1121 lines
22 KiB
C
1121 lines
22 KiB
C
/* $NetBSD: apprentice.c,v 1.33 2002/07/10 16:15:52 pooka Exp $ */
|
|
|
|
/*
|
|
* apprentice - make one pass through /etc/magic, learning its secrets.
|
|
*
|
|
* Copyright (c) Ian F. Darwin, 1987.
|
|
* Written by Ian F. Darwin.
|
|
*
|
|
* This software is not subject to any license of the American Telephone
|
|
* and Telegraph Company or of the Regents of the University of California.
|
|
*
|
|
* Permission is granted to anyone to use this software for any purpose on
|
|
* any computer system, and to alter it and redistribute it freely, subject
|
|
* to the following restrictions:
|
|
*
|
|
* 1. The author is not responsible for the consequences of use of this
|
|
* software, no matter how awful, even if they arise from flaws in it.
|
|
*
|
|
* 2. The origin of this software must not be misrepresented, either by
|
|
* explicit claim or by omission. Since few users ever read sources,
|
|
* credits must appear in the documentation.
|
|
*
|
|
* 3. Altered versions must be plainly marked as such, and must not be
|
|
* misrepresented as being the original software. Since few users
|
|
* ever read sources, credits must appear in the documentation.
|
|
*
|
|
* 4. This notice may not be removed or altered.
|
|
*/
|
|
|
|
#include "file.h"
|
|
#include <stdlib.h>
|
|
#ifdef HAVE_UNISTD_H
|
|
#include <unistd.h>
|
|
#endif
|
|
#include <string.h>
|
|
#include <ctype.h>
|
|
#include <fcntl.h>
|
|
#ifdef QUICK
|
|
#include <sys/mman.h>
|
|
#endif
|
|
|
|
#include <sys/cdefs.h>
|
|
#if defined(__RCSID) && !defined(lint)
|
|
#if 0
|
|
FILE_RCSID("@(#)Id: apprentice.c,v 1.49 2002/07/03 19:00:41 christos Exp ")
|
|
#else
|
|
__RCSID("$NetBSD: apprentice.c,v 1.33 2002/07/10 16:15:52 pooka Exp $");
|
|
#endif
|
|
#endif /* lint */
|
|
|
|
#define EATAB {while (isascii((unsigned char) *l) && \
|
|
isspace((unsigned char) *l)) ++l;}
|
|
#define LOWCASE(l) (isupper((unsigned char) (l)) ? \
|
|
tolower((unsigned char) (l)) : (l))
|
|
/*
|
|
* Work around a bug in headers on Digital Unix.
|
|
* At least confirmed for: OSF1 V4.0 878
|
|
*/
|
|
#if defined(__osf__) && defined(__DECC)
|
|
#ifdef MAP_FAILED
|
|
#undef MAP_FAILED
|
|
#endif
|
|
#endif
|
|
|
|
#ifndef MAP_FAILED
|
|
#define MAP_FAILED (void *) -1
|
|
#endif
|
|
|
|
#ifndef MAP_FILE
|
|
#define MAP_FILE 0
|
|
#endif
|
|
|
|
#ifdef __EMX__
|
|
char PATHSEP=';';
|
|
#else
|
|
char PATHSEP=':';
|
|
#endif
|
|
|
|
|
|
static int getvalue(struct magic *, char **);
|
|
static int hextoint(int);
|
|
static char *getstr(char *, char *, int, int *);
|
|
static int parse(struct magic **, uint32_t *, char *, int);
|
|
static void eatsize(char **);
|
|
static int apprentice_1(const char *, int);
|
|
static int apprentice_file(struct magic **, uint32_t *, const char *, int);
|
|
static void byteswap(struct magic *, uint32_t);
|
|
static void bs1(struct magic *);
|
|
static uint16_t swap2(uint16_t);
|
|
static uint32_t swap4(uint32_t);
|
|
static char *mkdbname(const char *);
|
|
static int apprentice_map(struct magic **, uint32_t *, const char *, int);
|
|
static int apprentice_compile(struct magic **, uint32_t *, const char *, int);
|
|
|
|
static int maxmagic = 0;
|
|
|
|
struct mlist mlist;
|
|
|
|
#ifdef COMPILE_ONLY
|
|
const char *magicfile;
|
|
char *progname;
|
|
int lineno;
|
|
|
|
int main(int, char *[]);
|
|
|
|
int
|
|
main(int argc, char *argv[])
|
|
{
|
|
int ret;
|
|
|
|
if ((progname = strrchr(argv[0], '/')) != NULL)
|
|
progname++;
|
|
else
|
|
progname = argv[0];
|
|
|
|
if (argc != 2) {
|
|
(void)fprintf(stderr, "usage: %s file\n", progname);
|
|
exit(1);
|
|
}
|
|
magicfile = argv[1];
|
|
|
|
exit(apprentice(magicfile, COMPILE));
|
|
}
|
|
#endif /* COMPILE_ONLY */
|
|
|
|
|
|
/*
|
|
* Handle one file.
