Annotation of src/usr.bin/mandoc/mdoc.c, Revision 1.59
1.59 ! schwarze 1: /* $Id: mdoc.c,v 1.58 2010/06/27 21:54:42 schwarze Exp $ */
1.1 kristaps 2: /*
1.56 schwarze 3: * Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@bsd.lv>
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:
1.54 schwarze 27: #include "mandoc.h"
1.58 schwarze 28: #include "regs.h"
1.1 kristaps 29: #include "libmdoc.h"
1.32 schwarze 30: #include "libmandoc.h"
1.1 kristaps 31:
32: const char *const __mdoc_macronames[MDOC_MAX] = {
1.7 schwarze 33: "Ap", "Dd", "Dt", "Os",
1.1 kristaps 34: "Sh", "Ss", "Pp", "D1",
35: "Dl", "Bd", "Ed", "Bl",
36: "El", "It", "Ad", "An",
37: "Ar", "Cd", "Cm", "Dv",
38: "Er", "Ev", "Ex", "Fa",
39: "Fd", "Fl", "Fn", "Ft",
40: "Ic", "In", "Li", "Nd",
41: "Nm", "Op", "Ot", "Pa",
42: "Rv", "St", "Va", "Vt",
43: /* LINTED */
1.33 schwarze 44: "Xr", "%A", "%B", "%D",
1.1 kristaps 45: /* LINTED */
1.33 schwarze 46: "%I", "%J", "%N", "%O",
1.1 kristaps 47: /* LINTED */
1.33 schwarze 48: "%P", "%R", "%T", "%V",
1.1 kristaps 49: "Ac", "Ao", "Aq", "At",
50: "Bc", "Bf", "Bo", "Bq",
51: "Bsx", "Bx", "Db", "Dc",
52: "Do", "Dq", "Ec", "Ef",
53: "Em", "Eo", "Fx", "Ms",
54: "No", "Ns", "Nx", "Ox",
55: "Pc", "Pf", "Po", "Pq",
56: "Qc", "Ql", "Qo", "Qq",
57: "Re", "Rs", "Sc", "So",
58: "Sq", "Sm", "Sx", "Sy",
59: "Tn", "Ux", "Xc", "Xo",
60: "Fo", "Fc", "Oo", "Oc",
61: "Bk", "Ek", "Bt", "Hf",
1.7 schwarze 62: "Fr", "Ud", "Lb", "Lp",
63: "Lk", "Mt", "Brq", "Bro",
1.1 kristaps 64: /* LINTED */
1.33 schwarze 65: "Brc", "%C", "Es", "En",
1.1 kristaps 66: /* LINTED */
1.33 schwarze 67: "Dx", "%Q", "br", "sp",
1.31 schwarze 68: /* LINTED */
1.56 schwarze 69: "%U", "Ta"
1.1 kristaps 70: };
71:
72: const char *const __mdoc_argnames[MDOC_ARG_MAX] = {
73: "split", "nosplit", "ragged",
74: "unfilled", "literal", "file",
75: "offset", "bullet", "dash",
76: "hyphen", "item", "enum",
77: "tag", "diag", "hang",
78: "ohang", "inset", "column",
79: "width", "compact", "std",
80: "filled", "words", "emphasis",
1.30 schwarze 81: "symbolic", "nested", "centered"
1.1 kristaps 82: };
83:
84: const char * const *mdoc_macronames = __mdoc_macronames;
85: const char * const *mdoc_argnames = __mdoc_argnames;
86:
1.40 schwarze 87: static void mdoc_node_free(struct mdoc_node *);
88: static void mdoc_node_unlink(struct mdoc *,
89: struct mdoc_node *);
1.1 kristaps 90: static void mdoc_free1(struct mdoc *);
1.32 schwarze 91: static void mdoc_alloc1(struct mdoc *);
1.1 kristaps 92: static struct mdoc_node *node_alloc(struct mdoc *, int, int,
1.37 schwarze 93: enum mdoct, enum mdoc_type);
1.1 kristaps 94: static int node_append(struct mdoc *,
95: struct mdoc_node *);
1.53 schwarze 96: static int mdoc_ptext(struct mdoc *, int, char *, int);
97: static int mdoc_pmacro(struct mdoc *, int, char *, int);
98: static int macrowarn(struct mdoc *, int,
99: const char *, int);
1.46 schwarze 100:
1.1 kristaps 101:
102: const struct mdoc_node *
103: mdoc_node(const struct mdoc *m)
104: {
105:
106: return(MDOC_HALT & m->flags ? NULL : m->first);
107: }
108:
109:
110: const struct mdoc_meta *
111: mdoc_meta(const struct mdoc *m)
112: {
113:
114: return(MDOC_HALT & m->flags ? NULL : &m->meta);
115: }
116:
117:
1.8 schwarze 118: /*
119: * Frees volatile resources (parse tree, meta-data, fields).
120: */
1.1 kristaps 121: static void
122: mdoc_free1(struct mdoc *mdoc)
123: {
124:
125: if (mdoc->first)
1.40 schwarze 126: mdoc_node_delete(mdoc, mdoc->first);
1.1 kristaps 127: if (mdoc->meta.title)
128: free(mdoc->meta.title);
129: if (mdoc->meta.os)
130: free(mdoc->meta.os);
131: if (mdoc->meta.name)
132: free(mdoc->meta.name);
133: if (mdoc->meta.arch)
134: free(mdoc->meta.arch);
135: if (mdoc->meta.vol)
136: free(mdoc->meta.vol);
1.50 schwarze 137: if (mdoc->meta.msec)
138: free(mdoc->meta.msec);
1.1 kristaps 139: }
140:
141:
1.8 schwarze 142: /*
143: * Allocate all volatile resources (parse tree, meta-data, fields).
144: */
1.32 schwarze 145: static void
1.1 kristaps 146: mdoc_alloc1(struct mdoc *mdoc)
147: {
148:
1.32 schwarze 149: memset(&mdoc->meta, 0, sizeof(struct mdoc_meta));
1.1 kristaps 150: mdoc->flags = 0;
1.9 schwarze 151: mdoc->lastnamed = mdoc->lastsec = SEC_NONE;
1.32 schwarze 152: mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node));
1.1 kristaps 153: mdoc->first = mdoc->last;
154: mdoc->last->type = MDOC_ROOT;
155: mdoc->next = MDOC_NEXT_CHILD;
156: }
157:
158:
159: /*
1.8 schwarze 160: * Free up volatile resources (see mdoc_free1()) then re-initialises the
161: * data with mdoc_alloc1(). After invocation, parse data has been reset
162: * and the parser is ready for re-invocation on a new tree; however,
163: * cross-parse non-volatile data is kept intact.
1.1 kristaps 164: */
1.32 schwarze 165: void
1.1 kristaps 166: mdoc_reset(struct mdoc *mdoc)
167: {
168:
169: mdoc_free1(mdoc);
1.32 schwarze 170: mdoc_alloc1(mdoc);
1.1 kristaps 171: }
172:
173:
174: /*
1.8 schwarze 175: * Completely free up all volatile and non-volatile parse resources.
176: * After invocation, the pointer is no longer usable.
1.1 kristaps 177: */
178: void
179: mdoc_free(struct mdoc *mdoc)
180: {
181:
182: mdoc_free1(mdoc);
183: free(mdoc);
184: }
185:
186:
1.8 schwarze 187: /*
188: * Allocate volatile and non-volatile parse resources.
189: */
1.1 kristaps 190: struct mdoc *
1.58 schwarze 191: mdoc_alloc(struct regset *regs, void *data,
192: int pflags, mandocmsg msg)
1.1 kristaps 193: {
194: struct mdoc *p;
195:
1.32 schwarze 196: p = mandoc_calloc(1, sizeof(struct mdoc));
197:
1.54 schwarze 198: p->msg = msg;
1.1 kristaps 199: p->data = data;
200: p->pflags = pflags;
1.58 schwarze 201: p->regs = regs;
1.1 kristaps 202:
1.32 schwarze 203: mdoc_hash_init();
204: mdoc_alloc1(p);
205: return(p);
1.1 kristaps 206: }
207:
208:
209: /*
210: * Climb back up the parse tree, validating open scopes. Mostly calls
1.8 schwarze 211: * through to macro_end() in macro.c.
1.1 kristaps 212: */
213: int
214: mdoc_endparse(struct mdoc *m)
215: {
216:
217: if (MDOC_HALT & m->flags)
218: return(0);
219: else if (mdoc_macroend(m))
220: return(1);
221: m->flags |= MDOC_HALT;
222: return(0);
223: }
224:
225:
226: /*
227: * Main parse routine. Parses a single line -- really just hands off to
1.45 schwarze 228: * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
1.1 kristaps 229: */
230: int
1.53 schwarze 231: mdoc_parseln(struct mdoc *m, int ln, char *buf, int offs)
1.1 kristaps 232: {
233:
234: if (MDOC_HALT & m->flags)
235: return(0);
236:
1.47 schwarze 237: m->flags |= MDOC_NEWLINE;
1.53 schwarze 238: return(('.' == buf[offs] || '\'' == buf[offs]) ?
239: mdoc_pmacro(m, ln, buf, offs) :
240: mdoc_ptext(m, ln, buf, offs));
1.1 kristaps 241: }
242:
243:
244: int
1.54 schwarze 245: mdoc_vmsg(struct mdoc *mdoc, enum mandocerr t,
246: int ln, int pos, const char *fmt, ...)
1.1 kristaps 247: {
248: char buf[256];
249: va_list ap;
250:
251: va_start(ap, fmt);
1.54 schwarze 252: vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
1.1 kristaps 253: va_end(ap);
1.17 schwarze 254:
1.54 schwarze 255: return((*mdoc->msg)(t, mdoc->data, ln, pos, buf));
1.5 schwarze 256: }
257:
258:
259: int
1.58 schwarze 260: mdoc_macro(MACRO_PROT_ARGS)
1.1 kristaps 261: {
1.46 schwarze 262: assert(tok < MDOC_MAX);
263:
264: /* If we're in the body, deny prologue calls. */
1.37 schwarze 265:
1.1 kristaps 266: if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
1.9 schwarze 267: MDOC_PBODY & m->flags)
1.58 schwarze 268: return(mdoc_pmsg(m, line, ppos, MANDOCERR_BADBODY));
1.46 schwarze 269:
270: /* If we're in the prologue, deny "body" macros. */
271:
1.1 kristaps 272: if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
1.39 schwarze 273: ! (MDOC_PBODY & m->flags)) {
1.58 schwarze 274: if ( ! mdoc_pmsg(m, line, ppos, MANDOCERR_BADPROLOG))
1.39 schwarze 275: return(0);
276: if (NULL == m->meta.title)
1.56 schwarze 277: m->meta.title = mandoc_strdup("UNKNOWN");
1.39 schwarze 278: if (NULL == m->meta.vol)
1.56 schwarze 279: m->meta.vol = mandoc_strdup("LOCAL");
1.39 schwarze 280: if (NULL == m->meta.os)
1.56 schwarze 281: m->meta.os = mandoc_strdup("LOCAL");
1.39 schwarze 282: if (0 == m->meta.date)
283: m->meta.date = time(NULL);
284: m->flags |= MDOC_PBODY;
285: }
1.1 kristaps 286:
1.58 schwarze 287: return((*mdoc_macros[tok].fp)(m, tok, line, ppos, pos, buf));
1.1 kristaps 288: }
289:
290:
291: static int
292: node_append(struct mdoc *mdoc, struct mdoc_node *p)
293: {
294:
295: assert(mdoc->last);
296: assert(mdoc->first);
297: assert(MDOC_ROOT != p->type);
298:
