Annotation of mandoc/mdoc.c, Revision 1.171
1.171 ! kristaps 1: /* $Id: mdoc.c,v 1.170 2010/12/16 17:14:48 kristaps Exp $ */
1.1 kristaps 2: /*
1.157 schwarze 3: * Copyright (c) 2008, 2009, 2010 Kristaps Dzonsons <kristaps@bsd.lv>
4: * Copyright (c) 2010 Ingo Schwarze <schwarze@openbsd.org>
1.1 kristaps 5: *
6: * Permission to use, copy, modify, and distribute this software for any
1.75 kristaps 7: * purpose with or without fee is hereby granted, provided that the above
8: * copyright notice and this permission notice appear in all copies.
1.1 kristaps 9: *
1.75 kristaps 10: * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
11: * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
12: * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
13: * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
14: * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
15: * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
16: * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
1.1 kristaps 17: */
1.114 kristaps 18: #ifdef HAVE_CONFIG_H
19: #include "config.h"
20: #endif
21:
1.106 kristaps 22: #include <sys/types.h>
23:
1.1 kristaps 24: #include <assert.h>
25: #include <stdarg.h>
1.73 kristaps 26: #include <stdio.h>
1.1 kristaps 27: #include <stdlib.h>
28: #include <string.h>
1.120 kristaps 29: #include <time.h>
1.1 kristaps 30:
1.136 kristaps 31: #include "mandoc.h"
1.70 kristaps 32: #include "libmdoc.h"
1.113 kristaps 33: #include "libmandoc.h"
1.1 kristaps 34:
35: const char *const __mdoc_macronames[MDOC_MAX] = {
1.82 kristaps 36: "Ap", "Dd", "Dt", "Os",
1.1 kristaps 37: "Sh", "Ss", "Pp", "D1",
38: "Dl", "Bd", "Ed", "Bl",
39: "El", "It", "Ad", "An",
40: "Ar", "Cd", "Cm", "Dv",
41: "Er", "Ev", "Ex", "Fa",
42: "Fd", "Fl", "Fn", "Ft",
43: "Ic", "In", "Li", "Nd",
44: "Nm", "Op", "Ot", "Pa",
45: "Rv", "St", "Va", "Vt",
46: /* LINTED */
1.114 kristaps 47: "Xr", "%A", "%B", "%D",
1.1 kristaps 48: /* LINTED */
1.114 kristaps 49: "%I", "%J", "%N", "%O",
1.1 kristaps 50: /* LINTED */
1.114 kristaps 51: "%P", "%R", "%T", "%V",
1.1 kristaps 52: "Ac", "Ao", "Aq", "At",
53: "Bc", "Bf", "Bo", "Bq",
54: "Bsx", "Bx", "Db", "Dc",
55: "Do", "Dq", "Ec", "Ef",
56: "Em", "Eo", "Fx", "Ms",
57: "No", "Ns", "Nx", "Ox",
58: "Pc", "Pf", "Po", "Pq",
59: "Qc", "Ql", "Qo", "Qq",
60: "Re", "Rs", "Sc", "So",
61: "Sq", "Sm", "Sx", "Sy",
62: "Tn", "Ux", "Xc", "Xo",
63: "Fo", "Fc", "Oo", "Oc",
64: "Bk", "Ek", "Bt", "Hf",
1.82 kristaps 65: "Fr", "Ud", "Lb", "Lp",
66: "Lk", "Mt", "Brq", "Bro",
1.64 kristaps 67: /* LINTED */
1.114 kristaps 68: "Brc", "%C", "Es", "En",
1.69 kristaps 69: /* LINTED */
1.114 kristaps 70: "Dx", "%Q", "br", "sp",
1.110 kristaps 71: /* LINTED */
1.141 kristaps 72: "%U", "Ta"
1.1 kristaps 73: };
74:
75: const char *const __mdoc_argnames[MDOC_ARG_MAX] = {
76: "split", "nosplit", "ragged",
77: "unfilled", "literal", "file",
78: "offset", "bullet", "dash",
79: "hyphen", "item", "enum",
80: "tag", "diag", "hang",
81: "ohang", "inset", "column",
82: "width", "compact", "std",
1.52 kristaps 83: "filled", "words", "emphasis",
1.108 kristaps 84: "symbolic", "nested", "centered"
1.1 kristaps 85: };
86:
87: const char * const *mdoc_macronames = __mdoc_macronames;
88: const char * const *mdoc_argnames = __mdoc_argnames;
89:
1.121 kristaps 90: static void mdoc_node_free(struct mdoc_node *);
91: static void mdoc_node_unlink(struct mdoc *,
92: struct mdoc_node *);
1.73 kristaps 93: static void mdoc_free1(struct mdoc *);
1.113 kristaps 94: static void mdoc_alloc1(struct mdoc *);
1.73 kristaps 95: static struct mdoc_node *node_alloc(struct mdoc *, int, int,
1.117 kristaps 96: enum mdoct, enum mdoc_type);
1.73 kristaps 97: static int node_append(struct mdoc *,
1.71 kristaps 98: struct mdoc_node *);
1.149 kristaps 99: static int mdoc_ptext(struct mdoc *, int, char *, int);
100: static int mdoc_pmacro(struct mdoc *, int, char *, int);
1.124 kristaps 101:
1.88 kristaps 102:
1.1 kristaps 103: const struct mdoc_node *
1.71 kristaps 104: mdoc_node(const struct mdoc *m)
1.1 kristaps 105: {
106:
1.71 kristaps 107: return(MDOC_HALT & m->flags ? NULL : m->first);
1.1 kristaps 108: }
109:
110:
1.37 kristaps 111: const struct mdoc_meta *
1.71 kristaps 112: mdoc_meta(const struct mdoc *m)
1.37 kristaps 113: {
114:
1.71 kristaps 115: return(MDOC_HALT & m->flags ? NULL : &m->meta);
1.37 kristaps 116: }
117:
118:
1.85 kristaps 119: /*
120: * Frees volatile resources (parse tree, meta-data, fields).
121: */
1.73 kristaps 122: static void
123: mdoc_free1(struct mdoc *mdoc)
1.67 kristaps 124: {
125:
126: if (mdoc->first)
1.121 kristaps 127: mdoc_node_delete(mdoc, mdoc->first);
1.67 kristaps 128: if (mdoc->meta.title)
129: free(mdoc->meta.title);
130: if (mdoc->meta.os)
131: free(mdoc->meta.os);
132: if (mdoc->meta.name)
133: free(mdoc->meta.name);
134: if (mdoc->meta.arch)
135: free(mdoc->meta.arch);
136: if (mdoc->meta.vol)
137: free(mdoc->meta.vol);
1.133 kristaps 138: if (mdoc->meta.msec)
139: free(mdoc->meta.msec);
1.73 kristaps 140: }
141:
142:
1.85 kristaps 143: /*
144: * Allocate all volatile resources (parse tree, meta-data, fields).
145: */
1.113 kristaps 146: static void
1.73 kristaps 147: mdoc_alloc1(struct mdoc *mdoc)
148: {
1.67 kristaps 149:
1.112 kristaps 150: memset(&mdoc->meta, 0, sizeof(struct mdoc_meta));
1.67 kristaps 151: mdoc->flags = 0;
1.85 kristaps 152: mdoc->lastnamed = mdoc->lastsec = SEC_NONE;
1.113 kristaps 153: mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node));
1.70 kristaps 154: mdoc->first = mdoc->last;
1.67 kristaps 155: mdoc->last->type = MDOC_ROOT;
156: mdoc->next = MDOC_NEXT_CHILD;
1.73 kristaps 157: }
158:
159:
160: /*
1.85 kristaps 161: * Free up volatile resources (see mdoc_free1()) then re-initialises the
162: * data with mdoc_alloc1(). After invocation, parse data has been reset
163: * and the parser is ready for re-invocation on a new tree; however,
164: * cross-parse non-volatile data is kept intact.
1.73 kristaps 165: */
1.113 kristaps 166: void
1.73 kristaps 167: mdoc_reset(struct mdoc *mdoc)
168: {
169:
170: mdoc_free1(mdoc);
1.113 kristaps 171: mdoc_alloc1(mdoc);
1.67 kristaps 172: }
173:
174:
1.68 kristaps 175: /*
1.85 kristaps 176: * Completely free up all volatile and non-volatile parse resources.
177: * After invocation, the pointer is no longer usable.
1.68 kristaps 178: */
1.67 kristaps 179: void
1.38 kristaps 180: mdoc_free(struct mdoc *mdoc)
1.34 kristaps 181: {
182:
1.73 kristaps 183: mdoc_free1(mdoc);
1.1 kristaps 184: free(mdoc);
185: }
186:
187:
1.85 kristaps 188: /*
189: * Allocate volatile and non-volatile parse resources.
190: */
1.1 kristaps 191: struct mdoc *
1.163 schwarze 192: mdoc_alloc(struct regset *regs, void *data, mandocmsg msg)
1.1 kristaps 193: {
194: struct mdoc *p;
195:
1.113 kristaps 196: p = mandoc_calloc(1, sizeof(struct mdoc));
197:
1.136 kristaps 198: p->msg = msg;
1.1 kristaps 199: p->data = data;
1.149 kristaps 200: p->regs = regs;
1.73 kristaps 201:
1.113 kristaps 202: mdoc_hash_init();
203: mdoc_alloc1(p);
204: return(p);
1.1 kristaps 205: }
206:
207:
1.68 kristaps 208: /*
209: * Climb back up the parse tree, validating open scopes. Mostly calls
1.85 kristaps 210: * through to macro_end() in macro.c.
1.68 kristaps 211: */
1.1 kristaps 212: int
1.72 kristaps 213: mdoc_endparse(struct mdoc *m)
1.20 kristaps 214: {
215:
1.72 kristaps 216: if (MDOC_HALT & m->flags)
1.20 kristaps 217: return(0);
1.72 kristaps 218: else if (mdoc_macroend(m))
1.20 kristaps 219: return(1);
1.72 kristaps 220: m->flags |= MDOC_HALT;
221: return(0);
1.20 kristaps 222: }
223:
224:
1.50 kristaps 225: /*
1.53 kristaps 226: * Main parse routine. Parses a single line -- really just hands off to
1.123 kristaps 227: * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
1.50 kristaps 228: */
1.20 kristaps 229: int
1.149 kristaps 230: mdoc_parseln(struct mdoc *m, int ln, char *buf, int offs)
1.1 kristaps 231: {
232:
1.53 kristaps 233: if (MDOC_HALT & m->flags)
1.20 kristaps 234: return(0);
1.50 kristaps 235:
1.130 kristaps 236: m->flags |= MDOC_NEWLINE;
1.153 schwarze 237:
238: /*
239: * Let the roff nS register switch SYNOPSIS mode early,
240: * such that the parser knows at all times
241: * whether this mode is on or off.
242: * Note that this mode is also switched by the Sh macro.
243: */
244: if (m->regs->regs[(int)REG_nS].set) {
245: if (m->regs->regs[(int)REG_nS].v.u)
246: m->flags |= MDOC_SYNOPSIS;
247: else
248: m->flags &= ~MDOC_SYNOPSIS;
249: }
250:
1.135 kristaps 251: return(('.' == buf[offs] || '\'' == buf[offs]) ?
1.149 kristaps 252: mdoc_pmacro(m, ln, buf, offs) :
253: mdoc_ptext(m, ln, buf, offs));
1.1 kristaps 254: }
255:
256:
257: int
1.136 kristaps 258: mdoc_vmsg(struct mdoc *mdoc, enum mandocerr t,
259: int ln, int pos, const char *fmt, ...)
1.1 kristaps 260: {
1.31 kristaps 261: char buf[256];
262: va_list ap;
1.1 kristaps 263:
1.31 kristaps 264: va_start(ap, fmt);
1.136 kristaps 265: vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
1.31 kristaps 266: va_end(ap);
1.88 kristaps 267:
1.136 kristaps 268: return((*mdoc->msg)(t, mdoc->data, ln, pos, buf));
1.88 kristaps 269: }
270:
271:
272: int
1.148 kristaps 273: mdoc_macro(MACRO_PROT_ARGS)
1.88 kristaps 274: {
1.122 kristaps 275: assert(tok < MDOC_MAX);
276:
277: /* If we're in the body, deny prologue calls. */
1.117 kristaps 278:
1.88 kristaps 279: if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
1.122 kristaps 280: MDOC_PBODY & m->flags)
1.148 kristaps 281: return(mdoc_pmsg(m, line, ppos, MANDOCERR_BADBODY));
1.122 kristaps 282:
283: /* If we're in the prologue, deny "body" macros. */
284:
285: if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
286: ! (MDOC_PBODY & m->flags)) {
1.148 kristaps 287: if ( ! mdoc_pmsg(m, line, ppos, MANDOCERR_BADPROLOG))
1.120 kristaps 288: return(0);
1.167 kristaps 289: if (NULL == m->meta.msec)
290: m->meta.msec = mandoc_strdup("1");
1.120 kristaps 291: if (NULL == m->meta.title)
1.140 kristaps 292: m->meta.title = mandoc_strdup("UNKNOWN");
1.120 kristaps 293: if (NULL == m->meta.vol)
1.140 kristaps 294: m->meta.vol = mandoc_strdup("LOCAL");
1.120 kristaps 295: if (NULL == m->meta.os)
1.140 kristaps 296: m->meta.os = mandoc_strdup("LOCAL");
1.120 kristaps 297: if (0 == m->meta.date)
298: m->meta.date = time(NULL);
299: m->flags |= MDOC_PBODY;
300: }
1.88 kristaps 301:
1.149 kristaps 302: return((*mdoc_macros[tok].fp)(m, tok, line, ppos, pos, buf));
1.73 kristaps 303: }
304:
305:
306: static int
307: node_append(struct mdoc *mdoc, struct mdoc_node *p)
1.1 kristaps 308: {
309:
1.25 kristaps 310: assert(mdoc->last);
311: assert(mdoc->first);
312: assert(MDOC_ROOT != p->type);
1.1 kristaps 313:
1.13 kristaps 314: switch (mdoc->next) {
315: case (MDOC_NEXT_SIBLING):
1.6 kristaps 316: mdoc->last->next = p;
317: p->prev = mdoc->last;
1.13 kristaps 318: p->parent = mdoc->last->parent;
1.1 kristaps 319: break;
1.13 kristaps 320: case (MDOC_NEXT_CHILD):
321: mdoc->last->child = p;
1.1 kristaps 322: p->parent = mdoc->last;
323: break;
324: default:
1.13 kristaps 325: abort();
326: /* NOTREACHED */
1.1 kristaps 327: }
328:
1.86 kristaps 329: p->parent->nchild++;
330:
1.23 kristaps 331: if ( ! mdoc_valid_pre(mdoc, p))
332: return(0);
1.27 kristaps 333:
334: switch (p->type) {
335: case (MDOC_HEAD):
336: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 337: p->parent->head = p;
1.27 kristaps 338: break;
339: case (MDOC_TAIL):
340: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 341: p->parent->tail = p;
1.27 kristaps 342: break;
343: case (MDOC_BODY):
1.152 schwarze 344: if (p->end)
345: break;
1.27 kristaps 346: assert(MDOC_BLOCK == p->parent->type);
1.53 kristaps 347: p->parent->body = p;
1.27 kristaps 348: break;
349: default:
350: break;
351: }
352:
1.1 kristaps 353: mdoc->last = p;
1.71 kristaps 354:
355: switch (p->type) {
356: case (MDOC_TEXT):
357: if ( ! mdoc_valid_post(mdoc))
358: return(0);
359: break;
360: default:
361: break;
362: }
363:
1.23 kristaps 364: return(1);
1.1 kristaps 365: }
366:
367:
1.45 kristaps 368: static struct mdoc_node *
1.117 kristaps 369: node_alloc(struct mdoc *m, int line, int pos,
370: enum mdoct tok, enum mdoc_type type)
1.45 kristaps 371: {
1.46 kristaps 372: struct mdoc_node *p;
373:
1.113 kristaps 374: p = mandoc_calloc(1, sizeof(struct mdoc_node));
1.91 kristaps 375: p->sec = m->lastsec;
1.73 kristaps 376: p->line = line;
377: p->pos = pos;
378: p->tok = tok;
1.118 kristaps 379: p->type = type;
1.150 kristaps 380:
381: /* Flag analysis. */
382:
1.153 schwarze 383: if (MDOC_SYNOPSIS & m->flags)
384: p->flags |= MDOC_SYNPRETTY;
385: else
386: p->flags &= ~MDOC_SYNPRETTY;
1.130 kristaps 387: if (MDOC_NEWLINE & m->flags)
388: p->flags |= MDOC_LINE;
389: m->flags &= ~MDOC_NEWLINE;
1.150 kristaps 390:
1.46 kristaps 391: return(p);
1.45 kristaps 392: }
393:
394:
1.23 kristaps 395: int
1.117 kristaps 396: mdoc_tail_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
1.17 kristaps 397: {
398: struct mdoc_node *p;
399:
1.91 kristaps 400: p = node_alloc(m, line, pos, tok, MDOC_TAIL);
1.102 kristaps 401: if ( ! node_append(m, p))
402: return(0);
403: m->next = MDOC_NEXT_CHILD;
404: return(1);
1.17 kristaps 405: }
406:
407:
1.23 kristaps 408: int
1.117 kristaps 409: mdoc_head_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
1.1 kristaps 410: {
411: struct mdoc_node *p;
412:
1.91 kristaps 413: assert(m->first);
414: assert(m->last);
1.1 kristaps 415:
1.91 kristaps 416: p = node_alloc(m, line, pos, tok, MDOC_HEAD);
1.102 kristaps 417: if ( ! node_append(m, p))
418: return(0);
419: m->next = MDOC_NEXT_CHILD;
420: return(1);
1.1 kristaps 421: }
422:
423:
1.23 kristaps 424: int
1.117 kristaps 425: mdoc_body_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
1.1 kristaps 426: {
427: struct mdoc_node *p;
428:
1.91 kristaps 429: p = node_alloc(m, line, pos, tok, MDOC_BODY);
1.102 kristaps 430: if ( ! node_append(m, p))
431: return(0);
432: m->next = MDOC_NEXT_CHILD;
1.152 schwarze 433: return(1);
434: }
435:
436:
437: int
438: mdoc_endbody_alloc(struct mdoc *m, int line, int pos, enum mdoct tok,
439: struct mdoc_node *body, enum mdoc_endbody end)
440: {
441: struct mdoc_node *p;
442:
443: p = node_alloc(m, line, pos, tok, MDOC_BODY);
444: p->pending = body;
445: p->end = end;
446: if ( ! node_append(m, p))
447: return(0);
448: m->next = MDOC_NEXT_SIBLING;
1.102 kristaps 449: return(1);
1.1 kristaps 450: }
451:
452:
1.23 kristaps 453: int
1.91 kristaps 454: mdoc_block_alloc(struct mdoc *m, int line, int pos,
1.117 kristaps 455: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 456: {
457: struct mdoc_node *p;
458:
1.91 kristaps 459: p = node_alloc(m, line, pos, tok, MDOC_BLOCK);
1.77 kristaps 460: p->args = args;
461: if (p->args)
1.53 kristaps 462: (args->refcnt)++;
1.102 kristaps 463: if ( ! node_append(m, p))
464: return(0);
465: m->next = MDOC_NEXT_CHILD;
466: return(1);
1.1 kristaps 467: }
468:
469:
1.23 kristaps 470: int
1.91 kristaps 471: mdoc_elem_alloc(struct mdoc *m, int line, int pos,
1.117 kristaps 472: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 473: {
474: struct mdoc_node *p;
475:
1.91 kristaps 476: p = node_alloc(m, line, pos, tok, MDOC_ELEM);
1.77 kristaps 477: p->args = args;
478: if (p->args)
1.53 kristaps 479: (args->refcnt)++;
1.102 kristaps 480: if ( ! node_append(m, p))
481: return(0);
482: m->next = MDOC_NEXT_CHILD;
483: return(1);
1.1 kristaps 484: }
485:
486:
1.124 kristaps 487: int
488: mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p)
1.1 kristaps 489: {
1.91 kristaps 490: struct mdoc_node *n;
1.124 kristaps 491: size_t sv, len;
492:
493: len = strlen(p);
1.1 kristaps 494:
1.125 kristaps 495: n = node_alloc(m, line, pos, MDOC_MAX, MDOC_TEXT);
1.113 kristaps 496: n->string = mandoc_malloc(len + 1);
1.91 kristaps 497: sv = strlcpy(n->string, p, len + 1);
498:
499: /* Prohibit truncation. */
500: assert(sv < len + 1);
501:
1.101 kristaps 502: if ( ! node_append(m, n))
503: return(0);
1.124 kristaps 504:
1.101 kristaps 505: m->next = MDOC_NEXT_SIBLING;
506: return(1);
1.91 kristaps 507: }
508:
509:
1.155 kristaps 510: static void
1.53 kristaps 511: mdoc_node_free(struct mdoc_node *p)
1.1 kristaps 512: {
513:
1.171 ! kristaps 514: if (p->norm && 0 == --(p->norm->refcnt))
! 515: free(p->norm);
1.154 kristaps 516:
1.53 kristaps 517: if (p->string)
518: free(p->string);
519: if (p->args)
520: mdoc_argv_free(p->args);
1.1 kristaps 521: free(p);
522: }
523:
524:
1.121 kristaps 525: static void
526: mdoc_node_unlink(struct mdoc *m, struct mdoc_node *n)
527: {
528:
529: /* Adjust siblings. */
530:
531: if (n->prev)
532: n->prev->next = n->next;
533: if (n->next)
534: n->next->prev = n->prev;
535:
536: /* Adjust parent. */
537:
538: if (n->parent) {
539: n->parent->nchild--;
540: if (n->parent->child == n)
541: n->parent->child = n->prev ? n->prev : n->next;
1.169 kristaps 542: if (n->parent->last == n)
543: n->parent->last = n->prev ? n->prev : NULL;
1.121 kristaps 544: }
545:
546: /* Adjust parse point, if applicable. */
547:
548: if (m && m->last == n) {
549: if (n->prev) {
550: m->last = n->prev;
551: m->next = MDOC_NEXT_SIBLING;
552: } else {
553: m->last = n->parent;
554: m->next = MDOC_NEXT_CHILD;
555: }
556: }
557:
558: if (m && m->first == n)
559: m->first = NULL;
560: }
561:
562:
1.53 kristaps 563: void
1.121 kristaps 564: mdoc_node_delete(struct mdoc *m, struct mdoc_node *p)
1.1 kristaps 565: {
566:
1.121 kristaps 567: while (p->child) {
568: assert(p->nchild);
569: mdoc_node_delete(m, p->child);
570: }
571: assert(0 == p->nchild);
1.1 kristaps 572:
1.121 kristaps 573: mdoc_node_unlink(m, p);
1.53 kristaps 574: mdoc_node_free(p);
1.1 kristaps 575: }
576:
577:
1.53 kristaps 578: /*
579: * Parse free-form text, that is, a line that does not begin with the
580: * control character.
581: */
582: static int
1.149 kristaps 583: mdoc_ptext(struct mdoc *m, int line, char *buf, int offs)
1.1 kristaps 584: {
1.142 kristaps 585: char *c, *ws, *end;
586: struct mdoc_node *n;
1.1 kristaps 587:
1.123 kristaps 588: /* Ignore bogus comments. */
589:
1.135 kristaps 590: if ('\\' == buf[offs] &&
591: '.' == buf[offs + 1] &&
592: '"' == buf[offs + 2])
1.136 kristaps 593: return(mdoc_pmsg(m, line, offs, MANDOCERR_BADCOMMENT));
1.123 kristaps 594:
1.124 kristaps 595: /* No text before an initial macro. */
596:
1.85 kristaps 597: if (SEC_NONE == m->lastnamed)
1.136 kristaps 598: return(mdoc_pmsg(m, line, offs, MANDOCERR_NOTEXT));
1.142 kristaps 599:
600: assert(m->last);
601: n = m->last;
602:
603: /*
1.144 kristaps 604: * Divert directly to list processing if we're encountering a
1.142 kristaps 605: * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
1.144 kristaps 606: * (a MDOC_BODY means it's already open, in which case we should
607: * process within its context in the normal way).
1.142 kristaps 608: */
609:
1.143 kristaps 610: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.171 ! kristaps 611: LIST_column == n->norm->d.Bl.type) {
1.144 kristaps 612: /* `Bl' is open without any children. */
1.142 kristaps 613: m->flags |= MDOC_FREECOL;
1.149 kristaps 614: return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
1.142 kristaps 615: }
616:
617: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
618: NULL != n->parent &&
619: MDOC_Bl == n->parent->tok &&
1.171 ! kristaps 620: LIST_column == n->parent->norm->d.Bl.type) {
1.144 kristaps 621: /* `Bl' has block-level `It' children. */
1.142 kristaps 622: m->flags |= MDOC_FREECOL;
1.149 kristaps 623: return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
1.142 kristaps 624: }
1.124 kristaps 625:
1.137 schwarze 626: /*
627: * Search for the beginning of unescaped trailing whitespace (ws)
628: * and for the first character not to be output (end).
629: */
1.139 kristaps 630:
631: /* FIXME: replace with strcspn(). */
1.137 schwarze 632: ws = NULL;
633: for (c = end = buf + offs; *c; c++) {
634: switch (*c) {
1.138 kristaps 635: case '-':
636: if (mandoc_hyph(buf + offs, c))
637: *c = ASCII_HYPH;
1.145 kristaps 638: ws = NULL;
1.138 kristaps 639: break;
1.137 schwarze 640: case ' ':
641: if (NULL == ws)
642: ws = c;
643: continue;
644: case '\t':
645: /*
646: * Always warn about trailing tabs,
647: * even outside literal context,
648: * where they should be put on the next line.
649: */
650: if (NULL == ws)
651: ws = c;
652: /*
653: * Strip trailing tabs in literal context only;
654: * outside, they affect the next line.
655: */
656: if (MDOC_LITERAL & m->flags)
657: continue;
658: break;
659: case '\\':
660: /* Skip the escaped character, too, if any. */
661: if (c[1])
662: c++;
663: /* FALLTHROUGH */
664: default:
665: ws = NULL;
666: break;
667: }
668: end = c + 1;
669: }
670: *end = '\0';
1.91 kristaps 671:
1.137 schwarze 672: if (ws)
673: if ( ! mdoc_pmsg(m, line, (int)(ws-buf), MANDOCERR_EOLNSPACE))
674: return(0);
1.115 kristaps 675:
1.137 schwarze 676: if ('\0' == buf[offs] && ! (MDOC_LITERAL & m->flags)) {
677: if ( ! mdoc_pmsg(m, line, (int)(c-buf), MANDOCERR_NOBLANKLN))
1.119 kristaps 678: return(0);
1.124 kristaps 679:
1.119 kristaps 680: /*
1.165 schwarze 681: * Insert a `sp' in the case of a blank line. Technically,
1.124 kristaps 682: * blank lines aren't allowed, but enough manuals assume this
683: * behaviour that we want to work around it.
1.119 kristaps 684: */
1.165 schwarze 685: if ( ! mdoc_elem_alloc(m, line, offs, MDOC_sp, NULL))
1.119 kristaps 686: return(0);
1.124 kristaps 687:
688: m->next = MDOC_NEXT_SIBLING;
689: return(1);
1.119 kristaps 690: }
1.68 kristaps 691:
1.137 schwarze 692: if ( ! mdoc_word_alloc(m, line, offs, buf+offs))
693: return(0);
1.91 kristaps 694:
1.137 schwarze 695: if (MDOC_LITERAL & m->flags)
696: return(1);
1.128 kristaps 697:
698: /*
699: * End-of-sentence check. If the last character is an unescaped
700: * EOS character, then flag the node as being the end of a
701: * sentence. The front-end will know how to interpret this.
702: */
1.132 kristaps 703:
1.137 schwarze 704: assert(buf < end);
705:
1.159 schwarze 706: if (mandoc_eos(buf+offs, (size_t)(end-buf-offs), 0))
1.128 kristaps 707: m->last->flags |= MDOC_EOS;
708:
709: return(1);
1.1 kristaps 710: }
711:
712:
1.53 kristaps 713: /*
714: * Parse a macro line, that is, a line beginning with the control
715: * character.
716: */
1.155 kristaps 717: static int
1.149 kristaps 718: mdoc_pmacro(struct mdoc *m, int ln, char *buf, int offs)
1.1 kristaps 719: {
1.144 kristaps 720: enum mdoct tok;
721: int i, j, sv;
722: char mac[5];
723: struct mdoc_node *n;
1.1 kristaps 724:
1.81 kristaps 725: /* Empty lines are ignored. */
1.63 kristaps 726:
1.135 kristaps 727: offs++;
728:
729: if ('\0' == buf[offs])
1.63 kristaps 730: return(1);
731:
1.135 kristaps 732: i = offs;
1.100 kristaps 733:
1.160 kristaps 734: /* Accept tabs/whitespace after the initial control char. */
1.100 kristaps 735:
1.160 kristaps 736: if (' ' == buf[i] || '\t' == buf[i]) {
1.100 kristaps 737: i++;
1.160 kristaps 738: while (buf[i] && (' ' == buf[i] || '\t' == buf[i]))
1.63 kristaps 739: i++;
1.115 kristaps 740: if ('\0' == buf[i])
1.63 kristaps 741: return(1);
742: }
1.1 kristaps 743:
1.130 kristaps 744: sv = i;
745:
1.160 kristaps 746: /*
1.162 schwarze 747: * Copy the first word into a nil-terminated buffer.
1.165 schwarze 748: * Stop copying when a tab, space, or eoln is encountered.
1.160 kristaps 749: */
1.1 kristaps 750:
1.162 schwarze 751: j = 0;
1.165 schwarze 752: while (j < 4 && '\0' != buf[i] && ' ' != buf[i] && '\t' != buf[i])
1.162 schwarze 753: mac[j++] = buf[i++];
1.135 kristaps 754: mac[j] = '\0';
1.1 kristaps 755:
1.163 schwarze 756: tok = (j > 1 || j < 4) ? mdoc_hash_find(mac) : MDOC_MAX;
757: if (MDOC_MAX == tok) {
1.168 kristaps 758: mdoc_vmsg(m, MANDOCERR_MACRO, ln, sv, "%s", buf + sv - 1);
1.58 kristaps 759: return(1);
1.53 kristaps 760: }
1.1 kristaps 761:
1.160 kristaps 762: /* Disregard the first trailing tab, if applicable. */
763:
764: if ('\t' == buf[i])
765: i++;
766:
767: /* Jump to the next non-whitespace word. */
1.1 kristaps 768:
1.65 kristaps 769: while (buf[i] && ' ' == buf[i])
1.53 kristaps 770: i++;
1.1 kristaps 771:
1.125 kristaps 772: /*
773: * Trailing whitespace. Note that tabs are allowed to be passed
774: * into the parser as "text", so we only warn about spaces here.
775: */
1.115 kristaps 776:
777: if ('\0' == buf[i] && ' ' == buf[i - 1])
1.136 kristaps 778: if ( ! mdoc_pmsg(m, ln, i - 1, MANDOCERR_EOLNSPACE))
1.115 kristaps 779: goto err;
780:
1.144 kristaps 781: /*
782: * If an initial macro or a list invocation, divert directly
783: * into macro processing.
784: */
785:
786: if (NULL == m->last || MDOC_It == tok || MDOC_El == tok) {
1.149 kristaps 787: if ( ! mdoc_macro(m, tok, ln, sv, &i, buf))
1.144 kristaps 788: goto err;
789: return(1);
790: }
791:
792: n = m->last;
793: assert(m->last);
794:
795: /*
796: * If the first macro of a `Bl -column', open an `It' block
797: * context around the parsed macro.
798: */
799:
800: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.171 ! kristaps 801: LIST_column == n->norm->d.Bl.type) {
1.144 kristaps 802: m->flags |= MDOC_FREECOL;
1.149 kristaps 803: if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
1.144 kristaps 804: goto err;
805: return(1);
806: }
807:
808: /*
809: * If we're following a block-level `It' within a `Bl -column'
810: * context (perhaps opened in the above block or in ptext()),
811: * then open an `It' block context around the parsed macro.
1.98 kristaps 812: */
1.144 kristaps 813:
814: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
815: NULL != n->parent &&
816: MDOC_Bl == n->parent->tok &&
1.171 ! kristaps 817: LIST_column == n->parent->norm->d.Bl.type) {
1.144 kristaps 818: m->flags |= MDOC_FREECOL;
1.149 kristaps 819: if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
1.144 kristaps 820: goto err;
821: return(1);
822: }
823:
824: /* Normal processing of a macro. */
825:
1.149 kristaps 826: if ( ! mdoc_macro(m, tok, ln, sv, &i, buf))
1.53 kristaps 827: goto err;
1.1 kristaps 828:
1.53 kristaps 829: return(1);
1.1 kristaps 830:
1.53 kristaps 831: err: /* Error out. */
1.1 kristaps 832:
1.53 kristaps 833: m->flags |= MDOC_HALT;
834: return(0);
1.1 kristaps 835: }
1.100 kristaps 836:
837:
CVSweb