Annotation of src/usr.bin/mandoc/mdoc.c, Revision 1.46
1.46 ! schwarze 1: /* $Id: mdoc.c,v 1.45 2010/05/08 01:57:33 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 */
1.44 schwarze 83: "deprecated comment style", /* EBADCOMMENT */
1.1 kristaps 84: };
85:
86: const char *const __mdoc_macronames[MDOC_MAX] = {
1.7 schwarze 87: "Ap", "Dd", "Dt", "Os",
1.1 kristaps 88: "Sh", "Ss", "Pp", "D1",
89: "Dl", "Bd", "Ed", "Bl",
90: "El", "It", "Ad", "An",
91: "Ar", "Cd", "Cm", "Dv",
92: "Er", "Ev", "Ex", "Fa",
93: "Fd", "Fl", "Fn", "Ft",
94: "Ic", "In", "Li", "Nd",
95: "Nm", "Op", "Ot", "Pa",
96: "Rv", "St", "Va", "Vt",
97: /* LINTED */
1.33 schwarze 98: "Xr", "%A", "%B", "%D",
1.1 kristaps 99: /* LINTED */
1.33 schwarze 100: "%I", "%J", "%N", "%O",
1.1 kristaps 101: /* LINTED */
1.33 schwarze 102: "%P", "%R", "%T", "%V",
1.1 kristaps 103: "Ac", "Ao", "Aq", "At",
104: "Bc", "Bf", "Bo", "Bq",
105: "Bsx", "Bx", "Db", "Dc",
106: "Do", "Dq", "Ec", "Ef",
107: "Em", "Eo", "Fx", "Ms",
108: "No", "Ns", "Nx", "Ox",
109: "Pc", "Pf", "Po", "Pq",
110: "Qc", "Ql", "Qo", "Qq",
111: "Re", "Rs", "Sc", "So",
112: "Sq", "Sm", "Sx", "Sy",
113: "Tn", "Ux", "Xc", "Xo",
114: "Fo", "Fc", "Oo", "Oc",
115: "Bk", "Ek", "Bt", "Hf",
1.7 schwarze 116: "Fr", "Ud", "Lb", "Lp",
117: "Lk", "Mt", "Brq", "Bro",
1.1 kristaps 118: /* LINTED */
1.33 schwarze 119: "Brc", "%C", "Es", "En",
1.1 kristaps 120: /* LINTED */
1.33 schwarze 121: "Dx", "%Q", "br", "sp",
1.31 schwarze 122: /* LINTED */
1.35 schwarze 123: "%U", "eos"
1.1 kristaps 124: };
125:
126: const char *const __mdoc_argnames[MDOC_ARG_MAX] = {
127: "split", "nosplit", "ragged",
128: "unfilled", "literal", "file",
129: "offset", "bullet", "dash",
130: "hyphen", "item", "enum",
131: "tag", "diag", "hang",
132: "ohang", "inset", "column",
133: "width", "compact", "std",
134: "filled", "words", "emphasis",
1.30 schwarze 135: "symbolic", "nested", "centered"
1.1 kristaps 136: };
137:
138: const char * const *mdoc_macronames = __mdoc_macronames;
139: const char * const *mdoc_argnames = __mdoc_argnames;
140:
1.40 schwarze 141: static void mdoc_node_free(struct mdoc_node *);
142: static void mdoc_node_unlink(struct mdoc *,
143: struct mdoc_node *);
1.1 kristaps 144: static void mdoc_free1(struct mdoc *);
1.32 schwarze 145: static void mdoc_alloc1(struct mdoc *);
1.1 kristaps 146: static struct mdoc_node *node_alloc(struct mdoc *, int, int,
1.37 schwarze 147: enum mdoct, enum mdoc_type);
1.1 kristaps 148: static int node_append(struct mdoc *,
149: struct mdoc_node *);
1.45 schwarze 150: static int mdoc_ptext(struct mdoc *, int, char *);
151: static int mdoc_pmacro(struct mdoc *, int, char *);
1.1 kristaps 152: static int macrowarn(struct mdoc *, int, const char *);
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);
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
1.45 schwarze 279: * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
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:
1.45 schwarze 288: return('.' == *buf ? mdoc_pmacro(m, ln, buf) :
289: mdoc_ptext(m, ln, buf));
1.1 kristaps 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: {
1.46 ! schwarze 347: assert(tok < MDOC_MAX);
! 348:
! 349: /* If we're in the body, deny prologue calls. */
1.37 schwarze 350:
1.1 kristaps 351: if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
