Annotation of src/usr.bin/mandoc/mdoc.c, Revision 1.99
1.99 ! schwarze 1: /* $Id: mdoc.c,v 1.98 2014/01/05 20:26:27 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>
21: #include <stdarg.h>
22: #include <stdio.h>
23: #include <stdlib.h>
24: #include <string.h>
1.39 schwarze 25: #include <time.h>
1.1 kristaps 26:
1.83 schwarze 27: #include "mdoc.h"
1.54 schwarze 28: #include "mandoc.h"
1.99 ! schwarze 29: #include "mandoc_aux.h"
1.1 kristaps 30: #include "libmdoc.h"
1.32 schwarze 31: #include "libmandoc.h"
1.1 kristaps 32:
33: const char *const __mdoc_macronames[MDOC_MAX] = {
1.7 schwarze 34: "Ap", "Dd", "Dt", "Os",
1.1 kristaps 35: "Sh", "Ss", "Pp", "D1",
36: "Dl", "Bd", "Ed", "Bl",
37: "El", "It", "Ad", "An",
38: "Ar", "Cd", "Cm", "Dv",
39: "Er", "Ev", "Ex", "Fa",
40: "Fd", "Fl", "Fn", "Ft",
41: "Ic", "In", "Li", "Nd",
42: "Nm", "Op", "Ot", "Pa",
43: "Rv", "St", "Va", "Vt",
44: /* LINTED */
1.33 schwarze 45: "Xr", "%A", "%B", "%D",
1.1 kristaps 46: /* LINTED */
1.33 schwarze 47: "%I", "%J", "%N", "%O",
1.1 kristaps 48: /* LINTED */
1.33 schwarze 49: "%P", "%R", "%T", "%V",
1.1 kristaps 50: "Ac", "Ao", "Aq", "At",
51: "Bc", "Bf", "Bo", "Bq",
52: "Bsx", "Bx", "Db", "Dc",
53: "Do", "Dq", "Ec", "Ef",
54: "Em", "Eo", "Fx", "Ms",
55: "No", "Ns", "Nx", "Ox",
56: "Pc", "Pf", "Po", "Pq",
57: "Qc", "Ql", "Qo", "Qq",
58: "Re", "Rs", "Sc", "So",
59: "Sq", "Sm", "Sx", "Sy",
60: "Tn", "Ux", "Xc", "Xo",
61: "Fo", "Fc", "Oo", "Oc",
62: "Bk", "Ek", "Bt", "Hf",
1.7 schwarze 63: "Fr", "Ud", "Lb", "Lp",
64: "Lk", "Mt", "Brq", "Bro",
1.1 kristaps 65: /* LINTED */
1.33 schwarze 66: "Brc", "%C", "Es", "En",
1.1 kristaps 67: /* LINTED */
1.33 schwarze 68: "Dx", "%Q", "br", "sp",
1.31 schwarze 69: /* LINTED */
1.77 schwarze 70: "%U", "Ta"
1.1 kristaps 71: };
72:
73: const char *const __mdoc_argnames[MDOC_ARG_MAX] = {
74: "split", "nosplit", "ragged",
75: "unfilled", "literal", "file",
76: "offset", "bullet", "dash",
77: "hyphen", "item", "enum",
78: "tag", "diag", "hang",
79: "ohang", "inset", "column",
80: "width", "compact", "std",
81: "filled", "words", "emphasis",
1.30 schwarze 82: "symbolic", "nested", "centered"
1.1 kristaps 83: };
84:
85: const char * const *mdoc_macronames = __mdoc_macronames;
86: const char * const *mdoc_argnames = __mdoc_argnames;
87:
1.40 schwarze 88: static void mdoc_node_free(struct mdoc_node *);
89: static void mdoc_node_unlink(struct mdoc *,
90: struct mdoc_node *);
1.1 kristaps 91: static void mdoc_free1(struct mdoc *);
1.32 schwarze 92: static void mdoc_alloc1(struct mdoc *);
1.1 kristaps 93: static struct mdoc_node *node_alloc(struct mdoc *, int, int,
1.37 schwarze 94: enum mdoct, enum mdoc_type);
1.1 kristaps 95: static int node_append(struct mdoc *,
96: struct mdoc_node *);
1.84 schwarze 97: #if 0
98: static int mdoc_preptext(struct mdoc *, int, char *, int);
99: #endif
1.53 schwarze 100: static int mdoc_ptext(struct mdoc *, int, char *, int);
101: static int mdoc_pmacro(struct mdoc *, int, char *, int);
1.1 kristaps 102:
103: const struct mdoc_node *
1.93 schwarze 104: mdoc_node(const struct mdoc *mdoc)
1.1 kristaps 105: {
106:
1.93 schwarze 107: assert( ! (MDOC_HALT & mdoc->flags));
108: return(mdoc->first);
1.1 kristaps 109: }
110:
111:
112: const struct mdoc_meta *
1.93 schwarze 113: mdoc_meta(const struct mdoc *mdoc)
1.1 kristaps 114: {
115:
1.93 schwarze 116: assert( ! (MDOC_HALT & mdoc->flags));
117: return(&mdoc->meta);
1.1 kristaps 118: }
119:
120:
1.8 schwarze 121: /*
122: * Frees volatile resources (parse tree, meta-data, fields).
