps 525: static void
1.203 schwarze 526: mdoc_node_unlink(struct mdoc *mdoc, struct mdoc_node *n)
1.121 kristaps 527: {
528:
529: /* Adjust siblings. */
530:
531: if (n->prev)
532: n->prev->next = n->next;
533: if (n->next)
534: n->next->prev = n->prev;
535:
536: /* Adjust parent. */
537:
538: if (n->parent) {
539: n->parent->nchild--;
540: if (n->parent->child == n)
541: n->parent->child = n->prev ? n->prev : n->next;
1.169 kristaps 542: if (n->parent->last == n)
543: n->parent->last = n->prev ? n->prev : NULL;
1.121 kristaps 544: }
545:
546: /* Adjust parse point, if applicable. */
547:
1.203 schwarze 548: if (mdoc && mdoc->last == n) {
1.121 kristaps 549: if (n->prev) {
1.203 schwarze 550: mdoc->last = n->prev;
551: mdoc->next = MDOC_NEXT_SIBLING;
1.121 kristaps 552: } else {
1.203 schwarze 553: mdoc->last = n->parent;
554: mdoc->next = MDOC_NEXT_CHILD;
1.121 kristaps 555: }
556: }
557:
1.203 schwarze 558: if (mdoc && mdoc->first == n)
559: mdoc->first = NULL;
1.121 kristaps 560: }
561:
1.53 kristaps 562: void
1.203 schwarze 563: mdoc_node_delete(struct mdoc *mdoc, struct mdoc_node *p)
1.1 kristaps 564: {
565:
1.121 kristaps 566: while (p->child) {
567: assert(p->nchild);
1.203 schwarze 568: mdoc_node_delete(mdoc, p->child);
1.121 kristaps 569: }
570: assert(0 == p->nchild);
1.1 kristaps 571:
1.203 schwarze 572: mdoc_node_unlink(mdoc, p);
1.53 kristaps 573: mdoc_node_free(p);
1.201 schwarze 574: }
575:
1.231 schwarze 576: void
1.203 schwarze 577: mdoc_node_relink(struct mdoc *mdoc, struct mdoc_node *p)
1.201 schwarze 578: {
579:
1.203 schwarze 580: mdoc_node_unlink(mdoc, p);
1.230 schwarze 581: node_append(mdoc, p);
1.1 kristaps 582: }
583:
1.53 kristaps 584: /*
585: * Parse free-form text, that is, a line that does not begin with the
586: * control character.
587: */
588: static int
1.203 schwarze 589: mdoc_ptext(struct mdoc *mdoc, int line, char *buf, int offs)
1.1 kristaps 590: {
1.142 kristaps 591: char *c, *ws, *end;
592: struct mdoc_node *n;
593:
1.203 schwarze 594: assert(mdoc->last);
595: n = mdoc->last;
1.142 kristaps 596:
597: /*
1.144 kristaps 598: * Divert directly to list processing if we're encountering a
1.142 kristaps 599: * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
1.144 kristaps 600: * (a MDOC_BODY means it's already open, in which case we should
601: * process within its context in the normal way).
1.142 kristaps 602: */
603:
1.238 ! schwarze 604: if (n->tok == MDOC_Bl && n->type == MDOC_BODY &&
! 605: n->end == ENDBODY_NOT && n->norm->Bl.type == LIST_column) {
1.144 kristaps 606: /* `Bl' is open without any children. */
1.203 schwarze 607: mdoc->flags |= MDOC_FREECOL;
1.232 schwarze 608: mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf);
609: return(1);
1.142 kristaps 610: }
611:
612: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
1.213 schwarze 613: NULL != n->parent &&
614: MDOC_Bl == n->parent->tok &&
615: LIST_column == n->parent->norm->Bl.type) {
1.144 kristaps 616: /* `Bl' has block-level `It' children. */
1.203 schwarze 617: mdoc->flags |= MDOC_FREECOL;
1.232 schwarze 618: mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf);
619: return(1);
1.142 kristaps 620: }
1.124 kristaps 621:
1.137 schwarze 622: /*
623: * Search for the beginning of unescaped trailing whitespace (ws)
624: * and for the first character not to be output (end).
625: */
1.139 kristaps 626:
627: /* FIXME: replace with strcspn(). */
1.137 schwarze 628: ws = NULL;
629: for (c = end = buf + offs; *c; c++) {
630: switch (*c) {
631: case ' ':
632: if (NULL == ws)
633: ws = c;
634: continue;
635: case '\t':
636: /*
637: * Always warn about trailing tabs,
638: * even outside literal context,
639: * where they should be put on the next line.
640: */
641: if (NULL == ws)
642: ws = c;
643: /*
644: * Strip trailing tabs in literal context only;
645: * outside, they affect the next line.
646: */
1.203 schwarze 647: if (MDOC_LITERAL & mdoc->flags)
1.137 schwarze 648: continue;
649: break;
650: case '\\':
651: /* Skip the escaped character, too, if any. */
652: if (c[1])
653: c++;
654: /* FALLTHROUGH */
655: default:
656: ws = NULL;
657: break;
658: }
659: end = c + 1;
660: }
661: *end = '\0';
1.91 kristaps 662:
1.137 schwarze 663: if (ws)
1.218 schwarze 664: mandoc_msg(MANDOCERR_SPACE_EOL, mdoc->parse,
665: line, (int)(ws-buf), NULL);
1.115 kristaps 666:
1.231 schwarze 667: if (buf[offs] == '\0' && ! (mdoc->flags & MDOC_LITERAL)) {
1.218 schwarze 668: mandoc_msg(MANDOCERR_FI_BLANK, mdoc->parse,
669: line, (int)(c - buf), NULL);
1.124 kristaps 670:
1.119 kristaps 671: /*
1.165 schwarze 672: * Insert a `sp' in the case of a blank line. Technically,
1.124 kristaps 673: * blank lines aren't allowed, but enough manuals assume this
674: * behaviour that we want to work around it.
1.119 kristaps 675: */
1.231 schwarze 676: mdoc_elem_alloc(mdoc, line, offs, MDOC_sp, NULL);
1.203 schwarze 677: mdoc->next = MDOC_NEXT_SIBLING;
1.230 schwarze 678: mdoc_valid_post(mdoc);
679: return(1);
1.119 kristaps 680: }
1.68 kristaps 681:
1.231 schwarze 682: mdoc_word_alloc(mdoc, line, offs, buf+offs);
1.91 kristaps 683:
1.231 schwarze 684: if (mdoc->flags & MDOC_LITERAL)
1.137 schwarze 685: return(1);
1.128 kristaps 686:
687: /*
688: * End-of-sentence check. If the last character is an unescaped
689: * EOS character, then flag the node as being the end of a
690: * sentence. The front-end will know how to interpret this.
691: */
1.132 kristaps 692:
1.137 schwarze 693: assert(buf < end);
694:
1.207 schwarze 695: if (mandoc_eos(buf+offs, (size_t)(end-buf-offs)))
1.203 schwarze 696: mdoc->last->flags |= MDOC_EOS;
1.128 kristaps 697: return(1);
1.1 kristaps 698: }
699:
1.53 kristaps 700: /*
701: * Parse a macro line, that is, a line beginning with the control
702: * character.
703: */
1.155 kristaps 704: static int
1.203 schwarze 705: mdoc_pmacro(struct mdoc *mdoc, int ln, char *buf, int offs)
1.1 kristaps 706: {
1.229 schwarze 707: struct mdoc_node *n;
708: const char *cp;
1.144 kristaps 709: enum mdoct tok;
1.188 kristaps 710: int i, sv;
1.144 kristaps 711: char mac[5];
1.63 kristaps 712:
1.188 kristaps 713: sv = offs;
1.130 kristaps 714:
1.213 schwarze 715: /*
1.162 schwarze 716: * Copy the first word into a nil-terminated buffer.
1.229 schwarze 717: * Stop when a space, tab, escape, or eoln is encountered.
1.160 kristaps 718: */
1.1 kristaps 719:
1.188 kristaps 720: i = 0;
1.229 schwarze 721: while (i < 4 && strchr(" \t\\", buf[offs]) == NULL)
1.188 kristaps 722: mac[i++] = buf[offs++];
723:
724: mac[i] = '\0';
725:
1.214 schwarze 726: tok = (i > 1 && i < 4) ? mdoc_hash_find(mac) : MDOC_MAX;
1.1 kristaps 727:
1.229 schwarze 728: if (tok == MDOC_MAX) {
1.222 schwarze 729: mandoc_msg(MANDOCERR_MACRO, mdoc->parse,
730: ln, sv, buf + sv - 1);
1.58 kristaps 731: return(1);
1.53 kristaps 732: }
1.1 kristaps 733:
1.229 schwarze 734: /* Skip a leading escape sequence or tab. */
1.160 kristaps 735:
1.229 schwarze 736: switch (buf[offs]) {
737: case '\\':
738: cp = buf + offs + 1;
739: mandoc_escape(&cp, NULL, NULL);
740: offs = cp - buf;
741: break;
742: case '\t':
1.188 kristaps 743: offs++;
1.229 schwarze 744: break;
745: default:
746: break;
747: }
1.160 kristaps 748:
749: /* Jump to the next non-whitespace word. */
1.1 kristaps 750:
1.188 kristaps 751: while (buf[offs] && ' ' == buf[offs])
752: offs++;
1.1 kristaps 753:
1.213 schwarze 754: /*
1.125 kristaps 755: * Trailing whitespace. Note that tabs are allowed to be passed
756: * into the parser as "text", so we only warn about spaces here.
757: */
1.115 kristaps 758:
1.188 kristaps 759: if ('\0' == buf[offs] && ' ' == buf[offs - 1])
1.218 schwarze 760: mandoc_msg(MANDOCERR_SPACE_EOL, mdoc->parse,
761: ln, offs - 1, NULL);
1.115 kristaps 762:
1.144 kristaps 763: /*
764: * If an initial macro or a list invocation, divert directly
765: * into macro processing.
