Annotation of src/usr.bin/mandoc/mdoc.c, Revision 1.41
1.41 ! schwarze 1: /* $Id: mdoc.c,v 1.40 2010/04/07 23:15:05 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 */
67: "invalid manual section", /* EBADMSEC */
68: "invalid section", /* EBADSEC */
69: "invalid font mode", /* EFONT */
70: "invalid date syntax", /* EBADDATE */
71: "invalid number format", /* ENUMFMT */
72: "superfluous width argument", /* ENOWIDTH */
73: "system: utsname error", /* EUTSNAME */
74: "obsolete macro", /* EOBS */
75: "end-of-line scope violation", /* EIMPBRK */
76: "empty macro ignored", /* EIGNE */
77: "unclosed explicit scope", /* EOPEN */
78: "unterminated quoted phrase", /* EQUOTPHR */
79: "closure macro without prior context", /* ENOCTX */
1.29 schwarze 80: "no description found for library", /* ELIB */
81: "bad child for parent context", /* EBADCHILD */
1.30 schwarze 82: "list arguments preceding type", /* ENOTYPE */
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 */
1.35 schwarze 122: "%U", "eos"
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 *);
149: static int parsetext(struct mdoc *, int, char *);
150: static int parsemacro(struct mdoc *, int, char *);
151: static int macrowarn(struct mdoc *, int, const char *);
1.19 schwarze 152: static int pstring(struct mdoc *, int, int,
153: const char *, size_t);
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);
190: }
191:
192:
1.8 schwarze 193: /*
194: * Allocate all volatile resources (parse tree, meta-data, fields).
195: */
1.32 schwarze 196: static void
1.1 kristaps 197: mdoc_alloc1(struct mdoc *mdoc)
198: {
199:
1.32 schwarze 200: memset(&mdoc->meta, 0, sizeof(struct mdoc_meta));
1.1 kristaps 201: mdoc->flags = 0;
1.9 schwarze 202: mdoc->lastnamed = mdoc->lastsec = SEC_NONE;
1.32 schwarze 203: mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node));
1.1 kristaps 204: mdoc->first = mdoc->last;
205: mdoc->last->type = MDOC_ROOT;
206: mdoc->next = MDOC_NEXT_CHILD;
207: }
208:
209:
210: /*
1.8 schwarze 211: * Free up volatile resources (see mdoc_free1()) then re-initialises the
212: * data with mdoc_alloc1(). After invocation, parse data has been reset
213: * and the parser is ready for re-invocation on a new tree; however,
214: * cross-parse non-volatile data is kept intact.
1.1 kristaps 215: */
1.32 schwarze 216: void
1.1 kristaps 217: mdoc_reset(struct mdoc *mdoc)
218: {
219:
220: mdoc_free1(mdoc);
1.32 schwarze 221: mdoc_alloc1(mdoc);
1.1 kristaps 222: }
223:
224:
225: /*
1.8 schwarze 226: * Completely free up all volatile and non-volatile parse resources.
227: * After invocation, the pointer is no longer usable.
1.1 kristaps 228: */
229: void
230: mdoc_free(struct mdoc *mdoc)
231: {
232:
233: mdoc_free1(mdoc);
234: free(mdoc);
235: }
236:
237:
1.8 schwarze 238: /*
239: * Allocate volatile and non-volatile parse resources.
240: */
1.1 kristaps 241: struct mdoc *
242: mdoc_alloc(void *data, int pflags, const struct mdoc_cb *cb)
243: {
244: struct mdoc *p;
245:
1.32 schwarze 246: p = mandoc_calloc(1, sizeof(struct mdoc));
247:
1.1 kristaps 248: if (cb)
1.32 schwarze 249: memcpy(&p->cb, cb, sizeof(struct mdoc_cb));
1.27 schwarze 250:
1.1 kristaps 251: p->data = data;
252: p->pflags = pflags;
253:
1.32 schwarze 254: mdoc_hash_init();
255: mdoc_alloc1(p);
256: return(p);
1.1 kristaps 257: }
258:
259:
260: /*
261: * Climb back up the parse tree, validating open scopes. Mostly calls
1.8 schwarze 262: * through to macro_end() in macro.c.
1.1 kristaps 263: */
264: int
265: mdoc_endparse(struct mdoc *m)
266: {
267:
268: if (MDOC_HALT & m->flags)
269: return(0);
270: else if (mdoc_macroend(m))
271: return(1);
272: m->flags |= MDOC_HALT;
273: return(0);
274: }
275:
276:
277: /*
278: * Main parse routine. Parses a single line -- really just hands off to
1.8 schwarze 279: * the macro (parsemacro()) or text parser (parsetext()).
1.1 kristaps 280: */
281: int
282: mdoc_parseln(struct mdoc *m, int ln, char *buf)
283: {
284:
285: if (MDOC_HALT & m->flags)
286: return(0);
287:
288: return('.' == *buf ? parsemacro(m, ln, buf) :
289: parsetext(m, ln, buf));
290: }
291:
292:
293: int
1.4 schwarze 294: mdoc_verr(struct mdoc *mdoc, int ln, int pos,
295: const char *fmt, ...)
1.1 kristaps 296: {
297: char buf[256];
298: va_list ap;
299:
300: if (NULL == mdoc->cb.mdoc_err)
301: return(0);
302:
303: va_start(ap, fmt);
304: (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
305: va_end(ap);
1.12 schwarze 306:
1.1 kristaps 307: return((*mdoc->cb.mdoc_err)(mdoc->data, ln, pos, buf));
308: }
309:
310:
311: int
1.12 schwarze 312: mdoc_vwarn(struct mdoc *mdoc, int ln, int pos, const char *fmt, ...)
1.1 kristaps 313: {
314: char buf[256];
315: va_list ap;
316:
317: if (NULL == mdoc->cb.mdoc_warn)
318: return(0);
319:
320: va_start(ap, fmt);
321: (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
322: va_end(ap);
1.17 schwarze 323:
1.11 schwarze 324: return((*mdoc->cb.mdoc_warn)(mdoc->data, ln, pos, buf));
1.5 schwarze 325: }
326:
327:
328: int
1.14 schwarze 329: mdoc_err(struct mdoc *m, int line, int pos, int iserr, enum merr type)
1.5 schwarze 330: {
1.14 schwarze 331: const char *p;
1.5 schwarze 332:
1.14 schwarze 333: p = __mdoc_merrnames[(int)type];
334: assert(p);
1.5 schwarze 335:
1.14 schwarze 336: if (iserr)
337: return(mdoc_verr(m, line, pos, p));
1.16 schwarze 338:
1.14 schwarze 339: return(mdoc_vwarn(m, line, pos, p));
1.5 schwarze 340: }
341:
342:
343: int
1.37 schwarze 344: mdoc_macro(struct mdoc *m, enum mdoct tok,
1.1 kristaps 345: int ln, int pp, int *pos, char *buf)
346: {
1.37 schwarze 347:
348: assert(tok < MDOC_MAX);
1.24 schwarze 349: /*
350: * If we're in the prologue, deny "body" macros. Similarly, if
351: * we're in the body, deny prologue calls.
