Annotation of mandoc/mdoc.c, Revision 1.212
1.212 ! schwarze 1: /* $Id: mdoc.c,v 1.211 2014/03/23 12:44:56 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:
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",
49: /* LINTED */
1.114 kristaps 50: "Xr", "%A", "%B", "%D",
1.1 kristaps 51: /* LINTED */
1.114 kristaps 52: "%I", "%J", "%N", "%O",
1.1 kristaps 53: /* LINTED */
1.114 kristaps 54: "%P", "%R", "%T", "%V",
1.1 kristaps 55: "Ac", "Ao", "Aq", "At",
56: "Bc", "Bf", "Bo", "Bq",
57: "Bsx", "Bx", "Db", "Dc",
58: "Do", "Dq", "Ec", "Ef",
59: "Em", "Eo", "Fx", "Ms",
60: "No", "Ns", "Nx", "Ox",
61: "Pc", "Pf", "Po", "Pq",
62: "Qc", "Ql", "Qo", "Qq",
63: "Re", "Rs", "Sc", "So",
64: "Sq", "Sm", "Sx", "Sy",
65: "Tn", "Ux", "Xc", "Xo",
66: "Fo", "Fc", "Oo", "Oc",
67: "Bk", "Ek", "Bt", "Hf",
1.82 kristaps 68: "Fr", "Ud", "Lb", "Lp",
69: "Lk", "Mt", "Brq", "Bro",
1.64 kristaps 70: /* LINTED */
1.114 kristaps 71: "Brc", "%C", "Es", "En",
1.69 kristaps 72: /* LINTED */
1.114 kristaps 73: "Dx", "%Q", "br", "sp",
1.110 kristaps 74: /* LINTED */
1.212 ! schwarze 75: "%U", "Ta", "ll",
1.1 kristaps 76: };
77:
78: const char *const __mdoc_argnames[MDOC_ARG_MAX] = {
79: "split", "nosplit", "ragged",
80: "unfilled", "literal", "file",
81: "offset", "bullet", "dash",
82: "hyphen", "item", "enum",
83: "tag", "diag", "hang",
84: "ohang", "inset", "column",
85: "width", "compact", "std",
1.52 kristaps 86: "filled", "words", "emphasis",
1.108 kristaps 87: "symbolic", "nested", "centered"
1.1 kristaps 88: };
89:
90: const char * const *mdoc_macronames = __mdoc_macronames;
91: const char * const *mdoc_argnames = __mdoc_argnames;
92:
1.121 kristaps 93: static void mdoc_node_free(struct mdoc_node *);
94: static void mdoc_node_unlink(struct mdoc *,
95: struct mdoc_node *);
1.73 kristaps 96: static void mdoc_free1(struct mdoc *);
1.113 kristaps 97: static void mdoc_alloc1(struct mdoc *);
1.73 kristaps 98: static struct mdoc_node *node_alloc(struct mdoc *, int, int,
1.117 kristaps 99: enum mdoct, enum mdoc_type);
1.73 kristaps 100: static int node_append(struct mdoc *,
1.71 kristaps 101: struct mdoc_node *);
1.193 kristaps 102: #if 0
1.191 kristaps 103: static int mdoc_preptext(struct mdoc *, int, char *, int);
1.193 kristaps 104: #endif
1.149 kristaps 105: static int mdoc_ptext(struct mdoc *, int, char *, int);
106: static int mdoc_pmacro(struct mdoc *, int, char *, int);
1.88 kristaps 107:
1.1 kristaps 108: const struct mdoc_node *
1.203 schwarze 109: mdoc_node(const struct mdoc *mdoc)
1.1 kristaps 110: {
111:
1.203 schwarze 112: assert( ! (MDOC_HALT & mdoc->flags));
113: return(mdoc->first);
1.1 kristaps 114: }
115:
116:
1.37 kristaps 117: const struct mdoc_meta *
1.203 schwarze 118: mdoc_meta(const struct mdoc *mdoc)
1.37 kristaps 119: {
120:
1.203 schwarze 121: assert( ! (MDOC_HALT & mdoc->flags));
122: return(&mdoc->meta);
1.37 kristaps 123: }
124:
125:
1.85 kristaps 126: /*
127: * Frees volatile resources (parse tree, meta-data, fields).
128: */
1.73 kristaps 129: static void
130: mdoc_free1(struct mdoc *mdoc)
1.67 kristaps 131: {
132:
133: if (mdoc->first)
1.121 kristaps 134: mdoc_node_delete(mdoc, mdoc->first);
1.67 kristaps 135: if (mdoc->meta.title)
136: free(mdoc->meta.title);
137: if (mdoc->meta.os)
138: free(mdoc->meta.os);
139: if (mdoc->meta.name)
140: free(mdoc->meta.name);
141: if (mdoc->meta.arch)
142: free(mdoc->meta.arch);
143: if (mdoc->meta.vol)
144: free(mdoc->meta.vol);
1.133 kristaps 145: if (mdoc->meta.msec)
146: free(mdoc->meta.msec);
1.183 kristaps 147: if (mdoc->meta.date)
148: free(mdoc->meta.date);
1.73 kristaps 149: }
150:
151:
1.85 kristaps 152: /*
153: * Allocate all volatile resources (parse tree, meta-data, fields).
154: */
1.113 kristaps 155: static void
1.73 kristaps 156: mdoc_alloc1(struct mdoc *mdoc)
157: {
1.67 kristaps 158:
1.112 kristaps 159: memset(&mdoc->meta, 0, sizeof(struct mdoc_meta));
1.67 kristaps 160: mdoc->flags = 0;
1.85 kristaps 161: mdoc->lastnamed = mdoc->lastsec = SEC_NONE;
1.113 kristaps 162: mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node));
1.70 kristaps 163: mdoc->first = mdoc->last;
1.67 kristaps 164: mdoc->last->type = MDOC_ROOT;
1.196 schwarze 165: mdoc->last->tok = MDOC_MAX;
1.67 kristaps 166: mdoc->next = MDOC_NEXT_CHILD;
1.73 kristaps 167: }
168:
169:
170: /*
1.85 kristaps 171: * Free up volatile resources (see mdoc_free1()) then re-initialises the
172: * data with mdoc_alloc1(). After invocation, parse data has been reset
173: * and the parser is ready for re-invocation on a new tree; however,
174: * cross-parse non-volatile data is kept intact.
1.73 kristaps 175: */
1.113 kristaps 176: void
1.73 kristaps 177: mdoc_reset(struct mdoc *mdoc)
178: {
179:
180: mdoc_free1(mdoc);
1.113 kristaps 181: mdoc_alloc1(mdoc);
1.67 kristaps 182: }
183:
184:
1.68 kristaps 185: /*
1.85 kristaps 186: * Completely free up all volatile and non-volatile parse resources.
187: * After invocation, the pointer is no longer usable.
1.68 kristaps 188: */
1.67 kristaps 189: void
1.38 kristaps 190: mdoc_free(struct mdoc *mdoc)
1.34 kristaps 191: {
192:
1.73 kristaps 193: mdoc_free1(mdoc);
1.1 kristaps 194: free(mdoc);
195: }
196:
197:
1.85 kristaps 198: /*
199: * Allocate volatile and non-volatile parse resources.
200: */
1.1 kristaps 201: struct mdoc *
1.208 schwarze 202: mdoc_alloc(struct roff *roff, struct mparse *parse,
203: char *defos, int quick)
1.1 kristaps 204: {
205: struct mdoc *p;
206:
1.113 kristaps 207: p = mandoc_calloc(1, sizeof(struct mdoc));
208:
1.185 kristaps 209: p->parse = parse;
1.197 schwarze 210: p->defos = defos;
1.208 schwarze 211: p->quick = quick;
1.189 kristaps 212: p->roff = roff;
1.73 kristaps 213:
1.113 kristaps 214: mdoc_hash_init();
215: mdoc_alloc1(p);
216: return(p);
1.1 kristaps 217: }
218:
219:
1.68 kristaps 220: /*
221: * Climb back up the parse tree, validating open scopes. Mostly calls
1.85 kristaps 222: * through to macro_end() in macro.c.
1.68 kristaps 223: */
1.1 kristaps 224: int
1.203 schwarze 225: mdoc_endparse(struct mdoc *mdoc)
1.20 kristaps 226: {
227:
1.203 schwarze 228: assert( ! (MDOC_HALT & mdoc->flags));
229: if (mdoc_macroend(mdoc))
1.20 kristaps 230: return(1);
1.203 schwarze 231: mdoc->flags |= MDOC_HALT;
1.72 kristaps 232: return(0);
1.181 kristaps 233: }
234:
235: int
1.203 schwarze 236: mdoc_addeqn(struct mdoc *mdoc, const struct eqn *ep)
1.181 kristaps 237: {
238: struct mdoc_node *n;
239:
1.203 schwarze 240: assert( ! (MDOC_HALT & mdoc->flags));
1.181 kristaps 241:
242: /* No text before an initial macro. */
243:
1.203 schwarze 244: if (SEC_NONE == mdoc->lastnamed) {
245: mdoc_pmsg(mdoc, ep->ln, ep->pos, MANDOCERR_NOTEXT);
1.181 kristaps 246: return(1);
247: }
248:
1.203 schwarze 249: n = node_alloc(mdoc, ep->ln, ep->pos, MDOC_MAX, MDOC_EQN);
1.181 kristaps 250: n->eqn = ep;
251:
1.203 schwarze 252: if ( ! node_append(mdoc, n))
1.181 kristaps 253: return(0);
254:
1.203 schwarze 255: mdoc->next = MDOC_NEXT_SIBLING;
1.181 kristaps 256: return(1);
1.20 kristaps 257: }
258:
1.175 kristaps 259: int
1.203 schwarze 260: mdoc_addspan(struct mdoc *mdoc, const struct tbl_span *sp)
1.175 kristaps 261: {
1.180 kristaps 262: struct mdoc_node *n;
1.175 kristaps 263:
1.203 schwarze 264: assert( ! (MDOC_HALT & mdoc->flags));
1.175 kristaps 265:
266: /* No text before an initial macro. */
267:
1.203 schwarze 268: if (SEC_NONE == mdoc->lastnamed) {
269: mdoc_pmsg(mdoc, sp->line, 0, MANDOCERR_NOTEXT);
1.175 kristaps 270: return(1);
271: }
272:
1.203 schwarze 273: n = node_alloc(mdoc, sp->line, 0, MDOC_MAX, MDOC_TBL);
1.180 kristaps 274: n->span = sp;
275:
1.203 schwarze 276: if ( ! node_append(mdoc, n))
1.180 kristaps 277: return(0);
278:
1.203 schwarze 279: mdoc->next = MDOC_NEXT_SIBLING;
1.180 kristaps 280: return(1);
1.175 kristaps 281: }
282:
1.20 kristaps 283:
1.50 kristaps 284: /*
1.53 kristaps 285: * Main parse routine. Parses a single line -- really just hands off to
1.123 kristaps 286: * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
1.50 kristaps 287: */
1.20 kristaps 288: int
1.203 schwarze 289: mdoc_parseln(struct mdoc *mdoc, int ln, char *buf, int offs)
1.1 kristaps 290: {
291:
1.203 schwarze 292: assert( ! (MDOC_HALT & mdoc->flags));
1.50 kristaps 293:
1.203 schwarze 294: mdoc->flags |= MDOC_NEWLINE;
1.153 schwarze 295:
296: /*
297: * Let the roff nS register switch SYNOPSIS mode early,
298: * such that the parser knows at all times
299: * whether this mode is on or off.
300: * Note that this mode is also switched by the Sh macro.
301: */
1.204 schwarze 302: if (roff_getreg(mdoc->roff, "nS"))
303: mdoc->flags |= MDOC_SYNOPSIS;
304: else
305: mdoc->flags &= ~MDOC_SYNOPSIS;
1.153 schwarze 306:
1.203 schwarze 307: return(roff_getcontrol(mdoc->roff, buf, &offs) ?
308: mdoc_pmacro(mdoc, ln, buf, offs) :
309: mdoc_ptext(mdoc, ln, buf, offs));
1.1 kristaps 310: }
311:
1.88 kristaps 312: int
1.148 kristaps 313: mdoc_macro(MACRO_PROT_ARGS)
1.88 kristaps 314: {
1.122 kristaps 315: assert(tok < MDOC_MAX);
316:
317: /* If we're in the body, deny prologue calls. */
1.117 kristaps 318:
1.88 kristaps 319: if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
1.203 schwarze 320: MDOC_PBODY & mdoc->flags) {
321: mdoc_pmsg(mdoc, line, ppos, MANDOCERR_BADBODY);
1.174 kristaps 322: return(1);
323: }
1.122 kristaps 324:
325: /* If we're in the prologue, deny "body" macros. */
326:
327: if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
1.203 schwarze 328: ! (MDOC_PBODY & mdoc->flags)) {
329: mdoc_pmsg(mdoc, line, ppos, MANDOCERR_BADPROLOG);
330: if (NULL == mdoc->meta.msec)
331: mdoc->meta.msec = mandoc_strdup("1");
332: if (NULL == mdoc->meta.title)
333: mdoc->meta.title = mandoc_strdup("UNKNOWN");
334: if (NULL == mdoc->meta.vol)
335: mdoc->meta.vol = mandoc_strdup("LOCAL");
336: if (NULL == mdoc->meta.os)
337: mdoc->meta.os = mandoc_strdup("LOCAL");
338: if (NULL == mdoc->meta.date)
339: mdoc->meta.date = mandoc_normdate
340: (mdoc->parse, NULL, line, ppos);
341: mdoc->flags |= MDOC_PBODY;
1.120 kristaps 342: }
1.88 kristaps 343:
1.203 schwarze 344: return((*mdoc_macros[tok].fp)(mdoc, tok, line, ppos, pos, buf));
1.73 kristaps 345: }
346:
347:
348: static int
349: node_append(struct mdoc *mdoc, struct mdoc_node *p)
1.1 kristaps 350: {
351:
1.25 kristaps 352: assert(mdoc->last);
353: assert(mdoc->first);
354: assert(MDOC_ROOT != p->type);
1.1 kristaps 355:
1.13 kristaps 356: switch (mdoc->next) {
357: case (MDOC_NEXT_SIBLING):
1.6 kristaps 358: mdoc->last->next = p;
359: p->prev = mdoc->last;
1.13 kristaps 360: p->parent = mdoc->last->parent;
1.1 kristaps 361: break;
1.13 kristaps 362: case (MDOC_NEXT_CHILD):
363: mdoc->last->child = p;
1.1 kristaps 364: p->parent = mdoc->last;
365: break;
366: default:
1.13 kristaps 367: abort();
368: /* NOTREACHED */
1.1 kristaps 369: }
370:
1.86 kristaps 371: p->parent->nchild++;
372:
1.172 kristaps 373: /*
374: * Copy over the normalised-data pointer of our parent. Not
375: * everybody has one, but copying a null pointer is fine.
376: */
377:
378: switch (p->type) {
379: case (MDOC_BODY):
1.202 schwarze 380: if (ENDBODY_NOT != p->end)
381: break;
1.172 kristaps 382: /* FALLTHROUGH */
383: case (MDOC_TAIL):
384: /* FALLTHROUGH */
385: case (MDOC_HEAD):
386: p->norm = p->parent->norm;
387: break;
388: default:
389: break;
390: }
391:
1.23 kristaps 392: if ( ! mdoc_valid_pre(mdoc, p))
393: return(0);
1.27 kristaps 394:
395: switch (p->type) {
396: case (MDOC_HEAD):
397: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 398: p->parent->head = p;
1.27 kristaps 399: break;
400: case (MDOC_TAIL):
401: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 402: p->parent->tail = p;
1.27 kristaps 403: break;
404: case (MDOC_BODY):
1.152 schwarze 405: if (p->end)
406: break;
1.27 kristaps 407: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 408: p->parent->body = p;
1.27 kristaps 409: break;
410: default:
411: break;
412: }
413:
1.1 kristaps 414: mdoc->last = p;
1.71 kristaps 415:
416: switch (p->type) {
1.176 kristaps 417: case (MDOC_TBL):
418: /* FALLTHROUGH */
1.71 kristaps 419: case (MDOC_TEXT):
420: if ( ! mdoc_valid_post(mdoc))
421: return(0);
422: break;
423: default:
424: break;
425: }
426:
1.23 kristaps 427: return(1);
1.1 kristaps 428: }
429:
430:
1.45 kristaps 431: static struct mdoc_node *
1.203 schwarze 432: node_alloc(struct mdoc *mdoc, int line, int pos,
1.117 kristaps 433: enum mdoct tok, enum mdoc_type type)
1.45 kristaps 434: {
1.46 kristaps 435: struct mdoc_node *p;
436:
1.113 kristaps 437: p = mandoc_calloc(1, sizeof(struct mdoc_node));
1.203 schwarze 438: p->sec = mdoc->lastsec;
1.73 kristaps 439: p->line = line;
440: p->pos = pos;
1.206 schwarze 441: p->lastline = line;
1.73 kristaps 442: p->tok = tok;
1.118 kristaps 443: p->type = type;
1.150 kristaps 444:
445: /* Flag analysis. */
446:
1.203 schwarze 447: if (MDOC_SYNOPSIS & mdoc->flags)
1.153 schwarze 448: p->flags |= MDOC_SYNPRETTY;
449: else
450: p->flags &= ~MDOC_SYNPRETTY;
1.203 schwarze 451: if (MDOC_NEWLINE & mdoc->flags)
1.130 kristaps 452: p->flags |= MDOC_LINE;
1.203 schwarze 453: mdoc->flags &= ~MDOC_NEWLINE;
1.150 kristaps 454:
1.46 kristaps 455: return(p);
1.45 kristaps 456: }
457:
458:
1.23 kristaps 459: int
1.203 schwarze 460: mdoc_tail_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.17 kristaps 461: {
462: struct mdoc_node *p;
463:
1.203 schwarze 464: p = node_alloc(mdoc, line, pos, tok, MDOC_TAIL);
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);
1.17 kristaps 469: }
470:
471:
1.23 kristaps 472: int
1.203 schwarze 473: mdoc_head_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.1 kristaps 474: {
475: struct mdoc_node *p;
476:
1.203 schwarze 477: assert(mdoc->first);
478: assert(mdoc->last);
1.1 kristaps 479:
1.203 schwarze 480: p = node_alloc(mdoc, line, pos, tok, MDOC_HEAD);
481: if ( ! node_append(mdoc, p))
1.102 kristaps 482: return(0);
1.203 schwarze 483: mdoc->next = MDOC_NEXT_CHILD;
1.102 kristaps 484: return(1);
1.1 kristaps 485: }
486:
487:
1.23 kristaps 488: int
1.203 schwarze 489: mdoc_body_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.1 kristaps 490: {
491: struct mdoc_node *p;
492:
1.203 schwarze 493: p = node_alloc(mdoc, line, pos, tok, MDOC_BODY);
494: if ( ! node_append(mdoc, p))
1.102 kristaps 495: return(0);
1.203 schwarze 496: mdoc->next = MDOC_NEXT_CHILD;
1.152 schwarze 497: return(1);
498: }
499:
500:
501: int
1.203 schwarze 502: mdoc_endbody_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok,
1.152 schwarze 503: struct mdoc_node *body, enum mdoc_endbody end)
504: {
505: struct mdoc_node *p;
506:
1.203 schwarze 507: p = node_alloc(mdoc, line, pos, tok, MDOC_BODY);
1.152 schwarze 508: p->pending = body;
1.202 schwarze 509: p->norm = body->norm;
1.152 schwarze 510: p->end = end;
1.203 schwarze 511: if ( ! node_append(mdoc, p))
1.152 schwarze 512: return(0);
1.203 schwarze 513: mdoc->next = MDOC_NEXT_SIBLING;
1.102 kristaps 514: return(1);
1.1 kristaps 515: }
516:
517:
1.23 kristaps 518: int
1.203 schwarze 519: mdoc_block_alloc(struct mdoc *mdoc, int line, int pos,
1.117 kristaps 520: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 521: {
522: struct mdoc_node *p;
523:
1.203 schwarze 524: p = node_alloc(mdoc, line, pos, tok, MDOC_BLOCK);
1.77 kristaps 525: p->args = args;
526: if (p->args)
1.53 kristaps 527: (args->refcnt)++;
1.172 kristaps 528:
529: switch (tok) {
530: case (MDOC_Bd):
531: /* FALLTHROUGH */
532: case (MDOC_Bf):
533: /* FALLTHROUGH */
534: case (MDOC_Bl):
1.173 kristaps 535: /* FALLTHROUGH */
536: case (MDOC_Rs):
1.172 kristaps 537: p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
538: break;
539: default:
540: break;
541: }
542:
1.203 schwarze 543: if ( ! node_append(mdoc, p))
1.102 kristaps 544: return(0);
1.203 schwarze 545: mdoc->next = MDOC_NEXT_CHILD;
1.102 kristaps 546: return(1);
1.1 kristaps 547: }
548:
549:
1.23 kristaps 550: int
1.203 schwarze 551: mdoc_elem_alloc(struct mdoc *mdoc, int line, int pos,
1.117 kristaps 552: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 553: {
554: struct mdoc_node *p;
555:
1.203 schwarze 556: p = node_alloc(mdoc, line, pos, tok, MDOC_ELEM);
1.77 kristaps 557: p->args = args;
558: if (p->args)
1.53 kristaps 559: (args->refcnt)++;
1.172 kristaps 560:
561: switch (tok) {
562: case (MDOC_An):
563: p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
564: break;
565: default:
566: break;
567: }
568:
1.203 schwarze 569: if ( ! node_append(mdoc, p))
1.102 kristaps 570: return(0);
1.203 schwarze 571: mdoc->next = MDOC_NEXT_CHILD;
1.175 kristaps 572: return(1);
573: }
1.1 kristaps 574:
1.124 kristaps 575: int
1.203 schwarze 576: mdoc_word_alloc(struct mdoc *mdoc, int line, int pos, const char *p)
1.1 kristaps 577: {
1.91 kristaps 578: struct mdoc_node *n;
1.1 kristaps 579:
1.203 schwarze 580: n = node_alloc(mdoc, line, pos, MDOC_MAX, MDOC_TEXT);
581: n->string = roff_strdup(mdoc->roff, p);
1.91 kristaps 582:
1.203 schwarze 583: if ( ! node_append(mdoc, n))
1.101 kristaps 584: return(0);
1.124 kristaps 585:
1.203 schwarze 586: mdoc->next = MDOC_NEXT_SIBLING;
1.101 kristaps 587: return(1);
1.91 kristaps 588: }
589:
1.205 schwarze 590: void
591: mdoc_word_append(struct mdoc *mdoc, const char *p)
592: {
593: struct mdoc_node *n;
594: char *addstr, *newstr;
595:
596: n = mdoc->last;
597: addstr = roff_strdup(mdoc->roff, p);
1.210 schwarze 598: mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
1.205 schwarze 599: free(addstr);
600: free(n->string);
601: n->string = newstr;
602: mdoc->next = MDOC_NEXT_SIBLING;
603: }
1.91 kristaps 604:
1.155 kristaps 605: static void
1.53 kristaps 606: mdoc_node_free(struct mdoc_node *p)
1.1 kristaps 607: {
608:
1.172 kristaps 609: if (MDOC_BLOCK == p->type || MDOC_ELEM == p->type)
1.171 kristaps 610: free(p->norm);
1.53 kristaps 611: if (p->string)
612: free(p->string);
613: if (p->args)
614: mdoc_argv_free(p->args);
1.1 kristaps 615: free(p);
616: }
617:
618:
1.121 kristaps 619: static void
1.203 schwarze 620: mdoc_node_unlink(struct mdoc *mdoc, struct mdoc_node *n)
1.121 kristaps 621: {
622:
623: /* Adjust siblings. */
624:
625: if (n->prev)
626: n->prev->next = n->next;
627: if (n->next)
628: n->next->prev = n->prev;
629:
630: /* Adjust parent. */
631:
632: if (n->parent) {
633: n->parent->nchild--;
634: if (n->parent->child == n)
635: n->parent->child = n->prev ? n->prev : n->next;
1.169 kristaps 636: if (n->parent->last == n)
637: n->parent->last = n->prev ? n->prev : NULL;
1.121 kristaps 638: }
639:
640: /* Adjust parse point, if applicable. */
641:
1.203 schwarze 642: if (mdoc && mdoc->last == n) {
1.121 kristaps 643: if (n->prev) {
1.203 schwarze 644: mdoc->last = n->prev;
645: mdoc->next = MDOC_NEXT_SIBLING;
1.121 kristaps 646: } else {
1.203 schwarze 647: mdoc->last = n->parent;
648: mdoc->next = MDOC_NEXT_CHILD;
1.121 kristaps 649: }
650: }
651:
1.203 schwarze 652: if (mdoc && mdoc->first == n)
653: mdoc->first = NULL;
1.121 kristaps 654: }
655:
656:
1.53 kristaps 657: void
1.203 schwarze 658: mdoc_node_delete(struct mdoc *mdoc, struct mdoc_node *p)
1.1 kristaps 659: {
660:
1.121 kristaps 661: while (p->child) {
662: assert(p->nchild);
1.203 schwarze 663: mdoc_node_delete(mdoc, p->child);
1.121 kristaps 664: }
665: assert(0 == p->nchild);
1.1 kristaps 666:
1.203 schwarze 667: mdoc_node_unlink(mdoc, p);
1.53 kristaps 668: mdoc_node_free(p);
1.201 schwarze 669: }
670:
671: int
1.203 schwarze 672: mdoc_node_relink(struct mdoc *mdoc, struct mdoc_node *p)
1.201 schwarze 673: {
674:
1.203 schwarze 675: mdoc_node_unlink(mdoc, p);
676: return(node_append(mdoc, p));
1.1 kristaps 677: }
678:
1.193 kristaps 679: #if 0
1.191 kristaps 680: /*
681: * Pre-treat a text line.
682: * Text lines can consist of equations, which must be handled apart from
683: * the regular text.
684: * Thus, use this function to step through a line checking if it has any
685: * equations embedded in it.
686: * This must handle multiple equations AND equations that do not end at
687: * the end-of-line, i.e., will re-enter in the next roff parse.
688: */
689: static int
1.203 schwarze 690: mdoc_preptext(struct mdoc *mdoc, int line, char *buf, int offs)
1.191 kristaps 691: {
692: char *start, *end;
693: char delim;
694:
695: while ('\0' != buf[offs]) {
696: /* Mark starting position if eqn is set. */
697: start = NULL;
1.203 schwarze 698: if ('\0' != (delim = roff_eqndelim(mdoc->roff)))
1.191 kristaps 699: if (NULL != (start = strchr(buf + offs, delim)))
700: *start++ = '\0';
701:
702: /* Parse text as normal. */
1.203 schwarze 703: if ( ! mdoc_ptext(mdoc, line, buf, offs))
1.191 kristaps 704: return(0);
705:
706: /* Continue only if an equation exists. */
707: if (NULL == start)
708: break;
709:
710: /* Read past the end of the equation. */
711: offs += start - (buf + offs);
712: assert(start == &buf[offs]);
713: if (NULL != (end = strchr(buf + offs, delim))) {
714: *end++ = '\0';
715: while (' ' == *end)
716: end++;
717: }
718:
719: /* Parse the equation itself. */
1.203 schwarze 720: roff_openeqn(mdoc->roff, NULL, line, offs, buf);
1.191 kristaps 721:
722: /* Process a finished equation? */
1.203 schwarze 723: if (roff_closeeqn(mdoc->roff))
724: if ( ! mdoc_addeqn(mdoc, roff_eqn(mdoc->roff)))
1.191 kristaps 725: return(0);
726: offs += (end - (buf + offs));
727: }
728:
729: return(1);
730: }
1.193 kristaps 731: #endif
1.1 kristaps 732:
1.53 kristaps 733: /*
734: * Parse free-form text, that is, a line that does not begin with the
735: * control character.
736: */
737: static int
1.203 schwarze 738: mdoc_ptext(struct mdoc *mdoc, int line, char *buf, int offs)
1.1 kristaps 739: {
1.142 kristaps 740: char *c, *ws, *end;
741: struct mdoc_node *n;
1.1 kristaps 742:
1.124 kristaps 743: /* No text before an initial macro. */
744:
1.203 schwarze 745: if (SEC_NONE == mdoc->lastnamed) {
746: mdoc_pmsg(mdoc, line, offs, MANDOCERR_NOTEXT);
1.174 kristaps 747: return(1);
748: }
1.142 kristaps 749:
1.203 schwarze 750: assert(mdoc->last);
751: n = mdoc->last;
1.142 kristaps 752:
753: /*
1.144 kristaps 754: * Divert directly to list processing if we're encountering a
1.142 kristaps 755: * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
1.144 kristaps 756: * (a MDOC_BODY means it's already open, in which case we should
757: * process within its context in the normal way).
1.142 kristaps 758: */
759:
1.143 kristaps 760: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.172 kristaps 761: LIST_column == n->norm->Bl.type) {
1.144 kristaps 762: /* `Bl' is open without any children. */
1.203 schwarze 763: mdoc->flags |= MDOC_FREECOL;
764: return(mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf));
1.142 kristaps 765: }
766:
767: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
768: NULL != n->parent &&
769: MDOC_Bl == n->parent->tok &&
1.172 kristaps 770: LIST_column == n->parent->norm->Bl.type) {
1.144 kristaps 771: /* `Bl' has block-level `It' children. */
1.203 schwarze 772: mdoc->flags |= MDOC_FREECOL;
773: return(mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf));
1.142 kristaps 774: }
1.124 kristaps 775:
1.137 schwarze 776: /*
777: * Search for the beginning of unescaped trailing whitespace (ws)
778: * and for the first character not to be output (end).
779: */
1.139 kristaps 780:
781: /* FIXME: replace with strcspn(). */
1.137 schwarze 782: ws = NULL;
783: for (c = end = buf + offs; *c; c++) {
784: switch (*c) {
785: case ' ':
786: if (NULL == ws)
787: ws = c;
788: continue;
789: case '\t':
790: /*
791: * Always warn about trailing tabs,
792: * even outside literal context,
793: * where they should be put on the next line.
794: */
795: if (NULL == ws)
796: ws = c;
797: /*
798: * Strip trailing tabs in literal context only;
799: * outside, they affect the next line.
800: */
1.203 schwarze 801: if (MDOC_LITERAL & mdoc->flags)
1.137 schwarze 802: continue;
803: break;
804: case '\\':
805: /* Skip the escaped character, too, if any. */
806: if (c[1])
807: c++;
808: /* FALLTHROUGH */
809: default:
810: ws = NULL;
811: break;
812: }
813: end = c + 1;
814: }
815: *end = '\0';
1.91 kristaps 816:
1.137 schwarze 817: if (ws)
1.203 schwarze 818: mdoc_pmsg(mdoc, line, (int)(ws-buf), MANDOCERR_EOLNSPACE);
1.115 kristaps 819:
1.203 schwarze 820: if ('\0' == buf[offs] && ! (MDOC_LITERAL & mdoc->flags)) {
821: mdoc_pmsg(mdoc, line, (int)(c-buf), MANDOCERR_NOBLANKLN);
1.124 kristaps 822:
1.119 kristaps 823: /*
1.165 schwarze 824: * Insert a `sp' in the case of a blank line. Technically,
1.124 kristaps 825: * blank lines aren't allowed, but enough manuals assume this
826: * behaviour that we want to work around it.
1.119 kristaps 827: */
1.203 schwarze 828: if ( ! mdoc_elem_alloc(mdoc, line, offs, MDOC_sp, NULL))
1.119 kristaps 829: return(0);
1.124 kristaps 830:
1.203 schwarze 831: mdoc->next = MDOC_NEXT_SIBLING;
1.199 schwarze 832:
1.203 schwarze 833: return(mdoc_valid_post(mdoc));
1.119 kristaps 834: }
1.68 kristaps 835:
1.203 schwarze 836: if ( ! mdoc_word_alloc(mdoc, line, offs, buf+offs))
1.137 schwarze 837: return(0);
1.91 kristaps 838:
1.203 schwarze 839: if (MDOC_LITERAL & mdoc->flags)
1.137 schwarze 840: return(1);
1.128 kristaps 841:
842: /*
843: * End-of-sentence check. If the last character is an unescaped
844: * EOS character, then flag the node as being the end of a
845: * sentence. The front-end will know how to interpret this.
846: */
1.132 kristaps 847:
1.137 schwarze 848: assert(buf < end);
849:
1.207 schwarze 850: if (mandoc_eos(buf+offs, (size_t)(end-buf-offs)))
1.203 schwarze 851: mdoc->last->flags |= MDOC_EOS;
1.128 kristaps 852:
853: return(1);
1.1 kristaps 854: }
855:
856:
1.53 kristaps 857: /*
858: * Parse a macro line, that is, a line beginning with the control
859: * character.
860: */
1.155 kristaps 861: static int
1.203 schwarze 862: mdoc_pmacro(struct mdoc *mdoc, int ln, char *buf, int offs)
1.1 kristaps 863: {
1.144 kristaps 864: enum mdoct tok;
1.188 kristaps 865: int i, sv;
1.144 kristaps 866: char mac[5];
867: struct mdoc_node *n;
1.1 kristaps 868:
1.188 kristaps 869: /* Empty post-control lines are ignored. */
1.63 kristaps 870:
1.188 kristaps 871: if ('"' == buf[offs]) {
1.203 schwarze 872: mdoc_pmsg(mdoc, ln, offs, MANDOCERR_BADCOMMENT);
1.188 kristaps 873: return(1);
874: } else if ('\0' == buf[offs])
1.63 kristaps 875: return(1);
876:
1.188 kristaps 877: sv = offs;
1.130 kristaps 878:
1.160 kristaps 879: /*
1.162 schwarze 880: * Copy the first word into a nil-terminated buffer.
1.165 schwarze 881: * Stop copying when a tab, space, or eoln is encountered.
1.160 kristaps 882: */
1.1 kristaps 883:
1.188 kristaps 884: i = 0;
885: while (i < 4 && '\0' != buf[offs] &&
886: ' ' != buf[offs] && '\t' != buf[offs])
887: mac[i++] = buf[offs++];
888:
889: mac[i] = '\0';
890:
891: tok = (i > 1 || i < 4) ? mdoc_hash_find(mac) : MDOC_MAX;
1.1 kristaps 892:
1.163 schwarze 893: if (MDOC_MAX == tok) {
1.203 schwarze 894: mandoc_vmsg(MANDOCERR_MACRO, mdoc->parse,
1.185 kristaps 895: ln, sv, "%s", buf + sv - 1);
1.58 kristaps 896: return(1);
1.53 kristaps 897: }
1.1 kristaps 898:
1.160 kristaps 899: /* Disregard the first trailing tab, if applicable. */
900:
1.188 kristaps 901: if ('\t' == buf[offs])
902: offs++;
1.160 kristaps 903:
904: /* Jump to the next non-whitespace word. */
1.1 kristaps 905:
1.188 kristaps 906: while (buf[offs] && ' ' == buf[offs])
907: offs++;
1.1 kristaps 908:
1.125 kristaps 909: /*
910: * Trailing whitespace. Note that tabs are allowed to be passed
911: * into the parser as "text", so we only warn about spaces here.
912: */
1.115 kristaps 913:
1.188 kristaps 914: if ('\0' == buf[offs] && ' ' == buf[offs - 1])
1.203 schwarze 915: mdoc_pmsg(mdoc, ln, offs - 1, MANDOCERR_EOLNSPACE);
1.115 kristaps 916:
1.144 kristaps 917: /*
918: * If an initial macro or a list invocation, divert directly
919: * into macro processing.
920: */
921:
1.203 schwarze 922: if (NULL == mdoc->last || MDOC_It == tok || MDOC_El == tok) {
923: if ( ! mdoc_macro(mdoc, tok, ln, sv, &offs, buf))
1.144 kristaps 924: goto err;
925: return(1);
926: }
927:
1.203 schwarze 928: n = mdoc->last;
929: assert(mdoc->last);
1.144 kristaps 930:
931: /*
932: * If the first macro of a `Bl -column', open an `It' block
933: * context around the parsed macro.
934: */
935:
936: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.172 kristaps 937: LIST_column == n->norm->Bl.type) {
1.203 schwarze 938: mdoc->flags |= MDOC_FREECOL;
939: if ( ! mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf))
1.144 kristaps 940: goto err;
941: return(1);
942: }
943:
944: /*
945: * If we're following a block-level `It' within a `Bl -column'
946: * context (perhaps opened in the above block or in ptext()),
947: * then open an `It' block context around the parsed macro.
1.98 kristaps 948: */
1.144 kristaps 949:
950: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
951: NULL != n->parent &&
952: MDOC_Bl == n->parent->tok &&
1.172 kristaps 953: LIST_column == n->parent->norm->Bl.type) {
1.203 schwarze 954: mdoc->flags |= MDOC_FREECOL;
955: if ( ! mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf))
1.144 kristaps 956: goto err;
957: return(1);
958: }
959:
960: /* Normal processing of a macro. */
961:
1.203 schwarze 962: if ( ! mdoc_macro(mdoc, tok, ln, sv, &offs, buf))
1.53 kristaps 963: goto err;
1.208 schwarze 964:
965: /* In quick mode (for mandocdb), abort after the NAME section. */
966:
967: if (mdoc->quick && MDOC_Sh == tok &&
968: SEC_NAME != mdoc->last->sec)
969: return(2);
1.1 kristaps 970:
1.53 kristaps 971: return(1);
1.1 kristaps 972:
1.53 kristaps 973: err: /* Error out. */
1.1 kristaps 974:
1.203 schwarze 975: mdoc->flags |= MDOC_HALT;
1.53 kristaps 976: return(0);
1.1 kristaps 977: }
1.100 kristaps 978:
1.186 kristaps 979: enum mdelim
980: mdoc_isdelim(const char *p)
981: {
982:
983: if ('\0' == p[0])
984: return(DELIM_NONE);
985:
986: if ('\0' == p[1])
987: switch (p[0]) {
988: case('('):
989: /* FALLTHROUGH */
990: case('['):
991: return(DELIM_OPEN);
992: case('|'):
993: return(DELIM_MIDDLE);
994: case('.'):
995: /* FALLTHROUGH */
996: case(','):
997: /* FALLTHROUGH */
998: case(';'):
999: /* FALLTHROUGH */
1000: case(':'):
1001: /* FALLTHROUGH */
1002: case('?'):
1003: /* FALLTHROUGH */
1004: case('!'):
1005: /* FALLTHROUGH */
1006: case(')'):
1007: /* FALLTHROUGH */
1008: case(']'):
1009: return(DELIM_CLOSE);
1010: default:
1011: return(DELIM_NONE);
1012: }
1013:
1014: if ('\\' != p[0])
1015: return(DELIM_NONE);
1.100 kristaps 1016:
1.186 kristaps 1017: if (0 == strcmp(p + 1, "."))
1018: return(DELIM_CLOSE);
1.200 schwarze 1019: if (0 == strcmp(p + 1, "fR|\\fP"))
1.186 kristaps 1020: return(DELIM_MIDDLE);
1021:
1022: return(DELIM_NONE);
1.211 schwarze 1023: }
1024:
1025: void
1026: mdoc_deroff(char **dest, const struct mdoc_node *n)
1027: {
1028: char *cp;
1029: size_t sz;
1030:
1031: if (MDOC_TEXT != n->type) {
1032: for (n = n->child; n; n = n->next)
1033: mdoc_deroff(dest, n);
1034: return;
1035: }
1036:
1037: /* Skip leading whitespace. */
1038:
1039: for (cp = n->string; '\0' != *cp; cp++)
1040: if (0 == isspace((unsigned char)*cp))
1041: break;
1042:
1043: /* Skip trailing whitespace. */
1044:
1045: for (sz = strlen(cp); sz; sz--)
1046: if (0 == isspace((unsigned char)cp[sz-1]))
1047: break;
1048:
1049: /* Skip empty strings. */
1050:
1051: if (0 == sz)
1052: return;
1053:
1054: if (NULL == *dest) {
1055: *dest = mandoc_strndup(cp, sz);
1056: return;
1057: }
1058:
1059: mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
1060: free(*dest);
1061: *dest = cp;
1.186 kristaps 1062: }
CVSweb