Annotation of src/usr.bin/mandoc/mdoc.c, Revision 1.102
1.102 ! schwarze 1: /* $Id: mdoc.c,v 1.101 2014/03/23 12:44:18 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:
34: const char *const __mdoc_macronames[MDOC_MAX] = {
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",
45: /* LINTED */
1.33 schwarze 46: "Xr", "%A", "%B", "%D",
1.1 kristaps 47: /* LINTED */
1.33 schwarze 48: "%I", "%J", "%N", "%O",
1.1 kristaps 49: /* LINTED */
1.33 schwarze 50: "%P", "%R", "%T", "%V",
1.1 kristaps 51: "Ac", "Ao", "Aq", "At",
52: "Bc", "Bf", "Bo", "Bq",
53: "Bsx", "Bx", "Db", "Dc",
54: "Do", "Dq", "Ec", "Ef",
55: "Em", "Eo", "Fx", "Ms",
56: "No", "Ns", "Nx", "Ox",
57: "Pc", "Pf", "Po", "Pq",
58: "Qc", "Ql", "Qo", "Qq",
59: "Re", "Rs", "Sc", "So",
60: "Sq", "Sm", "Sx", "Sy",
61: "Tn", "Ux", "Xc", "Xo",
62: "Fo", "Fc", "Oo", "Oc",
63: "Bk", "Ek", "Bt", "Hf",
1.7 schwarze 64: "Fr", "Ud", "Lb", "Lp",
65: "Lk", "Mt", "Brq", "Bro",
1.1 kristaps 66: /* LINTED */
1.33 schwarze 67: "Brc", "%C", "Es", "En",
1.1 kristaps 68: /* LINTED */
1.33 schwarze 69: "Dx", "%Q", "br", "sp",
1.31 schwarze 70: /* LINTED */
1.102 ! schwarze 71: "%U", "Ta", "ll",
1.1 kristaps 72: };
73:
74: const char *const __mdoc_argnames[MDOC_ARG_MAX] = {
75: "split", "nosplit", "ragged",
76: "unfilled", "literal", "file",
77: "offset", "bullet", "dash",
78: "hyphen", "item", "enum",
79: "tag", "diag", "hang",
80: "ohang", "inset", "column",
81: "width", "compact", "std",
82: "filled", "words", "emphasis",
1.30 schwarze 83: "symbolic", "nested", "centered"
1.1 kristaps 84: };
85:
86: const char * const *mdoc_macronames = __mdoc_macronames;
87: const char * const *mdoc_argnames = __mdoc_argnames;
88:
1.40 schwarze 89: static void mdoc_node_free(struct mdoc_node *);
90: static void mdoc_node_unlink(struct mdoc *,
91: struct mdoc_node *);
1.1 kristaps 92: static void mdoc_free1(struct mdoc *);
1.32 schwarze 93: static void mdoc_alloc1(struct mdoc *);
1.1 kristaps 94: static struct mdoc_node *node_alloc(struct mdoc *, int, int,
1.37 schwarze 95: enum mdoct, enum mdoc_type);
1.1 kristaps 96: static int node_append(struct mdoc *,
97: struct mdoc_node *);
1.84 schwarze 98: #if 0
99: static int mdoc_preptext(struct mdoc *, int, char *, int);
100: #endif
1.53 schwarze 101: static int mdoc_ptext(struct mdoc *, int, char *, int);
102: static int mdoc_pmacro(struct mdoc *, int, char *, int);
1.1 kristaps 103:
104: const struct mdoc_node *
1.93 schwarze 105: mdoc_node(const struct mdoc *mdoc)
1.1 kristaps 106: {
107:
1.93 schwarze 108: assert( ! (MDOC_HALT & mdoc->flags));
109: return(mdoc->first);
1.1 kristaps 110: }
111:
112:
113: const struct mdoc_meta *
1.93 schwarze 114: mdoc_meta(const struct mdoc *mdoc)
1.1 kristaps 115: {
116:
1.93 schwarze 117: assert( ! (MDOC_HALT & mdoc->flags));
118: return(&mdoc->meta);
1.1 kristaps 119: }
120:
121:
1.8 schwarze 122: /*
123: * Frees volatile resources (parse tree, meta-data, fields).
124: */
1.1 kristaps 125: static void
126: mdoc_free1(struct mdoc *mdoc)
127: {
128:
129: if (mdoc->first)
1.40 schwarze 130: mdoc_node_delete(mdoc, mdoc->first);
1.1 kristaps 131: if (mdoc->meta.title)
132: free(mdoc->meta.title);
133: if (mdoc->meta.os)
134: free(mdoc->meta.os);
135: if (mdoc->meta.name)
136: free(mdoc->meta.name);
137: if (mdoc->meta.arch)
138: free(mdoc->meta.arch);
139: if (mdoc->meta.vol)
140: free(mdoc->meta.vol);
1.50 schwarze 141: if (mdoc->meta.msec)
142: free(mdoc->meta.msec);
1.82 schwarze 143: if (mdoc->meta.date)
144: free(mdoc->meta.date);
1.1 kristaps 145: }
146:
147:
1.8 schwarze 148: /*
149: * Allocate all volatile resources (parse tree, meta-data, fields).
150: */
1.32 schwarze 151: static void
1.1 kristaps 152: mdoc_alloc1(struct mdoc *mdoc)
153: {
154:
1.32 schwarze 155: memset(&mdoc->meta, 0, sizeof(struct mdoc_meta));
1.1 kristaps 156: mdoc->flags = 0;
1.9 schwarze 157: mdoc->lastnamed = mdoc->lastsec = SEC_NONE;
1.32 schwarze 158: mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node));
1.1 kristaps 159: mdoc->first = mdoc->last;
160: mdoc->last->type = MDOC_ROOT;
1.86 schwarze 161: mdoc->last->tok = MDOC_MAX;
1.1 kristaps 162: mdoc->next = MDOC_NEXT_CHILD;
163: }
164:
165:
166: /*
1.8 schwarze 167: * Free up volatile resources (see mdoc_free1()) then re-initialises the
168: * data with mdoc_alloc1(). After invocation, parse data has been reset
169: * and the parser is ready for re-invocation on a new tree; however,
170: * cross-parse non-volatile data is kept intact.
1.1 kristaps 171: */
1.32 schwarze 172: void
1.1 kristaps 173: mdoc_reset(struct mdoc *mdoc)
174: {
175:
176: mdoc_free1(mdoc);
1.32 schwarze 177: mdoc_alloc1(mdoc);
1.1 kristaps 178: }
179:
180:
181: /*
1.8 schwarze 182: * Completely free up all volatile and non-volatile parse resources.
183: * After invocation, the pointer is no longer usable.
1.1 kristaps 184: */
185: void
186: mdoc_free(struct mdoc *mdoc)
187: {
188:
189: mdoc_free1(mdoc);
190: free(mdoc);
191: }
192:
193:
1.8 schwarze 194: /*
195: * Allocate volatile and non-volatile parse resources.
196: */
1.1 kristaps 197: struct mdoc *
1.98 schwarze 198: mdoc_alloc(struct roff *roff, struct mparse *parse,
199: char *defos, int quick)
1.1 kristaps 200: {
201: struct mdoc *p;
202:
1.32 schwarze 203: p = mandoc_calloc(1, sizeof(struct mdoc));
204:
1.83 schwarze 205: p->parse = parse;
1.87 schwarze 206: p->defos = defos;
1.98 schwarze 207: p->quick = quick;
1.84 schwarze 208: p->roff = roff;
1.1 kristaps 209:
1.32 schwarze 210: mdoc_hash_init();
211: mdoc_alloc1(p);
212: return(p);
1.1 kristaps 213: }
214:
215:
216: /*
217: * Climb back up the parse tree, validating open scopes. Mostly calls
1.8 schwarze 218: * through to macro_end() in macro.c.
1.1 kristaps 219: */
220: int
1.93 schwarze 221: mdoc_endparse(struct mdoc *mdoc)
1.1 kristaps 222: {
223:
1.93 schwarze 224: assert( ! (MDOC_HALT & mdoc->flags));
225: if (mdoc_macroend(mdoc))
1.1 kristaps 226: return(1);
1.93 schwarze 227: mdoc->flags |= MDOC_HALT;
1.1 kristaps 228: return(0);
1.81 schwarze 229: }
230:
231: int
1.93 schwarze 232: mdoc_addeqn(struct mdoc *mdoc, const struct eqn *ep)
1.81 schwarze 233: {
234: struct mdoc_node *n;
235:
1.93 schwarze 236: assert( ! (MDOC_HALT & mdoc->flags));
1.81 schwarze 237:
238: /* No text before an initial macro. */
239:
1.93 schwarze 240: if (SEC_NONE == mdoc->lastnamed) {
241: mdoc_pmsg(mdoc, ep->ln, ep->pos, MANDOCERR_NOTEXT);
1.81 schwarze 242: return(1);
243: }
244:
1.93 schwarze 245: n = node_alloc(mdoc, ep->ln, ep->pos, MDOC_MAX, MDOC_EQN);
1.81 schwarze 246: n->eqn = ep;
247:
1.93 schwarze 248: if ( ! node_append(mdoc, n))
1.81 schwarze 249: return(0);
250:
1.93 schwarze 251: mdoc->next = MDOC_NEXT_SIBLING;
1.81 schwarze 252: return(1);
1.1 kristaps 253: }
254:
1.77 schwarze 255: int
1.93 schwarze 256: mdoc_addspan(struct mdoc *mdoc, const struct tbl_span *sp)
1.77 schwarze 257: {
1.79 schwarze 258: struct mdoc_node *n;
1.77 schwarze 259:
1.93 schwarze 260: assert( ! (MDOC_HALT & mdoc->flags));
1.77 schwarze 261:
262: /* No text before an initial macro. */
263:
1.93 schwarze 264: if (SEC_NONE == mdoc->lastnamed) {
265: mdoc_pmsg(mdoc, sp->line, 0, MANDOCERR_NOTEXT);
1.77 schwarze 266: return(1);
267: }
268:
1.93 schwarze 269: n = node_alloc(mdoc, sp->line, 0, MDOC_MAX, MDOC_TBL);
1.79 schwarze 270: n->span = sp;
271:
1.93 schwarze 272: if ( ! node_append(mdoc, n))
1.79 schwarze 273: return(0);
274:
1.93 schwarze 275: mdoc->next = MDOC_NEXT_SIBLING;
1.79 schwarze 276: return(1);
1.77 schwarze 277: }
278:
1.1 kristaps 279:
280: /*
281: * Main parse routine. Parses a single line -- really just hands off to
1.45 schwarze 282: * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
1.1 kristaps 283: */
284: int
1.93 schwarze 285: mdoc_parseln(struct mdoc *mdoc, int ln, char *buf, int offs)
1.1 kristaps 286: {
287:
1.93 schwarze 288: assert( ! (MDOC_HALT & mdoc->flags));
1.1 kristaps 289:
1.93 schwarze 290: mdoc->flags |= MDOC_NEWLINE;
1.60 schwarze 291:
292: /*
293: * Let the roff nS register switch SYNOPSIS mode early,
294: * such that the parser knows at all times
295: * whether this mode is on or off.
296: * Note that this mode is also switched by the Sh macro.
297: */
1.94 schwarze 298: if (roff_getreg(mdoc->roff, "nS"))
299: mdoc->flags |= MDOC_SYNOPSIS;
300: else
301: mdoc->flags &= ~MDOC_SYNOPSIS;
1.60 schwarze 302:
1.93 schwarze 303: return(roff_getcontrol(mdoc->roff, buf, &offs) ?
304: mdoc_pmacro(mdoc, ln, buf, offs) :
305: mdoc_ptext(mdoc, ln, buf, offs));
1.1 kristaps 306: }
307:
1.5 schwarze 308: int
1.58 schwarze 309: mdoc_macro(MACRO_PROT_ARGS)
1.1 kristaps 310: {
1.46 schwarze 311: assert(tok < MDOC_MAX);
312:
313: /* If we're in the body, deny prologue calls. */
1.37 schwarze 314:
1.1 kristaps 315: if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
1.93 schwarze 316: MDOC_PBODY & mdoc->flags) {
317: mdoc_pmsg(mdoc, line, ppos, MANDOCERR_BADBODY);
1.76 schwarze 318: return(1);
319: }
1.46 schwarze 320:
321: /* If we're in the prologue, deny "body" macros. */
322:
1.1 kristaps 323: if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
1.93 schwarze 324: ! (MDOC_PBODY & mdoc->flags)) {
325: mdoc_pmsg(mdoc, line, ppos, MANDOCERR_BADPROLOG);
326: if (NULL == mdoc->meta.msec)
327: mdoc->meta.msec = mandoc_strdup("1");
328: if (NULL == mdoc->meta.title)
329: mdoc->meta.title = mandoc_strdup("UNKNOWN");
330: if (NULL == mdoc->meta.vol)
331: mdoc->meta.vol = mandoc_strdup("LOCAL");
332: if (NULL == mdoc->meta.os)
333: mdoc->meta.os = mandoc_strdup("LOCAL");
334: if (NULL == mdoc->meta.date)
335: mdoc->meta.date = mandoc_normdate
336: (mdoc->parse, NULL, line, ppos);
337: mdoc->flags |= MDOC_PBODY;
1.39 schwarze 338: }
1.1 kristaps 339:
1.93 schwarze 340: return((*mdoc_macros[tok].fp)(mdoc, tok, line, ppos, pos, buf));
1.1 kristaps 341: }
342:
343:
344: static int
345: node_append(struct mdoc *mdoc, struct mdoc_node *p)
346: {
347:
348: assert(mdoc->last);
349: assert(mdoc->first);
350: assert(MDOC_ROOT != p->type);
351:
352: switch (mdoc->next) {
353: case (MDOC_NEXT_SIBLING):
354: mdoc->last->next = p;
355: p->prev = mdoc->last;
356: p->parent = mdoc->last->parent;
357: break;
358: case (MDOC_NEXT_CHILD):
359: mdoc->last->child = p;
360: p->parent = mdoc->last;
361: break;
362: default:
363: abort();
364: /* NOTREACHED */
365: }
366:
1.10 schwarze 367: p->parent->nchild++;
368:
1.75 schwarze 369: /*
370: * Copy over the normalised-data pointer of our parent. Not
371: * everybody has one, but copying a null pointer is fine.
372: */
373:
374: switch (p->type) {
375: case (MDOC_BODY):
1.92 schwarze 376: if (ENDBODY_NOT != p->end)
377: break;
1.75 schwarze 378: /* FALLTHROUGH */
379: case (MDOC_TAIL):
380: /* FALLTHROUGH */
381: case (MDOC_HEAD):
382: p->norm = p->parent->norm;
383: break;
384: default:
385: break;
386: }
387:
1.1 kristaps 388: if ( ! mdoc_valid_pre(mdoc, p))
389: return(0);
390:
391: switch (p->type) {
392: case (MDOC_HEAD):
393: assert(MDOC_BLOCK == p->parent->type);
394: p->parent->head = p;
395: break;
396: case (MDOC_TAIL):
397: assert(MDOC_BLOCK == p->parent->type);
398: p->parent->tail = p;
399: break;
400: case (MDOC_BODY):
1.59 schwarze 401: if (p->end)
402: break;
1.1 kristaps 403: assert(MDOC_BLOCK == p->parent->type);
404: p->parent->body = p;
405: break;
406: default:
407: break;
408: }
409:
410: mdoc->last = p;
411:
412: switch (p->type) {
1.77 schwarze 413: case (MDOC_TBL):
414: /* FALLTHROUGH */
1.1 kristaps 415: case (MDOC_TEXT):
416: if ( ! mdoc_valid_post(mdoc))
417: return(0);
418: break;
419: default:
420: break;
421: }
422:
423: return(1);
424: }
425:
426:
427: static struct mdoc_node *
1.93 schwarze 428: node_alloc(struct mdoc *mdoc, int line, int pos,
1.37 schwarze 429: enum mdoct tok, enum mdoc_type type)
1.1 kristaps 430: {
431: struct mdoc_node *p;
432:
1.32 schwarze 433: p = mandoc_calloc(1, sizeof(struct mdoc_node));
1.93 schwarze 434: p->sec = mdoc->lastsec;
1.1 kristaps 435: p->line = line;
436: p->pos = pos;
1.96 schwarze 437: p->lastline = line;
1.1 kristaps 438: p->tok = tok;
1.37 schwarze 439: p->type = type;
1.58 schwarze 440:
441: /* Flag analysis. */
442:
1.93 schwarze 443: if (MDOC_SYNOPSIS & mdoc->flags)
1.60 schwarze 444: p->flags |= MDOC_SYNPRETTY;
445: else
446: p->flags &= ~MDOC_SYNPRETTY;
1.93 schwarze 447: if (MDOC_NEWLINE & mdoc->flags)
1.47 schwarze 448: p->flags |= MDOC_LINE;
1.93 schwarze 449: mdoc->flags &= ~MDOC_NEWLINE;
1.58 schwarze 450:
1.1 kristaps 451: return(p);
452: }
453:
454:
455: int
1.93 schwarze 456: mdoc_tail_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.1 kristaps 457: {
458: struct mdoc_node *p;
459:
1.93 schwarze 460: p = node_alloc(mdoc, line, pos, tok, MDOC_TAIL);
461: if ( ! node_append(mdoc, p))
1.25 schwarze 462: return(0);
1.93 schwarze 463: mdoc->next = MDOC_NEXT_CHILD;
1.25 schwarze 464: return(1);
1.1 kristaps 465: }
466:
467:
468: int
1.93 schwarze 469: mdoc_head_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.1 kristaps 470: {
471: struct mdoc_node *p;
472:
1.93 schwarze 473: assert(mdoc->first);
474: assert(mdoc->last);
1.1 kristaps 475:
1.93 schwarze 476: p = node_alloc(mdoc, line, pos, tok, MDOC_HEAD);
477: if ( ! node_append(mdoc, p))
1.25 schwarze 478: return(0);
1.93 schwarze 479: mdoc->next = MDOC_NEXT_CHILD;
1.25 schwarze 480: return(1);
1.1 kristaps 481: }
482:
483:
484: int
1.93 schwarze 485: mdoc_body_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.1 kristaps 486: {
487: struct mdoc_node *p;
488:
1.93 schwarze 489: p = node_alloc(mdoc, line, pos, tok, MDOC_BODY);
490: if ( ! node_append(mdoc, p))
1.25 schwarze 491: return(0);
1.93 schwarze 492: mdoc->next = MDOC_NEXT_CHILD;
1.59 schwarze 493: return(1);
494: }
495:
496:
497: int
1.93 schwarze 498: mdoc_endbody_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok,
1.59 schwarze 499: struct mdoc_node *body, enum mdoc_endbody end)
500: {
501: struct mdoc_node *p;
502:
1.93 schwarze 503: p = node_alloc(mdoc, line, pos, tok, MDOC_BODY);
1.59 schwarze 504: p->pending = body;
1.92 schwarze 505: p->norm = body->norm;
1.59 schwarze 506: p->end = end;
1.93 schwarze 507: if ( ! node_append(mdoc, p))
1.59 schwarze 508: return(0);
1.93 schwarze 509: mdoc->next = MDOC_NEXT_SIBLING;
1.25 schwarze 510: return(1);
1.1 kristaps 511: }
512:
513:
514: int
1.93 schwarze 515: mdoc_block_alloc(struct mdoc *mdoc, int line, int pos,
1.37 schwarze 516: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 517: {
518: struct mdoc_node *p;
519:
1.93 schwarze 520: p = node_alloc(mdoc, line, pos, tok, MDOC_BLOCK);
1.4 schwarze 521: p->args = args;
522: if (p->args)
1.1 kristaps 523: (args->refcnt)++;
1.75 schwarze 524:
525: switch (tok) {
526: case (MDOC_Bd):
527: /* FALLTHROUGH */
528: case (MDOC_Bf):
529: /* FALLTHROUGH */
530: case (MDOC_Bl):
531: /* FALLTHROUGH */
532: case (MDOC_Rs):
533: p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
534: break;
535: default:
536: break;
537: }
538:
1.93 schwarze 539: if ( ! node_append(mdoc, p))
1.25 schwarze 540: return(0);
1.93 schwarze 541: mdoc->next = MDOC_NEXT_CHILD;
1.25 schwarze 542: return(1);
1.1 kristaps 543: }
544:
545:
546: int
1.93 schwarze 547: mdoc_elem_alloc(struct mdoc *mdoc, int line, int pos,
1.37 schwarze 548: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 549: {
550: struct mdoc_node *p;
551:
1.93 schwarze 552: p = node_alloc(mdoc, line, pos, tok, MDOC_ELEM);
1.4 schwarze 553: p->args = args;
554: if (p->args)
1.1 kristaps 555: (args->refcnt)++;
1.75 schwarze 556:
557: switch (tok) {
558: case (MDOC_An):
559: p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
560: break;
561: default:
562: break;
563: }
564:
1.93 schwarze 565: if ( ! node_append(mdoc, p))
1.25 schwarze 566: return(0);
1.93 schwarze 567: mdoc->next = MDOC_NEXT_CHILD;
1.25 schwarze 568: return(1);
1.1 kristaps 569: }
570:
1.46 schwarze 571: int
1.93 schwarze 572: mdoc_word_alloc(struct mdoc *mdoc, int line, int pos, const char *p)
1.1 kristaps 573: {
1.19 schwarze 574: struct mdoc_node *n;
1.1 kristaps 575:
1.93 schwarze 576: n = node_alloc(mdoc, line, pos, MDOC_MAX, MDOC_TEXT);
577: n->string = roff_strdup(mdoc->roff, p);
1.19 schwarze 578:
1.93 schwarze 579: if ( ! node_append(mdoc, n))
1.25 schwarze 580: return(0);
1.46 schwarze 581:
1.93 schwarze 582: mdoc->next = MDOC_NEXT_SIBLING;
1.25 schwarze 583: return(1);
1.19 schwarze 584: }
585:
1.95 schwarze 586: void
587: mdoc_word_append(struct mdoc *mdoc, const char *p)
588: {
589: struct mdoc_node *n;
590: char *addstr, *newstr;
591:
592: n = mdoc->last;
593: addstr = roff_strdup(mdoc->roff, p);
1.100 schwarze 594: mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
1.95 schwarze 595: free(addstr);
596: free(n->string);
597: n->string = newstr;
598: mdoc->next = MDOC_NEXT_SIBLING;
599: }
1.19 schwarze 600:
1.61 schwarze 601: static void
1.1 kristaps 602: mdoc_node_free(struct mdoc_node *p)
603: {
1.61 schwarze 604:
1.75 schwarze 605: if (MDOC_BLOCK == p->type || MDOC_ELEM == p->type)
606: free(p->norm);
1.1 kristaps 607: if (p->string)
608: free(p->string);
609: if (p->args)
610: mdoc_argv_free(p->args);
611: free(p);
612: }
613:
614:
1.40 schwarze 615: static void
1.93 schwarze 616: mdoc_node_unlink(struct mdoc *mdoc, struct mdoc_node *n)
1.40 schwarze 617: {
618:
619: /* Adjust siblings. */
620:
621: if (n->prev)
622: n->prev->next = n->next;
623: if (n->next)
624: n->next->prev = n->prev;
625:
626: /* Adjust parent. */
627:
628: if (n->parent) {
629: n->parent->nchild--;
630: if (n->parent->child == n)
631: n->parent->child = n->prev ? n->prev : n->next;
1.72 schwarze 632: if (n->parent->last == n)
633: n->parent->last = n->prev ? n->prev : NULL;
1.40 schwarze 634: }
635:
636: /* Adjust parse point, if applicable. */
637:
1.93 schwarze 638: if (mdoc && mdoc->last == n) {
1.40 schwarze 639: if (n->prev) {
1.93 schwarze 640: mdoc->last = n->prev;
641: mdoc->next = MDOC_NEXT_SIBLING;
1.40 schwarze 642: } else {
1.93 schwarze 643: mdoc->last = n->parent;
644: mdoc->next = MDOC_NEXT_CHILD;
1.40 schwarze 645: }
646: }
647:
1.93 schwarze 648: if (mdoc && mdoc->first == n)
649: mdoc->first = NULL;
1.40 schwarze 650: }
651:
652:
1.1 kristaps 653: void
1.93 schwarze 654: mdoc_node_delete(struct mdoc *mdoc, struct mdoc_node *p)
1.1 kristaps 655: {
656:
1.40 schwarze 657: while (p->child) {
658: assert(p->nchild);
1.93 schwarze 659: mdoc_node_delete(mdoc, p->child);
1.40 schwarze 660: }
661: assert(0 == p->nchild);
1.1 kristaps 662:
1.93 schwarze 663: mdoc_node_unlink(mdoc, p);
1.1 kristaps 664: mdoc_node_free(p);
1.91 schwarze 665: }
666:
667: int
1.93 schwarze 668: mdoc_node_relink(struct mdoc *mdoc, struct mdoc_node *p)
1.91 schwarze 669: {
670:
1.93 schwarze 671: mdoc_node_unlink(mdoc, p);
672: return(node_append(mdoc, p));
1.1 kristaps 673: }
674:
1.84 schwarze 675: #if 0
676: /*
677: * Pre-treat a text line.
678: * Text lines can consist of equations, which must be handled apart from
679: * the regular text.
680: * Thus, use this function to step through a line checking if it has any
681: * equations embedded in it.
682: * This must handle multiple equations AND equations that do not end at
683: * the end-of-line, i.e., will re-enter in the next roff parse.
684: */
685: static int
1.93 schwarze 686: mdoc_preptext(struct mdoc *mdoc, int line, char *buf, int offs)
1.84 schwarze 687: {
688: char *start, *end;
689: char delim;
690:
691: while ('\0' != buf[offs]) {
692: /* Mark starting position if eqn is set. */
693: start = NULL;
1.93 schwarze 694: if ('\0' != (delim = roff_eqndelim(mdoc->roff)))
1.84 schwarze 695: if (NULL != (start = strchr(buf + offs, delim)))
696: *start++ = '\0';
697:
698: /* Parse text as normal. */
1.93 schwarze 699: if ( ! mdoc_ptext(mdoc, line, buf, offs))
1.84 schwarze 700: return(0);
701:
702: /* Continue only if an equation exists. */
703: if (NULL == start)
704: break;
705:
706: /* Read past the end of the equation. */
707: offs += start - (buf + offs);
708: assert(start == &buf[offs]);
709: if (NULL != (end = strchr(buf + offs, delim))) {
710: *end++ = '\0';
711: while (' ' == *end)
712: end++;
713: }
714:
715: /* Parse the equation itself. */
1.93 schwarze 716: roff_openeqn(mdoc->roff, NULL, line, offs, buf);
1.84 schwarze 717:
718: /* Process a finished equation? */
1.93 schwarze 719: if (roff_closeeqn(mdoc->roff))
720: if ( ! mdoc_addeqn(mdoc, roff_eqn(mdoc->roff)))
1.84 schwarze 721: return(0);
722: offs += (end - (buf + offs));
723: }
724:
725: return(1);
726: }
727: #endif
1.1 kristaps 728:
729: /*
730: * Parse free-form text, that is, a line that does not begin with the
731: * control character.
732: */
733: static int
1.93 schwarze 734: mdoc_ptext(struct mdoc *mdoc, int line, char *buf, int offs)
1.1 kristaps 735: {
1.56 schwarze 736: char *c, *ws, *end;
737: struct mdoc_node *n;
1.44 schwarze 738:
1.46 schwarze 739: /* No text before an initial macro. */
740:
1.93 schwarze 741: if (SEC_NONE == mdoc->lastnamed) {
742: mdoc_pmsg(mdoc, line, offs, MANDOCERR_NOTEXT);
1.76 schwarze 743: return(1);
744: }
1.46 schwarze 745:
1.93 schwarze 746: assert(mdoc->last);
747: n = mdoc->last;
1.56 schwarze 748:
749: /*
750: * Divert directly to list processing if we're encountering a
751: * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
752: * (a MDOC_BODY means it's already open, in which case we should
753: * process within its context in the normal way).
754: */
755:
756: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.75 schwarze 757: LIST_column == n->norm->Bl.type) {
1.56 schwarze 758: /* `Bl' is open without any children. */
1.93 schwarze 759: mdoc->flags |= MDOC_FREECOL;
760: return(mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf));
1.56 schwarze 761: }
762:
763: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
764: NULL != n->parent &&
765: MDOC_Bl == n->parent->tok &&
1.75 schwarze 766: LIST_column == n->parent->norm->Bl.type) {
1.56 schwarze 767: /* `Bl' has block-level `It' children. */
1.93 schwarze 768: mdoc->flags |= MDOC_FREECOL;
769: return(mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf));
1.56 schwarze 770: }
771:
1.52 schwarze 772: /*
773: * Search for the beginning of unescaped trailing whitespace (ws)
774: * and for the first character not to be output (end).
775: */
1.56 schwarze 776:
777: /* FIXME: replace with strcspn(). */
1.52 schwarze 778: ws = NULL;
1.53 schwarze 779: for (c = end = buf + offs; *c; c++) {
1.52 schwarze 780: switch (*c) {
781: case ' ':
782: if (NULL == ws)
783: ws = c;
784: continue;
785: case '\t':
786: /*
787: * Always warn about trailing tabs,
788: * even outside literal context,
789: * where they should be put on the next line.
790: */
791: if (NULL == ws)
792: ws = c;
793: /*
794: * Strip trailing tabs in literal context only;
795: * outside, they affect the next line.
796: */
1.93 schwarze 797: if (MDOC_LITERAL & mdoc->flags)
1.52 schwarze 798: continue;
799: break;
800: case '\\':
801: /* Skip the escaped character, too, if any. */
802: if (c[1])
803: c++;
804: /* FALLTHROUGH */
805: default:
806: ws = NULL;
807: break;
808: }
809: end = c + 1;
810: }
811: *end = '\0';
1.19 schwarze 812:
1.52 schwarze 813: if (ws)
1.93 schwarze 814: mdoc_pmsg(mdoc, line, (int)(ws-buf), MANDOCERR_EOLNSPACE);
1.34 schwarze 815:
1.93 schwarze 816: if ('\0' == buf[offs] && ! (MDOC_LITERAL & mdoc->flags)) {
817: mdoc_pmsg(mdoc, line, (int)(c-buf), MANDOCERR_NOBLANKLN);
1.46 schwarze 818:
1.40 schwarze 819: /*
1.66 schwarze 820: * Insert a `sp' in the case of a blank line. Technically,
1.46 schwarze 821: * blank lines aren't allowed, but enough manuals assume this
822: * behaviour that we want to work around it.
1.40 schwarze 823: */
1.93 schwarze 824: if ( ! mdoc_elem_alloc(mdoc, line, offs, MDOC_sp, NULL))
1.42 schwarze 825: return(0);
1.46 schwarze 826:
1.93 schwarze 827: mdoc->next = MDOC_NEXT_SIBLING;
1.89 schwarze 828:
1.93 schwarze 829: return(mdoc_valid_post(mdoc));
1.38 schwarze 830: }
1.1 kristaps 831:
1.93 schwarze 832: if ( ! mdoc_word_alloc(mdoc, line, offs, buf+offs))
1.1 kristaps 833: return(0);
834:
1.93 schwarze 835: if (MDOC_LITERAL & mdoc->flags)
1.52 schwarze 836: return(1);
837:
1.35 schwarze 838: /*
1.48 schwarze 839: * End-of-sentence check. If the last character is an unescaped
840: * EOS character, then flag the node as being the end of a
841: * sentence. The front-end will know how to interpret this.
1.35 schwarze 842: */
843:
1.52 schwarze 844: assert(buf < end);
1.48 schwarze 845:
1.97 schwarze 846: if (mandoc_eos(buf+offs, (size_t)(end-buf-offs)))
1.93 schwarze 847: mdoc->last->flags |= MDOC_EOS;
1.48 schwarze 848:
1.1 kristaps 849: return(1);
850: }
1.19 schwarze 851:
1.1 kristaps 852:
853: /*
854: * Parse a macro line, that is, a line beginning with the control
855: * character.
856: */
1.61 schwarze 857: static int
1.93 schwarze 858: mdoc_pmacro(struct mdoc *mdoc, int ln, char *buf, int offs)
1.1 kristaps 859: {
1.56 schwarze 860: enum mdoct tok;
1.83 schwarze 861: int i, sv;
1.56 schwarze 862: char mac[5];
863: struct mdoc_node *n;
1.1 kristaps 864:
1.83 schwarze 865: /* Empty post-control lines are ignored. */
1.1 kristaps 866:
1.83 schwarze 867: if ('"' == buf[offs]) {
1.93 schwarze 868: mdoc_pmsg(mdoc, ln, offs, MANDOCERR_BADCOMMENT);
1.83 schwarze 869: return(1);
870: } else if ('\0' == buf[offs])
1.1 kristaps 871: return(1);
872:
1.83 schwarze 873: sv = offs;
1.47 schwarze 874:
1.63 schwarze 875: /*
1.64 schwarze 876: * Copy the first word into a nil-terminated buffer.
877: * Stop copying when a tab, space, or eoln is encountered.
1.63 schwarze 878: */
1.1 kristaps 879:
1.83 schwarze 880: i = 0;
881: while (i < 4 && '\0' != buf[offs] &&
882: ' ' != buf[offs] && '\t' != buf[offs])
883: mac[i++] = buf[offs++];
884:
885: mac[i] = '\0';
886:
887: tok = (i > 1 || i < 4) ? mdoc_hash_find(mac) : MDOC_MAX;
1.1 kristaps 888:
1.65 schwarze 889: if (MDOC_MAX == tok) {
1.93 schwarze 890: mandoc_vmsg(MANDOCERR_MACRO, mdoc->parse,
1.83 schwarze 891: ln, sv, "%s", buf + sv - 1);
1.1 kristaps 892: return(1);
893: }
894:
1.63 schwarze 895: /* Disregard the first trailing tab, if applicable. */
896:
1.83 schwarze 897: if ('\t' == buf[offs])
898: offs++;
1.63 schwarze 899:
900: /* Jump to the next non-whitespace word. */
1.1 kristaps 901:
1.83 schwarze 902: while (buf[offs] && ' ' == buf[offs])
903: offs++;
1.34 schwarze 904:
1.46 schwarze 905: /*
906: * Trailing whitespace. Note that tabs are allowed to be passed
907: * into the parser as "text", so we only warn about spaces here.
908: */
1.34 schwarze 909:
1.83 schwarze 910: if ('\0' == buf[offs] && ' ' == buf[offs - 1])
1.93 schwarze 911: mdoc_pmsg(mdoc, ln, offs - 1, MANDOCERR_EOLNSPACE);
1.1 kristaps 912:
1.56 schwarze 913: /*
914: * If an initial macro or a list invocation, divert directly
915: * into macro processing.
916: */
917:
1.93 schwarze 918: if (NULL == mdoc->last || MDOC_It == tok || MDOC_El == tok) {
919: if ( ! mdoc_macro(mdoc, tok, ln, sv, &offs, buf))
1.56 schwarze 920: goto err;
921: return(1);
922: }
923:
1.93 schwarze 924: n = mdoc->last;
925: assert(mdoc->last);
1.56 schwarze 926:
927: /*
928: * If the first macro of a `Bl -column', open an `It' block
929: * context around the parsed macro.
930: */
931:
932: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.75 schwarze 933: LIST_column == n->norm->Bl.type) {
1.93 schwarze 934: mdoc->flags |= MDOC_FREECOL;
935: if ( ! mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf))
1.56 schwarze 936: goto err;
937: return(1);
938: }
939:
940: /*
941: * If we're following a block-level `It' within a `Bl -column'
942: * context (perhaps opened in the above block or in ptext()),
943: * then open an `It' block context around the parsed macro.
1.24 schwarze 944: */
1.56 schwarze 945:
946: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
947: NULL != n->parent &&
948: MDOC_Bl == n->parent->tok &&
1.75 schwarze 949: LIST_column == n->parent->norm->Bl.type) {
1.93 schwarze 950: mdoc->flags |= MDOC_FREECOL;
951: if ( ! mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf))
1.56 schwarze 952: goto err;
953: return(1);
954: }
955:
956: /* Normal processing of a macro. */
957:
1.93 schwarze 958: if ( ! mdoc_macro(mdoc, tok, ln, sv, &offs, buf))
1.1 kristaps 959: goto err;
1.98 schwarze 960:
961: /* In quick mode (for mandocdb), abort after the NAME section. */
962:
963: if (mdoc->quick && MDOC_Sh == tok &&
964: SEC_NAME != mdoc->last->sec)
965: return(2);
1.1 kristaps 966:
967: return(1);
968:
969: err: /* Error out. */
970:
1.93 schwarze 971: mdoc->flags |= MDOC_HALT;
1.1 kristaps 972: return(0);
973: }
1.24 schwarze 974:
1.83 schwarze 975: enum mdelim
976: mdoc_isdelim(const char *p)
977: {
978:
979: if ('\0' == p[0])
980: return(DELIM_NONE);
981:
982: if ('\0' == p[1])
983: switch (p[0]) {
984: case('('):
985: /* FALLTHROUGH */
986: case('['):
987: return(DELIM_OPEN);
988: case('|'):
989: return(DELIM_MIDDLE);
990: case('.'):
991: /* FALLTHROUGH */
992: case(','):
993: /* FALLTHROUGH */
994: case(';'):
995: /* FALLTHROUGH */
996: case(':'):
997: /* FALLTHROUGH */
998: case('?'):
999: /* FALLTHROUGH */
1000: case('!'):
1001: /* FALLTHROUGH */
1002: case(')'):
1003: /* FALLTHROUGH */
1004: case(']'):
1005: return(DELIM_CLOSE);
1006: default:
1007: return(DELIM_NONE);
1008: }
1009:
1010: if ('\\' != p[0])
1011: return(DELIM_NONE);
1.24 schwarze 1012:
1.83 schwarze 1013: if (0 == strcmp(p + 1, "."))
1014: return(DELIM_CLOSE);
1.90 schwarze 1015: if (0 == strcmp(p + 1, "fR|\\fP"))
1.83 schwarze 1016: return(DELIM_MIDDLE);
1017:
1018: return(DELIM_NONE);
1.101 schwarze 1019: }
1020:
1021: void
1022: mdoc_deroff(char **dest, const struct mdoc_node *n)
1023: {
1024: char *cp;
1025: size_t sz;
1026:
1027: if (MDOC_TEXT != n->type) {
1028: for (n = n->child; n; n = n->next)
1029: mdoc_deroff(dest, n);
1030: return;
1031: }
1032:
1033: /* Skip leading whitespace. */
1034:
1035: for (cp = n->string; '\0' != *cp; cp++)
1036: if (0 == isspace((unsigned char)*cp))
1037: break;
1038:
1039: /* Skip trailing whitespace. */
1040:
1041: for (sz = strlen(cp); sz; sz--)
1042: if (0 == isspace((unsigned char)cp[sz-1]))
1043: break;
1044:
1045: /* Skip empty strings. */
1046:
1047: if (0 == sz)
1048: return;
1049:
1050: if (NULL == *dest) {
1051: *dest = mandoc_strndup(cp, sz);
1052: return;
1053: }
1054:
1055: mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
1056: free(*dest);
1057: *dest = cp;
1.83 schwarze 1058: }
![[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