Annotation of mandoc/mdoc.c, Revision 1.226
1.226 ! schwarze 1: /* $Id: mdoc.c,v 1.225 2014/09/06 23:24:32 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);
1.213 schwarze 92: static int node_append(struct mdoc *,
1.71 kristaps 93: struct mdoc_node *);
1.149 kristaps 94: static int mdoc_ptext(struct mdoc *, int, char *, int);
95: static int mdoc_pmacro(struct mdoc *, int, char *, int);
1.88 kristaps 96:
1.213 schwarze 97:
1.1 kristaps 98: const struct mdoc_node *
1.203 schwarze 99: mdoc_node(const struct mdoc *mdoc)
1.1 kristaps 100: {
101:
1.203 schwarze 102: return(mdoc->first);
1.1 kristaps 103: }
104:
1.37 kristaps 105: const struct mdoc_meta *
1.203 schwarze 106: mdoc_meta(const struct mdoc *mdoc)
1.37 kristaps 107: {
108:
1.203 schwarze 109: return(&mdoc->meta);
1.37 kristaps 110: }
111:
1.85 kristaps 112: /*
113: * Frees volatile resources (parse tree, meta-data, fields).
114: */
1.73 kristaps 115: static void
116: mdoc_free1(struct mdoc *mdoc)
1.67 kristaps 117: {
118:
119: if (mdoc->first)
1.121 kristaps 120: mdoc_node_delete(mdoc, mdoc->first);
1.223 schwarze 121: free(mdoc->meta.msec);
122: free(mdoc->meta.vol);
123: free(mdoc->meta.arch);
124: free(mdoc->meta.date);
125: free(mdoc->meta.title);
126: free(mdoc->meta.os);
127: free(mdoc->meta.name);
1.73 kristaps 128: }
129:
1.85 kristaps 130: /*
131: * Allocate all volatile resources (parse tree, meta-data, fields).
132: */
1.113 kristaps 133: static void
1.73 kristaps 134: mdoc_alloc1(struct mdoc *mdoc)
135: {
1.67 kristaps 136:
1.112 kristaps 137: memset(&mdoc->meta, 0, sizeof(struct mdoc_meta));
1.67 kristaps 138: mdoc->flags = 0;
1.85 kristaps 139: mdoc->lastnamed = mdoc->lastsec = SEC_NONE;
1.113 kristaps 140: mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node));
1.70 kristaps 141: mdoc->first = mdoc->last;
1.67 kristaps 142: mdoc->last->type = MDOC_ROOT;
1.196 schwarze 143: mdoc->last->tok = MDOC_MAX;
1.67 kristaps 144: mdoc->next = MDOC_NEXT_CHILD;
1.73 kristaps 145: }
146:
147: /*
1.85 kristaps 148: * Free up volatile resources (see mdoc_free1()) then re-initialises the
149: * data with mdoc_alloc1(). After invocation, parse data has been reset
150: * and the parser is ready for re-invocation on a new tree; however,
151: * cross-parse non-volatile data is kept intact.
1.73 kristaps 152: */
1.113 kristaps 153: void
1.73 kristaps 154: mdoc_reset(struct mdoc *mdoc)
155: {
156:
157: mdoc_free1(mdoc);
1.113 kristaps 158: mdoc_alloc1(mdoc);
1.67 kristaps 159: }
160:
1.68 kristaps 161: /*
1.85 kristaps 162: * Completely free up all volatile and non-volatile parse resources.
163: * After invocation, the pointer is no longer usable.
1.68 kristaps 164: */
1.67 kristaps 165: void
1.38 kristaps 166: mdoc_free(struct mdoc *mdoc)
1.34 kristaps 167: {
168:
1.73 kristaps 169: mdoc_free1(mdoc);
1.1 kristaps 170: free(mdoc);
171: }
172:
1.85 kristaps 173: /*
1.213 schwarze 174: * Allocate volatile and non-volatile parse resources.
1.85 kristaps 175: */
1.1 kristaps 176: struct mdoc *
1.208 schwarze 177: mdoc_alloc(struct roff *roff, struct mparse *parse,
1.220 schwarze 178: const char *defos, int quick)
1.1 kristaps 179: {
180: struct mdoc *p;
181:
1.113 kristaps 182: p = mandoc_calloc(1, sizeof(struct mdoc));
183:
1.185 kristaps 184: p->parse = parse;
1.197 schwarze 185: p->defos = defos;
1.208 schwarze 186: p->quick = quick;
1.189 kristaps 187: p->roff = roff;
1.73 kristaps 188:
1.113 kristaps 189: mdoc_hash_init();
190: mdoc_alloc1(p);
191: return(p);
1.1 kristaps 192: }
193:
194: int
1.203 schwarze 195: mdoc_endparse(struct mdoc *mdoc)
1.20 kristaps 196: {
197:
1.221 schwarze 198: return(mdoc_macroend(mdoc));
1.181 kristaps 199: }
200:
201: int
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;
208:
1.203 schwarze 209: if ( ! node_append(mdoc, n))
1.181 kristaps 210: return(0);
211:
1.203 schwarze 212: mdoc->next = MDOC_NEXT_SIBLING;
1.181 kristaps 213: return(1);
1.20 kristaps 214: }
215:
1.175 kristaps 216: int
1.203 schwarze 217: mdoc_addspan(struct mdoc *mdoc, const struct tbl_span *sp)
1.175 kristaps 218: {
1.180 kristaps 219: struct mdoc_node *n;
1.175 kristaps 220:
1.203 schwarze 221: n = node_alloc(mdoc, sp->line, 0, MDOC_MAX, MDOC_TBL);
1.180 kristaps 222: n->span = sp;
223:
1.203 schwarze 224: if ( ! node_append(mdoc, n))
1.180 kristaps 225: return(0);
226:
1.203 schwarze 227: mdoc->next = MDOC_NEXT_SIBLING;
1.180 kristaps 228: return(1);
1.175 kristaps 229: }
230:
1.50 kristaps 231: /*
1.53 kristaps 232: * Main parse routine. Parses a single line -- really just hands off to
1.123 kristaps 233: * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
1.50 kristaps 234: */
1.20 kristaps 235: int
1.203 schwarze 236: mdoc_parseln(struct mdoc *mdoc, int ln, char *buf, int offs)
1.1 kristaps 237: {
238:
1.203 schwarze 239: mdoc->flags |= MDOC_NEWLINE;
1.153 schwarze 240:
241: /*
242: * Let the roff nS register switch SYNOPSIS mode early,
243: * such that the parser knows at all times
244: * whether this mode is on or off.
245: * Note that this mode is also switched by the Sh macro.
246: */
1.204 schwarze 247: if (roff_getreg(mdoc->roff, "nS"))
248: mdoc->flags |= MDOC_SYNOPSIS;
249: else
250: mdoc->flags &= ~MDOC_SYNOPSIS;
1.153 schwarze 251:
1.203 schwarze 252: return(roff_getcontrol(mdoc->roff, buf, &offs) ?
1.213 schwarze 253: mdoc_pmacro(mdoc, ln, buf, offs) :
254: mdoc_ptext(mdoc, ln, buf, offs));
1.1 kristaps 255: }
256:
1.88 kristaps 257: int
1.148 kristaps 258: mdoc_macro(MACRO_PROT_ARGS)
1.88 kristaps 259: {
1.122 kristaps 260: assert(tok < MDOC_MAX);
261:
1.223 schwarze 262: if (mdoc->flags & MDOC_PBODY) {
263: if (tok == MDOC_Dt) {
264: mandoc_vmsg(MANDOCERR_DT_LATE,
265: mdoc->parse, line, ppos,
266: "Dt %s", buf + *pos);
267: return(1);
268: }
269: } else if ( ! (mdoc_macros[tok].flags & MDOC_PROLOGUE)) {
270: if (mdoc->meta.title == NULL) {
271: mandoc_vmsg(MANDOCERR_DT_NOTITLE,
272: mdoc->parse, line, ppos, "%s %s",
273: mdoc_macronames[tok], buf + *pos);
274: mdoc->meta.title = mandoc_strdup("UNTITLED");
275: }
1.203 schwarze 276: if (NULL == mdoc->meta.vol)
277: mdoc->meta.vol = mandoc_strdup("LOCAL");
278: mdoc->flags |= MDOC_PBODY;
1.120 kristaps 279: }
1.88 kristaps 280:
1.203 schwarze 281: return((*mdoc_macros[tok].fp)(mdoc, tok, line, ppos, pos, buf));
1.73 kristaps 282: }
283:
284:
285: static int
286: node_append(struct mdoc *mdoc, struct mdoc_node *p)
1.1 kristaps 287: {
288:
1.25 kristaps 289: assert(mdoc->last);
290: assert(mdoc->first);
291: assert(MDOC_ROOT != p->type);
1.1 kristaps 292:
1.13 kristaps 293: switch (mdoc->next) {
1.213 schwarze 294: case MDOC_NEXT_SIBLING:
1.6 kristaps 295: mdoc->last->next = p;
296: p->prev = mdoc->last;
1.13 kristaps 297: p->parent = mdoc->last->parent;
1.1 kristaps 298: break;
1.213 schwarze 299: case MDOC_NEXT_CHILD:
1.13 kristaps 300: mdoc->last->child = p;
1.1 kristaps 301: p->parent = mdoc->last;
302: break;
303: default:
1.13 kristaps 304: abort();
305: /* NOTREACHED */
1.1 kristaps 306: }
307:
1.86 kristaps 308: p->parent->nchild++;
309:
1.172 kristaps 310: /*
311: * Copy over the normalised-data pointer of our parent. Not
312: * everybody has one, but copying a null pointer is fine.
313: */
314:
315: switch (p->type) {
1.213 schwarze 316: case MDOC_BODY:
1.202 schwarze 317: if (ENDBODY_NOT != p->end)
318: break;
1.172 kristaps 319: /* FALLTHROUGH */
1.213 schwarze 320: case MDOC_TAIL:
1.172 kristaps 321: /* FALLTHROUGH */
1.213 schwarze 322: case MDOC_HEAD:
1.172 kristaps 323: p->norm = p->parent->norm;
324: break;
325: default:
326: break;
327: }
328:
1.23 kristaps 329: if ( ! mdoc_valid_pre(mdoc, p))
330: return(0);
1.27 kristaps 331:
332: switch (p->type) {
1.213 schwarze 333: case MDOC_HEAD:
1.27 kristaps 334: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 335: p->parent->head = p;
1.27 kristaps 336: break;
1.213 schwarze 337: case MDOC_TAIL:
1.27 kristaps 338: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 339: p->parent->tail = p;
1.27 kristaps 340: break;
1.213 schwarze 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) {
1.213 schwarze 354: case MDOC_TBL:
1.176 kristaps 355: /* FALLTHROUGH */
1.213 schwarze 356: case MDOC_TEXT:
1.71 kristaps 357: if ( ! mdoc_valid_post(mdoc))
358: return(0);
359: break;
360: default:
361: break;
362: }
363:
1.23 kristaps 364: return(1);
1.1 kristaps 365: }
366:
1.45 kristaps 367: static struct mdoc_node *
1.213 schwarze 368: node_alloc(struct mdoc *mdoc, int line, int pos,
1.117 kristaps 369: enum mdoct tok, enum mdoc_type type)
1.45 kristaps 370: {
1.46 kristaps 371: struct mdoc_node *p;
372:
1.113 kristaps 373: p = mandoc_calloc(1, sizeof(struct mdoc_node));
1.203 schwarze 374: p->sec = mdoc->lastsec;
1.73 kristaps 375: p->line = line;
376: p->pos = pos;
1.206 schwarze 377: p->lastline = line;
1.73 kristaps 378: p->tok = tok;
1.118 kristaps 379: p->type = type;
1.150 kristaps 380:
381: /* Flag analysis. */
382:
1.203 schwarze 383: if (MDOC_SYNOPSIS & mdoc->flags)
1.153 schwarze 384: p->flags |= MDOC_SYNPRETTY;
385: else
386: p->flags &= ~MDOC_SYNPRETTY;
1.203 schwarze 387: if (MDOC_NEWLINE & mdoc->flags)
1.130 kristaps 388: p->flags |= MDOC_LINE;
1.203 schwarze 389: mdoc->flags &= ~MDOC_NEWLINE;
1.150 kristaps 390:
1.46 kristaps 391: return(p);
1.45 kristaps 392: }
393:
1.23 kristaps 394: int
1.203 schwarze 395: mdoc_tail_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.17 kristaps 396: {
397: struct mdoc_node *p;
398:
1.203 schwarze 399: p = node_alloc(mdoc, line, pos, tok, MDOC_TAIL);