|
|
*/
|
|
static int
|
|
apprentice_1(const char *fn, int action)
|
|
{
|
|
struct magic *magic = NULL;
|
|
uint32_t nmagic = 0;
|
|
struct mlist *ml;
|
|
int rv = -1;
|
|
|
|
if (action == COMPILE) {
|
|
rv = apprentice_file(&magic, &nmagic, fn, action);
|
|
if (rv == 0)
|
|
return apprentice_compile(&magic, &nmagic, fn, action);
|
|
else
|
|
return rv;
|
|
}
|
|
#ifndef COMPILE_ONLY
|
|
if ((rv = apprentice_map(&magic, &nmagic, fn, action)) != 0)
|
|
(void)fprintf(stderr, "%s: Using regular magic file `%s'\n",
|
|
progname, fn);
|
|
|
|
if (rv != 0)
|
|
rv = apprentice_file(&magic, &nmagic, fn, action);
|
|
|
|
if (rv != 0)
|
|
return rv;
|
|
|
|
if ((ml = malloc(sizeof(*ml))) == NULL) {
|
|
(void) fprintf(stderr, "%s: Out of memory (%s).\n", progname,
|
|
strerror(errno));
|
|
if (action == CHECK)
|
|
return -1;
|
|
}
|
|
|
|
if (magic == NULL || nmagic == 0)
|
|
return rv;
|
|
|
|
ml->magic = magic;
|
|
ml->nmagic = nmagic;
|
|
|
|
mlist.prev->next = ml;
|
|
ml->prev = mlist.prev;
|
|
ml->next = &mlist;
|
|
mlist.prev = ml;
|
|
|
|
return rv;
|
|
#endif /* COMPILE_ONLY */
|
|
}
|
|
|
|
|
|
/* const char *fn: list of magic files */
|
|
int
|
|
apprentice(const char *fn, int action)
|
|
{
|
|
char *p, *mfn;
|
|
int file_err, errs = -1;
|
|
|
|
mlist.next = mlist.prev = &mlist;
|
|
mfn = malloc(strlen(fn)+1);
|
|
if (mfn == NULL) {
|
|
(void) fprintf(stderr, "%s: Out of memory (%s).\n", progname,
|
|
strerror(errno));
|
|
if (action == CHECK)
|
|
return -1;
|
|
else
|
|
exit(1);
|
|
}
|
|
fn = strcpy(mfn, fn);
|
|
|
|
while (fn) {
|
|
p = strchr(fn, PATHSEP);
|
|
if (p)
|
|
*p++ = '\0';
|
|
file_err = apprentice_1(fn, action);
|
|
if (file_err > errs)
|
|
errs = file_err;
|
|
fn = p;
|
|
}
|
|
if (errs == -1)
|
|
(void) fprintf(stderr, "%s: couldn't find any magic files!\n",
|
|
progname);
|
|
if (action == CHECK && errs)
|
|
exit(1);
|
|
|
|
free(mfn);
|
|
return errs;
|
|
}
|
|
|
|
/*
|
|
* parse from a file
|
|
* const char *fn: name of magic file
|
|
*/
|
|
static int
|
|
apprentice_file(struct magic **magicp, uint32_t *nmagicp, const char *fn,
|
|
int action)
|
|
{
|
|
static const char hdr[] =
|
|
"cont\toffset\ttype\topcode\tmask\tvalue\tdesc";
|
|
FILE *f;
|
|
char line[BUFSIZ+1];
|
|
int errs = 0;
|
|
|
|
f = fopen(fn, "r");
|
|
if (f == NULL) {
|
|
if (errno != ENOENT)
|
|
(void) fprintf(stderr,
|
|
"%s: can't read magic file %s (%s)\n",
|
|
progname, fn, strerror(errno));
|
|
return -1;
|
|
}
|
|
|
|
maxmagic = MAXMAGIS;
|
|
*magicp = (struct magic *) calloc(sizeof(struct magic), maxmagic);
|
|
if (*magicp == NULL) {
|
|
(void) fprintf(stderr, "%s: Out of memory (%s).\n", progname,
|
|
strerror(errno));
|
|
if (action == CHECK)
|
|
return -1;
|
|
}
|
|
|
|
/* parse it */
|
|
if (action == CHECK) /* print silly verbose header for USG compat. */
