Annotation of mandoc/mdoc.c, Revision 1.215
1.215 ! schwarze 1: /* $Id: mdoc.c,v 1.214 2014/04/25 14:11:30 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:
1.213 schwarze 38: const char *const __mdoc_macronames[MDOC_MAX] = {
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",
1.212 schwarze 69: "%U", "Ta", "ll",
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: assert( ! (MDOC_HALT & mdoc->flags));
108: return(mdoc->first);
1.1 kristaps 109: }
110:
1.37 kristaps 111: const struct mdoc_meta *
1.203 schwarze 112: mdoc_meta(const struct mdoc *mdoc)
1.37 kristaps 113: {
114:
1.203 schwarze 115: assert( ! (MDOC_HALT & mdoc->flags));
116: return(&mdoc->meta);
1.37 kristaps 117: }
118:
1.85 kristaps 119: /*
120: * Frees volatile resources (parse tree, meta-data, fields).
121: */
1.73 kristaps 122: static void
123: mdoc_free1(struct mdoc *mdoc)
1.67 kristaps 124: {
125:
126: if (mdoc->first)
1.121 kristaps 127: mdoc_node_delete(mdoc, mdoc->first);
1.67 kristaps 128: if (mdoc->meta.title)
129: free(mdoc->meta.title);
130: if (mdoc->meta.os)
131: free(mdoc->meta.os);
132: if (mdoc->meta.name)
133: free(mdoc->meta.name);
134: if (mdoc->meta.arch)
135: free(mdoc->meta.arch);
136: if (mdoc->meta.vol)
137: free(mdoc->meta.vol);
1.133 kristaps 138: if (mdoc->meta.msec)
139: free(mdoc->meta.msec);
1.183 kristaps 140: if (mdoc->meta.date)
141: free(mdoc->meta.date);
1.73 kristaps 142: }
143:
1.85 kristaps 144: /*
145: * Allocate all volatile resources (parse tree, meta-data, fields).
146: */
1.113 kristaps 147: static void
1.73 kristaps 148: mdoc_alloc1(struct mdoc *mdoc)
149: {
1.67 kristaps 150:
1.112 kristaps 151: memset(&mdoc->meta, 0, sizeof(struct mdoc_meta));
1.67 kristaps 152: mdoc->flags = 0;
1.85 kristaps 153: mdoc->lastnamed = mdoc->lastsec = SEC_NONE;
1.113 kristaps 154: mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node));
1.70 kristaps 155: mdoc->first = mdoc->last;
1.67 kristaps 156: mdoc->last->type = MDOC_ROOT;
1.196 schwarze 157: mdoc->last->tok = MDOC_MAX;
1.67 kristaps 158: mdoc->next = MDOC_NEXT_CHILD;
1.73 kristaps 159: }
160:
161: /*
1.85 kristaps 162: * Free up volatile resources (see mdoc_free1()) then re-initialises the
163: * data with mdoc_alloc1(). After invocation, parse data has been reset
164: * and the parser is ready for re-invocation on a new tree; however,
165: * cross-parse non-volatile data is kept intact.
1.73 kristaps 166: */
1.113 kristaps 167: void
1.73 kristaps 168: mdoc_reset(struct mdoc *mdoc)
169: {
170:
171: mdoc_free1(mdoc);
1.113 kristaps 172: mdoc_alloc1(mdoc);
1.67 kristaps 173: }
174:
1.68 kristaps 175: /*
1.85 kristaps 176: * Completely free up all volatile and non-volatile parse resources.
177: * After invocation, the pointer is no longer usable.
1.68 kristaps 178: */
1.67 kristaps 179: void
1.38 kristaps 180: mdoc_free(struct mdoc *mdoc)
1.34 kristaps 181: {
182:
1.73 kristaps 183: mdoc_free1(mdoc);
1.1 kristaps 184: free(mdoc);
185: }
186:
1.85 kristaps 187: /*
1.213 schwarze 188: * Allocate volatile and non-volatile parse resources.
1.85 kristaps 189: */
1.1 kristaps 190: struct mdoc *
1.208 schwarze 191: mdoc_alloc(struct roff *roff, struct mparse *parse,
192: char *defos, int quick)
1.1 kristaps 193: {
194: struct mdoc *p;
195:
1.113 kristaps 196: p = mandoc_calloc(1, sizeof(struct mdoc));
197:
1.185 kristaps 198: p->parse = parse;
1.197 schwarze 199: p->defos = defos;
1.208 schwarze 200: p->quick = quick;
1.189 kristaps 201: p->roff = roff;
1.73 kristaps 202:
1.113 kristaps 203: mdoc_hash_init();
204: mdoc_alloc1(p);
205: return(p);
1.1 kristaps 206: }
207:
1.68 kristaps 208: /*
209: * Climb back up the parse tree, validating open scopes. Mostly calls
1.85 kristaps 210: * through to macro_end() in macro.c.