123: */
1.1 kristaps 124: static void
125: mdoc_free1(struct mdoc *mdoc)
126: {
127:
128: if (mdoc->first)
1.40 schwarze 129: mdoc_node_delete(mdoc, mdoc->first);
1.1 kristaps 130: if (mdoc->meta.title)
131: free(mdoc->meta.title);
132: if (mdoc->meta.os)
133: free(mdoc->meta.os);
134: if (mdoc->meta.name)
135: free(mdoc->meta.name);
136: if (mdoc->meta.arch)
137: free(mdoc->meta.arch);
138: if (mdoc->meta.vol)
139: free(mdoc->meta.vol);
1.50 schwarze 140: if (mdoc->meta.msec)
141: free(mdoc->meta.msec);
1.82 schwarze 142: if (mdoc->meta.date)
143: free(mdoc->meta.date);
1.1 kristaps 144: }
145:
146:
1.8 schwarze 147: /*
148: * Allocate all volatile resources (parse tree, meta-data, fields).
149: */
1.32 schwarze 150: static void
1.1 kristaps 151: mdoc_alloc1(struct mdoc *mdoc)
152: {
153:
1.32 schwarze 154: memset(&mdoc->meta, 0, sizeof(struct mdoc_meta));
1.1 kristaps 155: mdoc->flags = 0;
1.9 schwarze 156: mdoc->lastnamed = mdoc->lastsec = SEC_NONE;
1.32 schwarze 157: mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node));
1.1 kristaps 158: mdoc->first = mdoc->last;
159: mdoc->last->type = MDOC_ROOT;
1.86 schwarze 160: mdoc->last->tok = MDOC_MAX;
1.1 kristaps 161: mdoc->next = MDOC_NEXT_CHILD;
162: }
163:
164:
165: /*
1.8 schwarze 166: * Free up volatile resources (see mdoc_free1()) then re-initialises the
167: * data with mdoc_alloc1(). After invocation, parse data has been reset
168: * and the parser is ready for re-invocation on a new tree; however,
169: * cross-parse non-volatile data is kept intact.
1.1 kristaps 170: */
1.32 schwarze 171: void
1.1 kristaps 172: mdoc_reset(struct mdoc *mdoc)
173: {
174:
175: mdoc_free1(mdoc);
1.32 schwarze 176: mdoc_alloc1(mdoc);
1.1 kristaps 177: }
178:
179:
180: /*
1.8 schwarze 181: * Completely free up all volatile and non-volatile parse resources.
182: * After invocation, the pointer is no longer usable.
1.1 kristaps 183: */
184: void
185: mdoc_free(struct mdoc *mdoc)
186: {
187:
188: mdoc_free1(mdoc);
189: free(mdoc);
190: }
191:
192:
1.8 schwarze 193: /*
194: * Allocate volatile and non-volatile parse resources.
195: */
1.1 kristaps 196: struct mdoc *
1.98 schwarze 197: mdoc_alloc(struct roff *roff, struct mparse *parse,
198: char *defos, int quick)
1.1 kristaps 199: {
200: struct mdoc *p;
201:
1.32 schwarze 202: p = mandoc_calloc(1, sizeof(struct mdoc));
203:
1.83 schwarze 204: p->parse = parse;
1.87 schwarze 205: p->defos = defos;
1.98 schwarze 206: p->quick = quick;
1.84 schwarze 207: p->roff = roff;
1.1 kristaps 208:
1.32 schwarze 209: mdoc_hash_init();
210: mdoc_alloc1(p);
211: return(p);
1.1 kristaps 212: }
213:
214:
215: /*
216: * Climb back up the parse tree, validating open scopes. Mostly calls
1.8 schwarze 217: * through to macro_end() in macro.c.
