Annotation of mandoc/mdoc.c, Revision 1.214
1.214 ! schwarze 1: /* $Id: mdoc.c,v 1.213 2014/04/20 16:46:05 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) {
1.203 schwarze 308: mdoc_pmsg(mdoc, line, ppos, MANDOCERR_BADBODY);
1.174 kristaps 309: return(1);
310: }
1.122 kristaps 311:
312: /* If we're in the prologue, deny "body" macros. */
313:
1.213 schwarze 314: if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
315: ! (MDOC_PBODY & mdoc->flags)) {
1.203 schwarze 316: mdoc_pmsg(mdoc, line, ppos, MANDOCERR_BADPROLOG);
317: if (NULL == mdoc->meta.msec)
318: mdoc->meta.msec = mandoc_strdup("1");
319: if (NULL == mdoc->meta.title)
320: mdoc->meta.title = mandoc_strdup("UNKNOWN");
321: if (NULL == mdoc->meta.vol)
322: mdoc->meta.vol = mandoc_strdup("LOCAL");
323: if (NULL == mdoc->meta.os)
324: mdoc->meta.os = mandoc_strdup("LOCAL");
325: if (NULL == mdoc->meta.date)
326: mdoc->meta.date = mandoc_normdate
327: (mdoc->parse, NULL, line, ppos);
328: mdoc->flags |= MDOC_PBODY;
1.120 kristaps 329: }
1.88 kristaps 330:
1.203 schwarze 331: return((*mdoc_macros[tok].fp)(mdoc, tok, line, ppos, pos, buf));
1.73 kristaps 332: }
333:
334:
335: static int
336: node_append(struct mdoc *mdoc, struct mdoc_node *p)
1.1 kristaps 337: {
338:
1.25 kristaps 339: assert(mdoc->last);
340: assert(mdoc->first);
341: assert(MDOC_ROOT != p->type);
1.1 kristaps 342:
1.13 kristaps 343: switch (mdoc->next) {
1.213 schwarze 344: case MDOC_NEXT_SIBLING:
1.6 kristaps 345: mdoc->last->next = p;
346: p->prev = mdoc->last;
1.13 kristaps 347: p->parent = mdoc->last->parent;
1.1 kristaps 348: break;
1.213 schwarze 349: case MDOC_NEXT_CHILD:
1.13 kristaps 350: mdoc->last->child = p;
1.1 kristaps 351: p->parent = mdoc->last;
352: break;
353: default:
1.13 kristaps 354: abort();
355: /* NOTREACHED */
1.1 kristaps 356: }
357:
1.86 kristaps 358: p->parent->nchild++;
359:
1.172 kristaps 360: /*
361: * Copy over the normalised-data pointer of our parent. Not
362: * everybody has one, but copying a null pointer is fine.
363: */
364:
365: switch (p->type) {
1.213 schwarze 366: case MDOC_BODY:
1.202 schwarze 367: if (ENDBODY_NOT != p->end)
368: break;
1.172 kristaps 369: /* FALLTHROUGH */
1.213 schwarze 370: case MDOC_TAIL:
1.172 kristaps 371: /* FALLTHROUGH */
1.213 schwarze 372: case MDOC_HEAD:
1.172 kristaps 373: p->norm = p->parent->norm;
374: break;
375: default:
376: break;
377: }
378:
1.23 kristaps 379: if ( ! mdoc_valid_pre(mdoc, p))
380: return(0);
1.27 kristaps 381:
382: switch (p->type) {
1.213 schwarze 383: case MDOC_HEAD:
1.27 kristaps 384: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 385: p->parent->head = p;
1.27 kristaps 386: break;
1.213 schwarze 387: case MDOC_TAIL:
1.27 kristaps 388: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 389: p->parent->tail = p;
1.27 kristaps 390: break;
1.213 schwarze 391: case MDOC_BODY:
1.152 schwarze 392: if (p->end)
393: break;
1.27 kristaps 394: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 395: p->parent->body = p;
1.27 kristaps 396: break;
397: default:
398: break;
399: }
400:
1.1 kristaps 401: mdoc->last = p;
1.71 kristaps 402:
403: switch (p->type) {
1.213 schwarze 404: case MDOC_TBL:
1.176 kristaps 405: /* FALLTHROUGH */
1.213 schwarze 406: case MDOC_TEXT:
1.71 kristaps 407: if ( ! mdoc_valid_post(mdoc))
408: return(0);
409: break;
410: default:
411: break;
412: }
413:
1.23 kristaps 414: return(1);
1.1 kristaps 415: }
416:
1.45 kristaps 417: static struct mdoc_node *
1.213 schwarze 418: node_alloc(struct mdoc *mdoc, int line, int pos,
1.117 kristaps 419: enum mdoct tok, enum mdoc_type type)
1.45 kristaps 420: {
1.46 kristaps 421: struct mdoc_node *p;
422:
1.113 kristaps 423: p = mandoc_calloc(1, sizeof(struct mdoc_node));
1.203 schwarze 424: p->sec = mdoc->lastsec;
1.73 kristaps 425: p->line = line;
426: p->pos = pos;
1.206 schwarze 427: p->lastline = line;
1.73 kristaps 428: p->tok = tok;
1.118 kristaps 429: p->type = type;
1.150 kristaps 430:
431: /* Flag analysis. */
432:
1.203 schwarze 433: if (MDOC_SYNOPSIS & mdoc->flags)
1.153 schwarze 434: p->flags |= MDOC_SYNPRETTY;
435: else
436: p->flags &= ~MDOC_SYNPRETTY;
1.203 schwarze 437: if (MDOC_NEWLINE & mdoc->flags)
1.130 kristaps 438: p->flags |= MDOC_LINE;
1.203 schwarze 439: mdoc->flags &= ~MDOC_NEWLINE;
1.150 kristaps 440:
1.46 kristaps 441: return(p);
1.45 kristaps 442: }
443:
1.23 kristaps 444: int
1.203 schwarze 445: mdoc_tail_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.17 kristaps 446: {
447: struct mdoc_node *p;
448:
1.203 schwarze 449: p = node_alloc(mdoc, line, pos, tok, MDOC_TAIL);
450: if ( ! node_append(mdoc, p))
1.102 kristaps 451: return(0);
1.203 schwarze 452: mdoc->next = MDOC_NEXT_CHILD;
1.102 kristaps 453: return(1);
1.17 kristaps 454: }
455:
1.23 kristaps 456: int
1.203 schwarze 457: mdoc_head_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.1 kristaps 458: {
459: struct mdoc_node *p;
460:
1.203 schwarze 461: assert(mdoc->first);
462: assert(mdoc->last);
1.1 kristaps 463:
1.203 schwarze 464: p = node_alloc(mdoc, line, pos, tok, MDOC_HEAD);
465: if ( ! node_append(mdoc, p))
1.102 kristaps 466: return(0);
1.203 schwarze 467: mdoc->next = MDOC_NEXT_CHILD;
1.102 kristaps 468: return(1);
1.1 kristaps 469: }
470:
1.23 kristaps 471: int
1.203 schwarze 472: mdoc_body_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.1 kristaps 473: {
474: struct mdoc_node *p;
475:
1.203 schwarze 476: p = node_alloc(mdoc, line, pos, tok, MDOC_BODY);
477: if ( ! node_append(mdoc, p))
1.102 kristaps 478: return(0);
1.203 schwarze 479: mdoc->next = MDOC_NEXT_CHILD;
1.152 schwarze 480: return(1);
481: }
482:
483: int
1.203 schwarze 484: mdoc_endbody_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok,
1.152 schwarze 485: struct mdoc_node *body, enum mdoc_endbody end)
486: {
487: struct mdoc_node *p;
488:
1.203 schwarze 489: p = node_alloc(mdoc, line, pos, tok, MDOC_BODY);
1.152 schwarze 490: p->pending = body;
1.202 schwarze 491: p->norm = body->norm;
1.152 schwarze 492: p->end = end;
1.203 schwarze 493: if ( ! node_append(mdoc, p))
1.152 schwarze 494: return(0);
1.203 schwarze 495: mdoc->next = MDOC_NEXT_SIBLING;
1.102 kristaps 496: return(1);
1.1 kristaps 497: }
498:
1.23 kristaps 499: int
1.213 schwarze 500: mdoc_block_alloc(struct mdoc *mdoc, int line, int pos,
1.117 kristaps 501: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 502: {
503: struct mdoc_node *p;
504:
1.203 schwarze 505: p = node_alloc(mdoc, line, pos, tok, MDOC_BLOCK);
1.77 kristaps 506: p->args = args;
507: if (p->args)
1.53 kristaps 508: (args->refcnt)++;
1.172 kristaps 509:
510: switch (tok) {
1.213 schwarze 511: case MDOC_Bd:
1.172 kristaps 512: /* FALLTHROUGH */
1.213 schwarze 513: case MDOC_Bf:
1.172 kristaps 514: /* FALLTHROUGH */
1.213 schwarze 515: case MDOC_Bl:
1.173 kristaps 516: /* FALLTHROUGH */
1.213 schwarze 517: case MDOC_Rs:
1.172 kristaps 518: p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
519: break;
520: default:
521: break;
522: }
523:
1.203 schwarze 524: if ( ! node_append(mdoc, p))
1.102 kristaps 525: return(0);
1.203 schwarze 526: mdoc->next = MDOC_NEXT_CHILD;
1.102 kristaps 527: return(1);
1.1 kristaps 528: }
529:
1.23 kristaps 530: int
1.213 schwarze 531: mdoc_elem_alloc(struct mdoc *mdoc, int line, int pos,
1.117 kristaps 532: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 533: {
534: struct mdoc_node *p;
535:
1.203 schwarze 536: p = node_alloc(mdoc, line, pos, tok, MDOC_ELEM);
1.77 kristaps 537: p->args = args;
538: if (p->args)
1.53 kristaps 539: (args->refcnt)++;
1.172 kristaps 540:
541: switch (tok) {
1.213 schwarze 542: case MDOC_An:
1.172 kristaps 543: p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
544: break;
545: default:
546: break;
547: }
548:
1.203 schwarze 549: if ( ! node_append(mdoc, p))
1.102 kristaps 550: return(0);
1.203 schwarze 551: mdoc->next = MDOC_NEXT_CHILD;
1.175 kristaps 552: return(1);
553: }
1.1 kristaps 554:
1.124 kristaps 555: int
1.203 schwarze 556: mdoc_word_alloc(struct mdoc *mdoc, int line, int pos, const char *p)
1.1 kristaps 557: {
1.91 kristaps 558: struct mdoc_node *n;
1.1 kristaps 559:
1.203 schwarze 560: n = node_alloc(mdoc, line, pos, MDOC_MAX, MDOC_TEXT);
561: n->string = roff_strdup(mdoc->roff, p);
1.91 kristaps 562:
1.203 schwarze 563: if ( ! node_append(mdoc, n))
1.101 kristaps 564: return(0);
1.124 kristaps 565:
1.203 schwarze 566: mdoc->next = MDOC_NEXT_SIBLING;
1.101 kristaps 567: return(1);
1.91 kristaps 568: }
569:
1.205 schwarze 570: void
571: mdoc_word_append(struct mdoc *mdoc, const char *p)
572: {
573: struct mdoc_node *n;
574: char *addstr, *newstr;
575:
576: n = mdoc->last;
577: addstr = roff_strdup(mdoc->roff, p);
1.210 schwarze 578: mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
1.205 schwarze 579: free(addstr);
580: free(n->string);
581: n->string = newstr;
582: mdoc->next = MDOC_NEXT_SIBLING;
583: }
1.91 kristaps 584:
1.155 kristaps 585: static void
1.53 kristaps 586: mdoc_node_free(struct mdoc_node *p)
1.1 kristaps 587: {
588:
1.172 kristaps 589: if (MDOC_BLOCK == p->type || MDOC_ELEM == p->type)
1.171 kristaps 590: free(p->norm);
1.53 kristaps 591: if (p->string)
592: free(p->string);
593: if (p->args)
594: mdoc_argv_free(p->args);
1.1 kristaps 595: free(p);
596: }
597:
1.121 kristaps 598: static void
1.203 schwarze 599: mdoc_node_unlink(struct mdoc *mdoc, struct mdoc_node *n)
1.121 kristaps 600: {
601:
602: /* Adjust siblings. */
603:
604: if (n->prev)
605: n->prev->next = n->next;
606: if (n->next)
607: n->next->prev = n->prev;
608:
609: /* Adjust parent. */
610:
611: if (n->parent) {
612: n->parent->nchild--;
613: if (n->parent->child == n)
614: n->parent->child = n->prev ? n->prev : n->next;
1.169 kristaps 615: if (n->parent->last == n)
616: n->parent->last = n->prev ? n->prev : NULL;
1.121 kristaps 617: }
618:
619: /* Adjust parse point, if applicable. */
620:
1.203 schwarze 621: if (mdoc && mdoc->last == n) {
1.121 kristaps 622: if (n->prev) {
1.203 schwarze 623: mdoc->last = n->prev;
624: mdoc->next = MDOC_NEXT_SIBLING;
1.121 kristaps 625: } else {
1.203 schwarze 626: mdoc->last = n->parent;
627: mdoc->next = MDOC_NEXT_CHILD;
1.121 kristaps 628: }
629: }
630:
1.203 schwarze 631: if (mdoc && mdoc->first == n)
632: mdoc->first = NULL;
1.121 kristaps 633: }
634:
1.53 kristaps 635: void
1.203 schwarze 636: mdoc_node_delete(struct mdoc *mdoc, struct mdoc_node *p)
1.1 kristaps 637: {
638:
1.121 kristaps 639: while (p->child) {
640: assert(p->nchild);
1.203 schwarze 641: mdoc_node_delete(mdoc, p->child);
1.121 kristaps 642: }
643: assert(0 == p->nchild);
1.1 kristaps 644:
1.203 schwarze 645: mdoc_node_unlink(mdoc, p);
1.53 kristaps 646: mdoc_node_free(p);
1.201 schwarze 647: }
648:
649: int
1.203 schwarze 650: mdoc_node_relink(struct mdoc *mdoc, struct mdoc_node *p)
1.201 schwarze 651: {
652:
1.203 schwarze 653: mdoc_node_unlink(mdoc, p);
654: return(node_append(mdoc, p));
1.1 kristaps 655: }
656:
1.193 kristaps 657: #if 0
1.191 kristaps 658: /*
659: * Pre-treat a text line.
660: * Text lines can consist of equations, which must be handled apart from
661: * the regular text.
662: * Thus, use this function to step through a line checking if it has any
663: * equations embedded in it.
664: * This must handle multiple equations AND equations that do not end at
665: * the end-of-line, i.e., will re-enter in the next roff parse.
666: */
667: static int
1.203 schwarze 668: mdoc_preptext(struct mdoc *mdoc, int line, char *buf, int offs)
1.191 kristaps 669: {
670: char *start, *end;
671: char delim;
672:
673: while ('\0' != buf[offs]) {
674: /* Mark starting position if eqn is set. */
675: start = NULL;
1.203 schwarze 676: if ('\0' != (delim = roff_eqndelim(mdoc->roff)))
1.191 kristaps 677: if (NULL != (start = strchr(buf + offs, delim)))
678: *start++ = '\0';
679:
680: /* Parse text as normal. */
1.203 schwarze 681: if ( ! mdoc_ptext(mdoc, line, buf, offs))
1.191 kristaps 682: return(0);
683:
684: /* Continue only if an equation exists. */
685: if (NULL == start)
686: break;
687:
688: /* Read past the end of the equation. */
689: offs += start - (buf + offs);
690: assert(start == &buf[offs]);
691: if (NULL != (end = strchr(buf + offs, delim))) {
692: *end++ = '\0';
693: while (' ' == *end)
694: end++;
695: }
696:
697: /* Parse the equation itself. */
1.203 schwarze 698: roff_openeqn(mdoc->roff, NULL, line, offs, buf);
1.191 kristaps 699:
700: /* Process a finished equation? */
1.203 schwarze 701: if (roff_closeeqn(mdoc->roff))
702: if ( ! mdoc_addeqn(mdoc, roff_eqn(mdoc->roff)))
1.191 kristaps 703: return(0);
704: offs += (end - (buf + offs));
1.213 schwarze 705: }
1.191 kristaps 706:
707: return(1);
708: }
1.193 kristaps 709: #endif
1.1 kristaps 710:
1.53 kristaps 711: /*
712: * Parse free-form text, that is, a line that does not begin with the
713: * control character.
714: */
715: static int
1.203 schwarze 716: mdoc_ptext(struct mdoc *mdoc, int line, char *buf, int offs)
1.1 kristaps 717: {
1.142 kristaps 718: char *c, *ws, *end;
719: struct mdoc_node *n;
1.1 kristaps 720:
1.124 kristaps 721: /* No text before an initial macro. */
722:
1.203 schwarze 723: if (SEC_NONE == mdoc->lastnamed) {
724: mdoc_pmsg(mdoc, line, offs, MANDOCERR_NOTEXT);
1.174 kristaps 725: return(1);
726: }
1.142 kristaps 727:
1.203 schwarze 728: assert(mdoc->last);
729: n = mdoc->last;
1.142 kristaps 730:
731: /*
1.144 kristaps 732: * Divert directly to list processing if we're encountering a
1.142 kristaps 733: * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
1.144 kristaps 734: * (a MDOC_BODY means it's already open, in which case we should
735: * process within its context in the normal way).
1.142 kristaps 736: */
737:
1.143 kristaps 738: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.213 schwarze 739: LIST_column == n->norm->Bl.type) {
1.144 kristaps 740: /* `Bl' is open without any children. */
1.203 schwarze 741: mdoc->flags |= MDOC_FREECOL;
742: return(mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf));
1.142 kristaps 743: }
744:
745: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
1.213 schwarze 746: NULL != n->parent &&
747: MDOC_Bl == n->parent->tok &&
748: LIST_column == n->parent->norm->Bl.type) {
1.144 kristaps 749: /* `Bl' has block-level `It' children. */
1.203 schwarze 750: mdoc->flags |= MDOC_FREECOL;
751: return(mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf));
1.142 kristaps 752: }
1.124 kristaps 753:
1.137 schwarze 754: /*
755: * Search for the beginning of unescaped trailing whitespace (ws)
756: * and for the first character not to be output (end).
757: */
1.139 kristaps 758:
759: /* FIXME: replace with strcspn(). */
1.137 schwarze 760: ws = NULL;
761: for (c = end = buf + offs; *c; c++) {
762: switch (*c) {
763: case ' ':
764: if (NULL == ws)
765: ws = c;
766: continue;
767: case '\t':
768: /*
769: * Always warn about trailing tabs,
770: * even outside literal context,
771: * where they should be put on the next line.
772: */
773: if (NULL == ws)
774: ws = c;
775: /*
776: * Strip trailing tabs in literal context only;
777: * outside, they affect the next line.
778: */
1.203 schwarze 779: if (MDOC_LITERAL & mdoc->flags)
1.137 schwarze 780: continue;
781: break;
782: case '\\':
783: /* Skip the escaped character, too, if any. */
784: if (c[1])
785: c++;
786: /* FALLTHROUGH */
787: default:
788: ws = NULL;
789: break;
790: }
791: end = c + 1;
792: }
793: *end = '\0';
1.91 kristaps 794:
1.137 schwarze 795: if (ws)
1.203 schwarze 796: mdoc_pmsg(mdoc, line, (int)(ws-buf), MANDOCERR_EOLNSPACE);
1.115 kristaps 797:
1.203 schwarze 798: if ('\0' == buf[offs] && ! (MDOC_LITERAL & mdoc->flags)) {
799: mdoc_pmsg(mdoc, line, (int)(c-buf), MANDOCERR_NOBLANKLN);
1.124 kristaps 800:
1.119 kristaps 801: /*
1.165 schwarze 802: * Insert a `sp' in the case of a blank line. Technically,
1.124 kristaps 803: * blank lines aren't allowed, but enough manuals assume this
804: * behaviour that we want to work around it.
1.119 kristaps 805: */
1.203 schwarze 806: if ( ! mdoc_elem_alloc(mdoc, line, offs, MDOC_sp, NULL))
1.119 kristaps 807: return(0);
1.124 kristaps 808:
1.203 schwarze 809: mdoc->next = MDOC_NEXT_SIBLING;
1.199 schwarze 810:
1.203 schwarze 811: return(mdoc_valid_post(mdoc));
1.119 kristaps 812: }
1.68 kristaps 813:
1.203 schwarze 814: if ( ! mdoc_word_alloc(mdoc, line, offs, buf+offs))
1.137 schwarze 815: return(0);
1.91 kristaps 816:
1.203 schwarze 817: if (MDOC_LITERAL & mdoc->flags)
1.137 schwarze 818: return(1);
1.128 kristaps 819:
820: /*
821: * End-of-sentence check. If the last character is an unescaped
822: * EOS character, then flag the node as being the end of a
823: * sentence. The front-end will know how to interpret this.
824: */
1.132 kristaps 825:
1.137 schwarze 826: assert(buf < end);
827:
1.207 schwarze 828: if (mandoc_eos(buf+offs, (size_t)(end-buf-offs)))
1.203 schwarze 829: mdoc->last->flags |= MDOC_EOS;
1.128 kristaps 830:
831: return(1);
1.1 kristaps 832: }
833:
1.53 kristaps 834: /*
835: * Parse a macro line, that is, a line beginning with the control
836: * character.
837: */
1.155 kristaps 838: static int
1.203 schwarze 839: mdoc_pmacro(struct mdoc *mdoc, int ln, char *buf, int offs)
1.1 kristaps 840: {
1.144 kristaps 841: enum mdoct tok;
1.188 kristaps 842: int i, sv;
1.144 kristaps 843: char mac[5];
844: struct mdoc_node *n;
1.1 kristaps 845:
1.188 kristaps 846: /* Empty post-control lines are ignored. */
1.63 kristaps 847:
1.188 kristaps 848: if ('"' == buf[offs]) {
1.203 schwarze 849: mdoc_pmsg(mdoc, ln, offs, MANDOCERR_BADCOMMENT);
1.188 kristaps 850: return(1);
851: } else if ('\0' == buf[offs])
1.63 kristaps 852: return(1);
853:
1.188 kristaps 854: sv = offs;
1.130 kristaps 855:
1.213 schwarze 856: /*
1.162 schwarze 857: * Copy the first word into a nil-terminated buffer.
1.165 schwarze 858: * Stop copying when a tab, space, or eoln is encountered.
1.160 kristaps 859: */
1.1 kristaps 860:
1.188 kristaps 861: i = 0;
1.213 schwarze 862: while (i < 4 && '\0' != buf[offs] && ' ' != buf[offs] &&
863: '\t' != buf[offs])
1.188 kristaps 864: mac[i++] = buf[offs++];
865:
866: mac[i] = '\0';
867:
1.214 ! schwarze 868: tok = (i > 1 && i < 4) ? mdoc_hash_find(mac) : MDOC_MAX;
1.1 kristaps 869:
1.163 schwarze 870: if (MDOC_MAX == tok) {
1.213 schwarze 871: mandoc_vmsg(MANDOCERR_MACRO, mdoc->parse,
872: ln, sv, "%s", buf + sv - 1);
1.58 kristaps 873: return(1);
1.53 kristaps 874: }
1.1 kristaps 875:
1.160 kristaps 876: /* Disregard the first trailing tab, if applicable. */
877:
1.188 kristaps 878: if ('\t' == buf[offs])
879: offs++;
1.160 kristaps 880:
881: /* Jump to the next non-whitespace word. */
1.1 kristaps 882:
1.188 kristaps 883: while (buf[offs] && ' ' == buf[offs])
884: offs++;
1.1 kristaps 885:
1.213 schwarze 886: /*
1.125 kristaps 887: * Trailing whitespace. Note that tabs are allowed to be passed
888: * into the parser as "text", so we only warn about spaces here.
889: */
1.115 kristaps 890:
1.188 kristaps 891: if ('\0' == buf[offs] && ' ' == buf[offs - 1])
1.203 schwarze 892: mdoc_pmsg(mdoc, ln, offs - 1, MANDOCERR_EOLNSPACE);
1.115 kristaps 893:
1.144 kristaps 894: /*
895: * If an initial macro or a list invocation, divert directly
896: * into macro processing.
897: */
898:
1.203 schwarze 899: if (NULL == mdoc->last || MDOC_It == tok || MDOC_El == tok) {
1.213 schwarze 900: if ( ! mdoc_macro(mdoc, tok, ln, sv, &offs, buf))
1.144 kristaps 901: goto err;
902: return(1);
903: }
904:
1.203 schwarze 905: n = mdoc->last;
906: assert(mdoc->last);
1.144 kristaps 907:
908: /*
909: * If the first macro of a `Bl -column', open an `It' block
910: * context around the parsed macro.
911: */
912:
913: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.213 schwarze 914: LIST_column == n->norm->Bl.type) {
1.203 schwarze 915: mdoc->flags |= MDOC_FREECOL;
916: if ( ! mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf))
1.144 kristaps 917: goto err;
918: return(1);
919: }
920:
921: /*
922: * If we're following a block-level `It' within a `Bl -column'
923: * context (perhaps opened in the above block or in ptext()),
924: * then open an `It' block context around the parsed macro.
1.98 kristaps 925: */
1.144 kristaps 926:
927: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
1.213 schwarze 928: NULL != n->parent &&
929: MDOC_Bl == n->parent->tok &&
930: LIST_column == n->parent->norm->Bl.type) {
1.203 schwarze 931: mdoc->flags |= MDOC_FREECOL;
1.213 schwarze 932: if ( ! mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf))
1.144 kristaps 933: goto err;
934: return(1);
935: }
936:
937: /* Normal processing of a macro. */
938:
1.213 schwarze 939: if ( ! mdoc_macro(mdoc, tok, ln, sv, &offs, buf))
1.53 kristaps 940: goto err;
1.208 schwarze 941:
942: /* In quick mode (for mandocdb), abort after the NAME section. */
943:
944: if (mdoc->quick && MDOC_Sh == tok &&
945: SEC_NAME != mdoc->last->sec)
946: return(2);
1.1 kristaps 947:
1.53 kristaps 948: return(1);
1.1 kristaps 949:
1.53 kristaps 950: err: /* Error out. */
1.1 kristaps 951:
1.203 schwarze 952: mdoc->flags |= MDOC_HALT;
1.53 kristaps 953: return(0);
1.1 kristaps 954: }
1.100 kristaps 955:
1.186 kristaps 956: enum mdelim
957: mdoc_isdelim(const char *p)
958: {
959:
960: if ('\0' == p[0])
961: return(DELIM_NONE);
962:
963: if ('\0' == p[1])
964: switch (p[0]) {
1.213 schwarze 965: case '(':
1.186 kristaps 966: /* FALLTHROUGH */
1.213 schwarze 967: case '[':
1.186 kristaps 968: return(DELIM_OPEN);
1.213 schwarze 969: case '|':
1.186 kristaps 970: return(DELIM_MIDDLE);
1.213 schwarze 971: case '.':
1.186 kristaps 972: /* FALLTHROUGH */
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: return(DELIM_CLOSE);
987: default:
988: return(DELIM_NONE);
989: }
990:
991: if ('\\' != p[0])
992: return(DELIM_NONE);
1.100 kristaps 993:
1.186 kristaps 994: if (0 == strcmp(p + 1, "."))
995: return(DELIM_CLOSE);
1.200 schwarze 996: if (0 == strcmp(p + 1, "fR|\\fP"))
1.186 kristaps 997: return(DELIM_MIDDLE);
998:
999: return(DELIM_NONE);
1.211 schwarze 1000: }
1001:
1002: void
1003: mdoc_deroff(char **dest, const struct mdoc_node *n)
1004: {
1005: char *cp;
1006: size_t sz;
1007:
1008: if (MDOC_TEXT != n->type) {
1009: for (n = n->child; n; n = n->next)
1010: mdoc_deroff(dest, n);
1011: return;
1012: }
1013:
1014: /* Skip leading whitespace. */
1015:
1016: for (cp = n->string; '\0' != *cp; cp++)
1017: if (0 == isspace((unsigned char)*cp))
1018: break;
1019:
1020: /* Skip trailing whitespace. */
1021:
1022: for (sz = strlen(cp); sz; sz--)
1023: if (0 == isspace((unsigned char)cp[sz-1]))
1024: break;
1025:
1026: /* Skip empty strings. */
1027:
1028: if (0 == sz)
1029: return;
1030:
1031: if (NULL == *dest) {
1032: *dest = mandoc_strndup(cp, sz);
1033: return;
1034: }
1035:
1036: mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
1037: free(*dest);
1038: *dest = cp;
1.186 kristaps 1039: }
CVSweb