1 /* $Id: mandoc.c,v 1.103 2017/07/03 13:40:19 schwarze Exp $ */
3 * Copyright (c) 2008-2011, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2011-2015, 2017 Ingo Schwarze <schwarze@openbsd.org>
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.
10 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
11 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
12 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
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.
20 #include <sys/types.h>
31 #include "mandoc_aux.h"
34 #include "libmandoc.h"
36 static int a2time(time_t *, const char *, const char *);
37 static char *time2a(time_t);
41 mandoc_escape(const char **end, const char **start, int *sz)
43 const char *local_start;
49 * When the caller doesn't provide return storage,
59 * Beyond the backslash, at least one input character
60 * is part of the escape sequence. With one exception
61 * (see below), that character won't be returned.
69 switch ((*start)[-1]) {
71 * First the glyphs. There are several different forms of
72 * these, but each eventually returns a substring of the glyph
92 * Escapes taking no arguments at all.
103 * The \z escape is supposed to output the following
104 * character without advancing the cursor position.
105 * Since we are mostly dealing with terminal mode,
106 * let us just skip the next character.
109 return ESCAPE_SKIPCHAR;
112 * Handle all triggers matching \X(xy, \Xx, and \X[xxxx], where
113 * 'X' is the trigger. These have opaque sub-strings.
126 if (ESCAPE_ERROR == gly)
144 * These escapes are of the form \X'Y', where 'X' is the trigger
145 * and 'Y' is any string. These have opaque sub-strings.
146 * The \B and \w escapes are handled in roff.c, roff_res().
159 if (gly == ESCAPE_ERROR)
160 gly = ESCAPE_OVERSTRIKE;
166 * These escapes are of the form \X'N', where 'X' is the trigger
167 * and 'N' resolves to a numerical expression.
176 if (strchr(" %&()*+-./0123456789:<=>", **start)) {
181 switch ((*start)[-1]) {
197 * Special handling for the numbered character escape.
198 * XXX Do any other escapes need similar handling?
204 if (isdigit((unsigned char)**start)) {
206 return ESCAPE_IGNORE;
209 while (isdigit((unsigned char)**end))
214 return ESCAPE_NUMBERED;
217 * Sizes get a special category of their own.
222 /* See +/- counts as a sign. */
223 if ('+' == **end || '-' == **end || ASCII_HYPH == **end)
242 *sz = (*end)[-1] == 's' &&
243 isdigit((unsigned char)(*end)[1]) ? 2 : 1;
253 * Anything else is assumed to be a glyph.
254 * In this case, pass back the character after the backslash.
257 gly = ESCAPE_SPECIAL;
263 assert(ESCAPE_ERROR != gly);
266 * Read up to the terminating character,
267 * paying attention to nested escapes.
271 while (**end != term) {
278 mandoc_escape(end, NULL, NULL))
286 *sz = (*end)++ - *start;
289 if ((size_t)*sz > strlen(*start))
294 /* Run post-processors. */
299 if ('C' == **start) {
301 * Treat constant-width font modes
302 * just like regular font modes.
307 if ('B' == (*start)[0] && 'I' == (*start)[1])
317 gly = ESCAPE_FONTBOLD;
321 gly = ESCAPE_FONTITALIC;
324 gly = ESCAPE_FONTPREV;
328 gly = ESCAPE_FONTROMAN;
333 if (1 == *sz && 'c' == **start)
334 gly = ESCAPE_NOSPACE;
336 * Unicode escapes are defined in groff as \[u0000]
337 * to \[u10FFFF], where the contained value must be
338 * a valid Unicode codepoint. Here, however, only
339 * check the length and range.
341 if (**start != 'u' || *sz < 5 || *sz > 7)
343 if (*sz == 7 && ((*start)[1] != '1' || (*start)[2] != '0'))
345 if (*sz == 6 && (*start)[1] == '0')
347 if (*sz == 5 && (*start)[1] == 'D' &&
348 strchr("89ABCDEF", (*start)[2]) != NULL)
350 if ((int)strspn(*start + 1, "0123456789ABCDEFabcdef")
352 gly = ESCAPE_UNICODE;
362 * Parse a quoted or unquoted roff-style request or macro argument.
363 * Return a pointer to the parsed argument, which is either the original
364 * pointer or advanced by one byte in case the argument is quoted.
365 * NUL-terminate the argument in place.
366 * Collapse pairs of quotes inside quoted arguments.
367 * Advance the argument pointer to the next argument,
368 * or to the NUL byte terminating the argument line.
371 mandoc_getarg(struct mparse *parse, char **cpp, int ln, int *pos)
374 int quoted, pairs, white;
376 /* Quoting can only start with a new word. */
386 for (cp = start; '\0' != *cp; cp++) {
389 * Move the following text left
390 * after quoted quotes and after "\\" and "\t".
397 * In copy mode, translate double to single
398 * backslashes and backslash-t to literal tabs.
409 /* Skip escaped blanks. */
416 } else if (0 == quoted) {
418 /* Unescaped blanks end unquoted args. */
422 } else if ('"' == cp[0]) {
424 /* Quoted quotes collapse. */
428 /* Unquoted quotes end quoted args. */
435 /* Quoted argument without a closing quote. */
437 mandoc_msg(MANDOCERR_ARG_QUOTE, parse, ln, *pos, NULL);
439 /* NUL-terminate this argument and move to the next one. */
447 *pos += (int)(cp - start) + (quoted ? 1 : 0);
450 if ('\0' == *cp && (white || ' ' == cp[-1]))
451 mandoc_msg(MANDOCERR_SPACE_EOL, parse, ln, *pos, NULL);
457 a2time(time_t *t, const char *fmt, const char *p)
462 memset(&tm, 0, sizeof(struct tm));
466 pp = strptime(p, fmt, &tm);
468 if (NULL != pp && '\0' == *pp) {
490 * up to 9 characters for the month (September) + blank
491 * up to 2 characters for the day + comma + blank
492 * 4 characters for the year and a terminating '\0'
495 p = buf = mandoc_malloc(10 + 4 + 4 + 1);
497 if ((ssz = strftime(p, 10 + 1, "%B ", tm)) == 0)
502 * The output format is just "%d" here, not "%2d" or "%02d".
503 * That's also the reason why we can't just format the
504 * date as a whole with "%B %e, %Y" or "%B %d, %Y".
505 * Besides, the present approach is less prone to buffer
506 * overflows, in case anybody should ever introduce the bug
507 * of looking at LC_TIME.
510 if ((isz = snprintf(p, 4 + 1, "%d, ", tm->tm_mday)) == -1)
514 if (strftime(p, 4 + 1, "%Y", tm) == 0)
524 mandoc_normdate(struct roff_man *man, char *in, int ln, int pos)
529 /* No date specified: use today's date. */
531 if (in == NULL || *in == '\0' || strcmp(in, "$" "Mdocdate$") == 0) {
532 mandoc_msg(MANDOCERR_DATE_MISSING, man->parse, ln, pos, NULL);
533 return time2a(time(NULL));
536 /* Valid mdoc(7) date format. */
538 if (a2time(&t, "$" "Mdocdate: %b %d %Y $", in) ||
539 a2time(&t, "%b %d, %Y", in)) {
541 if (t > time(NULL) + 86400)
542 mandoc_msg(MANDOCERR_DATE_FUTURE, man->parse,
547 /* In man(7), do not warn about the legacy format. */
549 if (a2time(&t, "%Y-%m-%d", in) == 0)
550 mandoc_msg(MANDOCERR_DATE_BAD, man->parse, ln, pos, in);
551 else if (t > time(NULL) + 86400)
552 mandoc_msg(MANDOCERR_DATE_FUTURE, man->parse, ln, pos, in);
553 else if (man->macroset == MACROSET_MDOC)
554 mandoc_vmsg(MANDOCERR_DATE_LEGACY, man->parse,
555 ln, pos, "Dd %s", in);
557 /* Use any non-mdoc(7) date verbatim. */
559 return mandoc_strdup(in);
563 mandoc_eos(const char *p, size_t sz)
572 * End-of-sentence recognition must include situations where
573 * some symbols, such as `)', allow prior EOS punctuation to
577 enclosed = found = 0;
578 for (q = p + (int)sz - 1; q >= p; q--) {
594 (!enclosed || isalnum((unsigned char)*q));
598 return found && !enclosed;
602 * Convert a string to a long that may not be <0.
603 * If the string is invalid, or is less than 0, return -1.
606 mandoc_strntoi(const char *p, size_t sz, int base)
619 v = strtol(buf, &ep, base);
621 if (buf[0] == '\0' || *ep != '\0')