1.1 kristaps 218: */
219: int
1.93 schwarze 220: mdoc_endparse(struct mdoc *mdoc)
1.1 kristaps 221: {
222:
1.93 schwarze 223: assert( ! (MDOC_HALT & mdoc->flags));
224: if (mdoc_macroend(mdoc))
1.1 kristaps 225: return(1);
1.93 schwarze 226: mdoc->flags |= MDOC_HALT;
1.1 kristaps 227: return(0);
1.81 schwarze 228: }
229:
230: int
1.93 schwarze 231: mdoc_addeqn(struct mdoc *mdoc, const struct eqn *ep)
1.81 schwarze 232: {
233: struct mdoc_node *n;
234:
1.93 schwarze 235: assert( ! (MDOC_HALT & mdoc->flags));
1.81 schwarze 236:
237: /* No text before an initial macro. */
238:
1.93 schwarze 239: if (SEC_NONE == mdoc->lastnamed) {
240: mdoc_pmsg(mdoc, ep->ln, ep->pos, MANDOCERR_NOTEXT);
1.81 schwarze 241: return(1);
242: }
243:
1.93 schwarze 244: n = node_alloc(mdoc, ep->ln, ep->pos, MDOC_MAX, MDOC_EQN);
1.81 schwarze 245: n->eqn = ep;
246:
1.93 schwarze 247: if ( ! node_append(mdoc, n))
1.81 schwarze 248: return(0);
249:
1.93 schwarze 250: mdoc->next = MDOC_NEXT_SIBLING;
1.81 schwarze 251: return(1);
1.1 kristaps 252: }
253:
1.77 schwarze 254: int
1.93 schwarze 255: mdoc_addspan(struct mdoc *mdoc, const struct tbl_span *sp)
1.77 schwarze 256: {
1.79 schwarze 257: struct mdoc_node *n;
1.77 schwarze 258:
1.93 schwarze 259: assert( ! (MDOC_HALT & mdoc->flags));
1.77 schwarze 260:
261: /* No text before an initial macro. */
262:
1.93 schwarze 263: if (SEC_NONE == mdoc->lastnamed) {
264: mdoc_pmsg(mdoc, sp->line, 0, MANDOCERR_NOTEXT);
1.77 schwarze 265: return(1);
266: }
267:
1.93 schwarze 268: n = node_alloc(mdoc, sp->line, 0, MDOC_MAX, MDOC_TBL);
1.79 schwarze 269: n->span = sp;
270:
1.93 schwarze 271: if ( ! node_append(mdoc, n))
1.79 schwarze 272: return(0);
273:
1.93 schwarze 274: mdoc->next = MDOC_NEXT_SIBLING;
1.79 schwarze 275: return(1);
1.77 schwarze 276: }
277:
1.1 kristaps 278:
279: /*
280: * Main parse routine. Parses a single line -- really just hands off to
1.45 schwarze 281: * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
1.1 kristaps 282: */
283: int
1.93 schwarze 284: mdoc_parseln(struct mdoc *mdoc, int ln, char *buf, int offs)
1.1 kristaps 285: {
286:
1.93 schwarze 287: assert( ! (MDOC_HALT & mdoc->flags));
1.1 kristaps 288:
1.93 schwarze 289: mdoc->flags |= MDOC_NEWLINE;
1.60 schwarze 290:
291: /*
292: * Let the roff nS register switch SYNOPSIS mode early,
293: * such that the parser knows at all times
294: * whether this mode is on or off.
295: * Note that this mode is also switched by the Sh macro.
296: */
1.94 schwarze 297: if (roff_getreg(mdoc->roff, "nS"))
298: mdoc->flags |= MDOC_SYNOPSIS;
299: else
300: mdoc->flags &= ~MDOC_SYNOPSIS;
1.60 schwarze 301:
1.93 schwarze 302: return(roff_getcontrol(mdoc->roff, buf, &offs) ?
303: mdoc_pmacro(mdoc, ln, buf, offs) :
304: mdoc_ptext(mdoc, ln, buf, offs));
1.1 kristaps 305: }
306:
1.5 schwarze 307: int
1.58 schwarze 308: mdoc_macro(MACRO_PROT_ARGS)
1.1 kristaps 309: {
1.46 schwarze 310: assert(tok < MDOC_MAX);
311:
312: /* If we're in the body, deny prologue calls. */
1.37 schwarze 313:
1.1 kristaps 314: if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
1.93 schwarze 315: MDOC_PBODY & mdoc->flags) {
316: mdoc_pmsg(mdoc, line, ppos, MANDOCERR_BADBODY);
1.76 schwarze 317: return(1);
318: }
1.46 schwarze 319:
320: /* If we're in the prologue, deny "body" macros. */
321:
1.1 kristaps 322: if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
1.93 schwarze 323: ! (MDOC_PBODY & mdoc->flags)) {
324: mdoc_pmsg(mdoc, line, ppos, MANDOCERR_BADPROLOG);
325: if (NULL == mdoc->meta.msec)
326: mdoc->meta.msec = mandoc_strdup("1");
327: if (NULL == mdoc->meta.title)
328: mdoc->meta.title = mandoc_strdup("UNKNOWN");
329: if (NULL == mdoc->meta.vol)
330: mdoc->meta.vol = mandoc_strdup("LOCAL");
331: if (NULL == mdoc->meta.os)
332: mdoc->meta.os = mandoc_strdup("LOCAL");
333: if (NULL == mdoc->meta.date)
334: mdoc->meta.date = mandoc_normdate
335: (mdoc->parse, NULL, line, ppos);
336: mdoc->flags |= MDOC_PBODY;
1.39 schwarze 337: }
1.1 kristaps 338:
1.93 schwarze 339: return((*mdoc_macros[tok].fp)(mdoc, tok, line, ppos, pos, buf));
1.1 kristaps 340: }
341:
342:
343: static int
344: node_append(struct mdoc *mdoc, struct mdoc_node *p)
345: {
346:
347: assert(mdoc->last);
348: assert(mdoc->first);
349: assert(MDOC_ROOT != p->type);
350:
351: switch (mdoc->next) {
352: case (MDOC_NEXT_SIBLING):
353: mdoc->last->next = p;
354: p->prev = mdoc->last;
355: p->parent = mdoc->last->parent;
356: break;
357: case (MDOC_NEXT_CHILD):
358: mdoc->last->child = p;
359: p->parent = mdoc->last;
360: break;
361: default:
362: abort();
363: /* NOTREACHED */
364: }
365:
1.10 schwarze 366: p->parent->nchild++;
367:
1.75 schwarze 368: /*
369: * Copy over the normalised-data pointer of our parent. Not
370: * everybody has one, but copying a null pointer is fine.