299: switch (mdoc->next) {
300: case (MDOC_NEXT_SIBLING):
301: mdoc->last->next = p;
302: p->prev = mdoc->last;
303: p->parent = mdoc->last->parent;
304: break;
305: case (MDOC_NEXT_CHILD):
306: mdoc->last->child = p;
307: p->parent = mdoc->last;
308: break;
309: default:
310: abort();
311: /* NOTREACHED */
312: }
313:
1.10 schwarze 314: p->parent->nchild++;
315:
1.1 kristaps 316: if ( ! mdoc_valid_pre(mdoc, p))
317: return(0);
318: if ( ! mdoc_action_pre(mdoc, p))
319: return(0);
320:
321: switch (p->type) {
322: case (MDOC_HEAD):
323: assert(MDOC_BLOCK == p->parent->type);
324: p->parent->head = p;
325: break;
326: case (MDOC_TAIL):
327: assert(MDOC_BLOCK == p->parent->type);
328: p->parent->tail = p;
329: break;
330: case (MDOC_BODY):
1.59 ! schwarze 331: if (p->end)
! 332: break;
1.1 kristaps 333: assert(MDOC_BLOCK == p->parent->type);
334: p->parent->body = p;
335: break;
336: default:
337: break;
338: }
339:
340: mdoc->last = p;
341:
342: switch (p->type) {
343: case (MDOC_TEXT):
344: if ( ! mdoc_valid_post(mdoc))
345: return(0);
346: if ( ! mdoc_action_post(mdoc))
347: return(0);
348: break;
349: default:
350: break;
351: }
352:
353: return(1);
354: }
355:
356:
357: static struct mdoc_node *
1.37 schwarze 358: node_alloc(struct mdoc *m, int line, int pos,
359: enum mdoct tok, enum mdoc_type type)
1.1 kristaps 360: {
361: struct mdoc_node *p;
362:
1.32 schwarze 363: p = mandoc_calloc(1, sizeof(struct mdoc_node));
1.19 schwarze 364: p->sec = m->lastsec;
1.1 kristaps 365: p->line = line;
366: p->pos = pos;
367: p->tok = tok;
1.37 schwarze 368: p->type = type;
1.58 schwarze 369:
370: /* Flag analysis. */
371:
1.47 schwarze 372: if (MDOC_NEWLINE & m->flags)
373: p->flags |= MDOC_LINE;
374: m->flags &= ~MDOC_NEWLINE;
1.58 schwarze 375:
376: /* Section analysis. */
377:
378: if (SEC_SYNOPSIS == p->sec)
379: p->flags |= MDOC_SYNPRETTY;
380:
381: /* Register analysis. */
382:
383: if (m->regs->regs[(int)REG_nS].set) {
384: if (m->regs->regs[(int)REG_nS].v.u)
385: p->flags |= MDOC_SYNPRETTY;
386: else
387: p->flags &= ~MDOC_SYNPRETTY;
388: }
389:
1.1 kristaps 390: return(p);
391: }
392:
393:
394: int
1.37 schwarze 395: mdoc_tail_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
1.1 kristaps 396: {
397: struct mdoc_node *p;
398:
1.19 schwarze 399: p = node_alloc(m, line, pos, tok, MDOC_TAIL);
1.25 schwarze 400: if ( ! node_append(m, p))
401: return(0);
402: m->next = MDOC_NEXT_CHILD;
403: return(1);
1.1 kristaps 404: }
405:
406:
407: int
1.37 schwarze 408: mdoc_head_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
1.1 kristaps 409: {
410: struct mdoc_node *p;
411:
1.19 schwarze 412: assert(m->first);
413: assert(m->last);
1.1 kristaps 414:
1.19 schwarze 415: p = node_alloc(m, line, pos, tok, MDOC_HEAD);
1.25 schwarze 416: if ( ! node_append(m, p))
417: return(0);
418: m->next = MDOC_NEXT_CHILD;
419: return(1);
1.1 kristaps 420: }
421:
422:
423: int
1.37 schwarze 424: mdoc_body_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
1.1 kristaps 425: {
426: struct mdoc_node *p;
427:
1.19 schwarze 428: p = node_alloc(m, line, pos, tok, MDOC_BODY);
1.25 schwarze 429: if ( ! node_append(m, p))
430: return(0);
431: m->next = MDOC_NEXT_CHILD;
1.59 ! schwarze 432: return(1);
! 433: }
! 434:
! 435:
! 436: int
! 437: mdoc_endbody_alloc(struct mdoc *m, int line, int pos, enum mdoct tok,
! 438: struct mdoc_node *body, enum mdoc_endbody end)
! 439: {
! 440: struct mdoc_node *p;
! 441:
! 442: p = node_alloc(m, line, pos, tok, MDOC_BODY);
! 443: p->pending = body;
! 444: p->end = end;
! 445: if ( ! node_append(m, p))
! 446: return(0);
! 447: m->next = MDOC_NEXT_SIBLING;
1.25 schwarze 448: return(1);
1.1 kristaps 449: }
450:
451:
452: int
1.19 schwarze 453: mdoc_block_alloc(struct mdoc *m, int line, int pos,
1.37 schwarze 454: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 455: {
456: struct mdoc_node *p;
457:
1.19 schwarze 458: p = node_alloc(m, line, pos, tok, MDOC_BLOCK);
1.4 schwarze 459: p->args = args;
460: if (p->args)
1.1 kristaps 461: (args->refcnt)++;
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
1.19 schwarze 470: mdoc_elem_alloc(struct mdoc *m, int line, int pos,
1.37 schwarze 471: enum mdoct tok, struct mdoc_arg *args)
1.1 kristaps 472: {
473: struct mdoc_node *p;
474:
1.19 schwarze 475: p = node_alloc(m, line, pos, tok, MDOC_ELEM);
1.4 schwarze 476: p->args = args;
477: if (p->args)
1.1 kristaps 478: (args->refcnt)++;
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:
1.46 schwarze 486: int
487: mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p)
1.1 kristaps 488: {
1.19 schwarze 489: struct mdoc_node *n;
1.46 schwarze 490: size_t sv, len;
491:
492: len = strlen(p);
1.1 kristaps 493:
1.46 schwarze 494: n = node_alloc(m, line, pos, MDOC_MAX, MDOC_TEXT);
1.32 schwarze 495: n->string = mandoc_malloc(len + 1);
1.19 schwarze 496: sv = strlcpy(n->string, p, len + 1);
497:
498: /* Prohibit truncation. */
499: assert(sv < len + 1);
500:
1.25 schwarze 501: if ( ! node_append(m, n))
502: return(0);
1.46 schwarze 503:
1.25 schwarze 504: m->next = MDOC_NEXT_SIBLING;
505: return(1);
1.19 schwarze 506: }
507:
508:
1.1 kristaps 509: void
510: mdoc_node_free(struct mdoc_node *p)
511: {
512:
513: if (p->string)
514: free(p->string);
515: if (p->args)
516: mdoc_argv_free(p->args);
517: free(p);
518: }
519:
520:
1.40 schwarze 521: static void
522: mdoc_node_unlink(struct mdoc *m, struct mdoc_node *n)
523: {
524:
525: /* Adjust siblings. */
526:
527: if (n->prev)
528: n->prev->next = n->next;
529: if (n->next)
530: n->next->prev = n->prev;
531:
532: /* Adjust parent. */
533:
534: if (n->parent) {
535: n->parent->nchild--;
536: if (n->parent->child == n)
537: n->parent->child = n->prev ? n->prev : n->next;
538: }
539:
540: /* Adjust parse point, if applicable. */
541:
542: if (m && m->last == n) {
543: if (n->prev) {
544: m->last = n->prev;
545: m->next = MDOC_NEXT_SIBLING;
546: } else {
547: m->last = n->parent;
548: m->next = MDOC_NEXT_CHILD;
549: }
550: }
551:
552: if (m && m->first == n)
553: m->first = NULL;
554: }
555:
556:
1.1 kristaps 557: void
1.40 schwarze 558: mdoc_node_delete(struct mdoc *m, struct mdoc_node *p)
1.1 kristaps 559: {
560:
1.40 schwarze 561: while (p->child) {
562: assert(p->nchild);
563: mdoc_node_delete(m, p->child);
564: }
565: assert(0 == p->nchild);
1.1 kristaps 566:
1.40 schwarze 567: mdoc_node_unlink(m, p);
1.1 kristaps 568: mdoc_node_free(p);
569: }
570:
571:
572: /*
573: * Parse free-form text, that is, a line that does not begin with the
574: * control character.
575: */
576: static int
1.53 schwarze 577: mdoc_ptext(struct mdoc *m, int line, char *buf, int offs)
1.1 kristaps 578: {
1.56 schwarze 579: char *c, *ws, *end;
580: struct mdoc_node *n;
1.44 schwarze 581:
582: /* Ignore bogus comments. */
583:
1.53 schwarze 584: if ('\\' == buf[offs] &&
585: '.' == buf[offs + 1] &&
586: '"' == buf[offs + 2])
1.54 schwarze 587: return(mdoc_pmsg(m, line, offs, MANDOCERR_BADCOMMENT));
1.1 kristaps 588:
1.46 schwarze 589: /* No text before an initial macro. */
590:
1.9 schwarze 591: if (SEC_NONE == m->lastnamed)
1.54 schwarze 592: return(mdoc_pmsg(m, line, offs, MANDOCERR_NOTEXT));
1.46 schwarze 593:
1.56 schwarze 594: assert(m->last);
595: n = m->last;
596:
597: /*
598: * Divert directly to list processing if we're encountering a
599: * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
600: * (a MDOC_BODY means it's already open, in which case we should
601: * process within its context in the normal way).
602: */
603:
604: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.57 schwarze 605: LIST_column == n->data.Bl.type) {
1.56 schwarze 606: /* `Bl' is open without any children. */
607: m->flags |= MDOC_FREECOL;
608: return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
609: }
610:
611: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
612: NULL != n->parent &&
613: MDOC_Bl == n->parent->tok &&
1.57 schwarze 614: LIST_column == n->parent->data.Bl.type) {
1.56 schwarze 615: /* `Bl' has block-level `It' children. */
616: m->flags |= MDOC_FREECOL;
617: return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
618: }
619:
1.52 schwarze 620: /*
621: * Search for the beginning of unescaped trailing whitespace (ws)
622: * and for the first character not to be output (end).
623: */
1.56 schwarze 624:
625: /* FIXME: replace with strcspn(). */
1.52 schwarze 626: ws = NULL;
1.53 schwarze 627: for (c = end = buf + offs; *c; c++) {
1.52 schwarze 628: switch (*c) {
1.55 schwarze 629: case '-':
630: if (mandoc_hyph(buf + offs, c))
631: *c = ASCII_HYPH;
1.56 schwarze 632: ws = NULL;
1.55 schwarze 633: break;
1.52 schwarze 634: case ' ':
635: if (NULL == ws)
636: ws = c;
637: continue;
638: case '\t':
639: /*
640: * Always warn about trailing tabs,
641: * even outside literal context,
642: * where they should be put on the next line.
643: */
644: if (NULL == ws)
645: ws = c;
646: /*
647: * Strip trailing tabs in literal context only;
648: * outside, they affect the next line.
649: */
650: if (MDOC_LITERAL & m->flags)
651: continue;
652: break;
653: case '\\':
654: /* Skip the escaped character, too, if any. */
655: if (c[1])
656: c++;
657: /* FALLTHROUGH */
658: default:
659: ws = NULL;
660: break;
661: }
662: end = c + 1;
663: }
664: *end = '\0';
1.19 schwarze 665:
1.52 schwarze 666: if (ws)
1.54 schwarze 667: if ( ! mdoc_pmsg(m, line, (int)(ws-buf), MANDOCERR_EOLNSPACE))
1.52 schwarze 668: return(0);
1.34 schwarze 669:
1.53 schwarze 670: if ('\0' == buf[offs] && ! (MDOC_LITERAL & m->flags)) {
1.54 schwarze 671: if ( ! mdoc_pmsg(m, line, (int)(c-buf), MANDOCERR_NOBLANKLN))
1.38 schwarze 672: return(0);
1.46 schwarze 673:
1.40 schwarze 674: /*
1.46 schwarze 675: * Insert a `Pp' in the case of a blank line. Technically,
676: * blank lines aren't allowed, but enough manuals assume this
677: * behaviour that we want to work around it.
1.40 schwarze 678: */
1.53 schwarze 679: if ( ! mdoc_elem_alloc(m, line, offs, MDOC_Pp, NULL))
1.42 schwarze 680: return(0);
1.46 schwarze 681:
1.42 schwarze 682: m->next = MDOC_NEXT_SIBLING;
683: return(1);
1.38 schwarze 684: }
1.1 kristaps 685:
1.53 schwarze 686: if ( ! mdoc_word_alloc(m, line, offs, buf+offs))
1.1 kristaps 687: return(0);
688:
1.52 schwarze 689: if (MDOC_LITERAL & m->flags)
690: return(1);
691:
1.35 schwarze 692: /*
1.48 schwarze 693: * End-of-sentence check. If the last character is an unescaped
694: * EOS character, then flag the node as being the end of a
695: * sentence. The front-end will know how to interpret this.
1.35 schwarze 696: */
697:
1.52 schwarze 698: assert(buf < end);
1.48 schwarze 699:
1.53 schwarze 700: if (mandoc_eos(buf+offs, (size_t)(end-buf-offs)))
1.48 schwarze 701: m->last->flags |= MDOC_EOS;
702:
1.1 kristaps 703: return(1);
704: }
1.19 schwarze 705:
1.1 kristaps 706:
707: static int
1.53 schwarze 708: macrowarn(struct mdoc *m, int ln, const char *buf, int offs)
1.1 kristaps 709: {
1.54 schwarze 710: int rc;
711:
712: rc = mdoc_vmsg(m, MANDOCERR_MACRO, ln, offs,
713: "unknown macro: %s%s",
714: buf, strlen(buf) > 3 ? "..." : "");
715:
716: /* FIXME: logic should be in driver. */
1.56 schwarze 717: /* FIXME: broken, will error out and not omit a message. */
1.54 schwarze 718: return(MDOC_IGN_MACRO & m->pflags ? rc : 0);
1.1 kristaps 719: }
720:
721:
722: /*
723: * Parse a macro line, that is, a line beginning with the control
724: * character.
725: */
726: int
1.53 schwarze 727: mdoc_pmacro(struct mdoc *m, int ln, char *buf, int offs)
1.1 kristaps 728: {
1.56 schwarze 729: enum mdoct tok;
730: int i, j, sv;
731: char mac[5];
732: struct mdoc_node *n;
1.1 kristaps 733:
1.7 schwarze 734: /* Empty lines are ignored. */
1.1 kristaps 735:
1.53 schwarze 736: offs++;
737:
738: if ('\0' == buf[offs])
1.1 kristaps 739: return(1);
740:
1.53 schwarze 741: i = offs;
1.24 schwarze 742:
743: /* Accept whitespace after the initial control char. */
744:
745: if (' ' == buf[i]) {
746: i++;
1.1 kristaps 747: while (buf[i] && ' ' == buf[i])
748: i++;
1.34 schwarze 749: if ('\0' == buf[i])
1.1 kristaps 750: return(1);
751: }
752:
1.47 schwarze 753: sv = i;
754:
1.1 kristaps 755: /* Copy the first word into a nil-terminated buffer. */
756:
1.24 schwarze 757: for (j = 0; j < 4; j++, i++) {
1.34 schwarze 758: if ('\0' == (mac[j] = buf[i]))
1.1 kristaps 759: break;
760: else if (' ' == buf[i])
761: break;
1.26 schwarze 762:
763: /* Check for invalid characters. */
764:
765: if (isgraph((u_char)buf[i]))
766: continue;
1.54 schwarze 767: if ( ! mdoc_pmsg(m, ln, i, MANDOCERR_BADCHAR))
768: return(0);
769: i--;
1.1 kristaps 770: }
771:
1.53 schwarze 772: mac[j] = '\0';
1.1 kristaps 773:
1.24 schwarze 774: if (j == 4 || j < 2) {
1.53 schwarze 775: if ( ! macrowarn(m, ln, mac, sv))
1.1 kristaps 776: goto err;
777: return(1);
778: }
779:
1.46 schwarze 780: if (MDOC_MAX == (tok = mdoc_hash_find(mac))) {
1.53 schwarze 781: if ( ! macrowarn(m, ln, mac, sv))
1.1 kristaps 782: goto err;
783: return(1);
784: }
785:
786: /* The macro is sane. Jump to the next word. */
787:
788: while (buf[i] && ' ' == buf[i])
789: i++;
1.34 schwarze 790:
1.46 schwarze 791: /*
792: * Trailing whitespace. Note that tabs are allowed to be passed
793: * into the parser as "text", so we only warn about spaces here.
794: */
1.34 schwarze 795:
796: if ('\0' == buf[i] && ' ' == buf[i - 1])
1.54 schwarze 797: if ( ! mdoc_pmsg(m, ln, i - 1, MANDOCERR_EOLNSPACE))
1.34 schwarze 798: goto err;
1.1 kristaps 799:
1.56 schwarze 800: /*
801: * If an initial macro or a list invocation, divert directly
802: * into macro processing.
803: */
804:
805: if (NULL == m->last || MDOC_It == tok || MDOC_El == tok) {
806: if ( ! mdoc_macro(m, tok, ln, sv, &i, buf))
807: goto err;
808: return(1);
809: }
810:
811: n = m->last;
812: assert(m->last);
813:
814: /*
815: * If the first macro of a `Bl -column', open an `It' block
816: * context around the parsed macro.
817: */
818:
819: if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
1.57 schwarze 820: LIST_column == n->data.Bl.type) {
1.56 schwarze 821: m->flags |= MDOC_FREECOL;
1.58 schwarze 822: if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
1.56 schwarze 823: goto err;
824: return(1);
825: }
826:
827: /*
828: * If we're following a block-level `It' within a `Bl -column'
829: * context (perhaps opened in the above block or in ptext()),
830: * then open an `It' block context around the parsed macro.
1.24 schwarze 831: */
1.56 schwarze 832:
833: if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
834: NULL != n->parent &&
835: MDOC_Bl == n->parent->tok &&
1.57 schwarze 836: LIST_column == n->parent->data.Bl.type) {
1.56 schwarze 837: m->flags |= MDOC_FREECOL;
838: if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
839: goto err;
840: return(1);
841: }
842:
843: /* Normal processing of a macro. */
844:
1.47 schwarze 845: if ( ! mdoc_macro(m, tok, ln, sv, &i, buf))
1.1 kristaps 846: goto err;
847:
848: return(1);
849:
850: err: /* Error out. */
851:
852: m->flags |= MDOC_HALT;
853: return(0);
854: }
1.24 schwarze 855:
856: