freebsd-skq/contrib/mdocml/mandoc.c

631 lines
12 KiB
C
Raw Normal View History

2017-07-31 19:17:54 +00:00
/* $Id: mandoc.c,v 1.103 2017/07/03 13:40:19 schwarze Exp $ */
/*
* Copyright (c) 2008-2011, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
2017-06-08 19:29:07 +00:00
* Copyright (c) 2011-2015, 2017 Ingo Schwarze <schwarze@openbsd.org>
*
* Permission to use, copy, modify, and distribute this software for any
* purpose with or without fee is hereby granted, provided that the above
* copyright notice and this permission notice appear in all copies.
*
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
*/
#include "config.h"
#include <sys/types.h>
#include <assert.h>
#include <ctype.h>
#include <errno.h>
#include <limits.h>
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include <time.h>
2014-11-22 18:08:25 +00:00
#include "mandoc_aux.h"
2017-07-31 19:17:54 +00:00
#include "mandoc.h"
#include "roff.h"
#include "libmandoc.h"
static int a2time(time_t *, const char *, const char *);
static char *time2a(time_t);
2014-01-31 19:59:03 +00:00
enum mandoc_esc
mandoc_escape(const char **end, const char **start, int *sz)
2014-01-31 19:59:03 +00:00
{
const char *local_start;
int local_sz;
char term;
2014-11-22 18:08:25 +00:00
enum mandoc_esc gly;
/*
2014-01-31 19:59:03 +00:00
* When the caller doesn't provide return storage,
* use local storage.
*/
2014-01-31 19:59:03 +00:00
if (NULL == start)
start = &local_start;
if (NULL == sz)
sz = &local_sz;
2014-01-31 19:59:03 +00:00
/*
* Beyond the backslash, at least one input character
* is part of the escape sequence. With one exception
* (see below), that character won't be returned.
*/
gly = ESCAPE_ERROR;
2014-01-31 19:59:03 +00:00
*start = ++*end;
*sz = 0;
term = '\0';
2014-01-31 19:59:03 +00:00
switch ((*start)[-1]) {
/*
* First the glyphs. There are several different forms of
* these, but each eventually returns a substring of the glyph
* name.
*/
2014-11-22 18:08:25 +00:00
case '(':
gly = ESCAPE_SPECIAL;
2014-01-31 19:59:03 +00:00
*sz = 2;
break;
2014-11-22 18:08:25 +00:00
case '[':
gly = ESCAPE_SPECIAL;
term = ']';
break;
2014-11-22 18:08:25 +00:00
case 'C':
2014-01-31 19:59:03 +00:00
if ('\'' != **start)
2016-01-15 23:08:59 +00:00
return ESCAPE_ERROR;
2014-01-31 19:59:03 +00:00
*start = ++*end;
gly = ESCAPE_SPECIAL;
term = '\'';
break;
2014-01-31 19:59:03 +00:00
/*
* Escapes taking no arguments at all.
*/
2014-11-22 18:08:25 +00:00
case 'd':
case 'u':
2016-01-15 23:08:59 +00:00
case ',':
case '/':
return ESCAPE_IGNORE;
2017-07-31 19:17:54 +00:00
case 'p':
return ESCAPE_BREAK;
2014-01-31 19:59:03 +00:00
/*
* The \z escape is supposed to output the following
2014-11-22 18:08:25 +00:00
* character without advancing the cursor position.
2014-01-31 19:59:03 +00:00
* Since we are mostly dealing with terminal mode,
* let us just skip the next character.
*/
2014-11-22 18:08:25 +00:00
case 'z':
2016-01-15 23:08:59 +00:00
return ESCAPE_SKIPCHAR;
2014-01-31 19:59:03 +00:00
/*
* Handle all triggers matching \X(xy, \Xx, and \X[xxxx], where
* 'X' is the trigger. These have opaque sub-strings.
*/
2014-11-22 18:08:25 +00:00
case 'F':
case 'g':
case 'k':
case 'M':
case 'm':
case 'n':
case 'V':
case 'Y':
gly = ESCAPE_IGNORE;
/* FALLTHROUGH */
2014-11-22 18:08:25 +00:00
case 'f':
if (ESCAPE_ERROR == gly)
gly = ESCAPE_FONT;
2014-01-31 19:59:03 +00:00
switch (**start) {
2014-11-22 18:08:25 +00:00
case '(':
2014-01-31 19:59:03 +00:00
*start = ++*end;
*sz = 2;
break;
2014-11-22 18:08:25 +00:00
case '[':
2014-01-31 19:59:03 +00:00
*start = ++*end;
term = ']';
break;
default:
2014-01-31 19:59:03 +00:00
*sz = 1;
break;
}
break;
/*
* These escapes are of the form \X'Y', where 'X' is the trigger
* and 'Y' is any string. These have opaque sub-strings.
2014-11-22 18:08:25 +00:00
* The \B and \w escapes are handled in roff.c, roff_res().
*/
2014-11-22 18:08:25 +00:00
case 'A':
case 'b':
case 'D':
case 'R':
case 'X':
case 'Z':
gly = ESCAPE_IGNORE;
/* FALLTHROUGH */
case 'o':
if (**start == '\0')
2016-01-15 23:08:59 +00:00
return ESCAPE_ERROR;
if (gly == ESCAPE_ERROR)
gly = ESCAPE_OVERSTRIKE;
2014-11-22 18:08:25 +00:00
term = **start;
2014-01-31 19:59:03 +00:00
*start = ++*end;
break;
/*
* These escapes are of the form \X'N', where 'X' is the trigger
* and 'N' resolves to a numerical expression.
*/
2014-11-22 18:08:25 +00:00
case 'h':
case 'H':
case 'L':
case 'l':
case 'S':
case 'v':
case 'x':
if (strchr(" %&()*+-./0123456789:<=>", **start)) {
if ('\0' != **start)
++*end;
2016-01-15 23:08:59 +00:00
return ESCAPE_ERROR;
2014-11-22 18:08:25 +00:00
}
2017-06-08 19:29:07 +00:00
switch ((*start)[-1]) {
case 'h':
gly = ESCAPE_HORIZ;
break;
case 'l':
gly = ESCAPE_HLINE;
break;
default:
gly = ESCAPE_IGNORE;
break;
}
2014-11-22 18:08:25 +00:00
term = **start;
2014-01-31 19:59:03 +00:00
*start = ++*end;
break;
/*
* Special handling for the numbered character escape.
* XXX Do any other escapes need similar handling?
*/
2014-11-22 18:08:25 +00:00
case 'N':
2014-01-31 19:59:03 +00:00
if ('\0' == **start)
2016-01-15 23:08:59 +00:00
return ESCAPE_ERROR;
2014-01-31 19:59:03 +00:00
(*end)++;
if (isdigit((unsigned char)**start)) {
*sz = 1;
2016-01-15 23:08:59 +00:00
return ESCAPE_IGNORE;
2014-01-31 19:59:03 +00:00
}
(*start)++;
while (isdigit((unsigned char)**end))
(*end)++;
2014-01-31 19:59:03 +00:00
*sz = *end - *start;
if ('\0' != **end)
(*end)++;
2016-01-15 23:08:59 +00:00
return ESCAPE_NUMBERED;
2014-11-22 18:08:25 +00:00
/*
* Sizes get a special category of their own.
*/
2014-11-22 18:08:25 +00:00
case 's':
gly = ESCAPE_IGNORE;
/* See +/- counts as a sign. */
2014-01-31 19:59:03 +00:00
if ('+' == **end || '-' == **end || ASCII_HYPH == **end)
*start = ++*end;
2014-01-31 19:59:03 +00:00
switch (**end) {
2014-11-22 18:08:25 +00:00
case '(':
2014-01-31 19:59:03 +00:00
*start = ++*end;
*sz = 2;
break;
2014-11-22 18:08:25 +00:00
case '[':
2014-01-31 19:59:03 +00:00
*start = ++*end;
term = ']';
break;
2014-11-22 18:08:25 +00:00
case '\'':
2014-01-31 19:59:03 +00:00
*start = ++*end;
term = '\'';
break;
case '3':
case '2':
case '1':
*sz = (*end)[-1] == 's' &&
isdigit((unsigned char)(*end)[1]) ? 2 : 1;
break;
default:
2014-01-31 19:59:03 +00:00
*sz = 1;
break;
}
break;
/*
* Anything else is assumed to be a glyph.
2014-01-31 19:59:03 +00:00
* In this case, pass back the character after the backslash.
*/
default:
gly = ESCAPE_SPECIAL;
2014-01-31 19:59:03 +00:00
*start = --*end;
*sz = 1;
break;
}
assert(ESCAPE_ERROR != gly);
/*
2014-01-31 19:59:03 +00:00
* Read up to the terminating character,
* paying attention to nested escapes.
*/
if ('\0' != term) {
2014-01-31 19:59:03 +00:00
while (**end != term) {
switch (**end) {
2014-11-22 18:08:25 +00:00
case '\0':
2016-01-15 23:08:59 +00:00
return ESCAPE_ERROR;
2014-11-22 18:08:25 +00:00
case '\\':
2014-01-31 19:59:03 +00:00
(*end)++;
if (ESCAPE_ERROR ==
mandoc_escape(end, NULL, NULL))
2016-01-15 23:08:59 +00:00
return ESCAPE_ERROR;
2014-01-31 19:59:03 +00:00
break;
default:
(*end)++;
break;
}
}
*sz = (*end)++ - *start;
} else {
assert(*sz > 0);
if ((size_t)*sz > strlen(*start))
2016-01-15 23:08:59 +00:00
return ESCAPE_ERROR;
2014-01-31 19:59:03 +00:00
*end += *sz;
}
/* Run post-processors. */
switch (gly) {
2014-11-22 18:08:25 +00:00
case ESCAPE_FONT:
2014-01-31 19:59:03 +00:00
if (2 == *sz) {
if ('C' == **start) {
/*
* Treat constant-width font modes
* just like regular font modes.
*/
(*start)++;
(*sz)--;
} else {
if ('B' == (*start)[0] && 'I' == (*start)[1])
gly = ESCAPE_FONTBI;
break;
}
} else if (1 != *sz)
break;
2014-01-31 19:59:03 +00:00
switch (**start) {
2014-11-22 18:08:25 +00:00
case '3':
case 'B':
gly = ESCAPE_FONTBOLD;
break;
2014-11-22 18:08:25 +00:00
case '2':
case 'I':
gly = ESCAPE_FONTITALIC;
break;
2014-11-22 18:08:25 +00:00
case 'P':
gly = ESCAPE_FONTPREV;
break;
2014-11-22 18:08:25 +00:00
case '1':
case 'R':
gly = ESCAPE_FONTROMAN;
break;
}
break;
2014-11-22 18:08:25 +00:00
case ESCAPE_SPECIAL:
2014-01-31 19:59:03 +00:00
if (1 == *sz && 'c' == **start)
gly = ESCAPE_NOSPACE;
/*
* Unicode escapes are defined in groff as \[u0000]
* to \[u10FFFF], where the contained value must be
* a valid Unicode codepoint. Here, however, only
* check the length and range.
*/
if (**start != 'u' || *sz < 5 || *sz > 7)
break;
if (*sz == 7 && ((*start)[1] != '1' || (*start)[2] != '0'))
break;
if (*sz == 6 && (*start)[1] == '0')
break;
2016-01-15 23:08:59 +00:00
if (*sz == 5 && (*start)[1] == 'D' &&
strchr("89ABCDEF", (*start)[2]) != NULL)
break;
if ((int)strspn(*start + 1, "0123456789ABCDEFabcdef")
+ 1 == *sz)
gly = ESCAPE_UNICODE;
break;
default:
break;
}
2016-01-15 23:08:59 +00:00
return gly;
}
/*
* Parse a quoted or unquoted roff-style request or macro argument.
* Return a pointer to the parsed argument, which is either the original
* pointer or advanced by one byte in case the argument is quoted.
2014-01-31 19:59:03 +00:00
* NUL-terminate the argument in place.
* Collapse pairs of quotes inside quoted arguments.
* Advance the argument pointer to the next argument,
2014-01-31 19:59:03 +00:00
* or to the NUL byte terminating the argument line.
*/
char *
mandoc_getarg(struct mparse *parse, char **cpp, int ln, int *pos)
{
char *start, *cp;
int quoted, pairs, white;
/* Quoting can only start with a new word. */
start = *cpp;
quoted = 0;
if ('"' == *start) {
quoted = 1;
start++;
2014-11-22 18:08:25 +00:00
}
pairs = 0;
white = 0;
for (cp = start; '\0' != *cp; cp++) {
2014-01-31 19:59:03 +00:00
/*
* Move the following text left
* after quoted quotes and after "\\" and "\t".
*/
if (pairs)
cp[-pairs] = cp[0];
2014-01-31 19:59:03 +00:00
if ('\\' == cp[0]) {
2014-01-31 19:59:03 +00:00
/*
* In copy mode, translate double to single
* backslashes and backslash-t to literal tabs.
*/
switch (cp[1]) {
2014-11-22 18:08:25 +00:00
case 't':
2014-01-31 19:59:03 +00:00
cp[0] = '\t';
/* FALLTHROUGH */
2014-11-22 18:08:25 +00:00
case '\\':
pairs++;
cp++;
2014-01-31 19:59:03 +00:00
break;
2014-11-22 18:08:25 +00:00
case ' ':
/* Skip escaped blanks. */
2014-01-31 19:59:03 +00:00
if (0 == quoted)
cp++;
break;
default:
break;
}
} else if (0 == quoted) {
if (' ' == cp[0]) {
/* Unescaped blanks end unquoted args. */
white = 1;
break;
}
} else if ('"' == cp[0]) {
if ('"' == cp[1]) {
/* Quoted quotes collapse. */
pairs++;
cp++;
} else {
/* Unquoted quotes end quoted args. */
quoted = 2;
break;
}
}
}
/* Quoted argument without a closing quote. */
if (1 == quoted)
2014-11-22 18:08:25 +00:00
mandoc_msg(MANDOCERR_ARG_QUOTE, parse, ln, *pos, NULL);
2014-01-31 19:59:03 +00:00
/* NUL-terminate this argument and move to the next one. */
if (pairs)
cp[-pairs] = '\0';
if ('\0' != *cp) {
*cp++ = '\0';
while (' ' == *cp)
cp++;
}
*pos += (int)(cp - start) + (quoted ? 1 : 0);
*cpp = cp;
if ('\0' == *cp && (white || ' ' == cp[-1]))
2014-11-22 18:08:25 +00:00
mandoc_msg(MANDOCERR_SPACE_EOL, parse, ln, *pos, NULL);
2016-01-15 23:08:59 +00:00
return start;
}
static int
a2time(time_t *t, const char *fmt, const char *p)
{
struct tm tm;
char *pp;
memset(&tm, 0, sizeof(struct tm));
pp = NULL;
#if HAVE_STRPTIME
pp = strptime(p, fmt, &tm);
#endif
if (NULL != pp && '\0' == *pp) {
*t = mktime(&tm);
2016-01-15 23:08:59 +00:00
return 1;
}
2016-01-15 23:08:59 +00:00
return 0;
}
static char *
time2a(time_t t)
{
struct tm *tm;
char *buf, *p;
size_t ssz;
int isz;
tm = localtime(&t);
if (tm == NULL)
2016-01-15 23:08:59 +00:00
return NULL;
/*
* Reserve space:
* up to 9 characters for the month (September) + blank
* up to 2 characters for the day + comma + blank
* 4 characters for the year and a terminating '\0'
*/
2016-01-15 23:08:59 +00:00
p = buf = mandoc_malloc(10 + 4 + 4 + 1);
2016-01-15 23:08:59 +00:00
if ((ssz = strftime(p, 10 + 1, "%B ", tm)) == 0)
goto fail;
p += (int)ssz;
2016-01-15 23:08:59 +00:00
/*
* The output format is just "%d" here, not "%2d" or "%02d".
* That's also the reason why we can't just format the
* date as a whole with "%B %e, %Y" or "%B %d, %Y".
* Besides, the present approach is less prone to buffer
* overflows, in case anybody should ever introduce the bug
* of looking at LC_TIME.
*/
if ((isz = snprintf(p, 4 + 1, "%d, ", tm->tm_mday)) == -1)
goto fail;
p += isz;
2016-01-15 23:08:59 +00:00
if (strftime(p, 4 + 1, "%Y", tm) == 0)
goto fail;
2016-01-15 23:08:59 +00:00
return buf;
fail:
free(buf);
2016-01-15 23:08:59 +00:00
return NULL;
}
char *
2017-07-31 19:17:54 +00:00
mandoc_normdate(struct roff_man *man, char *in, int ln, int pos)
{
2017-07-31 19:17:54 +00:00
char *cp;
time_t t;
2016-01-15 23:08:59 +00:00
/* No date specified: use today's date. */
if (in == NULL || *in == '\0' || strcmp(in, "$" "Mdocdate$") == 0) {
2017-07-31 19:17:54 +00:00
mandoc_msg(MANDOCERR_DATE_MISSING, man->parse, ln, pos, NULL);
2016-01-15 23:08:59 +00:00
return time2a(time(NULL));
}
2016-01-15 23:08:59 +00:00
/* Valid mdoc(7) date format. */
if (a2time(&t, "$" "Mdocdate: %b %d %Y $", in) ||
2017-07-31 19:17:54 +00:00
a2time(&t, "%b %d, %Y", in)) {
cp = time2a(t);
if (t > time(NULL) + 86400)
mandoc_msg(MANDOCERR_DATE_FUTURE, man->parse,
ln, pos, cp);
return cp;
}
2016-01-15 23:08:59 +00:00
2017-07-31 19:17:54 +00:00
/* In man(7), do not warn about the legacy format. */
2016-01-15 23:08:59 +00:00
2017-07-31 19:17:54 +00:00
if (a2time(&t, "%Y-%m-%d", in) == 0)
mandoc_msg(MANDOCERR_DATE_BAD, man->parse, ln, pos, in);
else if (t > time(NULL) + 86400)
mandoc_msg(MANDOCERR_DATE_FUTURE, man->parse, ln, pos, in);
else if (man->macroset == MACROSET_MDOC)
mandoc_vmsg(MANDOCERR_DATE_LEGACY, man->parse,
ln, pos, "Dd %s", in);
2016-01-15 23:08:59 +00:00
/* Use any non-mdoc(7) date verbatim. */
return mandoc_strdup(in);
}
int
2014-11-22 18:08:25 +00:00
mandoc_eos(const char *p, size_t sz)
{
2014-11-22 18:08:25 +00:00
const char *q;
int enclosed, found;
if (0 == sz)
2016-01-15 23:08:59 +00:00
return 0;
/*
* End-of-sentence recognition must include situations where
* some symbols, such as `)', allow prior EOS punctuation to
* propagate outward.
*/
2014-11-22 18:08:25 +00:00
enclosed = found = 0;
for (q = p + (int)sz - 1; q >= p; q--) {
switch (*q) {
2014-11-22 18:08:25 +00:00
case '\"':
case '\'':
case ']':
case ')':
if (0 == found)
enclosed = 1;
break;
2014-11-22 18:08:25 +00:00
case '.':
case '!':
case '?':
found = 1;
break;
default:
2016-01-15 23:08:59 +00:00
return found &&
(!enclosed || isalnum((unsigned char)*q));
}
}
2016-01-15 23:08:59 +00:00
return found && !enclosed;
}
/*
* Convert a string to a long that may not be <0.
* If the string is invalid, or is less than 0, return -1.
*/
int
mandoc_strntoi(const char *p, size_t sz, int base)
{
char buf[32];
char *ep;
long v;
if (sz > 31)
2016-01-15 23:08:59 +00:00
return -1;
memcpy(buf, p, sz);
buf[(int)sz] = '\0';
errno = 0;
v = strtol(buf, &ep, base);
if (buf[0] == '\0' || *ep != '\0')
2016-01-15 23:08:59 +00:00
return -1;
if (v > INT_MAX)
v = INT_MAX;
if (v < INT_MIN)
v = INT_MIN;
2016-01-15 23:08:59 +00:00
return (int)v;
}