352: */
1.1 kristaps 353: if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
1.9 schwarze 354: MDOC_PBODY & m->flags)
1.16 schwarze 355: return(mdoc_perr(m, ln, pp, EPROLBODY));
1.1 kristaps 356: if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
1.39 schwarze 357: ! (MDOC_PBODY & m->flags)) {
358: if ( ! mdoc_pwarn(m, ln, pp, EBODYPROL))
359: return(0);
360: if (NULL == m->meta.title)
361: m->meta.title = mandoc_strdup("unknown");
362: if (NULL == m->meta.vol)
363: m->meta.vol = mandoc_strdup("local");
364: if (NULL == m->meta.os)
365: m->meta.os = mandoc_strdup("local");
366: if (0 == m->meta.date)
367: m->meta.date = time(NULL);
368: m->flags |= MDOC_PBODY;
369: }
1.1 kristaps 370:
371: return((*mdoc_macros[tok].fp)(m, tok, ln, pp, pos, buf));
372: }
373:
374:
375: static int
376: node_append(struct mdoc *mdoc, struct mdoc_node *p)
377: {
378:
379: assert(mdoc->last);
380: assert(mdoc->first);
381: assert(MDOC_ROOT != p->type);
382:
383: switch (mdoc->next) {
384: case (MDOC_NEXT_SIBLING):
385: mdoc->last->next = p;
386: p->prev = mdoc->last;
387: p->parent = mdoc->last->parent;
388: break;
389: case (MDOC_NEXT_CHILD):
390: mdoc->last->child = p;
391: p->parent = mdoc->last;
392: break;
393: default:
394: abort();
395: /* NOTREACHED */
396: }
397:
1.10 schwarze 398: p->parent->nchild++;
399:
1.1 kristaps 400: if ( ! mdoc_valid_pre(mdoc, p))
401: return(0);
402: if ( ! mdoc_action_pre(mdoc, p))
403: return(0);
404:
405: switch (p->type) {
406: case (MDOC_HEAD):
407: assert(MDOC_BLOCK == p->parent->type);
408: p->parent->head = p;
409: break;
410: case (MDOC_TAIL):
411: assert(MDOC_BLOCK == p->parent->type);
412: p->parent->tail = p;
413: break;
414: case (MDOC_BODY):
415: assert(MDOC_BLOCK == p->parent->type);
416: p->parent->body = p;
417: break;
418: default:
419: break;
420: }
421:
422: mdoc->last = p;
423:
424: switch (p->type) {
425: case (MDOC_TEXT):
426: if ( ! mdoc_valid_post(mdoc))
427: return(0);
428: if ( ! mdoc_action_post(mdoc))
429: return(0);
430: break;
431: default:
432: break;
433: }
434:
435: return(1);
436: }
437:
438:
439: static struct mdoc_node *
1.37 schwarze 440: node_alloc(struct mdoc *m, int line, int pos,
441: enum mdoct tok, enum mdoc_type type)
1.1 kristaps 442: {
443: struct mdoc_node *p;
444:
1.32 schwarze 445: p = mandoc_calloc(1, sizeof(struct mdoc_node));
1.19 schwarze 446: p->sec = m->lastsec;
1.1 kristaps 447: p->line = line;
448: p->pos = pos;
449: p->tok = tok;
1.37 schwarze 450: p->type = type;
1.1 kristaps 451:
452: return(p);
453: }
454:
455:
456: int
1.37 schwarze 457: mdoc_tail_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
1.1 kristaps 458: {
459: struct mdoc_node *p;
460:
1.19 schwarze 461: p = node_alloc(m, line, pos, tok, MDOC_TAIL);
1.25 schwarze 462: if ( ! node_append(m, p))
463: return(0);
464: m->next = MDOC_NEXT_CHILD;
465: return(1);
1.1 kristaps 466: }
467:
468:
469: int
1.37 schwarze 470: mdoc_head_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
1.1 kristaps 471: {
472: struct mdoc_node *p;
473:
1.19 schwarze 474: assert(m->first);
475: assert(m->last);
1.1 kristaps 476:
1.19 schwarze 477: p = node_alloc(m, line, pos, tok, MDOC_HEAD);
1.25 schwarze 478: if ( ! node_append(m, p))
479: return(0);
480: m->next = MDOC_NEXT_CHILD;
481: return(1);
1.1 kristaps 482: }
483:
484:
485: int
1.37 schwarze 486: mdoc_body_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
1.1 kristaps 487: {
488: struct mdoc_node *p;
489:
1.19 schwarze 490: p = node_alloc(m, line, pos, tok, MDOC_BODY);
1.25 schwarze 491: if ( ! node_append(m, p))
492: return(0);
493: m->next = MDOC_NEXT_CHILD;
494: return(1);
1.1 kristaps 495: }
496:
497:
498: int
1.19 schwarze 499: mdoc_block_alloc(struct mdoc *m, int line, int pos,
1.37 schwarze 500: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 501: {
502: struct mdoc_node *p;
503:
1.19 schwarze 504: p = node_alloc(m, line, pos, tok, MDOC_BLOCK);
1.4 schwarze 505: p->args = args;
506: if (p->args)
1.1 kristaps 507: (args->refcnt)++;
1.25 schwarze 508: if ( ! node_append(m, p))
509: return(0);
510: m->next = MDOC_NEXT_CHILD;
511: return(1);
1.1 kristaps 512: }
513:
514:
515: int
1.19 schwarze 516: mdoc_elem_alloc(struct mdoc *m, int line, int pos,
1.37 schwarze 517: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 518: {
519: struct mdoc_node *p;
520:
1.19 schwarze 521: p = node_alloc(m, line, pos, tok, MDOC_ELEM);
1.4 schwarze 522: p->args = args;
523: if (p->args)
1.1 kristaps 524: (args->refcnt)++;
1.25 schwarze 525: if ( ! node_append(m, p))
526: return(0);
527: m->next = MDOC_NEXT_CHILD;
528: return(1);
1.1 kristaps 529: }
530:
531:
1.19 schwarze 532: static int
533: pstring(struct mdoc *m, int line, int pos, const char *p, size_t len)
1.1 kristaps 534: {
1.19 schwarze 535: struct mdoc_node *n;
536: size_t sv;
1.1 kristaps 537:
1.19 schwarze 538: n = node_alloc(m, line, pos, -1, MDOC_TEXT);
1.32 schwarze 539: n->string = mandoc_malloc(len + 1);
1.19 schwarze 540: sv = strlcpy(n->string, p, len + 1);
541:
542: /* Prohibit truncation. */
543: assert(sv < len + 1);
544:
1.25 schwarze 545: if ( ! node_append(m, n))
546: return(0);
547: m->next = MDOC_NEXT_SIBLING;
548: return(1);
1.19 schwarze 549: }
550:
551:
552: int
553: mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p)
554: {
555:
556: return(pstring(m, line, pos, p, strlen(p)));
1.1 kristaps 557: }
558:
559:
560: void
561: mdoc_node_free(struct mdoc_node *p)
562: {
563:
564: if (p->string)
565: free(p->string);
566: if (p->args)
567: mdoc_argv_free(p->args);
568: free(p);
569: }
570:
571:
1.40 schwarze 572: static void
573: mdoc_node_unlink(struct mdoc *m, struct mdoc_node *n)
574: {
575:
576: /* Adjust siblings. */
577:
578: if (n->prev)
579: n->prev->next = n->next;
580: if (n->next)
581: n->next->prev = n->prev;
582:
583: /* Adjust parent. */
584:
585: if (n->parent) {
586: n->parent->nchild--;
587: if (n->parent->child == n)
588: n->parent->child = n->prev ? n->prev : n->next;
589: }
590:
591: /* Adjust parse point, if applicable. */
592:
593: if (m && m->last == n) {
594: if (n->prev) {
595: m->last = n->prev;
596: m->next = MDOC_NEXT_SIBLING;
597: } else {
598: m->last = n->parent;
599: m->next = MDOC_NEXT_CHILD;
600: }
601: }
602:
603: if (m && m->first == n)
604: m->first = NULL;
605: }
606:
607:
1.1 kristaps 608: void
1.40 schwarze 609: mdoc_node_delete(struct mdoc *m, struct mdoc_node *p)
1.1 kristaps 610: {
611:
1.40 schwarze 612: while (p->child) {
613: assert(p->nchild);
614: mdoc_node_delete(m, p->child);
615: }
616: assert(0 == p->nchild);
1.1 kristaps 617:
1.40 schwarze 618: mdoc_node_unlink(m, p);
1.1 kristaps 619: mdoc_node_free(p);
620: }
621:
622:
623: /*
624: * Parse free-form text, that is, a line that does not begin with the
625: * control character.
626: */
627: static int
628: parsetext(struct mdoc *m, int line, char *buf)
629: {
1.19 schwarze 630: int i, j;
1.34 schwarze 631: char sv;
1.1 kristaps 632:
1.9 schwarze 633: if (SEC_NONE == m->lastnamed)
1.16 schwarze 634: return(mdoc_perr(m, line, 0, ETEXTPROL));
1.19 schwarze 635:
636: /*
637: * If in literal mode, then pass the buffer directly to the
638: * back-end, as it should be preserved as a single term.
639: */
1.1 kristaps 640:
1.25 schwarze 641: if (MDOC_LITERAL & m->flags)
642: return(mdoc_word_alloc(m, line, 0, buf));
1.19 schwarze 643:
644: /* Disallow blank/white-space lines in non-literal mode. */
645:
646: for (i = 0; ' ' == buf[i]; i++)
647: /* Skip leading whitespace. */ ;
1.34 schwarze 648:
1.38 schwarze 649: if ('\0' == buf[i]) {
650: if ( ! mdoc_pwarn(m, line, 0, ENOBLANK))
651: return(0);
1.40 schwarze 652: /*
653: * Assume that a `Pp' should be inserted in the case of
654: * a blank line. Technically, blank lines aren't
655: * allowed, but enough manuals assume this behaviour
656: * that we want to work around it.
657: */
1.41 ! schwarze 658: return(mdoc_elem_alloc(m, line, 0, MDOC_Pp, NULL));
1.38 schwarze 659: }
1.1 kristaps 660:
1.19 schwarze 661: /*
662: * Break apart a free-form line into tokens. Spaces are
663: * stripped out of the input.
664: */
665:
666: for (j = i; buf[i]; i++) {
667: if (' ' != buf[i])
668: continue;
669:
670: /* Escaped whitespace. */
671: if (i && ' ' == buf[i] && '\\' == buf[i - 1])
672: continue;
673:
1.34 schwarze 674: sv = buf[i];
675: buf[i++] = '\0';
676:
1.19 schwarze 677: if ( ! pstring(m, line, j, &buf[j], (size_t)(i - j)))
678: return(0);
679:
1.34 schwarze 680: /* Trailing whitespace? Check at overwritten byte. */
681:
682: if (' ' == sv && '\0' == buf[i])
683: if ( ! mdoc_pwarn(m, line, i - 1, ETAILWS))
684: return(0);
685:
1.19 schwarze 686: for ( ; ' ' == buf[i]; i++)
687: /* Skip trailing whitespace. */ ;
688:
689: j = i;
1.34 schwarze 690:
691: /* Trailing whitespace? */
692:
693: if (' ' == buf[i - 1] && '\0' == buf[i])
694: if ( ! mdoc_pwarn(m, line, i - 1, ETAILWS))
695: return(0);
696:
697: if ('\0' == buf[i])
1.19 schwarze 698: break;
699: }
700:
701: if (j != i && ! pstring(m, line, j, &buf[j], (size_t)(i - j)))
1.1 kristaps 702: return(0);
703:
1.35 schwarze 704: /*
705: * Mark the end of a sentence. Only works when you respect
706: * Jason's rule: "new sentence, new line".
707: */
708: if ('.' == buf[i-1] || '!' == buf[i-1] || '?' == buf[i-1]) {
709: m->next = MDOC_NEXT_SIBLING;
710: if ( ! mdoc_elem_alloc(m, line, i, MDOC_eos, NULL))
711: return(0);
712: }
713:
1.1 kristaps 714: m->next = MDOC_NEXT_SIBLING;
715: return(1);
716: }
1.19 schwarze 717:
1.1 kristaps 718:
719:
720: static int
721: macrowarn(struct mdoc *m, int ln, const char *buf)
722: {
723: if ( ! (MDOC_IGN_MACRO & m->pflags))
1.23 schwarze 724: return(mdoc_verr(m, ln, 0,
1.1 kristaps 725: "unknown macro: %s%s",
726: buf, strlen(buf) > 3 ? "..." : ""));
1.23 schwarze 727: return(mdoc_vwarn(m, ln, 0, "unknown macro: %s%s",
1.1 kristaps 728: buf, strlen(buf) > 3 ? "..." : ""));
729: }
730:
731:
732: /*
733: * Parse a macro line, that is, a line beginning with the control
734: * character.
735: */
736: int
737: parsemacro(struct mdoc *m, int ln, char *buf)
738: {
1.25 schwarze 739: int i, j, c;
1.1 kristaps 740: char mac[5];
1.35 schwarze 741: struct mdoc_node *n;
742: char *t;
1.1 kristaps 743:
1.7 schwarze 744: /* Empty lines are ignored. */
1.1 kristaps 745:
1.34 schwarze 746: if ('\0' == buf[1])
1.1 kristaps 747: return(1);
748:
1.24 schwarze 749: i = 1;
750:
751: /* Accept whitespace after the initial control char. */
752:
753: if (' ' == buf[i]) {
754: i++;
1.1 kristaps 755: while (buf[i] && ' ' == buf[i])
756: i++;
1.34 schwarze 757: if ('\0' == buf[i])
1.1 kristaps 758: return(1);
759: }
760:
761: /* Copy the first word into a nil-terminated buffer. */
762:
1.24 schwarze 763: for (j = 0; j < 4; j++, i++) {
1.34 schwarze 764: if ('\0' == (mac[j] = buf[i]))
1.1 kristaps 765: break;
766: else if (' ' == buf[i])
767: break;
1.26 schwarze 768:
769: /* Check for invalid characters. */
770:
771: if (isgraph((u_char)buf[i]))
772: continue;
773: return(mdoc_perr(m, ln, i, EPRINT));
1.1 kristaps 774: }
775:
1.24 schwarze 776: mac[j] = 0;
1.1 kristaps 777:
1.24 schwarze 778: if (j == 4 || j < 2) {
1.1 kristaps 779: if ( ! macrowarn(m, ln, mac))
780: goto err;
781: return(1);
782: }
783:
1.27 schwarze 784: if (MDOC_MAX == (c = mdoc_hash_find(mac))) {
1.1 kristaps 785: if ( ! macrowarn(m, ln, mac))
786: goto err;
787: return(1);
788: }
789:
790: /* The macro is sane. Jump to the next word. */
791:
792: while (buf[i] && ' ' == buf[i])
793: i++;
1.34 schwarze 794:
795: /* Trailing whitespace? */
796:
797: if ('\0' == buf[i] && ' ' == buf[i - 1])
798: if ( ! mdoc_pwarn(m, ln, i - 1, ETAILWS))
799: goto err;
1.1 kristaps 800:
1.24 schwarze 801: /*
802: * Begin recursive parse sequence. Since we're at the start of
803: * the line, we don't need to do callable/parseable checks.
804: */
1.25 schwarze 805: if ( ! mdoc_macro(m, c, ln, 1, &i, buf))
1.1 kristaps 806: goto err;
1.35 schwarze 807:
808: /*
809: * Mark the end of a sentence, but be careful not to insert
810: * markers into reference blocks.
811: */
812: n = m->last;
813: if (n->child)
814: n = n->child;
815: while (n->next)
816: n = n->next;
817: if (MDOC_TEXT == n->type && m->last->parent->tok != MDOC_Rs) {
818: t = n->string;
1.36 schwarze 819: while (t[0] && t[1])
1.35 schwarze 820: t++;
821: if ('.' == *t || '!' == *t || '?' == *t) {
822: if ( ! mdoc_elem_alloc(m, ln, i, MDOC_eos, NULL))
823: return(0);
824: m->next = MDOC_NEXT_SIBLING;
825: }
826: }
1.1 kristaps 827:
828: return(1);
829:
830: err: /* Error out. */
831:
832: m->flags |= MDOC_HALT;
833: return(0);
834: }
1.24 schwarze 835:
836: