Annotation of src/usr.bin/mandoc/mdoc.c, Revision 1.112
1.112 ! schwarze 1: /* $Id: mdoc.c,v 1.111 2014/08/08 15:32:17 schwarze Exp $ */
1.1 kristaps 2: /*
1.77 schwarze 3: * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
1.98 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
1.3 schwarze 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: *
1.3 schwarze 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: */
1.28 schwarze 18: #include <sys/types.h>
19:
1.1 kristaps 20: #include <assert.h>
1.101 schwarze 21: #include <ctype.h>
1.1 kristaps 22: #include <stdarg.h>
23: #include <stdio.h>
24: #include <stdlib.h>
25: #include <string.h>
1.39 schwarze 26: #include <time.h>
1.1 kristaps 27:
1.83 schwarze 28: #include "mdoc.h"
1.54 schwarze 29: #include "mandoc.h"
1.99 schwarze 30: #include "mandoc_aux.h"
1.1 kristaps 31: #include "libmdoc.h"
1.32 schwarze 32: #include "libmandoc.h"
1.1 kristaps 33:
1.106 schwarze 34: const char *const __mdoc_macronames[MDOC_MAX + 1] = {
1.7 schwarze 35: "Ap", "Dd", "Dt", "Os",
1.1 kristaps 36: "Sh", "Ss", "Pp", "D1",
37: "Dl", "Bd", "Ed", "Bl",
38: "El", "It", "Ad", "An",
39: "Ar", "Cd", "Cm", "Dv",
40: "Er", "Ev", "Ex", "Fa",
41: "Fd", "Fl", "Fn", "Ft",
42: "Ic", "In", "Li", "Nd",
43: "Nm", "Op", "Ot", "Pa",
44: "Rv", "St", "Va", "Vt",
1.33 schwarze 45: "Xr", "%A", "%B", "%D",
46: "%I", "%J", "%N", "%O",
47: "%P", "%R", "%T", "%V",
1.1 kristaps 48: "Ac", "Ao", "Aq", "At",
49: "Bc", "Bf", "Bo", "Bq",
50: "Bsx", "Bx", "Db", "Dc",
51: "Do", "Dq", "Ec", "Ef",
52: "Em", "Eo", "Fx", "Ms",
53: "No", "Ns", "Nx", "Ox",
54: "Pc", "Pf", "Po", "Pq",
55: "Qc", "Ql", "Qo", "Qq",
56: "Re", "Rs", "Sc", "So",
57: "Sq", "Sm", "Sx", "Sy",
58: "Tn", "Ux", "Xc", "Xo",
59: "Fo", "Fc", "Oo", "Oc",
60: "Bk", "Ek", "Bt", "Hf",
1.7 schwarze 61: "Fr", "Ud", "Lb", "Lp",
62: "Lk", "Mt", "Brq", "Bro",
1.33 schwarze 63: "Brc", "%C", "Es", "En",
64: "Dx", "%Q", "br", "sp",
1.106 schwarze 65: "%U", "Ta", "ll", "text",
1.1 kristaps 66: };
67:
1.103 schwarze 68: const char *const __mdoc_argnames[MDOC_ARG_MAX] = {
1.1 kristaps 69: "split", "nosplit", "ragged",
1.103 schwarze 70: "unfilled", "literal", "file",
71: "offset", "bullet", "dash",
72: "hyphen", "item", "enum",
73: "tag", "diag", "hang",
74: "ohang", "inset", "column",
75: "width", "compact", "std",
1.1 kristaps 76: "filled", "words", "emphasis",
1.30 schwarze 77: "symbolic", "nested", "centered"
1.1 kristaps 78: };
79:
80: const char * const *mdoc_macronames = __mdoc_macronames;
81: const char * const *mdoc_argnames = __mdoc_argnames;
82:
1.40 schwarze 83: static void mdoc_node_free(struct mdoc_node *);
1.103 schwarze 84: static void mdoc_node_unlink(struct mdoc *,
1.40 schwarze 85: struct mdoc_node *);
1.1 kristaps 86: static void mdoc_free1(struct mdoc *);
1.32 schwarze 87: static void mdoc_alloc1(struct mdoc *);
1.103 schwarze 88: static struct mdoc_node *node_alloc(struct mdoc *, int, int,
1.37 schwarze 89: enum mdoct, enum mdoc_type);
1.103 schwarze 90: static int node_append(struct mdoc *,
1.1 kristaps 91: struct mdoc_node *);
1.84 schwarze 92: #if 0
93: static int mdoc_preptext(struct mdoc *, int, char *, int);
94: #endif
1.53 schwarze 95: static int mdoc_ptext(struct mdoc *, int, char *, int);
96: static int mdoc_pmacro(struct mdoc *, int, char *, int);
1.1 kristaps 97:
1.103 schwarze 98:
1.1 kristaps 99: const struct mdoc_node *
1.93 schwarze 100: mdoc_node(const struct mdoc *mdoc)
1.1 kristaps 101: {
102:
1.93 schwarze 103: return(mdoc->first);
1.1 kristaps 104: }
105:
106: const struct mdoc_meta *
1.93 schwarze 107: mdoc_meta(const struct mdoc *mdoc)
1.1 kristaps 108: {
109:
1.93 schwarze 110: return(&mdoc->meta);
1.1 kristaps 111: }
112:
1.8 schwarze 113: /*
114: * Frees volatile resources (parse tree, meta-data, fields).
115: */
1.1 kristaps 116: static void
117: mdoc_free1(struct mdoc *mdoc)
118: {
119:
120: if (mdoc->first)
1.40 schwarze 121: mdoc_node_delete(mdoc, mdoc->first);
1.1 kristaps 122: if (mdoc->meta.title)
123: free(mdoc->meta.title);
124: if (mdoc->meta.os)
125: free(mdoc->meta.os);
126: if (mdoc->meta.name)
127: free(mdoc->meta.name);
128: if (mdoc->meta.arch)
129: free(mdoc->meta.arch);
130: if (mdoc->meta.vol)
131: free(mdoc->meta.vol);
1.50 schwarze 132: if (mdoc->meta.msec)
133: free(mdoc->meta.msec);
1.82 schwarze 134: if (mdoc->meta.date)
135: free(mdoc->meta.date);
1.1 kristaps 136: }
137:
1.8 schwarze 138: /*
139: * Allocate all volatile resources (parse tree, meta-data, fields).
140: */
1.32 schwarze 141: static void
1.1 kristaps 142: mdoc_alloc1(struct mdoc *mdoc)
143: {
144:
1.32 schwarze 145: memset(&mdoc->meta, 0, sizeof(struct mdoc_meta));
1.1 kristaps 146: mdoc->flags = 0;
1.9 schwarze 147: mdoc->lastnamed = mdoc->lastsec = SEC_NONE;
1.32 schwarze 148: mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node));
1.1 kristaps 149: mdoc->first = mdoc->last;
150: mdoc->last->type = MDOC_ROOT;
1.86 schwarze 151: mdoc->last->tok = MDOC_MAX;
1.1 kristaps 152: mdoc->next = MDOC_NEXT_CHILD;
153: }
154:
155: /*
1.8 schwarze 156: * Free up volatile resources (see mdoc_free1()) then re-initialises the
157: * data with mdoc_alloc1(). After invocation, parse data has been reset
158: * and the parser is ready for re-invocation on a new tree; however,
159: * cross-parse non-volatile data is kept intact.
1.1 kristaps 160: */
1.32 schwarze 161: void
1.1 kristaps 162: mdoc_reset(struct mdoc *mdoc)
163: {
164:
165: mdoc_free1(mdoc);
1.32 schwarze 166: mdoc_alloc1(mdoc);
1.1 kristaps 167: }
168:
169: /*
1.8 schwarze 170: * Completely free up all volatile and non-volatile parse resources.
171: * After invocation, the pointer is no longer usable.
1.1 kristaps 172: */
173: void
174: mdoc_free(struct mdoc *mdoc)
175: {
176:
177: mdoc_free1(mdoc);
178: free(mdoc);
179: }
180:
1.8 schwarze 181: /*
1.103 schwarze 182: * Allocate volatile and non-volatile parse resources.
1.8 schwarze 183: */
1.1 kristaps 184: struct mdoc *
1.98 schwarze 185: mdoc_alloc(struct roff *roff, struct mparse *parse,
1.110 schwarze 186: const char *defos, int quick)
1.1 kristaps 187: {
188: struct mdoc *p;
189:
1.32 schwarze 190: p = mandoc_calloc(1, sizeof(struct mdoc));
191:
1.83 schwarze 192: p->parse = parse;
1.87 schwarze 193: p->defos = defos;
1.98 schwarze 194: p->quick = quick;
1.84 schwarze 195: p->roff = roff;
1.1 kristaps 196:
1.32 schwarze 197: mdoc_hash_init();
198: mdoc_alloc1(p);
199: return(p);
1.1 kristaps 200: }
201:
202: int
1.93 schwarze 203: mdoc_endparse(struct mdoc *mdoc)
1.1 kristaps 204: {
205:
1.111 schwarze 206: return(mdoc_macroend(mdoc));
1.81 schwarze 207: }
208:
209: int
1.93 schwarze 210: mdoc_addeqn(struct mdoc *mdoc, const struct eqn *ep)
1.81 schwarze 211: {
212: struct mdoc_node *n;
213:
1.93 schwarze 214: n = node_alloc(mdoc, ep->ln, ep->pos, MDOC_MAX, MDOC_EQN);
1.81 schwarze 215: n->eqn = ep;
216:
1.93 schwarze 217: if ( ! node_append(mdoc, n))
1.81 schwarze 218: return(0);
219:
1.93 schwarze 220: mdoc->next = MDOC_NEXT_SIBLING;
1.81 schwarze 221: return(1);
1.1 kristaps 222: }
223:
1.77 schwarze 224: int
1.93 schwarze 225: mdoc_addspan(struct mdoc *mdoc, const struct tbl_span *sp)
1.77 schwarze 226: {
1.79 schwarze 227: struct mdoc_node *n;
1.77 schwarze 228:
1.93 schwarze 229: n = node_alloc(mdoc, sp->line, 0, MDOC_MAX, MDOC_TBL);
1.79 schwarze 230: n->span = sp;
231:
1.93 schwarze 232: if ( ! node_append(mdoc, n))
1.79 schwarze 233: return(0);
234:
1.93 schwarze 235: mdoc->next = MDOC_NEXT_SIBLING;
1.79 schwarze 236: return(1);
1.77 schwarze 237: }
238:
1.1 kristaps 239: /*
240: * Main parse routine. Parses a single line -- really just hands off to
1.45 schwarze 241: * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
1.1 kristaps 242: */
243: int
1.93 schwarze 244: mdoc_parseln(struct mdoc *mdoc, int ln, char *buf, int offs)
1.1 kristaps 245: {
246:
1.93 schwarze 247: mdoc->flags |= MDOC_NEWLINE;
1.60 schwarze 248:
249: /*
250: * Let the roff nS register switch SYNOPSIS mode early,
251: * such that the parser knows at all times
252: * whether this mode is on or off.
253: * Note that this mode is also switched by the Sh macro.
254: */
1.94 schwarze 255: if (roff_getreg(mdoc->roff, "nS"))
256: mdoc->flags |= MDOC_SYNOPSIS;
257: else
258: mdoc->flags &= ~MDOC_SYNOPSIS;
1.60 schwarze 259:
1.93 schwarze 260: return(roff_getcontrol(mdoc->roff, buf, &offs) ?
1.103 schwarze 261: mdoc_pmacro(mdoc, ln, buf, offs) :
262: mdoc_ptext(mdoc, ln, buf, offs));
1.1 kristaps 263: }
264:
1.5 schwarze 265: int
1.58 schwarze 266: mdoc_macro(MACRO_PROT_ARGS)
1.1 kristaps 267: {
1.46 schwarze 268: assert(tok < MDOC_MAX);
269:
270: /* If we're in the body, deny prologue calls. */
1.37 schwarze 271:
1.103 schwarze 272: if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
273: MDOC_PBODY & mdoc->flags) {
1.105 schwarze 274: mandoc_vmsg(MANDOCERR_PROLOG_ONLY, mdoc->parse,
275: line, ppos, "%s", mdoc_macronames[tok]);
1.76 schwarze 276: return(1);
277: }
1.46 schwarze 278:
279: /* If we're in the prologue, deny "body" macros. */
280:
1.103 schwarze 281: if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
282: ! (MDOC_PBODY & mdoc->flags)) {
1.105 schwarze 283: mandoc_vmsg(MANDOCERR_PROLOG_BAD, mdoc->parse,
284: line, ppos, "%s", mdoc_macronames[tok]);
1.93 schwarze 285: if (NULL == mdoc->meta.msec)
286: mdoc->meta.msec = mandoc_strdup("1");
287: if (NULL == mdoc->meta.title)
288: mdoc->meta.title = mandoc_strdup("UNKNOWN");
289: if (NULL == mdoc->meta.vol)
290: mdoc->meta.vol = mandoc_strdup("LOCAL");
291: if (NULL == mdoc->meta.os)
292: mdoc->meta.os = mandoc_strdup("LOCAL");
293: if (NULL == mdoc->meta.date)
294: mdoc->meta.date = mandoc_normdate
295: (mdoc->parse, NULL, line, ppos);
296: mdoc->flags |= MDOC_PBODY;
1.39 schwarze 297: }
1.1 kristaps 298:
1.93 schwarze 299: return((*mdoc_macros[tok].fp)(mdoc, tok, line, ppos, pos, buf));
1.1 kristaps 300: }
301:
302:
303: static int
304: node_append(struct mdoc *mdoc, struct mdoc_node *p)
305: {
306:
307: assert(mdoc->last);
308: assert(mdoc->first);
309: assert(MDOC_ROOT != p->type);
310:
311: switch (mdoc->next) {
1.103 schwarze 312: case MDOC_NEXT_SIBLING:
1.1 kristaps 313: mdoc->last->next = p;
314: p->prev = mdoc->last;
315: p->parent = mdoc->last->parent;
316: break;
1.103 schwarze 317: case MDOC_NEXT_CHILD:
1.1 kristaps 318: mdoc->last->child = p;
319: p->parent = mdoc->last;
320: break;
321: default:
322: abort();
323: /* NOTREACHED */
324: }
325:
1.10 schwarze 326: p->parent->nchild++;
327:
1.75 schwarze 328: /*
329: * Copy over the normalised-data pointer of our parent. Not
330: * everybody has one, but copying a null pointer is fine.
331: */
332:
333: switch (p->type) {
1.103 schwarze 334: case MDOC_BODY:
1.92 schwarze 335: if (ENDBODY_NOT != p->end)
336: break;
1.75 schwarze 337: /* FALLTHROUGH */
1.103 schwarze 338: case MDOC_TAIL:
1.75 schwarze 339: /* FALLTHROUGH */
1.103 schwarze 340: case MDOC_HEAD:
1.75 schwarze 341: p->norm = p->parent->norm;
342: break;
343: default:
344: break;
345: }
346:
1.1 kristaps 347: if ( ! mdoc_valid_pre(mdoc, p))
348: return(0);
349:
350: switch (p->type) {
1.103 schwarze 351: case MDOC_HEAD:
1.1 kristaps 352: assert(MDOC_BLOCK == p->parent->type);
353: p->parent->head = p;
354: break;
1.103 schwarze 355: case MDOC_TAIL:
1.1 kristaps 356: assert(MDOC_BLOCK == p->parent->type);
357: p->parent->tail = p;
358: break;
1.103 schwarze 359: case MDOC_BODY:
1.59 schwarze 360: if (p->end)
361: break;
1.1 kristaps 362: assert(MDOC_BLOCK == p->parent->type);
363: p->parent->body = p;
364: break;
365: default:
366: break;
367: }
368:
369: mdoc->last = p;
370:
371: switch (p->type) {
1.103 schwarze 372: case MDOC_TBL:
1.77 schwarze 373: /* FALLTHROUGH */
1.103 schwarze 374: case MDOC_TEXT:
1.1 kristaps 375: if ( ! mdoc_valid_post(mdoc))
376: return(0);
377: break;
378: default:
379: break;
380: }
381:
382: return(1);
383: }
384:
385: static struct mdoc_node *
1.103 schwarze 386: node_alloc(struct mdoc *mdoc, int line, int pos,
1.37 schwarze 387: enum mdoct tok, enum mdoc_type type)
1.1 kristaps 388: {
389: struct mdoc_node *p;
390:
1.32 schwarze 391: p = mandoc_calloc(1, sizeof(struct mdoc_node));
1.93 schwarze 392: p->sec = mdoc->lastsec;
1.1 kristaps 393: p->line = line;
394: p->pos = pos;
1.96 schwarze 395: p->lastline = line;
1.1 kristaps 396: p->tok = tok;
1.37 schwarze 397: p->type = type;
1.58 schwarze 398:
399: /* Flag analysis. */
400:
1.93 schwarze 401: if (MDOC_SYNOPSIS & mdoc->flags)
1.60 schwarze 402: p->flags |= MDOC_SYNPRETTY;
403: else
404: p->flags &= ~MDOC_SYNPRETTY;
1.93 schwarze 405: if (MDOC_NEWLINE & mdoc->flags)
1.47 schwarze 406: p->flags |= MDOC_LINE;
1.93 schwarze 407: mdoc->flags &= ~MDOC_NEWLINE;
1.58 schwarze 408:
1.1 kristaps 409: return(p);
410: }
411:
412: int
1.93 schwarze 413: mdoc_tail_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.1 kristaps 414: {
415: struct mdoc_node *p;
416:
1.93 schwarze 417: p = node_alloc(mdoc, line, pos, tok, MDOC_TAIL);
418: if ( ! node_append(mdoc, p))
1.25 schwarze 419: return(0);
1.93 schwarze 420: mdoc->next = MDOC_NEXT_CHILD;
1.25 schwarze 421: return(1);
1.1 kristaps 422: }
423:
424: int
1.93 schwarze 425: mdoc_head_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.1 kristaps 426: {
427: struct mdoc_node *p;
428:
1.93 schwarze 429: assert(mdoc->first);
430: assert(mdoc->last);
1.1 kristaps 431:
1.93 schwarze 432: p = node_alloc(mdoc, line, pos, tok, MDOC_HEAD);
433: if ( ! node_append(mdoc, p))
1.25 schwarze 434: return(0);
1.93 schwarze 435: mdoc->next = MDOC_NEXT_CHILD;
1.25 schwarze 436: return(1);
1.1 kristaps 437: }
438:
439: int
1.93 schwarze 440: mdoc_body_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.1 kristaps 441: {
442: struct mdoc_node *p;
443:
1.93 schwarze 444: p = node_alloc(mdoc, line, pos, tok, MDOC_BODY);
445: if ( ! node_append(mdoc, p))
1.25 schwarze 446: return(0);
1.93 schwarze 447: mdoc->next = MDOC_NEXT_CHILD;
1.59 schwarze 448: return(1);
449: }
450:
451: int
1.93 schwarze 452: mdoc_endbody_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok,
1.59 schwarze 453: struct mdoc_node *body, enum mdoc_endbody end)
454: {
455: struct mdoc_node *p;
456:
1.93 schwarze 457: p = node_alloc(mdoc, line, pos, tok, MDOC_BODY);
1.59 schwarze 458: p->pending = body;
1.92 schwarze 459: p->norm = body->norm;
1.59 schwarze 460: p->end = end;
1.93 schwarze 461: if ( ! node_append(mdoc, p))
1.59 schwarze 462: return(0);
1.93 schwarze 463: mdoc->next = MDOC_NEXT_SIBLING;
1.25 schwarze 464: return(1);
1.1 kristaps 465: }
466:
467: int
1.103 schwarze 468: mdoc_block_alloc(struct mdoc *mdoc, int line, int pos,
1.37 schwarze 469: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 470: {
471: struct mdoc_node *p;
472:
1.93 schwarze 473: p = node_alloc(mdoc, line, pos, tok, MDOC_BLOCK);
1.4 schwarze 474: p->args = args;
475: if (p->args)
1.1 kristaps 476: (args->refcnt)++;
1.75 schwarze 477:
478: switch (tok) {
1.103 schwarze 479: case MDOC_Bd:
1.75 schwarze 480: /* FALLTHROUGH */
1.103 schwarze 481: case MDOC_Bf:
1.75 schwarze 482: /* FALLTHROUGH */
1.103 schwarze 483: case MDOC_Bl:
1.107 schwarze 484: /* FALLTHROUGH */
485: case MDOC_En:
1.75 schwarze 486: /* FALLTHROUGH */
1.103 schwarze 487: case MDOC_Rs:
1.75 schwarze 488: p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
489: break;
490: default:
491: break;
492: }
493:
1.93 schwarze 494: if ( ! node_append(mdoc, p))
1.25 schwarze 495: return(0);
1.93 schwarze 496: mdoc->next = MDOC_NEXT_CHILD;
1.25 schwarze 497: return(1);
1.1 kristaps 498: }
499:
500: int
1.103 schwarze 501: mdoc_elem_alloc(struct mdoc *mdoc, int line, int pos,
1.37 schwarze 502: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 503: {
504: struct mdoc_node *p;
505:
1.93 schwarze 506: p = node_alloc(mdoc, line, pos, tok, MDOC_ELEM);
1.4 schwarze 507: p->args = args;
508: if (p->args)
1.1 kristaps 509: (args->refcnt)++;
1.75 schwarze 510:
511: switch (tok) {
1.103 schwarze 512: case MDOC_An:
1.75 schwarze 513: p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
514: break;
515: default:
516: break;
517: }
518:
1.93 schwarze 519: if ( ! node_append(mdoc, p))
1.25 schwarze 520: return(0);
1.93 schwarze 521: mdoc->next = MDOC_NEXT_CHILD;
1.25 schwarze 522: return(1);
1.1 kristaps 523: }
524:
1.46 schwarze 525: int
1.93 schwarze 526: mdoc_word_alloc(struct mdoc *mdoc, int line, int pos, const char *p)
1.1 kristaps 527: {
1.19 schwarze 528: struct mdoc_node *n;
1.1 kristaps 529:
1.93 schwarze 530: n = node_alloc(mdoc, line, pos, MDOC_MAX, MDOC_TEXT);
531: n->string = roff_strdup(mdoc->roff, p);
1.19 schwarze 532:
1.93 schwarze 533: if ( ! node_append(mdoc, n))
1.25 schwarze 534: return(0);
1.46 schwarze 535:
1.93 schwarze 536: mdoc->next = MDOC_NEXT_SIBLING;
1.25 schwarze 537: return(1);
1.19 schwarze 538: }
539:
1.95 schwarze 540: void
541: mdoc_word_append(struct mdoc *mdoc, const char *p)
542: {
543: struct mdoc_node *n;
544: char *addstr, *newstr;
545:
546: n = mdoc->last;
547: addstr = roff_strdup(mdoc->roff, p);
1.100 schwarze 548: mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
1.95 schwarze 549: free(addstr);
550: free(n->string);
551: n->string = newstr;
552: mdoc->next = MDOC_NEXT_SIBLING;
553: }
1.19 schwarze 554:
1.61 schwarze 555: static void
1.1 kristaps 556: mdoc_node_free(struct mdoc_node *p)
557: {
1.61 schwarze 558:
1.75 schwarze 559: if (MDOC_BLOCK == p->type || MDOC_ELEM == p->type)
560: free(p->norm);
1.1 kristaps 561: if (p->string)
562: free(p->string);
563: if (p->args)
564: mdoc_argv_free(p->args);
565: free(p);
566: }
567:
1.40 schwarze 568: static void
1.93 schwarze 569: mdoc_node_unlink(struct mdoc *mdoc, struct mdoc_node *n)
1.40 schwarze 570: {
571:
572: /* Adjust siblings. */
573:
574: if (n->prev)
575: n->prev->next = n->next;
576: if (n->next)
577: n->next->prev = n->prev;
578:
579: /* Adjust parent. */
580:
581: if (n->parent) {
582: n->parent->nchild--;
583: if (n->parent->child == n)
584: n->parent->child = n->prev ? n->prev : n->next;
1.72 schwarze 585: if (n->parent->last == n)
586: n->parent->last = n->prev ? n->prev : NULL;
1.40 schwarze 587: }
588:
589: /* Adjust parse point, if applicable. */
590:
1.93 schwarze 591: if (mdoc && mdoc->last == n) {
1.40 schwarze 592: if (n->prev) {
1.93 schwarze 593: mdoc->last = n->prev;
594: mdoc->next = MDOC_NEXT_SIBLING;
1.40 schwarze 595: } else {
1.93 schwarze 596: mdoc->last = n->parent;
597: mdoc->next = MDOC_NEXT_CHILD;
1.40 schwarze 598: }
599: }
600:
1.93 schwarze 601: if (mdoc && mdoc->first == n)
602: mdoc->first = NULL;
1.40 schwarze 603: }
604:
1.1 kristaps 605: void
1.93 schwarze 606: mdoc_node_delete(struct mdoc *mdoc, struct mdoc_node *p)
1.1 kristaps 607: {
608:
1.40 schwarze 609: while (p->child) {
610: assert(p->nchild);
1.93 schwarze 611: mdoc_node_delete(mdoc, p->child);
1.40 schwarze 612: }
613: assert(0 == p->nchild);
1.1 kristaps 614:
1.93 schwarze 615: mdoc_node_unlink(mdoc, p);
1.1 kristaps 616: mdoc_node_free(p);
1.91 schwarze 617: }
618:
619: int
1.93 schwarze 620: mdoc_node_relink(struct mdoc *mdoc, struct mdoc_node *p)
1.91 schwarze 621: {
622:
1.93 schwarze 623: mdoc_node_unlink(mdoc, p);
624: return(node_append(mdoc, p));
1.1 kristaps 625: }
626:
1.84 schwarze 627: #if 0
628: /*
629: * Pre-treat a text line.
630: * Text lines can consist of equations, which must be handled apart from
631: * the regular text.
632: * Thus, use this function to step through a line checking if it has any
633: * equations embedded in it.
634: * This must handle multiple equations AND equations that do not end at
635: * the end-of-line, i.e., will re-enter in the next roff parse.
636: */
637: static int
1.93 schwarze 638: mdoc_preptext(struct mdoc *mdoc, int line, char *buf, int offs)
1.84 schwarze 639: {
640: char *start, *end;
641: char delim;
642:
643: while ('\0' != buf[offs]) {
644: /* Mark starting position if eqn is set. */
645: start = NULL;
1.93 schwarze 646: if ('\0' != (delim = roff_eqndelim(mdoc->roff)))
1.84 schwarze 647: if (NULL != (start = strchr(buf + offs, delim)))
648: *start++ = '\0';
649:
650: /* Parse text as normal. */
1.93 schwarze 651: if ( ! mdoc_ptext(mdoc, line, buf, offs))
1.84 schwarze 652: return(0);
653:
654: /* Continue only if an equation exists. */
655: if (NULL == start)
656: break;
657:
658: /* Read past the end of the equation. */
659: offs += start - (buf + offs);
660: assert(start == &buf[offs]);
661: if (NULL != (end = strchr(buf + offs, delim))) {
662: *end++ = '\0';
663: while (' ' == *end)
664: end++;
665: }
666:
667: /* Parse the equation itself. */
1.93 schwarze 668: roff_openeqn(mdoc->roff, NULL, line, offs, buf);
1.84 schwarze 669:
670: /* Process a finished equation? */
1.93 schwarze 671: if (roff_closeeqn(mdoc->roff))
672: if ( ! mdoc_addeqn(mdoc, roff_eqn(mdoc->roff)))
1.84 schwarze 673: return(0);
674: offs += (end - (buf + offs));
1.103 schwarze 675: }
1.84 schwarze 676:
677: return(1);
678: }
679: #endif
1.1 kristaps 680:
681: /*
682: * Parse free-form text, that is, a line that does not begin with the
683: * control character.
684: */
685: static int
1.93 schwarze 686: mdoc_ptext(struct mdoc *mdoc, int line, char *buf, int offs)
1.1 kristaps 687: {
1.56 schwarze 688: char *c, *ws, *end;
689: struct mdoc_node *n;
1.46 schwarze 690:
1.93 schwarze 691: assert(mdoc->last);
692: n = mdoc->last;
1.56 schwarze 693:
694: /*
695: * Divert directly to list processing if we're encountering a
696: * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
697: * (a MDOC_BODY means it's already open, in which case we should
698: * process within its context in the normal way).
699: */
700:
701: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.103 schwarze 702: LIST_column == n->norm->Bl.type) {
1.56 schwarze 703: /* `Bl' is open without any children. */
1.93 schwarze 704: mdoc->flags |= MDOC_FREECOL;
705: return(mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf));
1.56 schwarze 706: }
707:
708: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
1.103 schwarze 709: NULL != n->parent &&
710: MDOC_Bl == n->parent->tok &&
711: LIST_column == n->parent->norm->Bl.type) {
1.56 schwarze 712: /* `Bl' has block-level `It' children. */
1.93 schwarze 713: mdoc->flags |= MDOC_FREECOL;
714: return(mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf));
1.56 schwarze 715: }
716:
1.52 schwarze 717: /*
718: * Search for the beginning of unescaped trailing whitespace (ws)
719: * and for the first character not to be output (end).
720: */
1.56 schwarze 721:
722: /* FIXME: replace with strcspn(). */
1.52 schwarze 723: ws = NULL;
1.53 schwarze 724: for (c = end = buf + offs; *c; c++) {
1.52 schwarze 725: switch (*c) {
726: case ' ':
727: if (NULL == ws)
728: ws = c;
729: continue;
730: case '\t':
731: /*
732: * Always warn about trailing tabs,
733: * even outside literal context,
734: * where they should be put on the next line.
735: */
736: if (NULL == ws)
737: ws = c;
738: /*
739: * Strip trailing tabs in literal context only;
740: * outside, they affect the next line.
741: */
1.93 schwarze 742: if (MDOC_LITERAL & mdoc->flags)
1.52 schwarze 743: continue;
744: break;
745: case '\\':
746: /* Skip the escaped character, too, if any. */
747: if (c[1])
748: c++;
749: /* FALLTHROUGH */
750: default:
751: ws = NULL;
752: break;
753: }
754: end = c + 1;
755: }
756: *end = '\0';
1.19 schwarze 757:
1.52 schwarze 758: if (ws)
1.108 schwarze 759: mandoc_msg(MANDOCERR_SPACE_EOL, mdoc->parse,
760: line, (int)(ws-buf), NULL);
1.34 schwarze 761:
1.93 schwarze 762: if ('\0' == buf[offs] && ! (MDOC_LITERAL & mdoc->flags)) {
1.108 schwarze 763: mandoc_msg(MANDOCERR_FI_BLANK, mdoc->parse,
764: line, (int)(c - buf), NULL);
1.46 schwarze 765:
1.40 schwarze 766: /*
1.66 schwarze 767: * Insert a `sp' in the case of a blank line. Technically,
1.46 schwarze 768: * blank lines aren't allowed, but enough manuals assume this
769: * behaviour that we want to work around it.
1.40 schwarze 770: */
1.93 schwarze 771: if ( ! mdoc_elem_alloc(mdoc, line, offs, MDOC_sp, NULL))
1.42 schwarze 772: return(0);
1.46 schwarze 773:
1.93 schwarze 774: mdoc->next = MDOC_NEXT_SIBLING;
1.89 schwarze 775:
1.93 schwarze 776: return(mdoc_valid_post(mdoc));
1.38 schwarze 777: }
1.1 kristaps 778:
1.93 schwarze 779: if ( ! mdoc_word_alloc(mdoc, line, offs, buf+offs))
1.1 kristaps 780: return(0);
781:
1.93 schwarze 782: if (MDOC_LITERAL & mdoc->flags)
1.52 schwarze 783: return(1);
784:
1.35 schwarze 785: /*
1.48 schwarze 786: * End-of-sentence check. If the last character is an unescaped
787: * EOS character, then flag the node as being the end of a
788: * sentence. The front-end will know how to interpret this.
1.35 schwarze 789: */
790:
1.52 schwarze 791: assert(buf < end);
1.48 schwarze 792:
1.97 schwarze 793: if (mandoc_eos(buf+offs, (size_t)(end-buf-offs)))
1.93 schwarze 794: mdoc->last->flags |= MDOC_EOS;
1.48 schwarze 795:
1.1 kristaps 796: return(1);
797: }
1.19 schwarze 798:
1.1 kristaps 799: /*
800: * Parse a macro line, that is, a line beginning with the control
801: * character.
802: */
1.61 schwarze 803: static int
1.93 schwarze 804: mdoc_pmacro(struct mdoc *mdoc, int ln, char *buf, int offs)
1.1 kristaps 805: {
1.56 schwarze 806: enum mdoct tok;
1.83 schwarze 807: int i, sv;
1.56 schwarze 808: char mac[5];
809: struct mdoc_node *n;
1.1 kristaps 810:
1.83 schwarze 811: /* Empty post-control lines are ignored. */
1.1 kristaps 812:
1.83 schwarze 813: if ('"' == buf[offs]) {
1.108 schwarze 814: mandoc_msg(MANDOCERR_COMMENT_BAD, mdoc->parse,
815: ln, offs, NULL);
1.83 schwarze 816: return(1);
817: } else if ('\0' == buf[offs])
1.1 kristaps 818: return(1);
819:
1.83 schwarze 820: sv = offs;
1.47 schwarze 821:
1.103 schwarze 822: /*
1.64 schwarze 823: * Copy the first word into a nil-terminated buffer.
824: * Stop copying when a tab, space, or eoln is encountered.
1.63 schwarze 825: */
1.1 kristaps 826:
1.83 schwarze 827: i = 0;
1.103 schwarze 828: while (i < 4 && '\0' != buf[offs] && ' ' != buf[offs] &&
829: '\t' != buf[offs])
1.83 schwarze 830: mac[i++] = buf[offs++];
831:
832: mac[i] = '\0';
833:
1.104 schwarze 834: tok = (i > 1 && i < 4) ? mdoc_hash_find(mac) : MDOC_MAX;
1.1 kristaps 835:
1.65 schwarze 836: if (MDOC_MAX == tok) {
1.112 ! schwarze 837: mandoc_msg(MANDOCERR_MACRO, mdoc->parse,
! 838: ln, sv, buf + sv - 1);
1.1 kristaps 839: return(1);
840: }
841:
1.63 schwarze 842: /* Disregard the first trailing tab, if applicable. */
843:
1.83 schwarze 844: if ('\t' == buf[offs])
845: offs++;
1.63 schwarze 846:
847: /* Jump to the next non-whitespace word. */
1.1 kristaps 848:
1.83 schwarze 849: while (buf[offs] && ' ' == buf[offs])
850: offs++;
1.34 schwarze 851:
1.103 schwarze 852: /*
1.46 schwarze 853: * Trailing whitespace. Note that tabs are allowed to be passed
854: * into the parser as "text", so we only warn about spaces here.
855: */
1.34 schwarze 856:
1.83 schwarze 857: if ('\0' == buf[offs] && ' ' == buf[offs - 1])
1.108 schwarze 858: mandoc_msg(MANDOCERR_SPACE_EOL, mdoc->parse,
859: ln, offs - 1, NULL);
1.1 kristaps 860:
1.56 schwarze 861: /*
862: * If an initial macro or a list invocation, divert directly
863: * into macro processing.
864: */
865:
1.111 schwarze 866: if (NULL == mdoc->last || MDOC_It == tok || MDOC_El == tok)
867: return(mdoc_macro(mdoc, tok, ln, sv, &offs, buf));
1.56 schwarze 868:
1.93 schwarze 869: n = mdoc->last;
870: assert(mdoc->last);
1.56 schwarze 871:
872: /*
873: * If the first macro of a `Bl -column', open an `It' block
874: * context around the parsed macro.
875: */
876:
877: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.103 schwarze 878: LIST_column == n->norm->Bl.type) {
1.93 schwarze 879: mdoc->flags |= MDOC_FREECOL;
1.111 schwarze 880: return(mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf));
1.56 schwarze 881: }
882:
883: /*
884: * If we're following a block-level `It' within a `Bl -column'
885: * context (perhaps opened in the above block or in ptext()),
886: * then open an `It' block context around the parsed macro.
1.24 schwarze 887: */
1.56 schwarze 888:
889: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
1.103 schwarze 890: NULL != n->parent &&
891: MDOC_Bl == n->parent->tok &&
892: LIST_column == n->parent->norm->Bl.type) {
1.93 schwarze 893: mdoc->flags |= MDOC_FREECOL;
1.111 schwarze 894: return(mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf));
1.56 schwarze 895: }
896:
897: /* Normal processing of a macro. */
898:
1.103 schwarze 899: if ( ! mdoc_macro(mdoc, tok, ln, sv, &offs, buf))
1.111 schwarze 900: return(0);
1.98 schwarze 901:
902: /* In quick mode (for mandocdb), abort after the NAME section. */
903:
904: if (mdoc->quick && MDOC_Sh == tok &&
905: SEC_NAME != mdoc->last->sec)
906: return(2);
1.1 kristaps 907:
908: return(1);
909: }
1.24 schwarze 910:
1.83 schwarze 911: enum mdelim
912: mdoc_isdelim(const char *p)
913: {
914:
915: if ('\0' == p[0])
916: return(DELIM_NONE);
917:
918: if ('\0' == p[1])
919: switch (p[0]) {
1.103 schwarze 920: case '(':
1.83 schwarze 921: /* FALLTHROUGH */
1.103 schwarze 922: case '[':
1.83 schwarze 923: return(DELIM_OPEN);
1.103 schwarze 924: case '|':
1.83 schwarze 925: return(DELIM_MIDDLE);
1.103 schwarze 926: case '.':
1.83 schwarze 927: /* FALLTHROUGH */
1.103 schwarze 928: case ',':
1.83 schwarze 929: /* FALLTHROUGH */
1.103 schwarze 930: case ';':
1.83 schwarze 931: /* FALLTHROUGH */
1.103 schwarze 932: case ':':
1.83 schwarze 933: /* FALLTHROUGH */
1.103 schwarze 934: case '?':
1.83 schwarze 935: /* FALLTHROUGH */
1.103 schwarze 936: case '!':
1.83 schwarze 937: /* FALLTHROUGH */
1.103 schwarze 938: case ')':
1.83 schwarze 939: /* FALLTHROUGH */
1.103 schwarze 940: case ']':
1.83 schwarze 941: return(DELIM_CLOSE);
942: default:
943: return(DELIM_NONE);
944: }
945:
946: if ('\\' != p[0])
947: return(DELIM_NONE);
1.24 schwarze 948:
1.83 schwarze 949: if (0 == strcmp(p + 1, "."))
950: return(DELIM_CLOSE);
1.90 schwarze 951: if (0 == strcmp(p + 1, "fR|\\fP"))
1.83 schwarze 952: return(DELIM_MIDDLE);
953:
954: return(DELIM_NONE);
1.101 schwarze 955: }
956:
957: void
958: mdoc_deroff(char **dest, const struct mdoc_node *n)
959: {
960: char *cp;
961: size_t sz;
962:
963: if (MDOC_TEXT != n->type) {
964: for (n = n->child; n; n = n->next)
965: mdoc_deroff(dest, n);
966: return;
967: }
968:
969: /* Skip leading whitespace. */
970:
971: for (cp = n->string; '\0' != *cp; cp++)
972: if (0 == isspace((unsigned char)*cp))
973: break;
974:
975: /* Skip trailing whitespace. */
976:
977: for (sz = strlen(cp); sz; sz--)
978: if (0 == isspace((unsigned char)cp[sz-1]))
979: break;
980:
981: /* Skip empty strings. */
982:
983: if (0 == sz)
984: return;
985:
986: if (NULL == *dest) {
987: *dest = mandoc_strndup(cp, sz);
988: return;
989: }
990:
991: mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
992: free(*dest);
993: *dest = cp;
1.83 schwarze 994: }
![[BACK]](/icons/back.gif){kind=icon}
Return to mdoc.c CVS log ![[TXT]](/icons/text.gif){kind=icon}
![[DIR]](/icons/dir.gif){kind=icon}
Up to [local] / src / usr.bin / mandoc