|
|
(void) printf("%s\n", hdr);
|
|
|
|
for (lineno = 1;fgets(line, BUFSIZ, f) != NULL; lineno++) {
|
|
if (line[0]=='#') /* comment, do not parse */
|
|
continue;
|
|
if (strlen(line) <= (unsigned)1) /* null line, garbage, etc */
|
|
continue;
|
|
line[strlen(line)-1] = '\0'; /* delete newline */
|
|
if (parse(magicp, nmagicp, line, action) != 0)
|
|
errs = 1;
|
|
}
|
|
|
|
(void) fclose(f);
|
|
if (errs) {
|
|
free(*magicp);
|
|
*magicp = NULL;
|
|
*nmagicp = 0;
|
|
}
|
|
return errs;
|
|
}
|
|
|
|
/*
|
|
* extend the sign bit if the comparison is to be signed
|
|
*/
|
|
uint32_t
|
|
signextend(struct magic *m, uint32_t v)
|
|
{
|
|
if (!(m->flag & UNSIGNED))
|
|
switch(m->type) {
|
|
/*
|
|
* Do not remove the casts below. They are
|
|
* vital. When later compared with the data,
|
|
* the sign extension must have happened.
|
|
*/
|
|
case BYTE:
|
|
v = (char) v;
|
|
break;
|
|
case SHORT:
|
|
case BESHORT:
|
|
case LESHORT:
|
|
v = (short) v;
|
|
break;
|
|
case DATE:
|
|
case BEDATE:
|
|
case LEDATE:
|
|
case LDATE:
|
|
case BELDATE:
|
|
case LELDATE:
|
|
case LONG:
|
|
case BELONG:
|
|
case LELONG:
|
|
v = (int32_t) v;
|
|
break;
|
|
case STRING:
|
|
case PSTRING:
|
|
break;
|
|
case REGEX:
|
|
break;
|
|
default:
|
|
magwarn("can't happen: m->type=%d\n",
|
|
m->type);
|
|
return -1;
|
|
}
|
|
return v;
|
|
}
|
|
|
|
/*
|
|
* parse one line from magic file, put into magic[index++] if valid
|
|
*/
|
|
static int
|
|
parse(struct magic **magicp, uint32_t *nmagicp, char *l, int action)
|
|
{
|
|
int i = 0;
|
|
struct magic *m;
|
|
char *t;
|
|
|
|
#define ALLOC_INCR 200
|
|
if (*nmagicp + 1 >= maxmagic){
|
|
maxmagic += ALLOC_INCR;
|
|
if ((m = (struct magic *) realloc(*magicp,
|
|
sizeof(struct magic) * maxmagic)) == NULL) {
|
|
(void) fprintf(stderr, "%s: Out of memory (%s).\n",
|
|
progname, strerror(errno));
|
|
if (*magicp)
|
|
free(*magicp);
|
|
if (action == CHECK)
|
|
return -1;
|
|
else
|
|
exit(1);
|
|
}
|
|
*magicp = m;
|
|
memset(&(*magicp)[*nmagicp], 0, sizeof(struct magic)
|
|
* ALLOC_INCR);
|
|
}
|
|
m = &(*magicp)[*nmagicp];
|
|
m->flag = 0;
|
|
m->cont_level = 0;
|
|
|
|
while (*l == '>') {
|
|
++l; /* step over */
|
|
m->cont_level++;
|
|
}
|
|
|
|
if (m->cont_level != 0 && *l == '(') {
|
|
++l; /* step over */
|
|
m->flag |= INDIR;
|
|
}
|
|
if (m->cont_level != 0 && *l == '&') {
|
|
++l; /* step over */
|
|
m->flag |= OFFADD;
|
|
}
|
|
|
|
/* get offset, then skip over it */
|
|
m->offset = (int) strtoul(l,&t,0);
|
|
if (l == t)
|
|
magwarn("offset %s invalid", l);
|
|
l = t;
|
|
|
|
if (m->flag & INDIR) {
|
|
m->in_type = LONG;
|
|
m->in_offset = 0;
|
|
/*
|
|
* read [.lbs][+-]nnnnn)
|
|
*/
|
|
if (*l == '.') {
|
|
l++;
|
|
switch (*l) {
|
|
case 'l':
|
|
m->in_type = LELONG;
|
|
break;
