Annotation of src/usr.bin/mandoc/mandoc.c, Revision 1.34
1.34 ! schwarze 1: /* $Id: mandoc.c,v 1.33 2012/05/28 17:08:48 schwarze Exp $ */
1.1 schwarze 2: /*
1.24 schwarze 3: * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
1.32 schwarze 4: * Copyright (c) 2011, 2012 Ingo Schwarze <schwarze@openbsd.org>
1.1 schwarze 5: *
6: * Permission to use, copy, modify, and distribute this software for any
7: * purpose with or without fee is hereby granted, provided that the above
8: * copyright notice and this permission notice appear in all copies.
9: *
1.21 schwarze 10: * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.1 schwarze 11: * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.21 schwarze 12: * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.1 schwarze 13: * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
14: * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
15: * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
16: * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
17: */
1.2 schwarze 18: #include <sys/types.h>
19:
1.1 schwarze 20: #include <assert.h>
21: #include <ctype.h>
1.26 schwarze 22: #include <errno.h>
23: #include <limits.h>
1.1 schwarze 24: #include <stdlib.h>
1.4 schwarze 25: #include <stdio.h>
26: #include <string.h>
1.5 schwarze 27: #include <time.h>
1.1 schwarze 28:
1.14 schwarze 29: #include "mandoc.h"
1.1 schwarze 30: #include "libmandoc.h"
31:
1.22 schwarze 32: #define DATESIZE 32
33:
1.14 schwarze 34: static int a2time(time_t *, const char *, const char *);
1.22 schwarze 35: static char *time2a(time_t);
1.5 schwarze 36:
1.26 schwarze 37:
38: enum mandoc_esc
39: mandoc_escape(const char **end, const char **start, int *sz)
40: {
1.34 ! schwarze 41: const char *local_start;
! 42: int local_sz;
! 43: char term;
1.26 schwarze 44: enum mandoc_esc gly;
45:
1.34 ! schwarze 46: /*
! 47: * When the caller doesn't provide return storage,
! 48: * use local storage.
! 49: */
! 50:
! 51: if (NULL == start)
! 52: start = &local_start;
! 53: if (NULL == sz)
! 54: sz = &local_sz;
! 55:
! 56: /*
! 57: * Beyond the backslash, at least one input character
! 58: * is part of the escape sequence. With one exception
! 59: * (see below), that character won't be returned.
! 60: */
! 61:
1.26 schwarze 62: gly = ESCAPE_ERROR;
1.34 ! schwarze 63: *start = ++*end;
! 64: *sz = 0;
1.33 schwarze 65: term = '\0';
1.26 schwarze 66:
1.34 ! schwarze 67: switch ((*start)[-1]) {
1.26 schwarze 68: /*
69: * First the glyphs. There are several different forms of
70: * these, but each eventually returns a substring of the glyph
71: * name.
72: */
73: case ('('):
74: gly = ESCAPE_SPECIAL;
1.34 ! schwarze 75: *sz = 2;
1.26 schwarze 76: break;
77: case ('['):
78: gly = ESCAPE_SPECIAL;
79: /*
80: * Unicode escapes are defined in groff as \[uXXXX] to
81: * \[u10FFFF], where the contained value must be a valid
82: * Unicode codepoint. Here, however, only check whether
83: * it's not a zero-width escape.
84: */
1.34 ! schwarze 85: if ('u' == (*start)[0] && ']' != (*start)[1])
1.26 schwarze 86: gly = ESCAPE_UNICODE;
87: term = ']';
88: break;
89: case ('C'):
1.34 ! schwarze 90: if ('\'' != **start)
1.26 schwarze 91: return(ESCAPE_ERROR);
92: gly = ESCAPE_SPECIAL;
1.34 ! schwarze 93: *start = ++*end;
1.26 schwarze 94: term = '\'';
95: break;
1.32 schwarze 96:
97: /*
98: * The \z escape is supposed to output the following
99: * character without advancing the cursor position.
100: * Since we are mostly dealing with terminal mode,
101: * let us just skip the next character.
102: */
103: case ('z'):
104: return(ESCAPE_SKIPCHAR);
1.1 schwarze 105:
1.26 schwarze 106: /*
107: * Handle all triggers matching \X(xy, \Xx, and \X[xxxx], where
108: * 'X' is the trigger. These have opaque sub-strings.
109: */
110: case ('F'):
1.16 schwarze 111: /* FALLTHROUGH */
1.26 schwarze 112: case ('g'):
1.16 schwarze 113: /* FALLTHROUGH */
1.26 schwarze 114: case ('k'):
1.1 schwarze 115: /* FALLTHROUGH */
1.26 schwarze 116: case ('M'):
1.14 schwarze 117: /* FALLTHROUGH */
1.26 schwarze 118: case ('m'):
1.1 schwarze 119: /* FALLTHROUGH */
1.26 schwarze 120: case ('n'):
1.1 schwarze 121: /* FALLTHROUGH */
1.26 schwarze 122: case ('V'):
1.1 schwarze 123: /* FALLTHROUGH */
1.26 schwarze 124: case ('Y'):
1.29 schwarze 125: gly = ESCAPE_IGNORE;
1.1 schwarze 126: /* FALLTHROUGH */
1.26 schwarze 127: case ('f'):
128: if (ESCAPE_ERROR == gly)
129: gly = ESCAPE_FONT;
1.34 ! schwarze 130: switch (**start) {
1.26 schwarze 131: case ('('):
1.34 ! schwarze 132: *start = ++*end;
! 133: *sz = 2;
1.26 schwarze 134: break;
135: case ('['):
1.34 ! schwarze 136: *start = ++*end;
1.26 schwarze 137: term = ']';
138: break;
139: default:
1.34 ! schwarze 140: *sz = 1;
1.26 schwarze 141: break;
142: }
143: break;
144:
145: /*
146: * These escapes are of the form \X'Y', where 'X' is the trigger
147: * and 'Y' is any string. These have opaque sub-strings.
148: */
149: case ('A'):
1.13 schwarze 150: /* FALLTHROUGH */
1.26 schwarze 151: case ('b'):
1.1 schwarze 152: /* FALLTHROUGH */
1.16 schwarze 153: case ('D'):
1.1 schwarze 154: /* FALLTHROUGH */
1.26 schwarze 155: case ('o'):
1.1 schwarze 156: /* FALLTHROUGH */
1.26 schwarze 157: case ('R'):
1.1 schwarze 158: /* FALLTHROUGH */
1.26 schwarze 159: case ('X'):
1.1 schwarze 160: /* FALLTHROUGH */
1.26 schwarze 161: case ('Z'):
1.34 ! schwarze 162: if ('\'' != **start)
1.26 schwarze 163: return(ESCAPE_ERROR);
164: gly = ESCAPE_IGNORE;
1.34 ! schwarze 165: *start = ++*end;
1.16 schwarze 166: term = '\'';
167: break;
1.26 schwarze 168:
169: /*
170: * These escapes are of the form \X'N', where 'X' is the trigger
171: * and 'N' resolves to a numerical expression.
172: */
173: case ('B'):
174: /* FALLTHROUGH */
1.17 schwarze 175: case ('h'):
176: /* FALLTHROUGH */
1.26 schwarze 177: case ('H'):
178: /* FALLTHROUGH */
179: case ('L'):
180: /* FALLTHROUGH */
181: case ('l'):
1.29 schwarze 182: gly = ESCAPE_NUMBERED;
1.26 schwarze 183: /* FALLTHROUGH */
184: case ('S'):
185: /* FALLTHROUGH */
1.17 schwarze 186: case ('v'):
187: /* FALLTHROUGH */
1.26 schwarze 188: case ('w'):
189: /* FALLTHROUGH */
190: case ('x'):
1.34 ! schwarze 191: if ('\'' != **start)
! 192: return(ESCAPE_ERROR);
1.26 schwarze 193: if (ESCAPE_ERROR == gly)
194: gly = ESCAPE_IGNORE;
1.34 ! schwarze 195: *start = ++*end;
1.33 schwarze 196: term = '\'';
1.26 schwarze 197: break;
1.29 schwarze 198:
199: /*
200: * Special handling for the numbered character escape.
201: * XXX Do any other escapes need similar handling?
202: */
203: case ('N'):
1.34 ! schwarze 204: if ('\0' == **start)
1.29 schwarze 205: return(ESCAPE_ERROR);
1.34 ! schwarze 206: (*end)++;
! 207: if (isdigit((unsigned char)**start)) {
! 208: *sz = 1;
1.29 schwarze 209: return(ESCAPE_IGNORE);
1.34 ! schwarze 210: }
! 211: (*start)++;
1.29 schwarze 212: while (isdigit((unsigned char)**end))
213: (*end)++;
1.34 ! schwarze 214: *sz = *end - *start;
1.29 schwarze 215: if ('\0' != **end)
216: (*end)++;
217: return(ESCAPE_NUMBERED);
1.26 schwarze 218:
219: /*
220: * Sizes get a special category of their own.
221: */
1.6 schwarze 222: case ('s'):
1.26 schwarze 223: gly = ESCAPE_IGNORE;
1.17 schwarze 224:
1.26 schwarze 225: /* See +/- counts as a sign. */
1.34 ! schwarze 226: if ('+' == **end || '-' == **end || ASCII_HYPH == **end)
! 227: (*end)++;
1.6 schwarze 228:
1.34 ! schwarze 229: switch (**end) {
1.16 schwarze 230: case ('('):
1.34 ! schwarze 231: *start = ++*end;
! 232: *sz = 2;
1.16 schwarze 233: break;
234: case ('['):
1.34 ! schwarze 235: *start = ++*end;
1.33 schwarze 236: term = ']';
1.16 schwarze 237: break;
238: case ('\''):
1.34 ! schwarze 239: *start = ++*end;
1.33 schwarze 240: term = '\'';
1.16 schwarze 241: break;
242: default:
1.34 ! schwarze 243: *sz = 1;
1.16 schwarze 244: break;
1.6 schwarze 245: }
246:
1.26 schwarze 247: break;
248:
249: /*
250: * Anything else is assumed to be a glyph.
1.34 ! schwarze 251: * In this case, pass back the character after the backslash.
1.26 schwarze 252: */
253: default:
254: gly = ESCAPE_SPECIAL;
1.34 ! schwarze 255: *start = --*end;
! 256: *sz = 1;
1.26 schwarze 257: break;
258: }
259:
260: assert(ESCAPE_ERROR != gly);
261:
262: /*
1.33 schwarze 263: * Read up to the terminating character,
264: * paying attention to nested escapes.
1.26 schwarze 265: */
266:
267: if ('\0' != term) {
1.33 schwarze 268: while (**end != term) {
269: switch (**end) {
270: case ('\0'):
271: return(ESCAPE_ERROR);
272: case ('\\'):
273: (*end)++;
274: if (ESCAPE_ERROR ==
275: mandoc_escape(end, NULL, NULL))
276: return(ESCAPE_ERROR);
277: break;
278: default:
279: (*end)++;
280: break;
281: }
282: }
1.34 ! schwarze 283: *sz = (*end)++ - *start;
1.33 schwarze 284: } else {
1.34 ! schwarze 285: assert(*sz > 0);
! 286: if ((size_t)*sz > strlen(*start))
1.26 schwarze 287: return(ESCAPE_ERROR);
1.34 ! schwarze 288: *end += *sz;
1.26 schwarze 289: }
1.19 schwarze 290:
1.26 schwarze 291: /* Run post-processors. */
1.19 schwarze 292:
1.26 schwarze 293: switch (gly) {
294: case (ESCAPE_FONT):
1.30 schwarze 295: /*
296: * Pretend that the constant-width font modes are the
297: * same as the regular font modes.
298: */
1.34 ! schwarze 299: if (2 == *sz && 'C' == **start) {
! 300: (*start)++;
! 301: (*sz)--;
! 302: } else if (1 != *sz)
1.26 schwarze 303: break;
1.30 schwarze 304:
1.34 ! schwarze 305: switch (**start) {
1.26 schwarze 306: case ('3'):
307: /* FALLTHROUGH */
308: case ('B'):
309: gly = ESCAPE_FONTBOLD;
310: break;
311: case ('2'):
312: /* FALLTHROUGH */
313: case ('I'):
314: gly = ESCAPE_FONTITALIC;
1.16 schwarze 315: break;
1.26 schwarze 316: case ('P'):
317: gly = ESCAPE_FONTPREV;
1.16 schwarze 318: break;
1.26 schwarze 319: case ('1'):
320: /* FALLTHROUGH */
321: case ('R'):
322: gly = ESCAPE_FONTROMAN;
1.1 schwarze 323: break;
324: }
1.16 schwarze 325: break;
1.26 schwarze 326: case (ESCAPE_SPECIAL):
1.34 ! schwarze 327: if (1 == *sz && 'c' == **start)
1.26 schwarze 328: gly = ESCAPE_NOSPACE;
1.16 schwarze 329: break;
1.1 schwarze 330: default:
1.16 schwarze 331: break;
1.1 schwarze 332: }
333:
1.26 schwarze 334: return(gly);
1.1 schwarze 335: }
336:
1.4 schwarze 337: void *
338: mandoc_calloc(size_t num, size_t size)
339: {
340: void *ptr;
341:
342: ptr = calloc(num, size);
343: if (NULL == ptr) {
344: perror(NULL);
1.20 schwarze 345: exit((int)MANDOCLEVEL_SYSERR);
1.4 schwarze 346: }
347:
348: return(ptr);
349: }
350:
351:
352: void *
353: mandoc_malloc(size_t size)
354: {
355: void *ptr;
356:
357: ptr = malloc(size);
358: if (NULL == ptr) {
359: perror(NULL);
1.20 schwarze 360: exit((int)MANDOCLEVEL_SYSERR);
1.4 schwarze 361: }
362:
363: return(ptr);
364: }
365:
366:
367: void *
368: mandoc_realloc(void *ptr, size_t size)
369: {
370:
371: ptr = realloc(ptr, size);
372: if (NULL == ptr) {
373: perror(NULL);
1.20 schwarze 374: exit((int)MANDOCLEVEL_SYSERR);
1.4 schwarze 375: }
376:
377: return(ptr);
378: }
379:
1.27 schwarze 380: char *
381: mandoc_strndup(const char *ptr, size_t sz)
382: {
383: char *p;
384:
385: p = mandoc_malloc(sz + 1);
386: memcpy(p, ptr, sz);
387: p[(int)sz] = '\0';
388: return(p);
389: }
1.4 schwarze 390:
391: char *
392: mandoc_strdup(const char *ptr)
393: {
394: char *p;
395:
396: p = strdup(ptr);
397: if (NULL == p) {
398: perror(NULL);
1.20 schwarze 399: exit((int)MANDOCLEVEL_SYSERR);
1.4 schwarze 400: }
401:
402: return(p);
1.21 schwarze 403: }
404:
405: /*
406: * Parse a quoted or unquoted roff-style request or macro argument.
407: * Return a pointer to the parsed argument, which is either the original
408: * pointer or advanced by one byte in case the argument is quoted.
409: * Null-terminate the argument in place.
410: * Collapse pairs of quotes inside quoted arguments.
411: * Advance the argument pointer to the next argument,
412: * or to the null byte terminating the argument line.
413: */
414: char *
1.25 schwarze 415: mandoc_getarg(struct mparse *parse, char **cpp, int ln, int *pos)
1.21 schwarze 416: {
417: char *start, *cp;
418: int quoted, pairs, white;
419:
420: /* Quoting can only start with a new word. */
421: start = *cpp;
1.26 schwarze 422: quoted = 0;
1.21 schwarze 423: if ('"' == *start) {
424: quoted = 1;
425: start++;
1.26 schwarze 426: }
1.21 schwarze 427:
428: pairs = 0;
429: white = 0;
430: for (cp = start; '\0' != *cp; cp++) {
431: /* Move left after quoted quotes and escaped backslashes. */
432: if (pairs)
433: cp[-pairs] = cp[0];
434: if ('\\' == cp[0]) {
435: if ('\\' == cp[1]) {
436: /* Poor man's copy mode. */
437: pairs++;
438: cp++;
439: } else if (0 == quoted && ' ' == cp[1])
440: /* Skip escaped blanks. */
441: cp++;
442: } else if (0 == quoted) {
443: if (' ' == cp[0]) {
444: /* Unescaped blanks end unquoted args. */
445: white = 1;
446: break;
447: }
448: } else if ('"' == cp[0]) {
449: if ('"' == cp[1]) {
450: /* Quoted quotes collapse. */
451: pairs++;
452: cp++;
453: } else {
454: /* Unquoted quotes end quoted args. */
455: quoted = 2;
456: break;
457: }
458: }
459: }
460:
461: /* Quoted argument without a closing quote. */
1.25 schwarze 462: if (1 == quoted)
463: mandoc_msg(MANDOCERR_BADQUOTE, parse, ln, *pos, NULL);
1.21 schwarze 464:
465: /* Null-terminate this argument and move to the next one. */
466: if (pairs)
467: cp[-pairs] = '\0';
468: if ('\0' != *cp) {
469: *cp++ = '\0';
470: while (' ' == *cp)
471: cp++;
472: }
1.24 schwarze 473: *pos += (int)(cp - start) + (quoted ? 1 : 0);
1.21 schwarze 474: *cpp = cp;
475:
1.25 schwarze 476: if ('\0' == *cp && (white || ' ' == cp[-1]))
477: mandoc_msg(MANDOCERR_EOLNSPACE, parse, ln, *pos, NULL);
1.21 schwarze 478:
479: return(start);
1.4 schwarze 480: }
1.5 schwarze 481:
482: static int
483: a2time(time_t *t, const char *fmt, const char *p)
484: {
485: struct tm tm;
486: char *pp;
487:
488: memset(&tm, 0, sizeof(struct tm));
489:
490: pp = strptime(p, fmt, &tm);
491: if (NULL != pp && '\0' == *pp) {
492: *t = mktime(&tm);
493: return(1);
494: }
495:
496: return(0);
497: }
498:
1.22 schwarze 499: static char *
500: time2a(time_t t)
501: {
1.28 schwarze 502: struct tm *tm;
1.23 schwarze 503: char *buf, *p;
504: size_t ssz;
1.22 schwarze 505: int isz;
506:
1.28 schwarze 507: tm = localtime(&t);
1.22 schwarze 508:
1.23 schwarze 509: /*
510: * Reserve space:
511: * up to 9 characters for the month (September) + blank
512: * up to 2 characters for the day + comma + blank
513: * 4 characters for the year and a terminating '\0'
514: */
515: p = buf = mandoc_malloc(10 + 4 + 4 + 1);
516:
1.28 schwarze 517: if (0 == (ssz = strftime(p, 10 + 1, "%B ", tm)))
1.23 schwarze 518: goto fail;
519: p += (int)ssz;
1.22 schwarze 520:
1.28 schwarze 521: if (-1 == (isz = snprintf(p, 4 + 1, "%d, ", tm->tm_mday)))
1.23 schwarze 522: goto fail;
1.22 schwarze 523: p += isz;
524:
1.28 schwarze 525: if (0 == strftime(p, 4 + 1, "%Y", tm))
1.23 schwarze 526: goto fail;
527: return(buf);
528:
529: fail:
530: free(buf);
531: return(NULL);
1.22 schwarze 532: }
533:
534: char *
1.25 schwarze 535: mandoc_normdate(struct mparse *parse, char *in, int ln, int pos)
1.5 schwarze 536: {
1.22 schwarze 537: char *out;
1.5 schwarze 538: time_t t;
539:
1.22 schwarze 540: if (NULL == in || '\0' == *in ||
541: 0 == strcmp(in, "$" "Mdocdate$")) {
1.25 schwarze 542: mandoc_msg(MANDOCERR_NODATE, parse, ln, pos, NULL);
1.22 schwarze 543: time(&t);
544: }
1.31 schwarze 545: else if (a2time(&t, "%Y-%m-%d", in))
546: t = 0;
1.22 schwarze 547: else if (!a2time(&t, "$" "Mdocdate: %b %d %Y $", in) &&
1.31 schwarze 548: !a2time(&t, "%b %d, %Y", in)) {
1.25 schwarze 549: mandoc_msg(MANDOCERR_BADDATE, parse, ln, pos, NULL);
1.22 schwarze 550: t = 0;
1.5 schwarze 551: }
1.22 schwarze 552: out = t ? time2a(t) : NULL;
1.23 schwarze 553: return(out ? out : mandoc_strdup(in));
1.5 schwarze 554: }
555:
1.9 schwarze 556: int
1.15 schwarze 557: mandoc_eos(const char *p, size_t sz, int enclosed)
1.9 schwarze 558: {
1.15 schwarze 559: const char *q;
1.16 schwarze 560: int found;
1.9 schwarze 561:
1.10 schwarze 562: if (0 == sz)
563: return(0);
1.9 schwarze 564:
1.11 schwarze 565: /*
566: * End-of-sentence recognition must include situations where
567: * some symbols, such as `)', allow prior EOS punctuation to
1.26 schwarze 568: * propagate outward.
1.11 schwarze 569: */
570:
1.16 schwarze 571: found = 0;
572: for (q = p + (int)sz - 1; q >= p; q--) {
1.15 schwarze 573: switch (*q) {
1.11 schwarze 574: case ('\"'):
575: /* FALLTHROUGH */
576: case ('\''):
577: /* FALLTHROUGH */
578: case (']'):
579: /* FALLTHROUGH */
580: case (')'):
1.15 schwarze 581: if (0 == found)
582: enclosed = 1;
1.11 schwarze 583: break;
584: case ('.'):
585: /* FALLTHROUGH */
586: case ('!'):
587: /* FALLTHROUGH */
588: case ('?'):
1.15 schwarze 589: found = 1;
590: break;
1.11 schwarze 591: default:
1.20 schwarze 592: return(found && (!enclosed || isalnum((unsigned char)*q)));
1.11 schwarze 593: }
1.9 schwarze 594: }
595:
1.15 schwarze 596: return(found && !enclosed);
1.24 schwarze 597: }
598:
599: /*
1.25 schwarze 600: * Find out whether a line is a macro line or not. If it is, adjust the
601: * current position and return one; if it isn't, return zero and don't
602: * change the current position.
1.24 schwarze 603: */
1.25 schwarze 604: int
605: mandoc_getcontrol(const char *cp, int *ppos)
1.24 schwarze 606: {
1.25 schwarze 607: int pos;
1.24 schwarze 608:
1.25 schwarze 609: pos = *ppos;
1.24 schwarze 610:
1.25 schwarze 611: if ('\\' == cp[pos] && '.' == cp[pos + 1])
612: pos += 2;
613: else if ('.' == cp[pos] || '\'' == cp[pos])
614: pos++;
615: else
616: return(0);
1.24 schwarze 617:
1.25 schwarze 618: while (' ' == cp[pos] || '\t' == cp[pos])
619: pos++;
1.24 schwarze 620:
1.25 schwarze 621: *ppos = pos;
622: return(1);
1.9 schwarze 623: }
1.26 schwarze 624:
625: /*
626: * Convert a string to a long that may not be <0.
627: * If the string is invalid, or is less than 0, return -1.
628: */
629: int
1.27 schwarze 630: mandoc_strntoi(const char *p, size_t sz, int base)
1.26 schwarze 631: {
632: char buf[32];
633: char *ep;
634: long v;
635:
636: if (sz > 31)
637: return(-1);
638:
639: memcpy(buf, p, sz);
640: buf[(int)sz] = '\0';
641:
642: errno = 0;
643: v = strtol(buf, &ep, base);
644:
645: if (buf[0] == '\0' || *ep != '\0')
646: return(-1);
647:
1.27 schwarze 648: if (v > INT_MAX)
649: v = INT_MAX;
650: if (v < INT_MIN)
651: v = INT_MIN;
1.26 schwarze 652:
653: return((int)v);
654: }
![[BACK]](/icons/back.gif){kind=icon}
Return to mandoc.c CVS log ![[TXT]](/icons/text.gif){kind=icon}
![[DIR]](/icons/dir.gif){kind=icon}
Up to [local] / src / usr.bin / mandoc