Annotation of src/usr.bin/mandoc/mdoc.c, Revision 1.53
1.53 ! schwarze 1: /* $Id: mdoc.c,v 1.52 2010/05/16 20:46:15 schwarze Exp $ */
1.1 kristaps 2: /*
1.3 schwarze 3: * Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@kth.se>
1.1 kristaps 4: *
5: * Permission to use, copy, modify, and distribute this software for any
1.3 schwarze 6: * purpose with or without fee is hereby granted, provided that the above
7: * copyright notice and this permission notice appear in all copies.
1.1 kristaps 8: *
1.3 schwarze 9: * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
10: * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
11: * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
12: * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
13: * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
14: * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
15: * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
1.1 kristaps 16: */
1.28 schwarze 17: #include <sys/types.h>
18:
1.1 kristaps 19: #include <assert.h>
20: #include <ctype.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:
27: #include "libmdoc.h"
1.32 schwarze 28: #include "libmandoc.h"
1.1 kristaps 29:
1.14 schwarze 30: const char *const __mdoc_merrnames[MERRMAX] = {
31: "trailing whitespace", /* ETAILWS */
32: "unexpected quoted parameter", /* EQUOTPARM */
33: "unterminated quoted parameter", /* EQUOTTERM */
34: "argument parameter suggested", /* EARGVAL */
35: "macro disallowed in prologue", /* EBODYPROL */
36: "macro disallowed in body", /* EPROLBODY */
37: "text disallowed in prologue", /* ETEXTPROL */
38: "blank line disallowed", /* ENOBLANK */
39: "text parameter too long", /* ETOOLONG */
40: "invalid escape sequence", /* EESCAPE */
41: "invalid character", /* EPRINT */
42: "document has no body", /* ENODAT */
43: "document has no prologue", /* ENOPROLOGUE */
44: "expected line arguments", /* ELINE */
45: "invalid AT&T argument", /* EATT */
46: "default name not yet set", /* ENAME */
47: "missing list type", /* ELISTTYPE */
48: "missing display type", /* EDISPTYPE */
49: "too many display types", /* EMULTIDISP */
50: "too many list types", /* EMULTILIST */
51: "NAME section must be first", /* ESECNAME */
52: "badly-formed NAME section", /* ENAMESECINC */
53: "argument repeated", /* EARGREP */
54: "expected boolean parameter", /* EBOOL */
55: "inconsistent column syntax", /* ECOLMIS */
56: "nested display invalid", /* ENESTDISP */
57: "width argument missing", /* EMISSWIDTH */
58: "invalid section for this manual section", /* EWRONGMSEC */
59: "section out of conventional order", /* ESECOOO */
60: "section repeated", /* ESECREP */
61: "invalid standard argument", /* EBADSTAND */
62: "multi-line arguments discouraged", /* ENOMULTILINE */
63: "multi-line arguments suggested", /* EMULTILINE */
64: "line arguments discouraged", /* ENOLINE */
65: "prologue macro out of conventional order", /* EPROLOOO */
66: "prologue macro repeated", /* EPROLREP */
1.50 schwarze 67: "invalid manual section", /* EBADMSEC */
1.14 schwarze 68: "invalid font mode", /* EFONT */
69: "invalid date syntax", /* EBADDATE */
70: "invalid number format", /* ENUMFMT */
71: "superfluous width argument", /* ENOWIDTH */
72: "system: utsname error", /* EUTSNAME */
73: "obsolete macro", /* EOBS */
74: "end-of-line scope violation", /* EIMPBRK */
75: "empty macro ignored", /* EIGNE */
76: "unclosed explicit scope", /* EOPEN */
77: "unterminated quoted phrase", /* EQUOTPHR */
78: "closure macro without prior context", /* ENOCTX */
1.29 schwarze 79: "no description found for library", /* ELIB */
80: "bad child for parent context", /* EBADCHILD */
1.30 schwarze 81: "list arguments preceding type", /* ENOTYPE */
1.44 schwarze 82: "deprecated comment style", /* EBADCOMMENT */
1.1 kristaps 83: };
84:
85: const char *const __mdoc_macronames[MDOC_MAX] = {
1.7 schwarze 86: "Ap", "Dd", "Dt", "Os",
1.1 kristaps 87: "Sh", "Ss", "Pp", "D1",
88: "Dl", "Bd", "Ed", "Bl",
89: "El", "It", "Ad", "An",
90: "Ar", "Cd", "Cm", "Dv",
91: "Er", "Ev", "Ex", "Fa",
92: "Fd", "Fl", "Fn", "Ft",
93: "Ic", "In", "Li", "Nd",
94: "Nm", "Op", "Ot", "Pa",
95: "Rv", "St", "Va", "Vt",
96: /* LINTED */
1.33 schwarze 97: "Xr", "%A", "%B", "%D",
1.1 kristaps 98: /* LINTED */
1.33 schwarze 99: "%I", "%J", "%N", "%O",
1.1 kristaps 100: /* LINTED */
1.33 schwarze 101: "%P", "%R", "%T", "%V",
1.1 kristaps 102: "Ac", "Ao", "Aq", "At",
103: "Bc", "Bf", "Bo", "Bq",
104: "Bsx", "Bx", "Db", "Dc",
105: "Do", "Dq", "Ec", "Ef",
106: "Em", "Eo", "Fx", "Ms",
107: "No", "Ns", "Nx", "Ox",
108: "Pc", "Pf", "Po", "Pq",
109: "Qc", "Ql", "Qo", "Qq",
110: "Re", "Rs", "Sc", "So",
111: "Sq", "Sm", "Sx", "Sy",
112: "Tn", "Ux", "Xc", "Xo",
113: "Fo", "Fc", "Oo", "Oc",
114: "Bk", "Ek", "Bt", "Hf",
1.7 schwarze 115: "Fr", "Ud", "Lb", "Lp",
116: "Lk", "Mt", "Brq", "Bro",
1.1 kristaps 117: /* LINTED */
1.33 schwarze 118: "Brc", "%C", "Es", "En",
1.1 kristaps 119: /* LINTED */
1.33 schwarze 120: "Dx", "%Q", "br", "sp",
1.31 schwarze 121: /* LINTED */
1.48 schwarze 122: "%U"
1.1 kristaps 123: };
124:
125: const char *const __mdoc_argnames[MDOC_ARG_MAX] = {
126: "split", "nosplit", "ragged",
127: "unfilled", "literal", "file",
128: "offset", "bullet", "dash",
129: "hyphen", "item", "enum",
130: "tag", "diag", "hang",
131: "ohang", "inset", "column",
132: "width", "compact", "std",
133: "filled", "words", "emphasis",
1.30 schwarze 134: "symbolic", "nested", "centered"
1.1 kristaps 135: };
136:
137: const char * const *mdoc_macronames = __mdoc_macronames;
138: const char * const *mdoc_argnames = __mdoc_argnames;
139:
1.40 schwarze 140: static void mdoc_node_free(struct mdoc_node *);
141: static void mdoc_node_unlink(struct mdoc *,
142: struct mdoc_node *);
1.1 kristaps 143: static void mdoc_free1(struct mdoc *);
1.32 schwarze 144: static void mdoc_alloc1(struct mdoc *);
1.1 kristaps 145: static struct mdoc_node *node_alloc(struct mdoc *, int, int,
1.37 schwarze 146: enum mdoct, enum mdoc_type);
1.1 kristaps 147: static int node_append(struct mdoc *,
148: struct mdoc_node *);
1.53 ! schwarze 149: static int mdoc_ptext(struct mdoc *, int, char *, int);
! 150: static int mdoc_pmacro(struct mdoc *, int, char *, int);
! 151: static int macrowarn(struct mdoc *, int,
! 152: const char *, int);
1.46 schwarze 153:
1.1 kristaps 154:
155: const struct mdoc_node *
156: mdoc_node(const struct mdoc *m)
157: {
158:
159: return(MDOC_HALT & m->flags ? NULL : m->first);
160: }
161:
162:
163: const struct mdoc_meta *
164: mdoc_meta(const struct mdoc *m)
165: {
166:
167: return(MDOC_HALT & m->flags ? NULL : &m->meta);
168: }
169:
170:
1.8 schwarze 171: /*
172: * Frees volatile resources (parse tree, meta-data, fields).
173: */
1.1 kristaps 174: static void
175: mdoc_free1(struct mdoc *mdoc)
176: {
177:
178: if (mdoc->first)
1.40 schwarze 179: mdoc_node_delete(mdoc, mdoc->first);
1.1 kristaps 180: if (mdoc->meta.title)
181: free(mdoc->meta.title);
182: if (mdoc->meta.os)
183: free(mdoc->meta.os);
184: if (mdoc->meta.name)
185: free(mdoc->meta.name);
186: if (mdoc->meta.arch)
187: free(mdoc->meta.arch);
188: if (mdoc->meta.vol)
189: free(mdoc->meta.vol);
1.50 schwarze 190: if (mdoc->meta.msec)
191: free(mdoc->meta.msec);
1.1 kristaps 192: }
193:
194:
1.8 schwarze 195: /*
196: * Allocate all volatile resources (parse tree, meta-data, fields).
197: */
1.32 schwarze 198: static void
1.1 kristaps 199: mdoc_alloc1(struct mdoc *mdoc)
200: {
201:
1.32 schwarze 202: memset(&mdoc->meta, 0, sizeof(struct mdoc_meta));
1.1 kristaps 203: mdoc->flags = 0;
1.9 schwarze 204: mdoc->lastnamed = mdoc->lastsec = SEC_NONE;
1.32 schwarze 205: mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node));
1.1 kristaps 206: mdoc->first = mdoc->last;
207: mdoc->last->type = MDOC_ROOT;
208: mdoc->next = MDOC_NEXT_CHILD;
209: }
210:
211:
212: /*
1.8 schwarze 213: * Free up volatile resources (see mdoc_free1()) then re-initialises the
214: * data with mdoc_alloc1(). After invocation, parse data has been reset
215: * and the parser is ready for re-invocation on a new tree; however,
216: * cross-parse non-volatile data is kept intact.
1.1 kristaps 217: */
1.32 schwarze 218: void
1.1 kristaps 219: mdoc_reset(struct mdoc *mdoc)
220: {
221:
222: mdoc_free1(mdoc);
1.32 schwarze 223: mdoc_alloc1(mdoc);
1.1 kristaps 224: }
225:
226:
227: /*
1.8 schwarze 228: * Completely free up all volatile and non-volatile parse resources.
229: * After invocation, the pointer is no longer usable.
1.1 kristaps 230: */
231: void
232: mdoc_free(struct mdoc *mdoc)
233: {
234:
235: mdoc_free1(mdoc);
236: free(mdoc);
237: }
238:
239:
1.8 schwarze 240: /*
241: * Allocate volatile and non-volatile parse resources.
242: */
1.1 kristaps 243: struct mdoc *
244: mdoc_alloc(void *data, int pflags, const struct mdoc_cb *cb)
245: {
246: struct mdoc *p;
247:
1.32 schwarze 248: p = mandoc_calloc(1, sizeof(struct mdoc));
249:
1.1 kristaps 250: if (cb)
1.32 schwarze 251: memcpy(&p->cb, cb, sizeof(struct mdoc_cb));
1.27 schwarze 252:
1.1 kristaps 253: p->data = data;
254: p->pflags = pflags;
255:
1.32 schwarze 256: mdoc_hash_init();
257: mdoc_alloc1(p);
258: return(p);
1.1 kristaps 259: }
260:
261:
262: /*
263: * Climb back up the parse tree, validating open scopes. Mostly calls
1.8 schwarze 264: * through to macro_end() in macro.c.
1.1 kristaps 265: */
266: int
267: mdoc_endparse(struct mdoc *m)
268: {
269:
270: if (MDOC_HALT & m->flags)
271: return(0);
272: else if (mdoc_macroend(m))
273: return(1);
274: m->flags |= MDOC_HALT;
275: return(0);
276: }
277:
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
1.53 ! schwarze 284: mdoc_parseln(struct mdoc *m, int ln, char *buf, int offs)
1.1 kristaps 285: {
286:
287: if (MDOC_HALT & m->flags)
288: return(0);
289:
1.47 schwarze 290: m->flags |= MDOC_NEWLINE;
1.53 ! schwarze 291: return(('.' == buf[offs] || '\'' == buf[offs]) ?
! 292: mdoc_pmacro(m, ln, buf, offs) :
! 293: mdoc_ptext(m, ln, buf, offs));
1.1 kristaps 294: }
295:
296:
297: int
1.4 schwarze 298: mdoc_verr(struct mdoc *mdoc, int ln, int pos,
299: const char *fmt, ...)
1.1 kristaps 300: {
301: char buf[256];
302: va_list ap;
303:
304: if (NULL == mdoc->cb.mdoc_err)
305: return(0);
306:
307: va_start(ap, fmt);
308: (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
309: va_end(ap);
1.12 schwarze 310:
1.1 kristaps 311: return((*mdoc->cb.mdoc_err)(mdoc->data, ln, pos, buf));
312: }
313:
314:
315: int
1.12 schwarze 316: mdoc_vwarn(struct mdoc *mdoc, int ln, int pos, const char *fmt, ...)
1.1 kristaps 317: {
318: char buf[256];
319: va_list ap;
320:
321: if (NULL == mdoc->cb.mdoc_warn)
322: return(0);
323:
324: va_start(ap, fmt);
325: (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
326: va_end(ap);
1.17 schwarze 327:
1.11 schwarze 328: return((*mdoc->cb.mdoc_warn)(mdoc->data, ln, pos, buf));
1.5 schwarze 329: }
330:
331:
332: int
1.14 schwarze 333: mdoc_err(struct mdoc *m, int line, int pos, int iserr, enum merr type)
1.5 schwarze 334: {
1.14 schwarze 335: const char *p;
1.5 schwarze 336:
1.14 schwarze 337: p = __mdoc_merrnames[(int)type];
338: assert(p);
1.5 schwarze 339:
1.14 schwarze 340: if (iserr)
341: return(mdoc_verr(m, line, pos, p));
1.16 schwarze 342:
1.14 schwarze 343: return(mdoc_vwarn(m, line, pos, p));
1.5 schwarze 344: }
345:
346:
347: int
1.37 schwarze 348: mdoc_macro(struct mdoc *m, enum mdoct tok,
1.1 kristaps 349: int ln, int pp, int *pos, char *buf)
350: {
1.46 schwarze 351: assert(tok < MDOC_MAX);
352:
353: /* If we're in the body, deny prologue calls. */
1.37 schwarze 354:
1.1 kristaps 355: if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
1.9 schwarze 356: MDOC_PBODY & m->flags)
1.16 schwarze 357: return(mdoc_perr(m, ln, pp, EPROLBODY));
1.46 schwarze 358:
359: /* If we're in the prologue, deny "body" macros. */
360:
1.1 kristaps 361: if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
1.39 schwarze 362: ! (MDOC_PBODY & m->flags)) {
363: if ( ! mdoc_pwarn(m, ln, pp, EBODYPROL))
364: return(0);
365: if (NULL == m->meta.title)
366: m->meta.title = mandoc_strdup("unknown");
367: if (NULL == m->meta.vol)
368: m->meta.vol = mandoc_strdup("local");
369: if (NULL == m->meta.os)
370: m->meta.os = mandoc_strdup("local");
371: if (0 == m->meta.date)
372: m->meta.date = time(NULL);
373: m->flags |= MDOC_PBODY;
374: }
1.1 kristaps 375:
376: return((*mdoc_macros[tok].fp)(m, tok, ln, pp, pos, buf));
377: }
378:
379:
380: static int
381: node_append(struct mdoc *mdoc, struct mdoc_node *p)
382: {
383:
384: assert(mdoc->last);
385: assert(mdoc->first);
386: assert(MDOC_ROOT != p->type);
387:
388: switch (mdoc->next) {
389: case (MDOC_NEXT_SIBLING):
390: mdoc->last->next = p;
391: p->prev = mdoc->last;
392: p->parent = mdoc->last->parent;
393: break;
394: case (MDOC_NEXT_CHILD):
395: mdoc->last->child = p;
396: p->parent = mdoc->last;
397: break;
398: default:
399: abort();
400: /* NOTREACHED */
401: }
402:
1.10 schwarze 403: p->parent->nchild++;
404:
1.1 kristaps 405: if ( ! mdoc_valid_pre(mdoc, p))
406: return(0);
407: if ( ! mdoc_action_pre(mdoc, p))
408: return(0);
409:
410: switch (p->type) {
411: case (MDOC_HEAD):
412: assert(MDOC_BLOCK == p->parent->type);
413: p->parent->head = p;
414: break;
415: case (MDOC_TAIL):
416: assert(MDOC_BLOCK == p->parent->type);
417: p->parent->tail = p;
418: break;
419: case (MDOC_BODY):
420: assert(MDOC_BLOCK == p->parent->type);
421: p->parent->body = p;
422: break;
423: default:
424: break;
425: }
426:
427: mdoc->last = p;
428:
429: switch (p->type) {
430: case (MDOC_TEXT):
431: if ( ! mdoc_valid_post(mdoc))
432: return(0);
433: if ( ! mdoc_action_post(mdoc))
434: return(0);
435: break;
436: default:
437: break;
438: }
439:
440: return(1);
441: }
442:
443:
444: static struct mdoc_node *
1.37 schwarze 445: node_alloc(struct mdoc *m, int line, int pos,
446: enum mdoct tok, enum mdoc_type type)
1.1 kristaps 447: {
448: struct mdoc_node *p;
449:
1.32 schwarze 450: p = mandoc_calloc(1, sizeof(struct mdoc_node));
1.19 schwarze 451: p->sec = m->lastsec;
1.1 kristaps 452: p->line = line;
453: p->pos = pos;
454: p->tok = tok;
1.37 schwarze 455: p->type = type;
1.47 schwarze 456: if (MDOC_NEWLINE & m->flags)
457: p->flags |= MDOC_LINE;
458: m->flags &= ~MDOC_NEWLINE;
1.1 kristaps 459: return(p);
460: }
461:
462:
463: int
1.37 schwarze 464: mdoc_tail_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
1.1 kristaps 465: {
466: struct mdoc_node *p;
467:
1.19 schwarze 468: p = node_alloc(m, line, pos, tok, MDOC_TAIL);
1.25 schwarze 469: if ( ! node_append(m, p))
470: return(0);
471: m->next = MDOC_NEXT_CHILD;
472: return(1);
1.1 kristaps 473: }
474:
475:
476: int
1.37 schwarze 477: mdoc_head_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
1.1 kristaps 478: {
479: struct mdoc_node *p;
480:
1.19 schwarze 481: assert(m->first);
482: assert(m->last);
1.1 kristaps 483:
1.19 schwarze 484: p = node_alloc(m, line, pos, tok, MDOC_HEAD);
1.25 schwarze 485: if ( ! node_append(m, p))
486: return(0);
487: m->next = MDOC_NEXT_CHILD;
488: return(1);
1.1 kristaps 489: }
490:
491:
492: int
1.37 schwarze 493: mdoc_body_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
1.1 kristaps 494: {
495: struct mdoc_node *p;
496:
1.19 schwarze 497: p = node_alloc(m, line, pos, tok, MDOC_BODY);
1.25 schwarze 498: if ( ! node_append(m, p))
499: return(0);
500: m->next = MDOC_NEXT_CHILD;
501: return(1);
1.1 kristaps 502: }
503:
504:
505: int
1.19 schwarze 506: mdoc_block_alloc(struct mdoc *m, int line, int pos,
1.37 schwarze 507: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 508: {
509: struct mdoc_node *p;
510:
1.19 schwarze 511: p = node_alloc(m, line, pos, tok, MDOC_BLOCK);
1.4 schwarze 512: p->args = args;
513: if (p->args)
1.1 kristaps 514: (args->refcnt)++;
1.25 schwarze 515: if ( ! node_append(m, p))
516: return(0);
517: m->next = MDOC_NEXT_CHILD;
518: return(1);
1.1 kristaps 519: }
520:
521:
522: int
1.19 schwarze 523: mdoc_elem_alloc(struct mdoc *m, int line, int pos,
1.37 schwarze 524: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 525: {
526: struct mdoc_node *p;
527:
1.19 schwarze 528: p = node_alloc(m, line, pos, tok, MDOC_ELEM);
1.4 schwarze 529: p->args = args;
530: if (p->args)
1.1 kristaps 531: (args->refcnt)++;
1.25 schwarze 532: if ( ! node_append(m, p))
533: return(0);
534: m->next = MDOC_NEXT_CHILD;
535: return(1);
1.1 kristaps 536: }
537:
538:
1.46 schwarze 539: int
540: mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p)
1.1 kristaps 541: {
1.19 schwarze 542: struct mdoc_node *n;
1.46 schwarze 543: size_t sv, len;
544:
545: len = strlen(p);
1.1 kristaps 546:
1.46 schwarze 547: n = node_alloc(m, line, pos, MDOC_MAX, MDOC_TEXT);
1.32 schwarze 548: n->string = mandoc_malloc(len + 1);
1.19 schwarze 549: sv = strlcpy(n->string, p, len + 1);
550:
551: /* Prohibit truncation. */
552: assert(sv < len + 1);
553:
1.25 schwarze 554: if ( ! node_append(m, n))
555: return(0);
1.46 schwarze 556:
1.25 schwarze 557: m->next = MDOC_NEXT_SIBLING;
558: return(1);
1.19 schwarze 559: }
560:
561:
1.1 kristaps 562: void
563: mdoc_node_free(struct mdoc_node *p)
564: {
565:
566: if (p->string)
567: free(p->string);
568: if (p->args)
569: mdoc_argv_free(p->args);
570: free(p);
571: }
572:
573:
1.40 schwarze 574: static void
575: mdoc_node_unlink(struct mdoc *m, struct mdoc_node *n)
576: {
577:
578: /* Adjust siblings. */
579:
580: if (n->prev)
581: n->prev->next = n->next;
582: if (n->next)
583: n->next->prev = n->prev;
584:
585: /* Adjust parent. */
586:
587: if (n->parent) {
588: n->parent->nchild--;
589: if (n->parent->child == n)
590: n->parent->child = n->prev ? n->prev : n->next;
591: }
592:
593: /* Adjust parse point, if applicable. */
594:
595: if (m && m->last == n) {
596: if (n->prev) {
597: m->last = n->prev;
598: m->next = MDOC_NEXT_SIBLING;
599: } else {
600: m->last = n->parent;
601: m->next = MDOC_NEXT_CHILD;
602: }
603: }
604:
605: if (m && m->first == n)
606: m->first = NULL;
607: }
608:
609:
1.1 kristaps 610: void
1.40 schwarze 611: mdoc_node_delete(struct mdoc *m, struct mdoc_node *p)
1.1 kristaps 612: {
613:
1.40 schwarze 614: while (p->child) {
615: assert(p->nchild);
616: mdoc_node_delete(m, p->child);
617: }
618: assert(0 == p->nchild);
1.1 kristaps 619:
1.40 schwarze 620: mdoc_node_unlink(m, p);
1.1 kristaps 621: mdoc_node_free(p);
622: }
623:
624:
625: /*
626: * Parse free-form text, that is, a line that does not begin with the
627: * control character.
628: */
629: static int
1.53 ! schwarze 630: mdoc_ptext(struct mdoc *m, int line, char *buf, int offs)
1.1 kristaps 631: {
1.52 schwarze 632: char *c, *ws, *end;
1.44 schwarze 633:
634: /* Ignore bogus comments. */
635:
1.53 ! schwarze 636: if ('\\' == buf[offs] &&
! 637: '.' == buf[offs + 1] &&
! 638: '"' == buf[offs + 2])
! 639: return(mdoc_pwarn(m, line, offs, EBADCOMMENT));
1.1 kristaps 640:
1.46 schwarze 641: /* No text before an initial macro. */
642:
1.9 schwarze 643: if (SEC_NONE == m->lastnamed)
1.53 ! schwarze 644: return(mdoc_perr(m, line, offs, ETEXTPROL));
1.46 schwarze 645:
1.52 schwarze 646: /*
647: * Search for the beginning of unescaped trailing whitespace (ws)
648: * and for the first character not to be output (end).
649: */
650: ws = NULL;
1.53 ! schwarze 651: for (c = end = buf + offs; *c; c++) {
1.52 schwarze 652: switch (*c) {
653: case ' ':
654: if (NULL == ws)
655: ws = c;
656: continue;
657: case '\t':
658: /*
659: * Always warn about trailing tabs,
660: * even outside literal context,
661: * where they should be put on the next line.
662: */
663: if (NULL == ws)
664: ws = c;
665: /*
666: * Strip trailing tabs in literal context only;
667: * outside, they affect the next line.
668: */
669: if (MDOC_LITERAL & m->flags)
670: continue;
671: break;
672: case '\\':
673: /* Skip the escaped character, too, if any. */
674: if (c[1])
675: c++;
676: /* FALLTHROUGH */
677: default:
678: ws = NULL;
679: break;
680: }
681: end = c + 1;
682: }
683: *end = '\0';
1.19 schwarze 684:
1.52 schwarze 685: if (ws)
686: if ( ! mdoc_pwarn(m, line, (int)(ws-buf), ETAILWS))
687: return(0);
1.34 schwarze 688:
1.53 ! schwarze 689: if ('\0' == buf[offs] && ! (MDOC_LITERAL & m->flags)) {
1.52 schwarze 690: if ( ! mdoc_pwarn(m, line, (int)(c-buf), ENOBLANK))
1.38 schwarze 691: return(0);
1.46 schwarze 692:
1.40 schwarze 693: /*
1.46 schwarze 694: * Insert a `Pp' in the case of a blank line. Technically,
695: * blank lines aren't allowed, but enough manuals assume this
696: * behaviour that we want to work around it.
1.40 schwarze 697: */
1.53 ! schwarze 698: if ( ! mdoc_elem_alloc(m, line, offs, MDOC_Pp, NULL))
1.42 schwarze 699: return(0);
1.46 schwarze 700:
1.42 schwarze 701: m->next = MDOC_NEXT_SIBLING;
702: return(1);
1.38 schwarze 703: }
1.1 kristaps 704:
1.53 ! schwarze 705: if ( ! mdoc_word_alloc(m, line, offs, buf+offs))
1.1 kristaps 706: return(0);
707:
1.52 schwarze 708: if (MDOC_LITERAL & m->flags)
709: return(1);
710:
1.35 schwarze 711: /*
1.48 schwarze 712: * End-of-sentence check. If the last character is an unescaped
713: * EOS character, then flag the node as being the end of a
714: * sentence. The front-end will know how to interpret this.
1.35 schwarze 715: */
716:
1.52 schwarze 717: assert(buf < end);
1.48 schwarze 718:
1.53 ! schwarze 719: if (mandoc_eos(buf+offs, (size_t)(end-buf-offs)))
1.48 schwarze 720: m->last->flags |= MDOC_EOS;
721:
1.1 kristaps 722: return(1);
723: }
1.19 schwarze 724:
1.1 kristaps 725:
726: static int
1.53 ! schwarze 727: macrowarn(struct mdoc *m, int ln, const char *buf, int offs)
1.1 kristaps 728: {
729: if ( ! (MDOC_IGN_MACRO & m->pflags))
1.53 ! schwarze 730: return(mdoc_verr(m, ln, offs, "unknown macro: %s%s",
1.1 kristaps 731: buf, strlen(buf) > 3 ? "..." : ""));
1.53 ! schwarze 732: return(mdoc_vwarn(m, ln, offs, "unknown macro: %s%s",
1.1 kristaps 733: buf, strlen(buf) > 3 ? "..." : ""));
734: }
735:
736:
737: /*
738: * Parse a macro line, that is, a line beginning with the control
739: * character.
740: */
741: int
1.53 ! schwarze 742: mdoc_pmacro(struct mdoc *m, int ln, char *buf, int offs)
1.1 kristaps 743: {
1.46 schwarze 744: enum mdoct tok;
1.47 schwarze 745: int i, j, sv;
1.46 schwarze 746: char mac[5];
1.1 kristaps 747:
1.7 schwarze 748: /* Empty lines are ignored. */
1.1 kristaps 749:
1.53 ! schwarze 750: offs++;
! 751:
! 752: if ('\0' == buf[offs])
1.1 kristaps 753: return(1);
754:
1.53 ! schwarze 755: i = offs;
1.24 schwarze 756:
757: /* Accept whitespace after the initial control char. */
758:
759: if (' ' == buf[i]) {
760: i++;
1.1 kristaps 761: while (buf[i] && ' ' == buf[i])
762: i++;
1.34 schwarze 763: if ('\0' == buf[i])
1.1 kristaps 764: return(1);
765: }
766:
1.47 schwarze 767: sv = i;
768:
1.1 kristaps 769: /* Copy the first word into a nil-terminated buffer. */
770:
1.24 schwarze 771: for (j = 0; j < 4; j++, i++) {
1.34 schwarze 772: if ('\0' == (mac[j] = buf[i]))
1.1 kristaps 773: break;
774: else if (' ' == buf[i])
775: break;
1.26 schwarze 776:
777: /* Check for invalid characters. */
778:
779: if (isgraph((u_char)buf[i]))
780: continue;
781: return(mdoc_perr(m, ln, i, EPRINT));
1.1 kristaps 782: }
783:
1.53 ! schwarze 784: mac[j] = '\0';
1.1 kristaps 785:
1.24 schwarze 786: if (j == 4 || j < 2) {
1.53 ! schwarze 787: if ( ! macrowarn(m, ln, mac, sv))
1.1 kristaps 788: goto err;
789: return(1);
790: }
791:
1.46 schwarze 792: if (MDOC_MAX == (tok = mdoc_hash_find(mac))) {
1.53 ! schwarze 793: if ( ! macrowarn(m, ln, mac, sv))
1.1 kristaps 794: goto err;
795: return(1);
796: }
797:
798: /* The macro is sane. Jump to the next word. */
799:
800: while (buf[i] && ' ' == buf[i])
801: i++;
1.34 schwarze 802:
1.46 schwarze 803: /*
804: * Trailing whitespace. Note that tabs are allowed to be passed
805: * into the parser as "text", so we only warn about spaces here.
806: */
1.34 schwarze 807:
808: if ('\0' == buf[i] && ' ' == buf[i - 1])
809: if ( ! mdoc_pwarn(m, ln, i - 1, ETAILWS))
810: goto err;
1.1 kristaps 811:
1.24 schwarze 812: /*
813: * Begin recursive parse sequence. Since we're at the start of
814: * the line, we don't need to do callable/parseable checks.
815: */
1.47 schwarze 816: if ( ! mdoc_macro(m, tok, ln, sv, &i, buf))
1.1 kristaps 817: goto err;
818:
819: return(1);
820:
821: err: /* Error out. */
822:
823: m->flags |= MDOC_HALT;
824: return(0);
825: }
1.24 schwarze 826:
827: