Annotation of src/usr.bin/mandoc/mandoc.c, Revision 1.63
1.63 ! schwarze 1: /* $OpenBSD: mandoc.c,v 1.62 2015/10/06 18:30:43 schwarze Exp $ */
1.1 schwarze 2: /*
1.58 schwarze 3: * Copyright (c) 2008-2011, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
4: * Copyright (c) 2011-2015 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"
1.45 schwarze 30: #include "mandoc_aux.h"
1.1 schwarze 31: #include "libmandoc.h"
32:
1.22 schwarze 33: #define DATESIZE 32
34:
1.14 schwarze 35: static int a2time(time_t *, const char *, const char *);
1.22 schwarze 36: static char *time2a(time_t);
1.5 schwarze 37:
1.26 schwarze 38:
39: enum mandoc_esc
1.44 schwarze 40: mandoc_escape(const char **end, const char **start, int *sz)
1.26 schwarze 41: {
1.34 schwarze 42: const char *local_start;
43: int local_sz;
44: char term;
1.48 schwarze 45: enum mandoc_esc gly;
1.26 schwarze 46:
1.34 schwarze 47: /*
48: * When the caller doesn't provide return storage,
49: * use local storage.
50: */
51:
52: if (NULL == start)
53: start = &local_start;
54: if (NULL == sz)
55: sz = &local_sz;
56:
57: /*
58: * Beyond the backslash, at least one input character
59: * is part of the escape sequence. With one exception
60: * (see below), that character won't be returned.
61: */
62:
1.26 schwarze 63: gly = ESCAPE_ERROR;
1.34 schwarze 64: *start = ++*end;
65: *sz = 0;
1.33 schwarze 66: term = '\0';
1.26 schwarze 67:
1.34 schwarze 68: switch ((*start)[-1]) {
1.26 schwarze 69: /*
70: * First the glyphs. There are several different forms of
71: * these, but each eventually returns a substring of the glyph
72: * name.
73: */
1.48 schwarze 74: case '(':
1.26 schwarze 75: gly = ESCAPE_SPECIAL;
1.34 schwarze 76: *sz = 2;
1.26 schwarze 77: break;
1.48 schwarze 78: case '[':
1.26 schwarze 79: gly = ESCAPE_SPECIAL;
80: term = ']';
81: break;
1.48 schwarze 82: case 'C':
1.34 schwarze 83: if ('\'' != **start)
1.62 schwarze 84: return ESCAPE_ERROR;
1.34 schwarze 85: *start = ++*end;
1.54 schwarze 86: gly = ESCAPE_SPECIAL;
1.26 schwarze 87: term = '\'';
88: break;
1.41 schwarze 89:
90: /*
91: * Escapes taking no arguments at all.
92: */
1.48 schwarze 93: case 'd':
94: case 'u':
1.61 schwarze 95: case ',':
96: case '/':
1.62 schwarze 97: return ESCAPE_IGNORE;
1.32 schwarze 98:
99: /*
100: * The \z escape is supposed to output the following
1.48 schwarze 101: * character without advancing the cursor position.
1.32 schwarze 102: * Since we are mostly dealing with terminal mode,
103: * let us just skip the next character.
104: */
1.48 schwarze 105: case 'z':
1.62 schwarze 106: return ESCAPE_SKIPCHAR;
1.1 schwarze 107:
1.26 schwarze 108: /*
109: * Handle all triggers matching \X(xy, \Xx, and \X[xxxx], where
110: * 'X' is the trigger. These have opaque sub-strings.
111: */
1.48 schwarze 112: case 'F':
113: case 'g':
114: case 'k':
115: case 'M':
116: case 'm':
117: case 'n':
118: case 'V':
119: case 'Y':
1.29 schwarze 120: gly = ESCAPE_IGNORE;
1.1 schwarze 121: /* FALLTHROUGH */
1.48 schwarze 122: case 'f':
1.26 schwarze 123: if (ESCAPE_ERROR == gly)
124: gly = ESCAPE_FONT;
1.34 schwarze 125: switch (**start) {
1.48 schwarze 126: case '(':
1.34 schwarze 127: *start = ++*end;
128: *sz = 2;
1.26 schwarze 129: break;
1.48 schwarze 130: case '[':
1.34 schwarze 131: *start = ++*end;
1.26 schwarze 132: term = ']';
133: break;
134: default:
1.34 schwarze 135: *sz = 1;
1.26 schwarze 136: break;
137: }
138: break;
139:
140: /*
141: * These escapes are of the form \X'Y', where 'X' is the trigger
142: * and 'Y' is any string. These have opaque sub-strings.
1.47 schwarze 143: * The \B and \w escapes are handled in roff.c, roff_res().
1.26 schwarze 144: */
1.48 schwarze 145: case 'A':
146: case 'b':
147: case 'D':
148: case 'R':
149: case 'X':
150: case 'Z':
1.59 schwarze 151: gly = ESCAPE_IGNORE;
152: /* FALLTHROUGH */
153: case 'o':
154: if (**start == '\0')
1.62 schwarze 155: return ESCAPE_ERROR;
1.59 schwarze 156: if (gly == ESCAPE_ERROR)
157: gly = ESCAPE_OVERSTRIKE;
1.46 schwarze 158: term = **start;
1.34 schwarze 159: *start = ++*end;
1.16 schwarze 160: break;
1.26 schwarze 161:
162: /*
163: * These escapes are of the form \X'N', where 'X' is the trigger
164: * and 'N' resolves to a numerical expression.
165: */
1.48 schwarze 166: case 'h':
167: case 'H':
168: case 'L':
169: case 'l':
170: case 'S':
171: case 'v':
172: case 'x':
1.51 schwarze 173: if (strchr(" %&()*+-./0123456789:<=>", **start)) {
1.53 schwarze 174: if ('\0' != **start)
175: ++*end;
1.62 schwarze 176: return ESCAPE_ERROR;
1.51 schwarze 177: }
1.42 schwarze 178: gly = ESCAPE_IGNORE;
1.46 schwarze 179: term = **start;
1.34 schwarze 180: *start = ++*end;
1.26 schwarze 181: break;
1.29 schwarze 182:
183: /*
184: * Special handling for the numbered character escape.
185: * XXX Do any other escapes need similar handling?
186: */
1.48 schwarze 187: case 'N':
1.34 schwarze 188: if ('\0' == **start)
1.62 schwarze 189: return ESCAPE_ERROR;
1.34 schwarze 190: (*end)++;
191: if (isdigit((unsigned char)**start)) {
192: *sz = 1;
1.62 schwarze 193: return ESCAPE_IGNORE;
1.34 schwarze 194: }
195: (*start)++;
1.29 schwarze 196: while (isdigit((unsigned char)**end))
197: (*end)++;
1.34 schwarze 198: *sz = *end - *start;
1.29 schwarze 199: if ('\0' != **end)
200: (*end)++;
1.62 schwarze 201: return ESCAPE_NUMBERED;
1.26 schwarze 202:
1.48 schwarze 203: /*
1.26 schwarze 204: * Sizes get a special category of their own.
205: */
1.48 schwarze 206: case 's':
1.26 schwarze 207: gly = ESCAPE_IGNORE;
1.17 schwarze 208:
1.26 schwarze 209: /* See +/- counts as a sign. */
1.34 schwarze 210: if ('+' == **end || '-' == **end || ASCII_HYPH == **end)
1.58 schwarze 211: *start = ++*end;
1.6 schwarze 212:
1.34 schwarze 213: switch (**end) {
1.48 schwarze 214: case '(':
1.34 schwarze 215: *start = ++*end;
216: *sz = 2;
1.16 schwarze 217: break;
1.48 schwarze 218: case '[':
1.34 schwarze 219: *start = ++*end;
1.33 schwarze 220: term = ']';
1.16 schwarze 221: break;
1.48 schwarze 222: case '\'':
1.34 schwarze 223: *start = ++*end;
1.33 schwarze 224: term = '\'';
1.60 schwarze 225: break;
226: case '3':
227: case '2':
228: case '1':
229: *sz = (*end)[-1] == 's' &&
230: isdigit((unsigned char)(*end)[1]) ? 2 : 1;
1.16 schwarze 231: break;
232: default:
1.34 schwarze 233: *sz = 1;
1.16 schwarze 234: break;
1.6 schwarze 235: }
236:
1.26 schwarze 237: break;
238:
239: /*
240: * Anything else is assumed to be a glyph.
1.34 schwarze 241: * In this case, pass back the character after the backslash.
1.26 schwarze 242: */
243: default:
244: gly = ESCAPE_SPECIAL;
1.34 schwarze 245: *start = --*end;
246: *sz = 1;
1.26 schwarze 247: break;
248: }
249:
250: assert(ESCAPE_ERROR != gly);
251:
252: /*
1.33 schwarze 253: * Read up to the terminating character,
254: * paying attention to nested escapes.
1.26 schwarze 255: */
256:
257: if ('\0' != term) {
1.33 schwarze 258: while (**end != term) {
259: switch (**end) {
1.48 schwarze 260: case '\0':
1.62 schwarze 261: return ESCAPE_ERROR;
1.48 schwarze 262: case '\\':
1.33 schwarze 263: (*end)++;
264: if (ESCAPE_ERROR ==
265: mandoc_escape(end, NULL, NULL))
1.62 schwarze 266: return ESCAPE_ERROR;
1.33 schwarze 267: break;
268: default:
269: (*end)++;
270: break;
271: }
272: }
1.34 schwarze 273: *sz = (*end)++ - *start;
1.33 schwarze 274: } else {
1.34 schwarze 275: assert(*sz > 0);
276: if ((size_t)*sz > strlen(*start))
1.62 schwarze 277: return ESCAPE_ERROR;
1.34 schwarze 278: *end += *sz;
1.26 schwarze 279: }
1.19 schwarze 280:
1.26 schwarze 281: /* Run post-processors. */
1.19 schwarze 282:
1.26 schwarze 283: switch (gly) {
1.48 schwarze 284: case ESCAPE_FONT:
1.37 schwarze 285: if (2 == *sz) {
286: if ('C' == **start) {
287: /*
288: * Treat constant-width font modes
289: * just like regular font modes.
290: */
291: (*start)++;
292: (*sz)--;
293: } else {
294: if ('B' == (*start)[0] && 'I' == (*start)[1])
295: gly = ESCAPE_FONTBI;
296: break;
297: }
1.34 schwarze 298: } else if (1 != *sz)
1.26 schwarze 299: break;
1.30 schwarze 300:
1.34 schwarze 301: switch (**start) {
1.48 schwarze 302: case '3':
303: case 'B':
1.26 schwarze 304: gly = ESCAPE_FONTBOLD;
305: break;
1.48 schwarze 306: case '2':
307: case 'I':
1.26 schwarze 308: gly = ESCAPE_FONTITALIC;
1.16 schwarze 309: break;
1.48 schwarze 310: case 'P':
1.26 schwarze 311: gly = ESCAPE_FONTPREV;
1.16 schwarze 312: break;
1.48 schwarze 313: case '1':
314: case 'R':
1.26 schwarze 315: gly = ESCAPE_FONTROMAN;
1.1 schwarze 316: break;
317: }
1.16 schwarze 318: break;
1.48 schwarze 319: case ESCAPE_SPECIAL:
1.34 schwarze 320: if (1 == *sz && 'c' == **start)
1.26 schwarze 321: gly = ESCAPE_NOSPACE;
1.54 schwarze 322: /*
1.55 schwarze 323: * Unicode escapes are defined in groff as \[u0000]
1.54 schwarze 324: * to \[u10FFFF], where the contained value must be
325: * a valid Unicode codepoint. Here, however, only
1.55 schwarze 326: * check the length and range.
1.54 schwarze 327: */
1.55 schwarze 328: if (**start != 'u' || *sz < 5 || *sz > 7)
329: break;
330: if (*sz == 7 && ((*start)[1] != '1' || (*start)[2] != '0'))
331: break;
332: if (*sz == 6 && (*start)[1] == '0')
333: break;
334: if ((int)strspn(*start + 1, "0123456789ABCDEFabcdef")
1.54 schwarze 335: + 1 == *sz)
336: gly = ESCAPE_UNICODE;
1.16 schwarze 337: break;
1.1 schwarze 338: default:
1.16 schwarze 339: break;
1.1 schwarze 340: }
341:
1.62 schwarze 342: return gly;
1.21 schwarze 343: }
344:
345: /*
346: * Parse a quoted or unquoted roff-style request or macro argument.
347: * Return a pointer to the parsed argument, which is either the original
348: * pointer or advanced by one byte in case the argument is quoted.
1.40 schwarze 349: * NUL-terminate the argument in place.
1.21 schwarze 350: * Collapse pairs of quotes inside quoted arguments.
351: * Advance the argument pointer to the next argument,
1.40 schwarze 352: * or to the NUL byte terminating the argument line.
1.21 schwarze 353: */
354: char *
1.25 schwarze 355: mandoc_getarg(struct mparse *parse, char **cpp, int ln, int *pos)
1.21 schwarze 356: {
357: char *start, *cp;
358: int quoted, pairs, white;
359:
360: /* Quoting can only start with a new word. */
361: start = *cpp;
1.26 schwarze 362: quoted = 0;
1.21 schwarze 363: if ('"' == *start) {
364: quoted = 1;
365: start++;
1.48 schwarze 366: }
1.21 schwarze 367:
368: pairs = 0;
369: white = 0;
370: for (cp = start; '\0' != *cp; cp++) {
1.36 schwarze 371:
372: /*
373: * Move the following text left
374: * after quoted quotes and after "\\" and "\t".
375: */
1.21 schwarze 376: if (pairs)
377: cp[-pairs] = cp[0];
1.36 schwarze 378:
1.21 schwarze 379: if ('\\' == cp[0]) {
1.36 schwarze 380: /*
381: * In copy mode, translate double to single
382: * backslashes and backslash-t to literal tabs.
383: */
384: switch (cp[1]) {
1.48 schwarze 385: case 't':
1.36 schwarze 386: cp[0] = '\t';
387: /* FALLTHROUGH */
1.48 schwarze 388: case '\\':
1.21 schwarze 389: pairs++;
390: cp++;
1.36 schwarze 391: break;
1.48 schwarze 392: case ' ':
1.21 schwarze 393: /* Skip escaped blanks. */
1.36 schwarze 394: if (0 == quoted)
395: cp++;
396: break;
397: default:
398: break;
399: }
1.21 schwarze 400: } else if (0 == quoted) {
401: if (' ' == cp[0]) {
402: /* Unescaped blanks end unquoted args. */
403: white = 1;
404: break;
405: }
406: } else if ('"' == cp[0]) {
407: if ('"' == cp[1]) {
408: /* Quoted quotes collapse. */
409: pairs++;
410: cp++;
411: } else {
412: /* Unquoted quotes end quoted args. */
413: quoted = 2;
414: break;
415: }
416: }
417: }
418:
419: /* Quoted argument without a closing quote. */
1.25 schwarze 420: if (1 == quoted)
1.52 schwarze 421: mandoc_msg(MANDOCERR_ARG_QUOTE, parse, ln, *pos, NULL);
1.21 schwarze 422:
1.40 schwarze 423: /* NUL-terminate this argument and move to the next one. */
1.21 schwarze 424: if (pairs)
425: cp[-pairs] = '\0';
426: if ('\0' != *cp) {
427: *cp++ = '\0';
428: while (' ' == *cp)
429: cp++;
430: }
1.24 schwarze 431: *pos += (int)(cp - start) + (quoted ? 1 : 0);
1.21 schwarze 432: *cpp = cp;
433:
1.25 schwarze 434: if ('\0' == *cp && (white || ' ' == cp[-1]))
1.52 schwarze 435: mandoc_msg(MANDOCERR_SPACE_EOL, parse, ln, *pos, NULL);
1.21 schwarze 436:
1.62 schwarze 437: return start;
1.4 schwarze 438: }
1.5 schwarze 439:
440: static int
441: a2time(time_t *t, const char *fmt, const char *p)
442: {
443: struct tm tm;
444: char *pp;
445:
446: memset(&tm, 0, sizeof(struct tm));
447:
448: pp = strptime(p, fmt, &tm);
449: if (NULL != pp && '\0' == *pp) {
450: *t = mktime(&tm);
1.62 schwarze 451: return 1;
1.5 schwarze 452: }
453:
1.62 schwarze 454: return 0;
1.5 schwarze 455: }
456:
1.22 schwarze 457: static char *
458: time2a(time_t t)
459: {
1.28 schwarze 460: struct tm *tm;
1.23 schwarze 461: char *buf, *p;
462: size_t ssz;
1.22 schwarze 463: int isz;
464:
1.28 schwarze 465: tm = localtime(&t);
1.57 schwarze 466: if (tm == NULL)
1.62 schwarze 467: return NULL;
1.22 schwarze 468:
1.23 schwarze 469: /*
470: * Reserve space:
471: * up to 9 characters for the month (September) + blank
472: * up to 2 characters for the day + comma + blank
473: * 4 characters for the year and a terminating '\0'
474: */
475: p = buf = mandoc_malloc(10 + 4 + 4 + 1);
476:
1.28 schwarze 477: if (0 == (ssz = strftime(p, 10 + 1, "%B ", tm)))
1.23 schwarze 478: goto fail;
479: p += (int)ssz;
1.22 schwarze 480:
1.28 schwarze 481: if (-1 == (isz = snprintf(p, 4 + 1, "%d, ", tm->tm_mday)))
1.23 schwarze 482: goto fail;
1.22 schwarze 483: p += isz;
484:
1.28 schwarze 485: if (0 == strftime(p, 4 + 1, "%Y", tm))
1.23 schwarze 486: goto fail;
1.62 schwarze 487: return buf;
1.23 schwarze 488:
489: fail:
490: free(buf);
1.62 schwarze 491: return NULL;
1.22 schwarze 492: }
493:
494: char *
1.25 schwarze 495: mandoc_normdate(struct mparse *parse, char *in, int ln, int pos)
1.5 schwarze 496: {
1.22 schwarze 497: char *out;
1.5 schwarze 498: time_t t;
499:
1.22 schwarze 500: if (NULL == in || '\0' == *in ||
501: 0 == strcmp(in, "$" "Mdocdate$")) {
1.49 schwarze 502: mandoc_msg(MANDOCERR_DATE_MISSING, parse, ln, pos, NULL);
1.22 schwarze 503: time(&t);
504: }
1.31 schwarze 505: else if (a2time(&t, "%Y-%m-%d", in))
506: t = 0;
1.22 schwarze 507: else if (!a2time(&t, "$" "Mdocdate: %b %d %Y $", in) &&
1.31 schwarze 508: !a2time(&t, "%b %d, %Y", in)) {
1.50 schwarze 509: mandoc_msg(MANDOCERR_DATE_BAD, parse, ln, pos, in);
1.22 schwarze 510: t = 0;
1.5 schwarze 511: }
1.22 schwarze 512: out = t ? time2a(t) : NULL;
1.62 schwarze 513: return out ? out : mandoc_strdup(in);
1.5 schwarze 514: }
515:
1.9 schwarze 516: int
1.43 schwarze 517: mandoc_eos(const char *p, size_t sz)
1.9 schwarze 518: {
1.43 schwarze 519: const char *q;
520: int enclosed, found;
1.9 schwarze 521:
1.10 schwarze 522: if (0 == sz)
1.62 schwarze 523: return 0;
1.9 schwarze 524:
1.11 schwarze 525: /*
526: * End-of-sentence recognition must include situations where
527: * some symbols, such as `)', allow prior EOS punctuation to
1.26 schwarze 528: * propagate outward.
1.11 schwarze 529: */
530:
1.43 schwarze 531: enclosed = found = 0;
1.16 schwarze 532: for (q = p + (int)sz - 1; q >= p; q--) {
1.15 schwarze 533: switch (*q) {
1.48 schwarze 534: case '\"':
535: case '\'':
536: case ']':
537: case ')':
1.15 schwarze 538: if (0 == found)
539: enclosed = 1;
1.11 schwarze 540: break;
1.48 schwarze 541: case '.':
542: case '!':
543: case '?':
1.15 schwarze 544: found = 1;
545: break;
1.11 schwarze 546: default:
1.62 schwarze 547: return found &&
548: (!enclosed || isalnum((unsigned char)*q));
1.11 schwarze 549: }
1.9 schwarze 550: }
551:
1.62 schwarze 552: return found && !enclosed;
1.9 schwarze 553: }
1.26 schwarze 554:
555: /*
556: * Convert a string to a long that may not be <0.
557: * If the string is invalid, or is less than 0, return -1.
558: */
559: int
1.27 schwarze 560: mandoc_strntoi(const char *p, size_t sz, int base)
1.26 schwarze 561: {
562: char buf[32];
563: char *ep;
564: long v;
565:
566: if (sz > 31)
1.62 schwarze 567: return -1;
1.26 schwarze 568:
569: memcpy(buf, p, sz);
570: buf[(int)sz] = '\0';
571:
572: errno = 0;
573: v = strtol(buf, &ep, base);
574:
575: if (buf[0] == '\0' || *ep != '\0')
1.62 schwarze 576: return -1;
1.26 schwarze 577:
1.27 schwarze 578: if (v > INT_MAX)
579: v = INT_MAX;
580: if (v < INT_MIN)
581: v = INT_MIN;
1.26 schwarze 582:
1.62 schwarze 583: return (int)v;
1.26 schwarze 584: }
![[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