1.9 schwarze 352: MDOC_PBODY & m->flags)
1.16 schwarze 353: return(mdoc_perr(m, ln, pp, EPROLBODY));
1.46 ! schwarze 354:
! 355: /* If we're in the prologue, deny "body" macros. */
! 356:
1.1 kristaps 357: if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
1.39 schwarze 358: ! (MDOC_PBODY & m->flags)) {
359: if ( ! mdoc_pwarn(m, ln, pp, EBODYPROL))
360: return(0);
361: if (NULL == m->meta.title)
362: m->meta.title = mandoc_strdup("unknown");
363: if (NULL == m->meta.vol)
364: m->meta.vol = mandoc_strdup("local");
365: if (NULL == m->meta.os)
366: m->meta.os = mandoc_strdup("local");
367: if (0 == m->meta.date)
368: m->meta.date = time(NULL);
369: m->flags |= MDOC_PBODY;
370: }
1.1 kristaps 371:
372: return((*mdoc_macros[tok].fp)(m, tok, ln, pp, pos, buf));
373: }
374:
375:
376: static int
377: node_append(struct mdoc *mdoc, struct mdoc_node *p)
378: {
379:
380: assert(mdoc->last);
381: assert(mdoc->first);
382: assert(MDOC_ROOT != p->type);
383:
384: switch (mdoc->next) {
385: case (MDOC_NEXT_SIBLING):
386: mdoc->last->next = p;
387: p->prev = mdoc->last;
388: p->parent = mdoc->last->parent;
389: break;
390: case (MDOC_NEXT_CHILD):
391: mdoc->last->child = p;
392: p->parent = mdoc->last;
393: break;
394: default:
395: abort();
396: /* NOTREACHED */
397: }
398:
1.10 schwarze 399: p->parent->nchild++;
400:
1.1 kristaps 401: if ( ! mdoc_valid_pre(mdoc, p))
402: return(0);
403: if ( ! mdoc_action_pre(mdoc, p))
404: return(0);
405:
406: switch (p->type) {
407: case (MDOC_HEAD):
408: assert(MDOC_BLOCK == p->parent->type);
409: p->parent->head = p;
410: break;
411: case (MDOC_TAIL):
412: assert(MDOC_BLOCK == p->parent->type);
413: p->parent->tail = p;
414: break;
415: case (MDOC_BODY):
416: assert(MDOC_BLOCK == p->parent->type);
417: p->parent->body = p;
418: break;
419: default:
420: break;
421: }
422:
423: mdoc->last = p;
424:
425: switch (p->type) {
426: case (MDOC_TEXT):
427: if ( ! mdoc_valid_post(mdoc))
428: return(0);
429: if ( ! mdoc_action_post(mdoc))
430: return(0);
431: break;
432: default:
433: break;
434: }
435:
436: return(1);
437: }
438:
439:
440: static struct mdoc_node *
1.37 schwarze 441: node_alloc(struct mdoc *m, int line, int pos,
442: enum mdoct tok, enum mdoc_type type)
1.1 kristaps 443: {
444: struct mdoc_node *p;
445:
1.32 schwarze 446: p = mandoc_calloc(1, sizeof(struct mdoc_node));
1.19 schwarze 447: p->sec = m->lastsec;
1.1 kristaps 448: p->line = line;
449: p->pos = pos;
450: p->tok = tok;
1.37 schwarze 451: p->type = type;
1.1 kristaps 452:
453: return(p);
454: }
455:
456:
457: int
1.37 schwarze 458: mdoc_tail_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
1.1 kristaps 459: {
460: struct mdoc_node *p;
461:
1.19 schwarze 462: p = node_alloc(m, line, pos, tok, MDOC_TAIL);
1.25 schwarze 463: if ( ! node_append(m, p))
464: return(0);
465: m->next = MDOC_NEXT_CHILD;
466: return(1);
1.1 kristaps 467: }
468:
469:
470: int
1.37 schwarze 471: mdoc_head_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
1.1 kristaps 472: {
473: struct mdoc_node *p;
474:
1.19 schwarze 475: assert(m->first);
476: assert(m->last);
1.1 kristaps 477:
1.19 schwarze 478: p = node_alloc(m, line, pos, tok, MDOC_HEAD);
1.25 schwarze 479: if ( ! node_append(m, p))
480: return(0);
481: m->next = MDOC_NEXT_CHILD;
482: return(1);
1.1 kristaps 483: }
484:
485:
486: int
1.37 schwarze 487: mdoc_body_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
1.1 kristaps 488: {
489: struct mdoc_node *p;
490:
1.19 schwarze 491: p = node_alloc(m, line, pos, tok, MDOC_BODY);
1.25 schwarze 492: if ( ! node_append(m, p))
493: return(0);
494: m->next = MDOC_NEXT_CHILD;
495: return(1);
1.1 kristaps 496: }
497:
498:
499: int
1.19 schwarze 500: mdoc_block_alloc(struct mdoc *m, int line, int pos,
1.37 schwarze 501: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 502: {
503: struct mdoc_node *p;
504:
1.19 schwarze 505: p = node_alloc(m, line, pos, tok, MDOC_BLOCK);
1.4 schwarze 506: p->args = args;
507: if (p->args)
1.1 kristaps 508: (args->refcnt)++;
1.25 schwarze 509: if ( ! node_append(m, p))
510: return(0);
511: m->next = MDOC_NEXT_CHILD;
512: return(1);
1.1 kristaps 513: }
514:
515:
516: int
1.19 schwarze 517: mdoc_elem_alloc(struct mdoc *m, int line, int pos,
1.37 schwarze 518: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 519: {
520: struct mdoc_node *p;
521:
1.19 schwarze 522: p = node_alloc(m, line, pos, tok, MDOC_ELEM);
1.4 schwarze 523: p->args = args;
524: if (p->args)
1.1 kristaps 525: (args->refcnt)++;
1.25 schwarze 526: if ( ! node_append(m, p))
527: return(0);
528: m->next = MDOC_NEXT_CHILD;
529: return(1);
1.1 kristaps 530: }
531:
532:
1.46 ! schwarze 533: int
! 534: mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p)
1.1 kristaps 535: {
1.19 schwarze 536: struct mdoc_node *n;
1.46 ! schwarze 537: size_t sv, len;
! 538:
! 539: len = strlen(p);
1.1 kristaps 540:
1.46 ! schwarze 541: n = node_alloc(m, line, pos, MDOC_MAX, MDOC_TEXT);
1.32 schwarze 542: n->string = mandoc_malloc(len + 1);
1.19 schwarze 543: sv = strlcpy(n->string, p, len + 1);
544:
545: /* Prohibit truncation. */
546: assert(sv < len + 1);
547:
1.25 schwarze 548: if ( ! node_append(m, n))
549: return(0);
1.46 ! schwarze 550:
1.25 schwarze 551: m->next = MDOC_NEXT_SIBLING;
552: return(1);
1.19 schwarze 553: }
554:
555:
1.1 kristaps 556: void
557: mdoc_node_free(struct mdoc_node *p)
558: {
559:
560: if (p->string)
561: free(p->string);
562: if (p->args)
563: mdoc_argv_free(p->args);
564: free(p);
565: }
566:
567:
1.40 schwarze 568: static void
569: mdoc_node_unlink(struct mdoc *m, struct mdoc_node *n)
570: {
571:
572: /* Adjust siblings. */
573:
574: if (n->prev)
575: n->prev->next = n->next;
576: if (n->next)
577: n->next->prev = n->prev;
578:
579: /* Adjust parent. */
580:
581: if (n->parent) {
582: n->parent->nchild--;
583: if (n->parent->child == n)
584: n->parent->child = n->prev ? n->prev : n->next;
585: }
586:
587: /* Adjust parse point, if applicable. */
588:
589: if (m && m->last == n) {
590: if (n->prev) {
591: m->last = n->prev;
592: m->next = MDOC_NEXT_SIBLING;
593: } else {
594: m->last = n->parent;
595: m->next = MDOC_NEXT_CHILD;
596: }
597: }
598:
599: if (m && m->first == n)
600: m->first = NULL;
601: }
602:
603:
1.1 kristaps 604: void
1.40 schwarze 605: mdoc_node_delete(struct mdoc *m, struct mdoc_node *p)
1.1 kristaps 606: {
607:
1.40 schwarze 608: while (p->child) {
609: assert(p->nchild);
610: mdoc_node_delete(m, p->child);
611: }
612: assert(0 == p->nchild);
1.1 kristaps 613:
1.40 schwarze 614: mdoc_node_unlink(m, p);
1.1 kristaps 615: mdoc_node_free(p);
616: }
617:
618:
619: /*
620: * Parse free-form text, that is, a line that does not begin with the
621: * control character.
622: */
623: static int
1.45 schwarze 624: mdoc_ptext(struct mdoc *m, int line, char *buf)
1.1 kristaps 625: {
1.46 ! schwarze 626: int i;
1.44 schwarze 627:
628: /* Ignore bogus comments. */
629:
630: if ('\\' == buf[0] && '.' == buf[1] && '\"' == buf[2])
631: return(mdoc_pwarn(m, line, 0, EBADCOMMENT));
1.1 kristaps 632:
1.46 ! schwarze 633: /* No text before an initial macro. */
! 634:
1.9 schwarze 635: if (SEC_NONE == m->lastnamed)
1.16 schwarze 636: return(mdoc_perr(m, line, 0, ETEXTPROL));
1.46 ! schwarze 637:
! 638: /* Literal just gets pulled in as-is. */
1.19 schwarze 639:
1.25 schwarze 640: if (MDOC_LITERAL & m->flags)
641: return(mdoc_word_alloc(m, line, 0, buf));
1.19 schwarze 642:
1.46 ! schwarze 643: /* Check for a blank line, which may also consist of spaces. */
1.19 schwarze 644:
645: for (i = 0; ' ' == buf[i]; i++)
1.46 ! schwarze 646: /* Skip to first non-space. */ ;
1.34 schwarze 647:
1.38 schwarze 648: if ('\0' == buf[i]) {
649: if ( ! mdoc_pwarn(m, line, 0, ENOBLANK))
650: return(0);
1.46 ! schwarze 651:
1.40 schwarze 652: /*
1.46 ! schwarze 653: * Insert a `Pp' in the case of a blank line. Technically,
! 654: * blank lines aren't allowed, but enough manuals assume this
! 655: * behaviour that we want to work around it.
1.40 schwarze 656: */
1.42 schwarze 657: if ( ! mdoc_elem_alloc(m, line, 0, MDOC_Pp, NULL))
658: return(0);
1.46 ! schwarze 659:
1.42 schwarze 660: m->next = MDOC_NEXT_SIBLING;
661: return(1);
1.38 schwarze 662: }
1.1 kristaps 663:
1.46 ! schwarze 664: /*
! 665: * Warn if the last un-escaped character is whitespace. Then
! 666: * strip away the remaining spaces (tabs stay!).
1.19 schwarze 667: */
668:
1.46 ! schwarze 669: i = (int)strlen(buf);
! 670: assert(i);
1.34 schwarze 671:
1.46 ! schwarze 672: if (' ' == buf[i - 1] || '\t' == buf[i - 1]) {
! 673: if (i > 1 && '\\' != buf[i - 2])
1.34 schwarze 674: if ( ! mdoc_pwarn(m, line, i - 1, ETAILWS))
675: return(0);
676:
1.46 ! schwarze 677: for (--i; i && ' ' == buf[i]; i--)
! 678: /* Spin back to non-space. */ ;
1.34 schwarze 679:
1.46 ! schwarze 680: /* Jump ahead of escaped whitespace. */
! 681: i += '\\' == buf[i] ? 2 : 1;
1.34 schwarze 682:
1.46 ! schwarze 683: buf[i] = '\0';
1.19 schwarze 684: }
685:
1.46 ! schwarze 686: /* Allocate the whole word. */
! 687: if ( ! mdoc_word_alloc(m, line, 0, buf))
1.1 kristaps 688: return(0);
689:
1.35 schwarze 690: /*
691: * Mark the end of a sentence. Only works when you respect
692: * Jason's rule: "new sentence, new line".
693: */
694: if ('.' == buf[i-1] || '!' == buf[i-1] || '?' == buf[i-1]) {
695: m->next = MDOC_NEXT_SIBLING;
696: if ( ! mdoc_elem_alloc(m, line, i, MDOC_eos, NULL))
697: return(0);
698: }
699:
1.1 kristaps 700: m->next = MDOC_NEXT_SIBLING;
701: return(1);
702: }
1.19 schwarze 703:
1.1 kristaps 704:
705: static int
706: macrowarn(struct mdoc *m, int ln, const char *buf)
707: {
708: if ( ! (MDOC_IGN_MACRO & m->pflags))
1.46 ! schwarze 709: return(mdoc_verr(m, ln, 0, "unknown macro: %s%s",
1.1 kristaps 710: buf, strlen(buf) > 3 ? "..." : ""));
1.23 schwarze 711: return(mdoc_vwarn(m, ln, 0, "unknown macro: %s%s",
1.1 kristaps 712: buf, strlen(buf) > 3 ? "..." : ""));
713: }
714:
715:
716: /*
717: * Parse a macro line, that is, a line beginning with the control
718: * character.
719: */
720: int
1.45 schwarze 721: mdoc_pmacro(struct mdoc *m, int ln, char *buf)
1.1 kristaps 722: {
1.46 ! schwarze 723: enum mdoct tok;
! 724: int i, j;
! 725: char mac[5];
1.35 schwarze 726: struct mdoc_node *n;
727: char *t;
1.1 kristaps 728:
1.7 schwarze 729: /* Empty lines are ignored. */
1.1 kristaps 730:
1.34 schwarze 731: if ('\0' == buf[1])
1.1 kristaps 732: return(1);
733:
1.24 schwarze 734: i = 1;
735:
736: /* Accept whitespace after the initial control char. */
737:
738: if (' ' == buf[i]) {
739: i++;
1.1 kristaps 740: while (buf[i] && ' ' == buf[i])
741: i++;
1.34 schwarze 742: if ('\0' == buf[i])
1.1 kristaps 743: return(1);
744: }
745:
746: /* Copy the first word into a nil-terminated buffer. */
747:
1.24 schwarze 748: for (j = 0; j < 4; j++, i++) {
1.34 schwarze 749: if ('\0' == (mac[j] = buf[i]))
1.1 kristaps 750: break;
751: else if (' ' == buf[i])
752: break;
1.26 schwarze 753:
754: /* Check for invalid characters. */
755:
756: if (isgraph((u_char)buf[i]))
757: continue;
758: return(mdoc_perr(m, ln, i, EPRINT));
1.1 kristaps 759: }
760:
1.24 schwarze 761: mac[j] = 0;
1.1 kristaps 762:
1.24 schwarze 763: if (j == 4 || j < 2) {
1.1 kristaps 764: if ( ! macrowarn(m, ln, mac))
765: goto err;
766: return(1);
767: }
768:
1.46 ! schwarze 769: if (MDOC_MAX == (tok = mdoc_hash_find(mac))) {
1.1 kristaps 770: if ( ! macrowarn(m, ln, mac))
771: goto err;
772: return(1);
773: }
774:
775: /* The macro is sane. Jump to the next word. */
776:
777: while (buf[i] && ' ' == buf[i])
778: i++;
1.34 schwarze 779:
1.46 ! schwarze 780: /*
! 781: * Trailing whitespace. Note that tabs are allowed to be passed
! 782: * into the parser as "text", so we only warn about spaces here.
! 783: */
1.34 schwarze 784:
785: if ('\0' == buf[i] && ' ' == buf[i - 1])
786: if ( ! mdoc_pwarn(m, ln, i - 1, ETAILWS))
787: goto err;
1.1 kristaps 788:
1.24 schwarze 789: /*
790: * Begin recursive parse sequence. Since we're at the start of
791: * the line, we don't need to do callable/parseable checks.
792: */
1.46 ! schwarze 793: if ( ! mdoc_macro(m, tok, ln, 1, &i, buf))
1.1 kristaps 794: goto err;
1.35 schwarze 795:
796: /*
797: * Mark the end of a sentence, but be careful not to insert
1.43 schwarze 798: * markers into reference blocks and after ellipses in
799: * function definitions.
1.35 schwarze 800: */
801: n = m->last;
802: if (n->child)
803: n = n->child;
804: while (n->next)
805: n = n->next;
1.43 schwarze 806: if (MDOC_TEXT == n->type &&
807: MDOC_Fn != n->parent->tok &&
808: MDOC_Rs != m->last->parent->tok) {
1.35 schwarze 809: t = n->string;
1.36 schwarze 810: while (t[0] && t[1])
1.35 schwarze 811: t++;
812: if ('.' == *t || '!' == *t || '?' == *t) {
813: if ( ! mdoc_elem_alloc(m, ln, i, MDOC_eos, NULL))
814: return(0);
815: m->next = MDOC_NEXT_SIBLING;
816: }
817: }
1.1 kristaps 818:
819: return(1);
820:
821: err: /* Error out. */
822:
823: m->flags |= MDOC_HALT;
824: return(0);
825: }
1.24 schwarze 826:
827: