Annotation of src/usr.bin/mandoc/mandoc.c, Revision 1.65
1.65 ! schwarze 1: /* $OpenBSD: mandoc.c,v 1.64 2015/10/13 23:30:42 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"
1.22 schwarze 32:
1.14 schwarze 33: static int a2time(time_t *, const char *, const char *);
1.22 schwarze 34: static char *time2a(time_t);
1.5 schwarze 35:
1.26 schwarze 36:
37: enum mandoc_esc
1.44 schwarze 38: mandoc_escape(const char **end, const char **start, int *sz)
1.26 schwarze 39: {
1.34 schwarze 40: const char *local_start;
41: int local_sz;
42: char term;
1.48 schwarze 43: enum mandoc_esc gly;
1.26 schwarze 44:
1.34 schwarze 45: /*
46: * When the caller doesn't provide return storage,
47: * use local storage.
48: */
49:
50: if (NULL == start)
51: start = &local_start;
52: if (NULL == sz)
53: sz = &local_sz;
54:
55: /*
56: * Beyond the backslash, at least one input character
57: * is part of the escape sequence. With one exception
58: * (see below), that character won't be returned.
59: */
60:
1.26 schwarze 61: gly = ESCAPE_ERROR;
1.34 schwarze 62: *start = ++*end;
63: *sz = 0;
1.33 schwarze 64: term = '\0';
1.26 schwarze 65:
1.34 schwarze 66: switch ((*start)[-1]) {
1.26 schwarze 67: /*
68: * First the glyphs. There are several different forms of
69: * these, but each eventually returns a substring of the glyph
70: * name.
71: */
1.48 schwarze 72: case '(':
1.26 schwarze 73: gly = ESCAPE_SPECIAL;
1.34 schwarze 74: *sz = 2;
1.26 schwarze 75: break;
1.48 schwarze 76: case '[':
1.26 schwarze 77: gly = ESCAPE_SPECIAL;
78: term = ']';
79: break;
1.48 schwarze 80: case 'C':
1.34 schwarze 81: if ('\'' != **start)
1.62 schwarze 82: return ESCAPE_ERROR;
1.34 schwarze 83: *start = ++*end;
1.54 schwarze 84: gly = ESCAPE_SPECIAL;
1.26 schwarze 85: term = '\'';
86: break;
1.41 schwarze 87:
88: /*
89: * Escapes taking no arguments at all.
90: */
1.48 schwarze 91: case 'd':
92: case 'u':
1.61 schwarze 93: case ',':
94: case '/':
1.62 schwarze 95: return ESCAPE_IGNORE;
1.32 schwarze 96:
97: /*
98: * The \z escape is supposed to output the following
1.48 schwarze 99: * character without advancing the cursor position.
1.32 schwarze 100: * Since we are mostly dealing with terminal mode,
101: * let us just skip the next character.
102: */
1.48 schwarze 103: case 'z':
1.62 schwarze 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: */
1.48 schwarze 110: case 'F':
111: case 'g':
112: case 'k':
113: case 'M':
114: case 'm':
115: case 'n':
116: case 'V':
117: case 'Y':
1.29 schwarze 118: gly = ESCAPE_IGNORE;
1.1 schwarze 119: /* FALLTHROUGH */
1.48 schwarze 120: case 'f':
1.26 schwarze 121: if (ESCAPE_ERROR == gly)
122: gly = ESCAPE_FONT;
1.34 schwarze 123: switch (**start) {
1.48 schwarze 124: case '(':
1.34 schwarze 125: *start = ++*end;
126: *sz = 2;
1.26 schwarze 127: break;
1.48 schwarze 128: case '[':
1.34 schwarze 129: *start = ++*end;
1.26 schwarze 130: term = ']';
131: break;
132: default:
1.34 schwarze 133: *sz = 1;
1.26 schwarze 134: break;
135: }
136: break;
137:
138: /*
139: * These escapes are of the form \X'Y', where 'X' is the trigger
140: * and 'Y' is any string. These have opaque sub-strings.
1.47 schwarze 141: * The \B and \w escapes are handled in roff.c, roff_res().
1.26 schwarze 142: */
1.48 schwarze 143: case 'A':
144: case 'b':
145: case 'D':
146: case 'R':
147: case 'X':
148: case 'Z':
1.59 schwarze 149: gly = ESCAPE_IGNORE;
150: /* FALLTHROUGH */
151: case 'o':
152: if (**start == '\0')
1.62 schwarze 153: return ESCAPE_ERROR;
1.59 schwarze 154: if (gly == ESCAPE_ERROR)
155: gly = ESCAPE_OVERSTRIKE;
1.46 schwarze 156: term = **start;
1.34 schwarze 157: *start = ++*end;
1.16 schwarze 158: break;
1.26 schwarze 159:
160: /*
161: * These escapes are of the form \X'N', where 'X' is the trigger
162: * and 'N' resolves to a numerical expression.
163: */
1.48 schwarze 164: case 'h':
165: case 'H':
166: case 'L':
167: case 'l':
168: case 'S':
169: case 'v':
170: case 'x':
1.51 schwarze 171: if (strchr(" %&()*+-./0123456789:<=>", **start)) {
1.53 schwarze 172: if ('\0' != **start)
173: ++*end;
1.62 schwarze 174: return ESCAPE_ERROR;
1.51 schwarze 175: }
1.42 schwarze 176: gly = ESCAPE_IGNORE;
1.46 schwarze 177: term = **start;
1.34 schwarze 178: *start = ++*end;
1.26 schwarze 179: break;
1.29 schwarze 180:
181: /*
182: * Special handling for the numbered character escape.
183: * XXX Do any other escapes need similar handling?
184: */
1.48 schwarze 185: case 'N':
1.34 schwarze 186: if ('\0' == **start)
1.62 schwarze 187: return ESCAPE_ERROR;
1.34 schwarze 188: (*end)++;
189: if (isdigit((unsigned char)**start)) {
190: *sz = 1;
1.62 schwarze 191: return ESCAPE_IGNORE;
1.34 schwarze 192: }
193: (*start)++;
1.29 schwarze 194: while (isdigit((unsigned char)**end))
195: (*end)++;
1.34 schwarze 196: *sz = *end - *start;
1.29 schwarze 197: if ('\0' != **end)
198: (*end)++;
1.62 schwarze 199: return ESCAPE_NUMBERED;
1.26 schwarze 200:
1.48 schwarze 201: /*
1.26 schwarze 202: * Sizes get a special category of their own.
203: */
1.48 schwarze 204: case 's':
1.26 schwarze 205: gly = ESCAPE_IGNORE;
1.17 schwarze 206:
1.26 schwarze 207: /* See +/- counts as a sign. */
1.34 schwarze 208: if ('+' == **end || '-' == **end || ASCII_HYPH == **end)
1.58 schwarze 209: *start = ++*end;
1.6 schwarze 210:
1.34 schwarze 211: switch (**end) {
1.48 schwarze 212: case '(':
1.34 schwarze 213: *start = ++*end;
214: *sz = 2;
1.16 schwarze 215: break;
1.48 schwarze 216: case '[':
1.34 schwarze 217: *start = ++*end;
1.33 schwarze 218: term = ']';
1.16 schwarze 219: break;
1.48 schwarze 220: case '\'':
1.34 schwarze 221: *start = ++*end;
1.33 schwarze 222: term = '\'';
1.60 schwarze 223: break;
224: case '3':
225: case '2':
226: case '1':
227: *sz = (*end)[-1] == 's' &&
228: isdigit((unsigned char)(*end)[1]) ? 2 : 1;
1.16 schwarze 229: break;
230: default:
1.34 schwarze 231: *sz = 1;
1.16 schwarze 232: break;
1.6 schwarze 233: }
234:
1.26 schwarze 235: break;
236:
237: /*
238: * Anything else is assumed to be a glyph.
1.34 schwarze 239: * In this case, pass back the character after the backslash.
1.26 schwarze 240: */
241: default:
242: gly = ESCAPE_SPECIAL;
1.34 schwarze 243: *start = --*end;
244: *sz = 1;
1.26 schwarze 245: break;
246: }
247:
248: assert(ESCAPE_ERROR != gly);
249:
250: /*
1.33 schwarze 251: * Read up to the terminating character,
252: * paying attention to nested escapes.
1.26 schwarze 253: */
254:
255: if ('\0' != term) {
1.33 schwarze 256: while (**end != term) {
257: switch (**end) {
1.48 schwarze 258: case '\0':
1.62 schwarze 259: return ESCAPE_ERROR;
1.48 schwarze 260: case '\\':
1.33 schwarze 261: (*end)++;
262: if (ESCAPE_ERROR ==
263: mandoc_escape(end, NULL, NULL))
1.62 schwarze 264: return ESCAPE_ERROR;
1.33 schwarze 265: break;
266: default:
267: (*end)++;
268: break;
269: }
270: }
1.34 schwarze 271: *sz = (*end)++ - *start;
1.33 schwarze 272: } else {
1.34 schwarze 273: assert(*sz > 0);
274: if ((size_t)*sz > strlen(*start))
1.62 schwarze 275: return ESCAPE_ERROR;
1.34 schwarze 276: *end += *sz;
1.26 schwarze 277: }
1.19 schwarze 278:
1.26 schwarze 279: /* Run post-processors. */
1.19 schwarze 280:
1.26 schwarze 281: switch (gly) {
1.48 schwarze 282: case ESCAPE_FONT:
1.37 schwarze 283: if (2 == *sz) {
284: if ('C' == **start) {
285: /*
286: * Treat constant-width font modes
287: * just like regular font modes.
288: */
289: (*start)++;
290: (*sz)--;
291: } else {
292: if ('B' == (*start)[0] && 'I' == (*start)[1])
293: gly = ESCAPE_FONTBI;
294: break;
295: }
1.34 schwarze 296: } else if (1 != *sz)
1.26 schwarze 297: break;
1.30 schwarze 298:
1.34 schwarze 299: switch (**start) {
1.48 schwarze 300: case '3':
301: case 'B':
1.26 schwarze 302: gly = ESCAPE_FONTBOLD;
303: break;
1.48 schwarze 304: case '2':
305: case 'I':
1.26 schwarze 306: gly = ESCAPE_FONTITALIC;
1.16 schwarze 307: break;
1.48 schwarze 308: case 'P':
1.26 schwarze 309: gly = ESCAPE_FONTPREV;
1.16 schwarze 310: break;
1.48 schwarze 311: case '1':
312: case 'R':
1.26 schwarze 313: gly = ESCAPE_FONTROMAN;
1.1 schwarze 314: break;
315: }
1.16 schwarze 316: break;
1.48 schwarze 317: case ESCAPE_SPECIAL:
1.34 schwarze 318: if (1 == *sz && 'c' == **start)
1.26 schwarze 319: gly = ESCAPE_NOSPACE;
1.54 schwarze 320: /*
1.55 schwarze 321: * Unicode escapes are defined in groff as \[u0000]
1.54 schwarze 322: * to \[u10FFFF], where the contained value must be
323: * a valid Unicode codepoint. Here, however, only
1.55 schwarze 324: * check the length and range.
1.54 schwarze 325: */
1.55 schwarze 326: if (**start != 'u' || *sz < 5 || *sz > 7)
327: break;
328: if (*sz == 7 && ((*start)[1] != '1' || (*start)[2] != '0'))
329: break;
330: if (*sz == 6 && (*start)[1] == '0')
1.64 schwarze 331: break;
332: if (*sz == 5 && (*start)[1] == 'D' &&
333: strchr("89ABCDEF", (*start)[2]) != NULL)
1.55 schwarze 334: break;
335: if ((int)strspn(*start + 1, "0123456789ABCDEFabcdef")
1.54 schwarze 336: + 1 == *sz)
337: gly = ESCAPE_UNICODE;
1.16 schwarze 338: break;
1.1 schwarze 339: default:
1.16 schwarze 340: break;
1.1 schwarze 341: }
342:
1.62 schwarze 343: return gly;
1.21 schwarze 344: }
345:
346: /*
347: * Parse a quoted or unquoted roff-style request or macro argument.
348: * Return a pointer to the parsed argument, which is either the original
349: * pointer or advanced by one byte in case the argument is quoted.
1.40 schwarze 350: * NUL-terminate the argument in place.
1.21 schwarze 351: * Collapse pairs of quotes inside quoted arguments.
352: * Advance the argument pointer to the next argument,
1.40 schwarze 353: * or to the NUL byte terminating the argument line.
1.21 schwarze 354: */
355: char *
1.25 schwarze 356: mandoc_getarg(struct mparse *parse, char **cpp, int ln, int *pos)
1.21 schwarze 357: {
358: char *start, *cp;
359: int quoted, pairs, white;
360:
361: /* Quoting can only start with a new word. */
362: start = *cpp;
1.26 schwarze 363: quoted = 0;
1.21 schwarze 364: if ('"' == *start) {
365: quoted = 1;
366: start++;
1.48 schwarze 367: }
1.21 schwarze 368:
369: pairs = 0;
370: white = 0;
371: for (cp = start; '\0' != *cp; cp++) {
1.36 schwarze 372:
373: /*
374: * Move the following text left
375: * after quoted quotes and after "\\" and "\t".
376: */
1.21 schwarze 377: if (pairs)
378: cp[-pairs] = cp[0];
1.36 schwarze 379:
1.21 schwarze 380: if ('\\' == cp[0]) {
1.36 schwarze 381: /*
382: * In copy mode, translate double to single
383: * backslashes and backslash-t to literal tabs.
384: */
385: switch (cp[1]) {
1.48 schwarze 386: case 't':
1.36 schwarze 387: cp[0] = '\t';
388: /* FALLTHROUGH */
1.48 schwarze 389: case '\\':
1.21 schwarze 390: pairs++;
391: cp++;
1.36 schwarze 392: break;
1.48 schwarze 393: case ' ':
1.21 schwarze 394: /* Skip escaped blanks. */
1.36 schwarze 395: if (0 == quoted)
396: cp++;
397: break;
398: default:
399: break;
400: }
1.21 schwarze 401: } else if (0 == quoted) {
402: if (' ' == cp[0]) {
403: /* Unescaped blanks end unquoted args. */
404: white = 1;
405: break;
406: }
407: } else if ('"' == cp[0]) {
408: if ('"' == cp[1]) {
409: /* Quoted quotes collapse. */
410: pairs++;
411: cp++;
412: } else {
413: /* Unquoted quotes end quoted args. */
414: quoted = 2;
415: break;
416: }
417: }
418: }
419:
420: /* Quoted argument without a closing quote. */
1.25 schwarze 421: if (1 == quoted)
1.52 schwarze 422: mandoc_msg(MANDOCERR_ARG_QUOTE, parse, ln, *pos, NULL);
1.21 schwarze 423:
1.40 schwarze 424: /* NUL-terminate this argument and move to the next one. */
1.21 schwarze 425: if (pairs)
426: cp[-pairs] = '\0';
427: if ('\0' != *cp) {
428: *cp++ = '\0';
429: while (' ' == *cp)
430: cp++;
431: }
1.24 schwarze 432: *pos += (int)(cp - start) + (quoted ? 1 : 0);
1.21 schwarze 433: *cpp = cp;
434:
1.25 schwarze 435: if ('\0' == *cp && (white || ' ' == cp[-1]))
1.52 schwarze 436: mandoc_msg(MANDOCERR_SPACE_EOL, parse, ln, *pos, NULL);
1.21 schwarze 437:
1.62 schwarze 438: return start;
1.4 schwarze 439: }
1.5 schwarze 440:
441: static int
442: a2time(time_t *t, const char *fmt, const char *p)
443: {
444: struct tm tm;
445: char *pp;
446:
447: memset(&tm, 0, sizeof(struct tm));
448:
449: pp = strptime(p, fmt, &tm);
450: if (NULL != pp && '\0' == *pp) {
451: *t = mktime(&tm);
1.62 schwarze 452: return 1;
1.5 schwarze 453: }
454:
1.62 schwarze 455: return 0;
1.5 schwarze 456: }
457:
1.22 schwarze 458: static char *
459: time2a(time_t t)
460: {
1.28 schwarze 461: struct tm *tm;
1.23 schwarze 462: char *buf, *p;
463: size_t ssz;
1.22 schwarze 464: int isz;
465:
1.28 schwarze 466: tm = localtime(&t);
1.57 schwarze 467: if (tm == NULL)
1.62 schwarze 468: return NULL;
1.22 schwarze 469:
1.23 schwarze 470: /*
471: * Reserve space:
472: * up to 9 characters for the month (September) + blank
473: * up to 2 characters for the day + comma + blank
474: * 4 characters for the year and a terminating '\0'
475: */
476: p = buf = mandoc_malloc(10 + 4 + 4 + 1);
477:
1.28 schwarze 478: if (0 == (ssz = strftime(p, 10 + 1, "%B ", tm)))
1.23 schwarze 479: goto fail;
480: p += (int)ssz;
1.22 schwarze 481:
1.28 schwarze 482: if (-1 == (isz = snprintf(p, 4 + 1, "%d, ", tm->tm_mday)))
1.23 schwarze 483: goto fail;
1.22 schwarze 484: p += isz;
485:
1.28 schwarze 486: if (0 == strftime(p, 4 + 1, "%Y", tm))
1.23 schwarze 487: goto fail;
1.62 schwarze 488: return buf;
1.23 schwarze 489:
490: fail:
491: free(buf);
1.62 schwarze 492: return NULL;
1.22 schwarze 493: }
494:
495: char *
1.25 schwarze 496: mandoc_normdate(struct mparse *parse, char *in, int ln, int pos)
1.5 schwarze 497: {
1.22 schwarze 498: char *out;
1.5 schwarze 499: time_t t;
500:
1.22 schwarze 501: if (NULL == in || '\0' == *in ||
502: 0 == strcmp(in, "$" "Mdocdate$")) {
1.49 schwarze 503: mandoc_msg(MANDOCERR_DATE_MISSING, parse, ln, pos, NULL);
1.22 schwarze 504: time(&t);
505: }
1.31 schwarze 506: else if (a2time(&t, "%Y-%m-%d", in))
507: t = 0;
1.22 schwarze 508: else if (!a2time(&t, "$" "Mdocdate: %b %d %Y $", in) &&
1.31 schwarze 509: !a2time(&t, "%b %d, %Y", in)) {
1.50 schwarze 510: mandoc_msg(MANDOCERR_DATE_BAD, parse, ln, pos, in);
1.22 schwarze 511: t = 0;
1.5 schwarze 512: }
1.22 schwarze 513: out = t ? time2a(t) : NULL;
1.62 schwarze 514: return out ? out : mandoc_strdup(in);
1.5 schwarze 515: }
516:
1.9 schwarze 517: int
1.43 schwarze 518: mandoc_eos(const char *p, size_t sz)
1.9 schwarze 519: {
1.43 schwarze 520: const char *q;
521: int enclosed, found;
1.9 schwarze 522:
1.10 schwarze 523: if (0 == sz)
1.62 schwarze 524: return 0;
1.9 schwarze 525:
1.11 schwarze 526: /*
527: * End-of-sentence recognition must include situations where
528: * some symbols, such as `)', allow prior EOS punctuation to
1.26 schwarze 529: * propagate outward.
1.11 schwarze 530: */
531:
1.43 schwarze 532: enclosed = found = 0;
1.16 schwarze 533: for (q = p + (int)sz - 1; q >= p; q--) {
1.15 schwarze 534: switch (*q) {
1.48 schwarze 535: case '\"':
536: case '\'':
537: case ']':
538: case ')':
1.15 schwarze 539: if (0 == found)
540: enclosed = 1;
1.11 schwarze 541: break;
1.48 schwarze 542: case '.':
543: case '!':
544: case '?':
1.15 schwarze 545: found = 1;
546: break;
1.11 schwarze 547: default:
1.62 schwarze 548: return found &&
549: (!enclosed || isalnum((unsigned char)*q));
1.11 schwarze 550: }
1.9 schwarze 551: }
552:
1.62 schwarze 553: return found && !enclosed;
1.9 schwarze 554: }
1.26 schwarze 555:
556: /*
557: * Convert a string to a long that may not be <0.
558: * If the string is invalid, or is less than 0, return -1.
559: */
560: int
1.27 schwarze 561: mandoc_strntoi(const char *p, size_t sz, int base)
1.26 schwarze 562: {
563: char buf[32];
564: char *ep;
565: long v;
566:
567: if (sz > 31)
1.62 schwarze 568: return -1;
1.26 schwarze 569:
570: memcpy(buf, p, sz);
571: buf[(int)sz] = '\0';
572:
573: errno = 0;
574: v = strtol(buf, &ep, base);
575:
576: if (buf[0] == '\0' || *ep != '\0')
1.62 schwarze 577: return -1;
1.26 schwarze 578:
1.27 schwarze 579: if (v > INT_MAX)
580: v = INT_MAX;
581: if (v < INT_MIN)
582: v = INT_MIN;
1.26 schwarze 583:
1.62 schwarze 584: return (int)v;
1.26 schwarze 585: }
![[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