371: */
372:
373: switch (p->type) {
374: case (MDOC_BODY):
1.92 schwarze 375: if (ENDBODY_NOT != p->end)
376: break;
1.75 schwarze 377: /* FALLTHROUGH */
378: case (MDOC_TAIL):
379: /* FALLTHROUGH */
380: case (MDOC_HEAD):
381: p->norm = p->parent->norm;
382: break;
383: default:
384: break;
385: }
386:
1.1 kristaps 387: if ( ! mdoc_valid_pre(mdoc, p))
388: return(0);
389:
390: switch (p->type) {
391: case (MDOC_HEAD):
392: assert(MDOC_BLOCK == p->parent->type);
393: p->parent->head = p;
394: break;
395: case (MDOC_TAIL):
396: assert(MDOC_BLOCK == p->parent->type);
397: p->parent->tail = p;
398: break;
399: case (MDOC_BODY):
1.59 schwarze 400: if (p->end)
401: break;
1.1 kristaps 402: assert(MDOC_BLOCK == p->parent->type);
403: p->parent->body = p;
404: break;
405: default:
406: break;
407: }
408:
409: mdoc->last = p;
410:
411: switch (p->type) {
1.77 schwarze 412: case (MDOC_TBL):
413: /* FALLTHROUGH */
1.1 kristaps 414: case (MDOC_TEXT):
415: if ( ! mdoc_valid_post(mdoc))
416: return(0);
417: break;
418: default:
419: break;
420: }
421:
422: return(1);
423: }
424:
425:
426: static struct mdoc_node *
1.93 schwarze 427: node_alloc(struct mdoc *mdoc, int line, int pos,
1.37 schwarze 428: enum mdoct tok, enum mdoc_type type)
1.1 kristaps 429: {
430: struct mdoc_node *p;
431:
1.32 schwarze 432: p = mandoc_calloc(1, sizeof(struct mdoc_node));
1.93 schwarze 433: p->sec = mdoc->lastsec;
1.1 kristaps 434: p->line = line;
435: p->pos = pos;
1.96 schwarze 436: p->lastline = line;
1.1 kristaps 437: p->tok = tok;
1.37 schwarze 438: p->type = type;
1.58 schwarze 439:
440: /* Flag analysis. */
441:
1.93 schwarze 442: if (MDOC_SYNOPSIS & mdoc->flags)
1.60 schwarze 443: p->flags |= MDOC_SYNPRETTY;
444: else
445: p->flags &= ~MDOC_SYNPRETTY;
1.93 schwarze 446: if (MDOC_NEWLINE & mdoc->flags)
1.47 schwarze 447: p->flags |= MDOC_LINE;
1.93 schwarze 448: mdoc->flags &= ~MDOC_NEWLINE;
1.58 schwarze 449:
1.1 kristaps 450: return(p);
451: }
452:
453:
454: int
1.93 schwarze 455: mdoc_tail_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.1 kristaps 456: {
457: struct mdoc_node *p;
458:
1.93 schwarze 459: p = node_alloc(mdoc, line, pos, tok, MDOC_TAIL);
460: if ( ! node_append(mdoc, p))
1.25 schwarze 461: return(0);
1.93 schwarze 462: mdoc->next = MDOC_NEXT_CHILD;
1.25 schwarze 463: return(1);
1.1 kristaps 464: }
465:
466:
467: int
1.93 schwarze 468: mdoc_head_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.1 kristaps 469: {
470: struct mdoc_node *p;
471:
1.93 schwarze 472: assert(mdoc->first);
473: assert(mdoc->last);
1.1 kristaps 474:
1.93 schwarze 475: p = node_alloc(mdoc, line, pos, tok, MDOC_HEAD);
476: if ( ! node_append(mdoc, p))
1.25 schwarze 477: return(0);
1.93 schwarze 478: mdoc->next = MDOC_NEXT_CHILD;
1.25 schwarze 479: return(1);
1.1 kristaps 480: }
481:
482:
483: int
1.93 schwarze 484: mdoc_body_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok)
1.1 kristaps 485: {
486: struct mdoc_node *p;
487:
1.93 schwarze 488: p = node_alloc(mdoc, line, pos, tok, MDOC_BODY);
489: if ( ! node_append(mdoc, p))
1.25 schwarze 490: return(0);
1.93 schwarze 491: mdoc->next = MDOC_NEXT_CHILD;
1.59 schwarze 492: return(1);
493: }
494:
495:
496: int
1.93 schwarze 497: mdoc_endbody_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok,
1.59 schwarze 498: struct mdoc_node *body, enum mdoc_endbody end)
499: {
500: struct mdoc_node *p;
501:
1.93 schwarze 502: p = node_alloc(mdoc, line, pos, tok, MDOC_BODY);
1.59 schwarze 503: p->pending = body;
1.92 schwarze 504: p->norm = body->norm;
1.59 schwarze 505: p->end = end;
1.93 schwarze 506: if ( ! node_append(mdoc, p))
1.59 schwarze 507: return(0);
1.93 schwarze 508: mdoc->next = MDOC_NEXT_SIBLING;
1.25 schwarze 509: return(1);
1.1 kristaps 510: }
511:
512:
513: int
1.93 schwarze 514: mdoc_block_alloc(struct mdoc *mdoc, int line, int pos,
1.37 schwarze 515: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 516: {
517: struct mdoc_node *p;
518:
1.93 schwarze 519: p = node_alloc(mdoc, line, pos, tok, MDOC_BLOCK);
1.4 schwarze 520: p->args = args;
521: if (p->args)
1.1 kristaps 522: (args->refcnt)++;
1.75 schwarze 523:
524: switch (tok) {
525: case (MDOC_Bd):
526: /* FALLTHROUGH */
527: case (MDOC_Bf):
528: /* FALLTHROUGH */
529: case (MDOC_Bl):
530: /* FALLTHROUGH */
531: case (MDOC_Rs):
532: p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
533: break;
534: default:
535: break;
536: }
537:
1.93 schwarze 538: if ( ! node_append(mdoc, p))
1.25 schwarze 539: return(0);
1.93 schwarze 540: mdoc->next = MDOC_NEXT_CHILD;
1.25 schwarze 541: return(1);
1.1 kristaps 542: }
543:
544:
545: int
1.93 schwarze 546: mdoc_elem_alloc(struct mdoc *mdoc, int line, int pos,
1.37 schwarze 547: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 548: {
549: struct mdoc_node *p;
550:
1.93 schwarze 551: p = node_alloc(mdoc, line, pos, tok, MDOC_ELEM);
1.4 schwarze 552: p->args = args;
553: if (p->args)
1.1 kristaps 554: (args->refcnt)++;
1.75 schwarze 555:
556: switch (tok) {
557: case (MDOC_An):
558: p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
559: break;
560: default:
561: break;
562: }
563:
1.93 schwarze 564: if ( ! node_append(mdoc, p))
1.25 schwarze 565: return(0);
1.93 schwarze 566: mdoc->next = MDOC_NEXT_CHILD;
1.25 schwarze 567: return(1);
1.1 kristaps 568: }
569:
1.46 schwarze 570: int
1.93 schwarze 571: mdoc_word_alloc(struct mdoc *mdoc, int line, int pos, const char *p)
1.1 kristaps 572: {
1.19 schwarze 573: struct mdoc_node *n;
1.1 kristaps 574:
1.93 schwarze 575: n = node_alloc(mdoc, line, pos, MDOC_MAX, MDOC_TEXT);
576: n->string = roff_strdup(mdoc->roff, p);
1.19 schwarze 577:
1.93 schwarze 578: if ( ! node_append(mdoc, n))
1.25 schwarze 579: return(0);
1.46 schwarze 580:
1.93 schwarze 581: mdoc->next = MDOC_NEXT_SIBLING;
1.25 schwarze 582: return(1);
1.19 schwarze 583: }
584:
1.95 schwarze 585: void
586: mdoc_word_append(struct mdoc *mdoc, const char *p)
587: {
588: struct mdoc_node *n;
589: char *addstr, *newstr;
590:
591: n = mdoc->last;
592: addstr = roff_strdup(mdoc->roff, p);
593: if (-1 == asprintf(&newstr, "%s %s", n->string, addstr)) {
594: perror(NULL);
595: exit((int)MANDOCLEVEL_SYSERR);
596: }
597: free(addstr);
598: free(n->string);
599: n->string = newstr;
600: mdoc->next = MDOC_NEXT_SIBLING;
601: }
1.19 schwarze 602:
1.61 schwarze 603: static void
1.1 kristaps 604: mdoc_node_free(struct mdoc_node *p)
605: {
1.61 schwarze 606:
1.75 schwarze 607: if (MDOC_BLOCK == p->type || MDOC_ELEM == p->type)
608: free(p->norm);
1.1 kristaps 609: if (p->string)
610: free(p->string);
611: if (p->args)
612: mdoc_argv_free(p->args);
613: free(p);
614: }
615:
616:
1.40 schwarze 617: static void
1.93 schwarze 618: mdoc_node_unlink(struct mdoc *mdoc, struct mdoc_node *n)
1.40 schwarze 619: {
620:
621: /* Adjust siblings. */
622:
623: if (n->prev)
624: n->prev->next = n->next;
625: if (n->next)
626: n->next->prev = n->prev;
627:
628: /* Adjust parent. */
629:
630: if (n->parent) {
631: n->parent->nchild--;
632: if (n->parent->child == n)
633: n->parent->child = n->prev ? n->prev : n->next;
1.72 schwarze 634: if (n->parent->last == n)
635: n->parent->last = n->prev ? n->prev : NULL;
1.40 schwarze 636: }
637:
638: /* Adjust parse point, if applicable. */
639:
1.93 schwarze 640: if (mdoc && mdoc->last == n) {
1.40 schwarze 641: if (n->prev) {
1.93 schwarze 642: mdoc->last = n->prev;
643: mdoc->next = MDOC_NEXT_SIBLING;
1.40 schwarze 644: } else {
1.93 schwarze 645: mdoc->last = n->parent;
646: mdoc->next = MDOC_NEXT_CHILD;
1.40 schwarze 647: }
648: }
649:
1.93 schwarze 650: if (mdoc && mdoc->first == n)
651: mdoc->first = NULL;
1.40 schwarze 652: }
653:
654:
1.1 kristaps 655: void
1.93 schwarze 656: mdoc_node_delete(struct mdoc *mdoc, struct mdoc_node *p)
1.1 kristaps 657: {
658:
1.40 schwarze 659: while (p->child) {
660: assert(p->nchild);
1.93 schwarze 661: mdoc_node_delete(mdoc, p->child);
1.40 schwarze 662: }
663: assert(0 == p->nchild);
1.1 kristaps 664:
1.93 schwarze 665: mdoc_node_unlink(mdoc, p);
1.1 kristaps 666: mdoc_node_free(p);
1.91 schwarze 667: }
668:
669: int
1.93 schwarze 670: mdoc_node_relink(struct mdoc *mdoc, struct mdoc_node *p)
1.91 schwarze 671: {
672:
1.93 schwarze 673: mdoc_node_unlink(mdoc, p);
674: return(node_append(mdoc, p));
1.1 kristaps 675: }
676:
1.84 schwarze 677: #if 0
678: /*
679: * Pre-treat a text line.
680: * Text lines can consist of equations, which must be handled apart from
681: * the regular text.
682: * Thus, use this function to step through a line checking if it has any
683: * equations embedded in it.
684: * This must handle multiple equations AND equations that do not end at
685: * the end-of-line, i.e., will re-enter in the next roff parse.
686: */
687: static int
1.93 schwarze 688: mdoc_preptext(struct mdoc *mdoc, int line, char *buf, int offs)
1.84 schwarze 689: {
690: char *start, *end;
691: char delim;
692:
693: while ('\0' != buf[offs]) {
694: /* Mark starting position if eqn is set. */
695: start = NULL;
1.93 schwarze 696: if ('\0' != (delim = roff_eqndelim(mdoc->roff)))
1.84 schwarze 697: if (NULL != (start = strchr(buf + offs, delim)))
698: *start++ = '\0';
699:
700: /* Parse text as normal. */
1.93 schwarze 701: if ( ! mdoc_ptext(mdoc, line, buf, offs))
1.84 schwarze 702: return(0);
703:
704: /* Continue only if an equation exists. */
705: if (NULL == start)
706: break;
707:
708: /* Read past the end of the equation. */
709: offs += start - (buf + offs);
710: assert(start == &buf[offs]);
711: if (NULL != (end = strchr(buf + offs, delim))) {
712: *end++ = '\0';
713: while (' ' == *end)
714: end++;
715: }
716:
717: /* Parse the equation itself. */
1.93 schwarze 718: roff_openeqn(mdoc->roff, NULL, line, offs, buf);
1.84 schwarze 719:
720: /* Process a finished equation? */
1.93 schwarze 721: if (roff_closeeqn(mdoc->roff))
722: if ( ! mdoc_addeqn(mdoc, roff_eqn(mdoc->roff)))
1.84 schwarze 723: return(0);
724: offs += (end - (buf + offs));
725: }
726:
727: return(1);
728: }
729: #endif
1.1 kristaps 730:
731: /*
732: * Parse free-form text, that is, a line that does not begin with the
733: * control character.
734: */
735: static int
1.93 schwarze 736: mdoc_ptext(struct mdoc *mdoc, int line, char *buf, int offs)
1.1 kristaps 737: {
1.56 schwarze 738: char *c, *ws, *end;
739: struct mdoc_node *n;
1.44 schwarze 740:
1.46 schwarze 741: /* No text before an initial macro. */
742:
1.93 schwarze 743: if (SEC_NONE == mdoc->lastnamed) {
744: mdoc_pmsg(mdoc, line, offs, MANDOCERR_NOTEXT);
1.76 schwarze 745: return(1);
746: }
1.46 schwarze 747:
1.93 schwarze 748: assert(mdoc->last);
749: n = mdoc->last;
1.56 schwarze 750:
751: /*
752: * Divert directly to list processing if we're encountering a
753: * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
754: * (a MDOC_BODY means it's already open, in which case we should
755: * process within its context in the normal way).
756: */
757:
758: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.75 schwarze 759: LIST_column == n->norm->Bl.type) {
1.56 schwarze 760: /* `Bl' is open without any children. */
1.93 schwarze 761: mdoc->flags |= MDOC_FREECOL;
762: return(mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf));
1.56 schwarze 763: }
764:
765: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
766: NULL != n->parent &&
767: MDOC_Bl == n->parent->tok &&
1.75 schwarze 768: LIST_column == n->parent->norm->Bl.type) {
1.56 schwarze 769: /* `Bl' has block-level `It' children. */
1.93 schwarze 770: mdoc->flags |= MDOC_FREECOL;
771: return(mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf));
1.56 schwarze 772: }
773:
1.52 schwarze 774: /*
775: * Search for the beginning of unescaped trailing whitespace (ws)
776: * and for the first character not to be output (end).
777: */
1.56 schwarze 778:
779: /* FIXME: replace with strcspn(). */
1.52 schwarze 780: ws = NULL;
1.53 schwarze 781: for (c = end = buf + offs; *c; c++) {
1.52 schwarze 782: switch (*c) {
783: case ' ':
784: if (NULL == ws)
785: ws = c;
786: continue;
787: case '\t':
788: /*
789: * Always warn about trailing tabs,
790: * even outside literal context,
791: * where they should be put on the next line.
792: */
793: if (NULL == ws)
794: ws = c;
795: /*
796: * Strip trailing tabs in literal context only;
797: * outside, they affect the next line.
798: */
1.93 schwarze 799: if (MDOC_LITERAL & mdoc->flags)
1.52 schwarze 800: continue;
801: break;
802: case '\\':
803: /* Skip the escaped character, too, if any. */
804: if (c[1])
805: c++;
806: /* FALLTHROUGH */
807: default:
808: ws = NULL;
809: break;
810: }
811: end = c + 1;
812: }
813: *end = '\0';
1.19 schwarze 814:
1.52 schwarze 815: if (ws)
1.93 schwarze 816: mdoc_pmsg(mdoc, line, (int)(ws-buf), MANDOCERR_EOLNSPACE);
1.34 schwarze 817:
1.93 schwarze 818: if ('\0' == buf[offs] && ! (MDOC_LITERAL & mdoc->flags)) {
819: mdoc_pmsg(mdoc, line, (int)(c-buf), MANDOCERR_NOBLANKLN);
1.46 schwarze 820:
1.40 schwarze 821: /*
1.66 schwarze 822: * Insert a `sp' in the case of a blank line. Technically,
1.46 schwarze 823: * blank lines aren't allowed, but enough manuals assume this
824: * behaviour that we want to work around it.
1.40 schwarze 825: */
1.93 schwarze 826: if ( ! mdoc_elem_alloc(mdoc, line, offs, MDOC_sp, NULL))
1.42 schwarze 827: return(0);
1.46 schwarze 828:
1.93 schwarze 829: mdoc->next = MDOC_NEXT_SIBLING;
1.89 schwarze 830:
1.93 schwarze 831: return(mdoc_valid_post(mdoc));
1.38 schwarze 832: }
1.1 kristaps 833:
1.93 schwarze 834: if ( ! mdoc_word_alloc(mdoc, line, offs, buf+offs))
1.1 kristaps 835: return(0);
836:
1.93 schwarze 837: if (MDOC_LITERAL & mdoc->flags)
1.52 schwarze 838: return(1);
839:
1.35 schwarze 840: /*
1.48 schwarze 841: * End-of-sentence check. If the last character is an unescaped
842: * EOS character, then flag the node as being the end of a
843: * sentence. The front-end will know how to interpret this.
1.35 schwarze 844: */
845:
1.52 schwarze 846: assert(buf < end);
1.48 schwarze 847:
1.97 schwarze 848: if (mandoc_eos(buf+offs, (size_t)(end-buf-offs)))
1.93 schwarze 849: mdoc->last->flags |= MDOC_EOS;
1.48 schwarze 850:
1.1 kristaps 851: return(1);
852: }
1.19 schwarze 853:
1.1 kristaps 854:
855: /*
856: * Parse a macro line, that is, a line beginning with the control
857: * character.
858: */
1.61 schwarze 859: static int
1.93 schwarze 860: mdoc_pmacro(struct mdoc *mdoc, int ln, char *buf, int offs)
1.1 kristaps 861: {
1.56 schwarze 862: enum mdoct tok;
1.83 schwarze 863: int i, sv;
1.56 schwarze 864: char mac[5];
865: struct mdoc_node *n;
1.1 kristaps 866:
1.83 schwarze 867: /* Empty post-control lines are ignored. */
1.1 kristaps 868:
1.83 schwarze 869: if ('"' == buf[offs]) {
1.93 schwarze 870: mdoc_pmsg(mdoc, ln, offs, MANDOCERR_BADCOMMENT);
1.83 schwarze 871: return(1);
872: } else if ('\0' == buf[offs])
1.1 kristaps 873: return(1);
874:
1.83 schwarze 875: sv = offs;
1.47 schwarze 876:
1.63 schwarze 877: /*
1.64 schwarze 878: * Copy the first word into a nil-terminated buffer.
879: * Stop copying when a tab, space, or eoln is encountered.
1.63 schwarze 880: */
1.1 kristaps 881:
1.83 schwarze 882: i = 0;
883: while (i < 4 && '\0' != buf[offs] &&
884: ' ' != buf[offs] && '\t' != buf[offs])
885: mac[i++] = buf[offs++];
886:
887: mac[i] = '\0';
888:
889: tok = (i > 1 || i < 4) ? mdoc_hash_find(mac) : MDOC_MAX;
1.1 kristaps 890:
1.65 schwarze 891: if (MDOC_MAX == tok) {
1.93 schwarze 892: mandoc_vmsg(MANDOCERR_MACRO, mdoc->parse,
1.83 schwarze 893: ln, sv, "%s", buf + sv - 1);
1.1 kristaps 894: return(1);
895: }
896:
1.63 schwarze 897: /* Disregard the first trailing tab, if applicable. */
898:
1.83 schwarze 899: if ('\t' == buf[offs])
900: offs++;
1.63 schwarze 901:
902: /* Jump to the next non-whitespace word. */
1.1 kristaps 903:
1.83 schwarze 904: while (buf[offs] && ' ' == buf[offs])
905: offs++;
1.34 schwarze 906:
1.46 schwarze 907: /*
908: * Trailing whitespace. Note that tabs are allowed to be passed
909: * into the parser as "text", so we only warn about spaces here.
910: */
1.34 schwarze 911:
1.83 schwarze 912: if ('\0' == buf[offs] && ' ' == buf[offs - 1])
1.93 schwarze 913: mdoc_pmsg(mdoc, ln, offs - 1, MANDOCERR_EOLNSPACE);
1.1 kristaps 914:
1.56 schwarze 915: /*
916: * If an initial macro or a list invocation, divert directly
917: * into macro processing.
918: */
919:
1.93 schwarze 920: if (NULL == mdoc->last || MDOC_It == tok || MDOC_El == tok) {
921: if ( ! mdoc_macro(mdoc, tok, ln, sv, &offs, buf))
1.56 schwarze 922: goto err;
923: return(1);
924: }
925:
1.93 schwarze 926: n = mdoc->last;
927: assert(mdoc->last);
1.56 schwarze 928:
929: /*
930: * If the first macro of a `Bl -column', open an `It' block
931: * context around the parsed macro.
932: */
933:
934: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.75 schwarze 935: LIST_column == n->norm->Bl.type) {
1.93 schwarze 936: mdoc->flags |= MDOC_FREECOL;
937: if ( ! mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf))
1.56 schwarze 938: goto err;
939: return(1);
940: }
941:
942: /*
943: * If we're following a block-level `It' within a `Bl -column'
944: * context (perhaps opened in the above block or in ptext()),
945: * then open an `It' block context around the parsed macro.
1.24 schwarze 946: */
1.56 schwarze 947:
948: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
949: NULL != n->parent &&
950: MDOC_Bl == n->parent->tok &&
1.75 schwarze 951: LIST_column == n->parent->norm->Bl.type) {
1.93 schwarze 952: mdoc->flags |= MDOC_FREECOL;
953: if ( ! mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf))
1.56 schwarze 954: goto err;
955: return(1);
956: }
957:
958: /* Normal processing of a macro. */
959:
1.93 schwarze 960: if ( ! mdoc_macro(mdoc, tok, ln, sv, &offs, buf))
1.1 kristaps 961: goto err;
1.98 schwarze 962:
963: /* In quick mode (for mandocdb), abort after the NAME section. */
964:
965: if (mdoc->quick && MDOC_Sh == tok &&
966: SEC_NAME != mdoc->last->sec)
967: return(2);
1.1 kristaps 968:
969: return(1);
970:
971: err: /* Error out. */
972:
1.93 schwarze 973: mdoc->flags |= MDOC_HALT;
1.1 kristaps 974: return(0);
975: }
1.24 schwarze 976:
1.83 schwarze 977: enum mdelim
978: mdoc_isdelim(const char *p)
979: {
980:
981: if ('\0' == p[0])
982: return(DELIM_NONE);
983:
984: if ('\0' == p[1])
985: switch (p[0]) {
986: case('('):
987: /* FALLTHROUGH */
988: case('['):
989: return(DELIM_OPEN);
990: case('|'):
991: return(DELIM_MIDDLE);
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: /* FALLTHROUGH */
1006: case(']'):
1007: return(DELIM_CLOSE);
1008: default:
1009: return(DELIM_NONE);
1010: }
1011:
1012: if ('\\' != p[0])
1013: return(DELIM_NONE);
1.24 schwarze 1014:
1.83 schwarze 1015: if (0 == strcmp(p + 1, "."))
1016: return(DELIM_CLOSE);
1.90 schwarze 1017: if (0 == strcmp(p + 1, "fR|\\fP"))
1.83 schwarze 1018: return(DELIM_MIDDLE);
1019:
1020: return(DELIM_NONE);
1021: }
![[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