400: if ( ! node_append(mdoc, p))
1.102 kristaps 401: return(0);
1.203 schwarze 402: mdoc->next = MDOC_NEXT_CHILD;
1.102 kristaps 403: return(1);
1.17 kristaps 404: }
405:
1.23 kristaps 406: int
1.203 schwarze 407: mdoc_head_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.1 kristaps 408: {
409: struct mdoc_node *p;
410:
1.203 schwarze 411: assert(mdoc->first);
412: assert(mdoc->last);
1.1 kristaps 413:
1.203 schwarze 414: p = node_alloc(mdoc, line, pos, tok, MDOC_HEAD);
415: if ( ! node_append(mdoc, p))
1.102 kristaps 416: return(0);
1.203 schwarze 417: mdoc->next = MDOC_NEXT_CHILD;
1.102 kristaps 418: return(1);
1.1 kristaps 419: }
420:
1.23 kristaps 421: int
1.203 schwarze 422: mdoc_body_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.1 kristaps 423: {
424: struct mdoc_node *p;
425:
1.203 schwarze 426: p = node_alloc(mdoc, line, pos, tok, MDOC_BODY);
427: if ( ! node_append(mdoc, p))
1.102 kristaps 428: return(0);
1.203 schwarze 429: mdoc->next = MDOC_NEXT_CHILD;
1.152 schwarze 430: return(1);
431: }
432:
433: int
1.203 schwarze 434: mdoc_endbody_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok,
1.152 schwarze 435: struct mdoc_node *body, enum mdoc_endbody end)
436: {
437: struct mdoc_node *p;
438:
1.203 schwarze 439: p = node_alloc(mdoc, line, pos, tok, MDOC_BODY);
1.152 schwarze 440: p->pending = body;
1.202 schwarze 441: p->norm = body->norm;
1.152 schwarze 442: p->end = end;
1.203 schwarze 443: if ( ! node_append(mdoc, p))
1.152 schwarze 444: return(0);
1.203 schwarze 445: mdoc->next = MDOC_NEXT_SIBLING;
1.102 kristaps 446: return(1);
1.1 kristaps 447: }
448:
1.23 kristaps 449: int
1.213 schwarze 450: mdoc_block_alloc(struct mdoc *mdoc, int line, int pos,
1.117 kristaps 451: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 452: {
453: struct mdoc_node *p;
454:
1.203 schwarze 455: p = node_alloc(mdoc, line, pos, tok, MDOC_BLOCK);
1.77 kristaps 456: p->args = args;
457: if (p->args)
1.53 kristaps 458: (args->refcnt)++;
1.172 kristaps 459:
460: switch (tok) {
1.213 schwarze 461: case MDOC_Bd:
1.172 kristaps 462: /* FALLTHROUGH */
1.213 schwarze 463: case MDOC_Bf:
1.172 kristaps 464: /* FALLTHROUGH */
1.213 schwarze 465: case MDOC_Bl:
1.217 schwarze 466: /* FALLTHROUGH */
467: case MDOC_En:
1.173 kristaps 468: /* FALLTHROUGH */
1.213 schwarze 469: case MDOC_Rs:
1.172 kristaps 470: p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
471: break;
472: default:
473: break;
474: }
475:
1.203 schwarze 476: if ( ! node_append(mdoc, p))
1.102 kristaps 477: return(0);
1.203 schwarze 478: mdoc->next = MDOC_NEXT_CHILD;
1.102 kristaps 479: return(1);
1.1 kristaps 480: }
481:
1.23 kristaps 482: int
1.213 schwarze 483: mdoc_elem_alloc(struct mdoc *mdoc, int line, int pos,
1.117 kristaps 484: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 485: {
486: struct mdoc_node *p;
487:
1.203 schwarze 488: p = node_alloc(mdoc, line, pos, tok, MDOC_ELEM);
1.77 kristaps 489: p->args = args;
490: if (p->args)
1.53 kristaps 491: (args->refcnt)++;
1.172 kristaps 492:
493: switch (tok) {
1.213 schwarze 494: case MDOC_An:
1.172 kristaps 495: p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
496: break;
497: default:
498: break;
499: }
500:
1.203 schwarze 501: if ( ! node_append(mdoc, p))
1.102 kristaps 502: return(0);
1.203 schwarze 503: mdoc->next = MDOC_NEXT_CHILD;
1.175 kristaps 504: return(1);
505: }
1.1 kristaps 506:
1.124 kristaps 507: int
1.203 schwarze 508: mdoc_word_alloc(struct mdoc *mdoc, int line, int pos, const char *p)
1.1 kristaps 509: {
1.91 kristaps 510: struct mdoc_node *n;
1.1 kristaps 511:
1.203 schwarze 512: n = node_alloc(mdoc, line, pos, MDOC_MAX, MDOC_TEXT);
513: n->string = roff_strdup(mdoc->roff, p);
1.91 kristaps 514:
1.203 schwarze 515: if ( ! node_append(mdoc, n))
1.101 kristaps 516: return(0);
1.124 kristaps 517:
1.203 schwarze 518: mdoc->next = MDOC_NEXT_SIBLING;
1.101 kristaps 519: return(1);
1.91 kristaps 520: }
521:
1.205 schwarze 522: void
523: mdoc_word_append(struct mdoc *mdoc, const char *p)
524: {
525: struct mdoc_node *n;
526: char *addstr, *newstr;
527:
528: n = mdoc->last;
529: addstr = roff_strdup(mdoc->roff, p);
1.210 schwarze 530: mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
1.205 schwarze 531: free(addstr);
532: free(n->string);
533: n->string = newstr;
534: mdoc->next = MDOC_NEXT_SIBLING;
535: }
1.91 kristaps 536:
1.155 kristaps 537: static void
1.53 kristaps 538: mdoc_node_free(struct mdoc_node *p)
1.1 kristaps 539: {
540:
1.172 kristaps 541: if (MDOC_BLOCK == p->type || MDOC_ELEM == p->type)
1.171 kristaps 542: free(p->norm);
1.53 kristaps 543: if (p->string)
544: free(p->string);
545: if (p->args)
546: mdoc_argv_free(p->args);
1.1 kristaps 547: free(p);
548: }
549:
1.121 kristaps 550: static void
1.203 schwarze 551: mdoc_node_unlink(struct mdoc *mdoc, struct mdoc_node *n)
1.121 kristaps 552: {
553:
554: /* Adjust siblings. */
555:
556: if (n->prev)
557: n->prev->next = n->next;
558: if (n->next)
559: n->next->prev = n->prev;
560:
561: /* Adjust parent. */
562:
563: if (n->parent) {
564: n->parent->nchild--;
565: if (n->parent->child == n)
566: n->parent->child = n->prev ? n->prev : n->next;
1.169 kristaps 567: if (n->parent->last == n)
568: n->parent->last = n->prev ? n->prev : NULL;
1.121 kristaps 569: }
570:
571: /* Adjust parse point, if applicable. */
572:
1.203 schwarze 573: if (mdoc && mdoc->last == n) {
1.121 kristaps 574: if (n->prev) {
1.203 schwarze 575: mdoc->last = n->prev;
576: mdoc->next = MDOC_NEXT_SIBLING;
1.121 kristaps 577: } else {
1.203 schwarze 578: mdoc->last = n->parent;
579: mdoc->next = MDOC_NEXT_CHILD;
1.121 kristaps 580: }
581: }
582:
1.203 schwarze 583: if (mdoc && mdoc->first == n)
584: mdoc->first = NULL;
1.121 kristaps 585: }
586:
1.53 kristaps 587: void
1.203 schwarze 588: mdoc_node_delete(struct mdoc *mdoc, struct mdoc_node *p)
1.1 kristaps 589: {
590:
1.121 kristaps 591: while (p->child) {
592: assert(p->nchild);
1.203 schwarze 593: mdoc_node_delete(mdoc, p->child);
1.121 kristaps 594: }
595: assert(0 == p->nchild);
1.1 kristaps 596:
1.203 schwarze 597: mdoc_node_unlink(mdoc, p);
1.53 kristaps 598: mdoc_node_free(p);
1.201 schwarze 599: }
600:
601: int
1.203 schwarze 602: mdoc_node_relink(struct mdoc *mdoc, struct mdoc_node *p)
1.201 schwarze 603: {
604:
1.203 schwarze 605: mdoc_node_unlink(mdoc, p);
606: return(node_append(mdoc, p));
1.1 kristaps 607: }
608:
1.53 kristaps 609: /*
610: * Parse free-form text, that is, a line that does not begin with the
611: * control character.
612: */
613: static int
1.203 schwarze 614: mdoc_ptext(struct mdoc *mdoc, int line, char *buf, int offs)
1.1 kristaps 615: {
1.142 kristaps 616: char *c, *ws, *end;
617: struct mdoc_node *n;
618:
1.203 schwarze 619: assert(mdoc->last);
620: n = mdoc->last;
1.142 kristaps 621:
622: /*
1.144 kristaps 623: * Divert directly to list processing if we're encountering a
1.142 kristaps 624: * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
1.144 kristaps 625: * (a MDOC_BODY means it's already open, in which case we should
626: * process within its context in the normal way).
1.142 kristaps 627: */
628:
1.143 kristaps 629: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.213 schwarze 630: LIST_column == n->norm->Bl.type) {
1.144 kristaps 631: /* `Bl' is open without any children. */
1.203 schwarze 632: mdoc->flags |= MDOC_FREECOL;
633: return(mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf));
1.142 kristaps 634: }
635:
636: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
1.213 schwarze 637: NULL != n->parent &&
638: MDOC_Bl == n->parent->tok &&
639: LIST_column == n->parent->norm->Bl.type) {
1.144 kristaps 640: /* `Bl' has block-level `It' children. */
1.203 schwarze 641: mdoc->flags |= MDOC_FREECOL;
642: return(mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf));
1.142 kristaps 643: }
1.124 kristaps 644:
1.137 schwarze 645: /*
646: * Search for the beginning of unescaped trailing whitespace (ws)
647: * and for the first character not to be output (end).
648: */
1.139 kristaps 649:
650: /* FIXME: replace with strcspn(). */
1.137 schwarze 651: ws = NULL;
652: for (c = end = buf + offs; *c; c++) {
653: switch (*c) {
654: case ' ':
655: if (NULL == ws)
656: ws = c;
657: continue;
658: case '\t':
659: /*
660: * Always warn about trailing tabs,
661: * even outside literal context,
662: * where they should be put on the next line.
663: */
664: if (NULL == ws)
665: ws = c;
666: /*
667: * Strip trailing tabs in literal context only;
668: * outside, they affect the next line.
669: */
1.203 schwarze 670: if (MDOC_LITERAL & mdoc->flags)
1.137 schwarze 671: continue;
672: break;
673: case '\\':
674: /* Skip the escaped character, too, if any. */
675: if (c[1])
676: c++;
677: /* FALLTHROUGH */
678: default:
679: ws = NULL;
680: break;
681: }
682: end = c + 1;
683: }
684: *end = '\0';
1.91 kristaps 685:
1.137 schwarze 686: if (ws)
1.218 schwarze 687: mandoc_msg(MANDOCERR_SPACE_EOL, mdoc->parse,
688: line, (int)(ws-buf), NULL);
1.115 kristaps 689:
1.203 schwarze 690: if ('\0' == buf[offs] && ! (MDOC_LITERAL & mdoc->flags)) {
1.218 schwarze 691: mandoc_msg(MANDOCERR_FI_BLANK, mdoc->parse,
692: line, (int)(c - buf), NULL);
1.124 kristaps 693:
1.119 kristaps 694: /*
1.165 schwarze 695: * Insert a `sp' in the case of a blank line. Technically,
1.124 kristaps 696: * blank lines aren't allowed, but enough manuals assume this
697: * behaviour that we want to work around it.
1.119 kristaps 698: */
1.203 schwarze 699: if ( ! mdoc_elem_alloc(mdoc, line, offs, MDOC_sp, NULL))
1.119 kristaps 700: return(0);
1.124 kristaps 701:
1.203 schwarze 702: mdoc->next = MDOC_NEXT_SIBLING;
1.199 schwarze 703:
1.203 schwarze 704: return(mdoc_valid_post(mdoc));
1.119 kristaps 705: }
1.68 kristaps 706:
1.203 schwarze 707: if ( ! mdoc_word_alloc(mdoc, line, offs, buf+offs))
1.137 schwarze 708: return(0);
1.91 kristaps 709:
1.203 schwarze 710: if (MDOC_LITERAL & mdoc->flags)
1.137 schwarze 711: return(1);
1.128 kristaps 712:
713: /*
714: * End-of-sentence check. If the last character is an unescaped
715: * EOS character, then flag the node as being the end of a
716: * sentence. The front-end will know how to interpret this.
717: */
1.132 kristaps 718:
1.137 schwarze 719: assert(buf < end);
720:
1.207 schwarze 721: if (mandoc_eos(buf+offs, (size_t)(end-buf-offs)))
1.203 schwarze 722: mdoc->last->flags |= MDOC_EOS;
1.128 kristaps 723:
724: return(1);
1.1 kristaps 725: }
726:
1.53 kristaps 727: /*
728: * Parse a macro line, that is, a line beginning with the control
729: * character.
730: */
1.155 kristaps 731: static int
1.203 schwarze 732: mdoc_pmacro(struct mdoc *mdoc, int ln, char *buf, int offs)
1.1 kristaps 733: {
1.144 kristaps 734: enum mdoct tok;
1.188 kristaps 735: int i, sv;
1.144 kristaps 736: char mac[5];
737: struct mdoc_node *n;
1.63 kristaps 738:
1.188 kristaps 739: sv = offs;
1.130 kristaps 740:
1.213 schwarze 741: /*
1.162 schwarze 742: * Copy the first word into a nil-terminated buffer.
1.165 schwarze 743: * Stop copying when a tab, space, or eoln is encountered.
1.160 kristaps 744: */
1.1 kristaps 745:
1.188 kristaps 746: i = 0;
1.213 schwarze 747: while (i < 4 && '\0' != buf[offs] && ' ' != buf[offs] &&
748: '\t' != buf[offs])
1.188 kristaps 749: mac[i++] = buf[offs++];
750:
751: mac[i] = '\0';
752:
1.214 schwarze 753: tok = (i > 1 && i < 4) ? mdoc_hash_find(mac) : MDOC_MAX;
1.1 kristaps 754:
1.163 schwarze 755: if (MDOC_MAX == tok) {
1.222 schwarze 756: mandoc_msg(MANDOCERR_MACRO, mdoc->parse,
757: ln, sv, buf + sv - 1);
1.58 kristaps 758: return(1);
1.53 kristaps 759: }
1.1 kristaps 760:
1.160 kristaps 761: /* Disregard the first trailing tab, if applicable. */
762:
1.188 kristaps 763: if ('\t' == buf[offs])
764: offs++;
1.160 kristaps 765:
766: /* Jump to the next non-whitespace word. */
1.1 kristaps 767:
1.188 kristaps 768: while (buf[offs] && ' ' == buf[offs])
769: offs++;
1.1 kristaps 770:
1.213 schwarze 771: /*
1.125 kristaps 772: * Trailing whitespace. Note that tabs are allowed to be passed
773: * into the parser as "text", so we only warn about spaces here.
774: */
1.115 kristaps 775:
1.188 kristaps 776: if ('\0' == buf[offs] && ' ' == buf[offs - 1])
1.218 schwarze 777: mandoc_msg(MANDOCERR_SPACE_EOL, mdoc->parse,
778: ln, offs - 1, NULL);
1.115 kristaps 779:
1.144 kristaps 780: /*
781: * If an initial macro or a list invocation, divert directly
782: * into macro processing.
783: */
784:
1.221 schwarze 785: if (NULL == mdoc->last || MDOC_It == tok || MDOC_El == tok)
786: return(mdoc_macro(mdoc, tok, ln, sv, &offs, buf));
1.144 kristaps 787:
1.203 schwarze 788: n = mdoc->last;
789: assert(mdoc->last);
1.144 kristaps 790:
791: /*
792: * If the first macro of a `Bl -column', open an `It' block
793: * context around the parsed macro.
794: */
795:
796: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.213 schwarze 797: LIST_column == n->norm->Bl.type) {
1.203 schwarze 798: mdoc->flags |= MDOC_FREECOL;
1.221 schwarze 799: return(mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf));
1.144 kristaps 800: }
801:
802: /*
803: * If we're following a block-level `It' within a `Bl -column'
804: * context (perhaps opened in the above block or in ptext()),
805: * then open an `It' block context around the parsed macro.
1.98 kristaps 806: */
1.144 kristaps 807:
808: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
1.213 schwarze 809: NULL != n->parent &&
810: MDOC_Bl == n->parent->tok &&
811: LIST_column == n->parent->norm->Bl.type) {
1.203 schwarze 812: mdoc->flags |= MDOC_FREECOL;
1.221 schwarze 813: return(mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf));
1.144 kristaps 814: }
815:
816: /* Normal processing of a macro. */
817:
1.213 schwarze 818: if ( ! mdoc_macro(mdoc, tok, ln, sv, &offs, buf))
1.221 schwarze 819: return(0);
1.208 schwarze 820:
821: /* In quick mode (for mandocdb), abort after the NAME section. */
822:
823: if (mdoc->quick && MDOC_Sh == tok &&
824: SEC_NAME != mdoc->last->sec)
825: return(2);
1.1 kristaps 826:
1.53 kristaps 827: return(1);
1.1 kristaps 828: }
1.100 kristaps 829:
1.186 kristaps 830: enum mdelim
831: mdoc_isdelim(const char *p)
832: {
833:
834: if ('\0' == p[0])
835: return(DELIM_NONE);
836:
837: if ('\0' == p[1])
838: switch (p[0]) {
1.213 schwarze 839: case '(':
1.186 kristaps 840: /* FALLTHROUGH */
1.213 schwarze 841: case '[':
1.186 kristaps 842: return(DELIM_OPEN);
1.213 schwarze 843: case '|':
1.186 kristaps 844: return(DELIM_MIDDLE);
1.213 schwarze 845: case '.':
1.186 kristaps 846: /* FALLTHROUGH */
1.213 schwarze 847: case ',':
1.186 kristaps 848: /* FALLTHROUGH */
1.213 schwarze 849: case ';':
1.186 kristaps 850: /* FALLTHROUGH */
1.213 schwarze 851: case ':':
1.186 kristaps 852: /* FALLTHROUGH */
1.213 schwarze 853: case '?':
1.186 kristaps 854: /* FALLTHROUGH */
1.213 schwarze 855: case '!':
1.186 kristaps 856: /* FALLTHROUGH */
1.213 schwarze 857: case ')':
1.186 kristaps 858: /* FALLTHROUGH */
1.213 schwarze 859: case ']':
1.186 kristaps 860: return(DELIM_CLOSE);
861: default:
862: return(DELIM_NONE);
863: }
864:
865: if ('\\' != p[0])
866: return(DELIM_NONE);
1.100 kristaps 867:
1.186 kristaps 868: if (0 == strcmp(p + 1, "."))
869: return(DELIM_CLOSE);
1.200 schwarze 870: if (0 == strcmp(p + 1, "fR|\\fP"))
1.186 kristaps 871: return(DELIM_MIDDLE);
872:
873: return(DELIM_NONE);
1.211 schwarze 874: }
875:
876: void
877: mdoc_deroff(char **dest, const struct mdoc_node *n)
878: {
879: char *cp;
880: size_t sz;
881:
882: if (MDOC_TEXT != n->type) {
883: for (n = n->child; n; n = n->next)
884: mdoc_deroff(dest, n);
885: return;
886: }
887:
888: /* Skip leading whitespace. */
889:
890: for (cp = n->string; '\0' != *cp; cp++)
891: if (0 == isspace((unsigned char)*cp))
892: break;
893:
894: /* Skip trailing whitespace. */
895:
896: for (sz = strlen(cp); sz; sz--)
897: if (0 == isspace((unsigned char)cp[sz-1]))
898: break;
899:
900: /* Skip empty strings. */
901:
902: if (0 == sz)
903: return;
904:
905: if (NULL == *dest) {
906: *dest = mandoc_strndup(cp, sz);
907: return;
908: }
909:
910: mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
911: free(*dest);
912: *dest = cp;
1.186 kristaps 913: }
CVSweb