766: */
767:
1.232 schwarze 768: if (NULL == mdoc->last || MDOC_It == tok || MDOC_El == tok) {
769: mdoc_macro(mdoc, tok, ln, sv, &offs, buf);
770: return(1);
771: }
1.144 kristaps 772:
1.203 schwarze 773: n = mdoc->last;
774: assert(mdoc->last);
1.144 kristaps 775:
776: /*
777: * If the first macro of a `Bl -column', open an `It' block
778: * context around the parsed macro.
779: */
780:
1.238 ! schwarze 781: if (n->tok == MDOC_Bl && n->type == MDOC_BODY &&
! 782: n->end == ENDBODY_NOT && n->norm->Bl.type == LIST_column) {
1.203 schwarze 783: mdoc->flags |= MDOC_FREECOL;
1.232 schwarze 784: mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf);
785: return(1);
1.144 kristaps 786: }
787:
788: /*
789: * If we're following a block-level `It' within a `Bl -column'
790: * context (perhaps opened in the above block or in ptext()),
791: * then open an `It' block context around the parsed macro.
1.98 kristaps 792: */
1.144 kristaps 793:
794: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
1.213 schwarze 795: NULL != n->parent &&
796: MDOC_Bl == n->parent->tok &&
797: LIST_column == n->parent->norm->Bl.type) {
1.203 schwarze 798: mdoc->flags |= MDOC_FREECOL;
1.232 schwarze 799: mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf);
800: return(1);
1.144 kristaps 801: }
802:
803: /* Normal processing of a macro. */
804:
1.232 schwarze 805: mdoc_macro(mdoc, tok, ln, sv, &offs, buf);
1.208 schwarze 806:
807: /* In quick mode (for mandocdb), abort after the NAME section. */
808:
809: if (mdoc->quick && MDOC_Sh == tok &&
810: SEC_NAME != mdoc->last->sec)
811: return(2);
1.1 kristaps 812:
1.53 kristaps 813: return(1);
1.1 kristaps 814: }
1.100 kristaps 815:
1.186 kristaps 816: enum mdelim
817: mdoc_isdelim(const char *p)
818: {
819:
820: if ('\0' == p[0])
821: return(DELIM_NONE);
822:
823: if ('\0' == p[1])
824: switch (p[0]) {
1.213 schwarze 825: case '(':
1.186 kristaps 826: /* FALLTHROUGH */
1.213 schwarze 827: case '[':
1.186 kristaps 828: return(DELIM_OPEN);
1.213 schwarze 829: case '|':
1.186 kristaps 830: return(DELIM_MIDDLE);
1.213 schwarze 831: case '.':
1.186 kristaps 832: /* FALLTHROUGH */
1.213 schwarze 833: case ',':
1.186 kristaps 834: /* FALLTHROUGH */
1.213 schwarze 835: case ';':
1.186 kristaps 836: /* FALLTHROUGH */
1.213 schwarze 837: case ':':
1.186 kristaps 838: /* FALLTHROUGH */
1.213 schwarze 839: case '?':
1.186 kristaps 840: /* FALLTHROUGH */
1.213 schwarze 841: case '!':
1.186 kristaps 842: /* FALLTHROUGH */
1.213 schwarze 843: case ')':
1.186 kristaps 844: /* FALLTHROUGH */
1.213 schwarze 845: case ']':
1.186 kristaps 846: return(DELIM_CLOSE);
847: default:
848: return(DELIM_NONE);
849: }
850:
851: if ('\\' != p[0])
852: return(DELIM_NONE);
1.100 kristaps 853:
1.186 kristaps 854: if (0 == strcmp(p + 1, "."))
855: return(DELIM_CLOSE);
1.200 schwarze 856: if (0 == strcmp(p + 1, "fR|\\fP"))
1.186 kristaps 857: return(DELIM_MIDDLE);
858:
859: return(DELIM_NONE);
1.211 schwarze 860: }
861:
862: void
863: mdoc_deroff(char **dest, const struct mdoc_node *n)
864: {
865: char *cp;
866: size_t sz;
867:
868: if (MDOC_TEXT != n->type) {
869: for (n = n->child; n; n = n->next)
870: mdoc_deroff(dest, n);
871: return;
872: }
873:
874: /* Skip leading whitespace. */
875:
876: for (cp = n->string; '\0' != *cp; cp++)
877: if (0 == isspace((unsigned char)*cp))
878: break;
879:
880: /* Skip trailing whitespace. */
881:
882: for (sz = strlen(cp); sz; sz--)
883: if (0 == isspace((unsigned char)cp[sz-1]))
884: break;
885:
886: /* Skip empty strings. */
887:
888: if (0 == sz)
889: return;
890:
891: if (NULL == *dest) {
892: *dest = mandoc_strndup(cp, sz);
893: return;
894: }
895:
896: mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
897: free(*dest);
898: *dest = cp;
1.186 kristaps 899: }
CVSweb