1.68 kristaps 211: */
1.1 kristaps 212: int
1.203 schwarze 213: mdoc_endparse(struct mdoc *mdoc)
1.20 kristaps 214: {
215:
1.203 schwarze 216: assert( ! (MDOC_HALT & mdoc->flags));
217: if (mdoc_macroend(mdoc))
1.20 kristaps 218: return(1);
1.203 schwarze 219: mdoc->flags |= MDOC_HALT;
1.72 kristaps 220: return(0);
1.181 kristaps 221: }
222:
223: int
1.203 schwarze 224: mdoc_addeqn(struct mdoc *mdoc, const struct eqn *ep)
1.181 kristaps 225: {
226: struct mdoc_node *n;
227:
1.203 schwarze 228: assert( ! (MDOC_HALT & mdoc->flags));
1.181 kristaps 229:
230: /* No text before an initial macro. */
231:
1.203 schwarze 232: if (SEC_NONE == mdoc->lastnamed) {
233: mdoc_pmsg(mdoc, ep->ln, ep->pos, MANDOCERR_NOTEXT);
1.181 kristaps 234: return(1);
235: }
236:
1.203 schwarze 237: n = node_alloc(mdoc, ep->ln, ep->pos, MDOC_MAX, MDOC_EQN);
1.181 kristaps 238: n->eqn = ep;
239:
1.203 schwarze 240: if ( ! node_append(mdoc, n))
1.181 kristaps 241: return(0);
242:
1.203 schwarze 243: mdoc->next = MDOC_NEXT_SIBLING;
1.181 kristaps 244: return(1);
1.20 kristaps 245: }
246:
1.175 kristaps 247: int
1.203 schwarze 248: mdoc_addspan(struct mdoc *mdoc, const struct tbl_span *sp)
1.175 kristaps 249: {
1.180 kristaps 250: struct mdoc_node *n;
1.175 kristaps 251:
1.203 schwarze 252: assert( ! (MDOC_HALT & mdoc->flags));
1.175 kristaps 253:
254: /* No text before an initial macro. */
255:
1.203 schwarze 256: if (SEC_NONE == mdoc->lastnamed) {
257: mdoc_pmsg(mdoc, sp->line, 0, MANDOCERR_NOTEXT);
1.175 kristaps 258: return(1);
259: }
260:
1.203 schwarze 261: n = node_alloc(mdoc, sp->line, 0, MDOC_MAX, MDOC_TBL);
1.180 kristaps 262: n->span = sp;
263:
1.203 schwarze 264: if ( ! node_append(mdoc, n))
1.180 kristaps 265: return(0);
266:
1.203 schwarze 267: mdoc->next = MDOC_NEXT_SIBLING;
1.180 kristaps 268: return(1);
1.175 kristaps 269: }
270:
1.50 kristaps 271: /*
1.53 kristaps 272: * Main parse routine. Parses a single line -- really just hands off to
1.123 kristaps 273: * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
1.50 kristaps 274: */
1.20 kristaps 275: int
1.203 schwarze 276: mdoc_parseln(struct mdoc *mdoc, int ln, char *buf, int offs)
1.1 kristaps 277: {
278:
1.203 schwarze 279: assert( ! (MDOC_HALT & mdoc->flags));
1.50 kristaps 280:
1.203 schwarze 281: mdoc->flags |= MDOC_NEWLINE;
1.153 schwarze 282:
283: /*
284: * Let the roff nS register switch SYNOPSIS mode early,
285: * such that the parser knows at all times
286: * whether this mode is on or off.
287: * Note that this mode is also switched by the Sh macro.
288: */
1.204 schwarze 289: if (roff_getreg(mdoc->roff, "nS"))
290: mdoc->flags |= MDOC_SYNOPSIS;
291: else
292: mdoc->flags &= ~MDOC_SYNOPSIS;
1.153 schwarze 293:
1.203 schwarze 294: return(roff_getcontrol(mdoc->roff, buf, &offs) ?
1.213 schwarze 295: mdoc_pmacro(mdoc, ln, buf, offs) :
296: mdoc_ptext(mdoc, ln, buf, offs));
1.1 kristaps 297: }
298:
1.88 kristaps 299: int
1.148 kristaps 300: mdoc_macro(MACRO_PROT_ARGS)
1.88 kristaps 301: {
1.122 kristaps 302: assert(tok < MDOC_MAX);
303:
304: /* If we're in the body, deny prologue calls. */
1.117 kristaps 305:
1.213 schwarze 306: if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
307: MDOC_PBODY & mdoc->flags) {
1.215 ! schwarze 308: mandoc_vmsg(MANDOCERR_PROLOG_ONLY, mdoc->parse,
! 309: line, ppos, "%s", mdoc_macronames[tok]);
1.174 kristaps 310: return(1);
311: }
1.122 kristaps 312:
313: /* If we're in the prologue, deny "body" macros. */
314:
1.213 schwarze 315: if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
316: ! (MDOC_PBODY & mdoc->flags)) {
1.215 ! schwarze 317: mandoc_vmsg(MANDOCERR_PROLOG_BAD, mdoc->parse,
! 318: line, ppos, "%s", mdoc_macronames[tok]);
1.203 schwarze 319: if (NULL == mdoc->meta.msec)
320: mdoc->meta.msec = mandoc_strdup("1");
321: if (NULL == mdoc->meta.title)
322: mdoc->meta.title = mandoc_strdup("UNKNOWN");
323: if (NULL == mdoc->meta.vol)
324: mdoc->meta.vol = mandoc_strdup("LOCAL");
325: if (NULL == mdoc->meta.os)
326: mdoc->meta.os = mandoc_strdup("LOCAL");
327: if (NULL == mdoc->meta.date)
328: mdoc->meta.date = mandoc_normdate
329: (mdoc->parse, NULL, line, ppos);
330: mdoc->flags |= MDOC_PBODY;
1.120 kristaps 331: }
1.88 kristaps 332:
1.203 schwarze 333: return((*mdoc_macros[tok].fp)(mdoc, tok, line, ppos, pos, buf));
1.73 kristaps 334: }
335:
336:
337: static int
338: node_append(struct mdoc *mdoc, struct mdoc_node *p)
1.1 kristaps 339: {
340:
1.25 kristaps 341: assert(mdoc->last);
342: assert(mdoc->first);
343: assert(MDOC_ROOT != p->type);
1.1 kristaps 344:
1.13 kristaps 345: switch (mdoc->next) {
1.213 schwarze 346: case MDOC_NEXT_SIBLING:
1.6 kristaps 347: mdoc->last->next = p;
348: p->prev = mdoc->last;
1.13 kristaps 349: p->parent = mdoc->last->parent;
1.1 kristaps 350: break;
1.213 schwarze 351: case MDOC_NEXT_CHILD:
1.13 kristaps 352: mdoc->last->child = p;
1.1 kristaps 353: p->parent = mdoc->last;
354: break;
355: default:
1.13 kristaps 356: abort();
357: /* NOTREACHED */
1.1 kristaps 358: }
359:
1.86 kristaps 360: p->parent->nchild++;
361:
1.172 kristaps 362: /*
363: * Copy over the normalised-data pointer of our parent. Not
364: * everybody has one, but copying a null pointer is fine.
365: */
366:
367: switch (p->type) {
1.213 schwarze 368: case MDOC_BODY:
1.202 schwarze 369: if (ENDBODY_NOT != p->end)
370: break;
1.172 kristaps 371: /* FALLTHROUGH */
1.213 schwarze 372: case MDOC_TAIL:
1.172 kristaps 373: /* FALLTHROUGH */
1.213 schwarze 374: case MDOC_HEAD:
1.172 kristaps 375: p->norm = p->parent->norm;
376: break;
377: default:
378: break;
379: }
380:
1.23 kristaps 381: if ( ! mdoc_valid_pre(mdoc, p))
382: return(0);
1.27 kristaps 383:
384: switch (p->type) {
1.213 schwarze 385: case MDOC_HEAD:
1.27 kristaps 386: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 387: p->parent->head = p;
1.27 kristaps 388: break;
1.213 schwarze 389: case MDOC_TAIL:
1.27 kristaps 390: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 391: p->parent->tail = p;
1.27 kristaps 392: break;
1.213 schwarze 393: case MDOC_BODY:
1.152 schwarze 394: if (p->end)
395: break;
1.27 kristaps 396: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 397: p->parent->body = p;
1.27 kristaps 398: break;
399: default:
400: break;
401: }
402:
1.1 kristaps 403: mdoc->last = p;
1.71 kristaps 404:
405: switch (p->type) {
1.213 schwarze 406: case MDOC_TBL:
1.176 kristaps 407: /* FALLTHROUGH */
1.213 schwarze 408: case MDOC_TEXT:
1.71 kristaps 409: if ( ! mdoc_valid_post(mdoc))
410: return(0);
411: break;
412: default:
413: break;
414: }
415:
1.23 kristaps 416: return(1);
1.1 kristaps 417: }
418:
1.45 kristaps 419: static struct mdoc_node *
1.213 schwarze 420: node_alloc(struct mdoc *mdoc, int line, int pos,
1.117 kristaps 421: enum mdoct tok, enum mdoc_type type)
1.45 kristaps 422: {
1.46 kristaps 423: struct mdoc_node *p;
424:
1.113 kristaps 425: p = mandoc_calloc(1, sizeof(struct mdoc_node));
1.203 schwarze 426: p->sec = mdoc->lastsec;
1.73 kristaps 427: p->line = line;
428: p->pos = pos;
1.206 schwarze 429: p->lastline = line;
1.73 kristaps 430: p->tok = tok;
1.118 kristaps 431: p->type = type;
1.150 kristaps 432:
433: /* Flag analysis. */
434:
1.203 schwarze 435: if (MDOC_SYNOPSIS & mdoc->flags)
1.153 schwarze 436: p->flags |= MDOC_SYNPRETTY;
437: else
438: p->flags &= ~MDOC_SYNPRETTY;
1.203 schwarze 439: if (MDOC_NEWLINE & mdoc->flags)
1.130 kristaps 440: p->flags |= MDOC_LINE;
1.203 schwarze 441: mdoc->flags &= ~MDOC_NEWLINE;
1.150 kristaps 442:
1.46 kristaps 443: return(p);
1.45 kristaps 444: }
445:
1.23 kristaps 446: int
1.203 schwarze 447: mdoc_tail_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.17 kristaps 448: {
449: struct mdoc_node *p;
450:
1.203 schwarze 451: p = node_alloc(mdoc, line, pos, tok, MDOC_TAIL);
452: if ( ! node_append(mdoc, p))
1.102 kristaps 453: return(0);
1.203 schwarze 454: mdoc->next = MDOC_NEXT_CHILD;
1.102 kristaps 455: return(1);
1.17 kristaps 456: }
457:
1.23 kristaps 458: int
1.203 schwarze 459: mdoc_head_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.1 kristaps 460: {
461: struct mdoc_node *p;
462:
1.203 schwarze 463: assert(mdoc->first);
464: assert(mdoc->last);
1.1 kristaps 465:
1.203 schwarze 466: p = node_alloc(mdoc, line, pos, tok, MDOC_HEAD);
467: if ( ! node_append(mdoc, p))
1.102 kristaps 468: return(0);
1.203 schwarze 469: mdoc->next = MDOC_NEXT_CHILD;
1.102 kristaps 470: return(1);
1.1 kristaps 471: }
472:
1.23 kristaps 473: int
1.203 schwarze 474: mdoc_body_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.1 kristaps 475: {
476: struct mdoc_node *p;
477:
1.203 schwarze 478: p = node_alloc(mdoc, line, pos, tok, MDOC_BODY);
479: if ( ! node_append(mdoc, p))
1.102 kristaps 480: return(0);
1.203 schwarze 481: mdoc->next = MDOC_NEXT_CHILD;
1.152 schwarze 482: return(1);
483: }
484:
485: int
1.203 schwarze 486: mdoc_endbody_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok,
1.152 schwarze 487: struct mdoc_node *body, enum mdoc_endbody end)
488: {
489: struct mdoc_node *p;
490:
1.203 schwarze 491: p = node_alloc(mdoc, line, pos, tok, MDOC_BODY);
1.152 schwarze 492: p->pending = body;
1.202 schwarze 493: p->norm = body->norm;
1.152 schwarze 494: p->end = end;
1.203 schwarze 495: if ( ! node_append(mdoc, p))
1.152 schwarze 496: return(0);
1.203 schwarze 497: mdoc->next = MDOC_NEXT_SIBLING;
1.102 kristaps 498: return(1);
1.1 kristaps 499: }
500:
1.23 kristaps 501: int
1.213 schwarze 502: mdoc_block_alloc(struct mdoc *mdoc, int line, int pos,
1.117 kristaps 503: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 504: {
505: struct mdoc_node *p;
506:
1.203 schwarze 507: p = node_alloc(mdoc, line, pos, tok, MDOC_BLOCK);
1.77 kristaps 508: p->args = args;
509: if (p->args)
1.53 kristaps 510: (args->refcnt)++;
1.172 kristaps 511:
512: switch (tok) {
1.213 schwarze 513: case MDOC_Bd:
1.172 kristaps 514: /* FALLTHROUGH */
1.213 schwarze 515: case MDOC_Bf:
1.172 kristaps 516: /* FALLTHROUGH */
1.213 schwarze 517: case MDOC_Bl:
1.173 kristaps 518: /* FALLTHROUGH */
1.213 schwarze 519: case MDOC_Rs:
1.172 kristaps 520: p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
521: break;
522: default:
523: break;
524: }
525:
1.203 schwarze 526: if ( ! node_append(mdoc, p))
1.102 kristaps 527: return(0);
1.203 schwarze 528: mdoc->next = MDOC_NEXT_CHILD;
1.102 kristaps 529: return(1);
1.1 kristaps 530: }
531:
1.23 kristaps 532: int
1.213 schwarze 533: mdoc_elem_alloc(struct mdoc *mdoc, int line, int pos,
1.117 kristaps 534: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 535: {
536: struct mdoc_node *p;
537:
1.203 schwarze 538: p = node_alloc(mdoc, line, pos, tok, MDOC_ELEM);
1.77 kristaps 539: p->args = args;
540: if (p->args)
1.53 kristaps 541: (args->refcnt)++;
1.172 kristaps 542:
543: switch (tok) {
1.213 schwarze 544: case MDOC_An:
1.172 kristaps 545: p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
546: break;
547: default:
548: break;
549: }
550:
1.203 schwarze 551: if ( ! node_append(mdoc, p))
1.102 kristaps 552: return(0);
1.203 schwarze 553: mdoc->next = MDOC_NEXT_CHILD;
1.175 kristaps 554: return(1);
555: }
1.1 kristaps 556:
1.124 kristaps 557: int
1.203 schwarze 558: mdoc_word_alloc(struct mdoc *mdoc, int line, int pos, const char *p)
1.1 kristaps 559: {
1.91 kristaps 560: struct mdoc_node *n;
1.1 kristaps 561:
1.203 schwarze 562: n = node_alloc(mdoc, line, pos, MDOC_MAX, MDOC_TEXT);
563: n->string = roff_strdup(mdoc->roff, p);
1.91 kristaps 564:
1.203 schwarze 565: if ( ! node_append(mdoc, n))
1.101 kristaps 566: return(0);
1.124 kristaps 567:
1.203 schwarze 568: mdoc->next = MDOC_NEXT_SIBLING;
1.101 kristaps 569: return(1);
1.91 kristaps 570: }
571:
1.205 schwarze 572: void
573: mdoc_word_append(struct mdoc *mdoc, const char *p)
574: {
575: struct mdoc_node *n;
576: char *addstr, *newstr;
577:
578: n = mdoc->last;
579: addstr = roff_strdup(mdoc->roff, p);
1.210 schwarze 580: mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
1.205 schwarze 581: free(addstr);
582: free(n->string);
583: n->string = newstr;
584: mdoc->next = MDOC_NEXT_SIBLING;
585: }
1.91 kristaps 586:
1.155 kristaps 587: static void
1.53 kristaps 588: mdoc_node_free(struct mdoc_node *p)
1.1 kristaps 589: {
590:
1.172 kristaps 591: if (MDOC_BLOCK == p->type || MDOC_ELEM == p->type)
1.171 kristaps 592: free(p->norm);
1.53 kristaps 593: if (p->string)
594: free(p->string);
595: if (p->args)
596: mdoc_argv_free(p->args);
1.1 kristaps 597: free(p);
598: }
599:
1.121 kristaps 600: static void
1.203 schwarze 601: mdoc_node_unlink(struct mdoc *mdoc, struct mdoc_node *n)
1.121 kristaps 602: {
603:
604: /* Adjust siblings. */
605:
606: if (n->prev)
607: n->prev->next = n->next;
608: if (n->next)
609: n->next->prev = n->prev;
610:
611: /* Adjust parent. */
612:
613: if (n->parent) {
614: n->parent->nchild--;
615: if (n->parent->child == n)
616: n->parent->child = n->prev ? n->prev : n->next;
1.169 kristaps 617: if (n->parent->last == n)
618: n->parent->last = n->prev ? n->prev : NULL;
1.121 kristaps 619: }
620:
621: /* Adjust parse point, if applicable. */
622:
1.203 schwarze 623: if (mdoc && mdoc->last == n) {
1.121 kristaps 624: if (n->prev) {
1.203 schwarze 625: mdoc->last = n->prev;
626: mdoc->next = MDOC_NEXT_SIBLING;
1.121 kristaps 627: } else {
1.203 schwarze 628: mdoc->last = n->parent;
629: mdoc->next = MDOC_NEXT_CHILD;
1.121 kristaps 630: }
631: }
632:
1.203 schwarze 633: if (mdoc && mdoc->first == n)
634: mdoc->first = NULL;
1.121 kristaps 635: }
636:
1.53 kristaps 637: void
1.203 schwarze 638: mdoc_node_delete(struct mdoc *mdoc, struct mdoc_node *p)
1.1 kristaps 639: {
640:
1.121 kristaps 641: while (p->child) {
642: assert(p->nchild);
1.203 schwarze 643: mdoc_node_delete(mdoc, p->child);
1.121 kristaps 644: }
645: assert(0 == p->nchild);
1.1 kristaps 646:
1.203 schwarze 647: mdoc_node_unlink(mdoc, p);
1.53 kristaps 648: mdoc_node_free(p);
1.201 schwarze 649: }
650:
651: int
1.203 schwarze 652: mdoc_node_relink(struct mdoc *mdoc, struct mdoc_node *p)
1.201 schwarze 653: {
654:
1.203 schwarze 655: mdoc_node_unlink(mdoc, p);
656: return(node_append(mdoc, p));
1.1 kristaps 657: }
658:
1.193 kristaps 659: #if 0
1.191 kristaps 660: /*
661: * Pre-treat a text line.
662: * Text lines can consist of equations, which must be handled apart from
663: * the regular text.
664: * Thus, use this function to step through a line checking if it has any
665: * equations embedded in it.
666: * This must handle multiple equations AND equations that do not end at
667: * the end-of-line, i.e., will re-enter in the next roff parse.
668: */
669: static int
1.203 schwarze 670: mdoc_preptext(struct mdoc *mdoc, int line, char *buf, int offs)
1.191 kristaps 671: {
672: char *start, *end;
673: char delim;
674:
675: while ('\0' != buf[offs]) {
676: /* Mark starting position if eqn is set. */
677: start = NULL;
1.203 schwarze 678: if ('\0' != (delim = roff_eqndelim(mdoc->roff)))
1.191 kristaps 679: if (NULL != (start = strchr(buf + offs, delim)))
680: *start++ = '\0';
681:
682: /* Parse text as normal. */
1.203 schwarze 683: if ( ! mdoc_ptext(mdoc, line, buf, offs))
1.191 kristaps 684: return(0);
685:
686: /* Continue only if an equation exists. */
687: if (NULL == start)
688: break;
689:
690: /* Read past the end of the equation. */
691: offs += start - (buf + offs);
692: assert(start == &buf[offs]);
693: if (NULL != (end = strchr(buf + offs, delim))) {
694: *end++ = '\0';
695: while (' ' == *end)
696: end++;
697: }
698:
699: /* Parse the equation itself. */
1.203 schwarze 700: roff_openeqn(mdoc->roff, NULL, line, offs, buf);
1.191 kristaps 701:
702: /* Process a finished equation? */
1.203 schwarze 703: if (roff_closeeqn(mdoc->roff))
704: if ( ! mdoc_addeqn(mdoc, roff_eqn(mdoc->roff)))
1.191 kristaps 705: return(0);
706: offs += (end - (buf + offs));
1.213 schwarze 707: }
1.191 kristaps 708:
709: return(1);
710: }
1.193 kristaps 711: #endif
1.1 kristaps 712:
1.53 kristaps 713: /*
714: * Parse free-form text, that is, a line that does not begin with the
715: * control character.
716: */
717: static int
1.203 schwarze 718: mdoc_ptext(struct mdoc *mdoc, int line, char *buf, int offs)
1.1 kristaps 719: {
1.142 kristaps 720: char *c, *ws, *end;
721: struct mdoc_node *n;
1.1 kristaps 722:
1.124 kristaps 723: /* No text before an initial macro. */
724:
1.203 schwarze 725: if (SEC_NONE == mdoc->lastnamed) {
726: mdoc_pmsg(mdoc, line, offs, MANDOCERR_NOTEXT);
1.174 kristaps 727: return(1);
728: }
1.142 kristaps 729:
1.203 schwarze 730: assert(mdoc->last);
731: n = mdoc->last;
1.142 kristaps 732:
733: /*
1.144 kristaps 734: * Divert directly to list processing if we're encountering a
1.142 kristaps 735: * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
1.144 kristaps 736: * (a MDOC_BODY means it's already open, in which case we should
737: * process within its context in the normal way).
1.142 kristaps 738: */
739:
1.143 kristaps 740: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.213 schwarze 741: LIST_column == n->norm->Bl.type) {
1.144 kristaps 742: /* `Bl' is open without any children. */
1.203 schwarze 743: mdoc->flags |= MDOC_FREECOL;
744: return(mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf));
1.142 kristaps 745: }
746:
747: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
1.213 schwarze 748: NULL != n->parent &&
749: MDOC_Bl == n->parent->tok &&
750: LIST_column == n->parent->norm->Bl.type) {
1.144 kristaps 751: /* `Bl' has block-level `It' children. */
1.203 schwarze 752: mdoc->flags |= MDOC_FREECOL;
753: return(mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf));
1.142 kristaps 754: }
1.124 kristaps 755:
1.137 schwarze 756: /*
757: * Search for the beginning of unescaped trailing whitespace (ws)
758: * and for the first character not to be output (end).
759: */
1.139 kristaps 760:
761: /* FIXME: replace with strcspn(). */
1.137 schwarze 762: ws = NULL;
763: for (c = end = buf + offs; *c; c++) {
764: switch (*c) {
765: case ' ':
766: if (NULL == ws)
767: ws = c;
768: continue;
769: case '\t':
770: /*
771: * Always warn about trailing tabs,
772: * even outside literal context,
773: * where they should be put on the next line.
774: */
775: if (NULL == ws)
776: ws = c;
777: /*
778: * Strip trailing tabs in literal context only;
779: * outside, they affect the next line.
780: */
1.203 schwarze 781: if (MDOC_LITERAL & mdoc->flags)
1.137 schwarze 782: continue;
783: break;
784: case '\\':
785: /* Skip the escaped character, too, if any. */
786: if (c[1])
787: c++;
788: /* FALLTHROUGH */
789: default:
790: ws = NULL;
791: break;
792: }
793: end = c + 1;
794: }
795: *end = '\0';
1.91 kristaps 796:
1.137 schwarze 797: if (ws)
1.203 schwarze 798: mdoc_pmsg(mdoc, line, (int)(ws-buf), MANDOCERR_EOLNSPACE);
1.115 kristaps 799:
1.203 schwarze 800: if ('\0' == buf[offs] && ! (MDOC_LITERAL & mdoc->flags)) {
801: mdoc_pmsg(mdoc, line, (int)(c-buf), MANDOCERR_NOBLANKLN);
1.124 kristaps 802:
1.119 kristaps 803: /*
1.165 schwarze 804: * Insert a `sp' in the case of a blank line. Technically,
1.124 kristaps 805: * blank lines aren't allowed, but enough manuals assume this
806: * behaviour that we want to work around it.
1.119 kristaps 807: */
1.203 schwarze 808: if ( ! mdoc_elem_alloc(mdoc, line, offs, MDOC_sp, NULL))
1.119 kristaps 809: return(0);
1.124 kristaps 810:
1.203 schwarze 811: mdoc->next = MDOC_NEXT_SIBLING;
1.199 schwarze 812:
1.203 schwarze 813: return(mdoc_valid_post(mdoc));
1.119 kristaps 814: }
1.68 kristaps 815:
1.203 schwarze 816: if ( ! mdoc_word_alloc(mdoc, line, offs, buf+offs))
1.137 schwarze 817: return(0);
1.91 kristaps 818:
1.203 schwarze 819: if (MDOC_LITERAL & mdoc->flags)
1.137 schwarze 820: return(1);
1.128 kristaps 821:
822: /*
823: * End-of-sentence check. If the last character is an unescaped
824: * EOS character, then flag the node as being the end of a
825: * sentence. The front-end will know how to interpret this.
826: */
1.132 kristaps 827:
1.137 schwarze 828: assert(buf < end);
829:
1.207 schwarze 830: if (mandoc_eos(buf+offs, (size_t)(end-buf-offs)))
1.203 schwarze 831: mdoc->last->flags |= MDOC_EOS;
1.128 kristaps 832:
833: return(1);
1.1 kristaps 834: }
835:
1.53 kristaps 836: /*
837: * Parse a macro line, that is, a line beginning with the control
838: * character.
839: */
1.155 kristaps 840: static int
1.203 schwarze 841: mdoc_pmacro(struct mdoc *mdoc, int ln, char *buf, int offs)
1.1 kristaps 842: {
1.144 kristaps 843: enum mdoct tok;
1.188 kristaps 844: int i, sv;
1.144 kristaps 845: char mac[5];
846: struct mdoc_node *n;
1.1 kristaps 847:
1.188 kristaps 848: /* Empty post-control lines are ignored. */
1.63 kristaps 849:
1.188 kristaps 850: if ('"' == buf[offs]) {
1.203 schwarze 851: mdoc_pmsg(mdoc, ln, offs, MANDOCERR_BADCOMMENT);
1.188 kristaps 852: return(1);
853: } else if ('\0' == buf[offs])
1.63 kristaps 854: return(1);
855:
1.188 kristaps 856: sv = offs;
1.130 kristaps 857:
1.213 schwarze 858: /*
1.162 schwarze 859: * Copy the first word into a nil-terminated buffer.
1.165 schwarze 860: * Stop copying when a tab, space, or eoln is encountered.
1.160 kristaps 861: */
1.1 kristaps 862:
1.188 kristaps 863: i = 0;
1.213 schwarze 864: while (i < 4 && '\0' != buf[offs] && ' ' != buf[offs] &&
865: '\t' != buf[offs])
1.188 kristaps 866: mac[i++] = buf[offs++];
867:
868: mac[i] = '\0';
869:
1.214 schwarze 870: tok = (i > 1 && i < 4) ? mdoc_hash_find(mac) : MDOC_MAX;
1.1 kristaps 871:
1.163 schwarze 872: if (MDOC_MAX == tok) {
1.213 schwarze 873: mandoc_vmsg(MANDOCERR_MACRO, mdoc->parse,
874: ln, sv, "%s", buf + sv - 1);
1.58 kristaps 875: return(1);
1.53 kristaps 876: }
1.1 kristaps 877:
1.160 kristaps 878: /* Disregard the first trailing tab, if applicable. */
879:
1.188 kristaps 880: if ('\t' == buf[offs])
881: offs++;
1.160 kristaps 882:
883: /* Jump to the next non-whitespace word. */
1.1 kristaps 884:
1.188 kristaps 885: while (buf[offs] && ' ' == buf[offs])
886: offs++;
1.1 kristaps 887:
1.213 schwarze 888: /*
1.125 kristaps 889: * Trailing whitespace. Note that tabs are allowed to be passed
890: * into the parser as "text", so we only warn about spaces here.
891: */
1.115 kristaps 892:
1.188 kristaps 893: if ('\0' == buf[offs] && ' ' == buf[offs - 1])
1.203 schwarze 894: mdoc_pmsg(mdoc, ln, offs - 1, MANDOCERR_EOLNSPACE);
1.115 kristaps 895:
1.144 kristaps 896: /*
897: * If an initial macro or a list invocation, divert directly
898: * into macro processing.
899: */
900:
1.203 schwarze 901: if (NULL == mdoc->last || MDOC_It == tok || MDOC_El == tok) {
1.213 schwarze 902: if ( ! mdoc_macro(mdoc, tok, ln, sv, &offs, buf))
1.144 kristaps 903: goto err;
904: return(1);
905: }
906:
1.203 schwarze 907: n = mdoc->last;
908: assert(mdoc->last);
1.144 kristaps 909:
910: /*
911: * If the first macro of a `Bl -column', open an `It' block
912: * context around the parsed macro.
913: */
914:
915: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.213 schwarze 916: LIST_column == n->norm->Bl.type) {
1.203 schwarze 917: mdoc->flags |= MDOC_FREECOL;
918: if ( ! mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf))
1.144 kristaps 919: goto err;
920: return(1);
921: }
922:
923: /*
924: * If we're following a block-level `It' within a `Bl -column'
925: * context (perhaps opened in the above block or in ptext()),
926: * then open an `It' block context around the parsed macro.
1.98 kristaps 927: */
1.144 kristaps 928:
929: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
1.213 schwarze 930: NULL != n->parent &&
931: MDOC_Bl == n->parent->tok &&
932: LIST_column == n->parent->norm->Bl.type) {
1.203 schwarze 933: mdoc->flags |= MDOC_FREECOL;
1.213 schwarze 934: if ( ! mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf))
1.144 kristaps 935: goto err;
936: return(1);
937: }
938:
939: /* Normal processing of a macro. */
940:
1.213 schwarze 941: if ( ! mdoc_macro(mdoc, tok, ln, sv, &offs, buf))
1.53 kristaps 942: goto err;
1.208 schwarze 943:
944: /* In quick mode (for mandocdb), abort after the NAME section. */
945:
946: if (mdoc->quick && MDOC_Sh == tok &&
947: SEC_NAME != mdoc->last->sec)
948: return(2);
1.1 kristaps 949:
1.53 kristaps 950: return(1);
1.1 kristaps 951:
1.53 kristaps 952: err: /* Error out. */
1.1 kristaps 953:
1.203 schwarze 954: mdoc->flags |= MDOC_HALT;
1.53 kristaps 955: return(0);
1.1 kristaps 956: }
1.100 kristaps 957:
1.186 kristaps 958: enum mdelim
959: mdoc_isdelim(const char *p)
960: {
961:
962: if ('\0' == p[0])
963: return(DELIM_NONE);
964:
965: if ('\0' == p[1])
966: switch (p[0]) {
1.213 schwarze 967: case '(':
1.186 kristaps 968: /* FALLTHROUGH */
1.213 schwarze 969: case '[':
1.186 kristaps 970: return(DELIM_OPEN);
1.213 schwarze 971: case '|':
1.186 kristaps 972: return(DELIM_MIDDLE);
1.213 schwarze 973: case '.':
1.186 kristaps 974: /* FALLTHROUGH */
1.213 schwarze 975: case ',':
1.186 kristaps 976: /* FALLTHROUGH */
1.213 schwarze 977: case ';':
1.186 kristaps 978: /* FALLTHROUGH */
1.213 schwarze 979: case ':':
1.186 kristaps 980: /* FALLTHROUGH */
1.213 schwarze 981: case '?':
1.186 kristaps 982: /* FALLTHROUGH */
1.213 schwarze 983: case '!':
1.186 kristaps 984: /* FALLTHROUGH */
1.213 schwarze 985: case ')':
1.186 kristaps 986: /* FALLTHROUGH */
1.213 schwarze 987: case ']':
1.186 kristaps 988: return(DELIM_CLOSE);
989: default:
990: return(DELIM_NONE);
991: }
992:
993: if ('\\' != p[0])
994: return(DELIM_NONE);
1.100 kristaps 995:
1.186 kristaps 996: if (0 == strcmp(p + 1, "."))
997: return(DELIM_CLOSE);
1.200 schwarze 998: if (0 == strcmp(p + 1, "fR|\\fP"))
1.186 kristaps 999: return(DELIM_MIDDLE);
1000:
1001: return(DELIM_NONE);
1.211 schwarze 1002: }
1003:
1004: void
1005: mdoc_deroff(char **dest, const struct mdoc_node *n)
1006: {
1007: char *cp;
1008: size_t sz;
1009:
1010: if (MDOC_TEXT != n->type) {
1011: for (n = n->child; n; n = n->next)
1012: mdoc_deroff(dest, n);
1013: return;
1014: }
1015:
1016: /* Skip leading whitespace. */
1017:
1018: for (cp = n->string; '\0' != *cp; cp++)
1019: if (0 == isspace((unsigned char)*cp))
1020: break;
1021:
1022: /* Skip trailing whitespace. */
1023:
1024: for (sz = strlen(cp); sz; sz--)
1025: if (0 == isspace((unsigned char)cp[sz-1]))
1026: break;
1027:
1028: /* Skip empty strings. */
1029:
1030: if (0 == sz)
1031: return;
1032:
1033: if (NULL == *dest) {
1034: *dest = mandoc_strndup(cp, sz);
1035: return;
1036: }
1037:
1038: mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
1039: free(*dest);
1040: *dest = cp;
1.186 kristaps 1041: }
CVSweb