|
|
case 'L':
|
|
m->in_type = BELONG;
|
|
break;
|
|
case 'h':
|
|
case 's':
|
|
m->in_type = LESHORT;
|
|
break;
|
|
case 'H':
|
|
case 'S':
|
|
m->in_type = BESHORT;
|
|
break;
|
|
case 'c':
|
|
case 'b':
|
|
case 'C':
|
|
case 'B':
|
|
m->in_type = BYTE;
|
|
break;
|
|
default:
|
|
magwarn("indirect offset type %c invalid", *l);
|
|
break;
|
|
}
|
|
l++;
|
|
}
|
|
if (*l == '~') {
|
|
m->in_op = OPINVERSE;
|
|
l++;
|
|
}
|
|
switch (*l) {
|
|
case '&':
|
|
m->in_op |= OPAND;
|
|
l++;
|
|
break;
|
|
case '|':
|
|
m->in_op |= OPOR;
|
|
l++;
|
|
break;
|
|
case '^':
|
|
m->in_op |= OPXOR;
|
|
l++;
|
|
break;
|
|
case '+':
|
|
m->in_op |= OPADD;
|
|
l++;
|
|
break;
|
|
case '-':
|
|
m->in_op |= OPMINUS;
|
|
l++;
|
|
break;
|
|
case '*':
|
|
m->in_op |= OPMULTIPLY;
|
|
l++;
|
|
break;
|
|
case '/':
|
|
m->in_op |= OPDIVIDE;
|
|
l++;
|
|
break;
|
|
case '%':
|
|
m->in_op |= OPMODULO;
|
|
l++;
|
|
break;
|
|
}
|
|
if (isdigit((unsigned char)*l))
|
|
m->in_offset = strtoul(l, &t, 0);
|
|
else
|
|
t = l;
|
|
if (*t++ != ')')
|
|
magwarn("missing ')' in indirect offset");
|
|
l = t;
|
|
}
|
|
|
|
|
|
while (isascii((unsigned char)*l) && isdigit((unsigned char)*l))
|
|
++l;
|
|
EATAB;
|
|
|
|
#define NBYTE 4
|
|
#define NSHORT 5
|
|
#define NLONG 4
|
|
#define NSTRING 6
|
|
#define NDATE 4
|
|
#define NBESHORT 7
|
|
#define NBELONG 6
|
|
#define NBEDATE 6
|
|
#define NLESHORT 7
|
|
#define NLELONG 6
|
|
#define NLEDATE 6
|
|
#define NPSTRING 7
|
|
#define NLDATE 5
|
|
#define NBELDATE 7
|
|
#define NLELDATE 7
|
|
#define NREGEX 5
|
|
|
|
if (*l == 'u') {
|
|
++l;
|
|
m->flag |= UNSIGNED;
|
|
}
|
|
|
|
/* get type, skip it */
|
|
if (strncmp(l, "char", NBYTE)==0) { /* HP/UX compat */
|
|
m->type = BYTE;
|
|
l += NBYTE;
|
|
} else if (strncmp(l, "byte", NBYTE)==0) {
|
|
m->type = BYTE;
|
|
l += NBYTE;
|
|
} else if (strncmp(l, "short", NSHORT)==0) {
|
|
m->type = SHORT;
|
|
l += NSHORT;
|
|
} else if (strncmp(l, "long", NLONG)==0) {
|
|
m->type = LONG;
|
|
l += NLONG;
|
|
} else if (strncmp(l, "string", NSTRING)==0) {
|
|
m->type = STRING;
|
|
l += NSTRING;
|
|
} else if (strncmp(l, "date", NDATE)==0) {
|
|
m->type = DATE;
|
|
l += NDATE;
|
|
} else if (strncmp(l, "beshort", NBESHORT)==0) {
|
|
m->type = BESHORT;
|
|
l += NBESHORT;
|
|
} else if (strncmp(l, "belong", NBELONG)==0) {
|
|
m->type = BELONG;
|
|
l += NBELONG;
|
|
} else if (strncmp(l, "bedate", NBEDATE)==0) {
|
|
m->type = BEDATE;
|
|
l += NBEDATE;
|
|
} else if (strncmp(l, "leshort", NLESHORT)==0) {
|
|
m->type = LESHORT;
|
|
l += NLESHORT;
|
|
} else if (strncmp(l, "lelong", NLELONG)==0) {
|
|
m->type = LELONG;
|
|
l += NLELONG;
|
|
} else if (strncmp(l, "ledate", NLEDATE)==0) {
|
|
m->type = LEDATE;
|
|
l += NLEDATE;
|
|
} else if (strncmp(l, "pstring", NPSTRING)==0) {
|
|
m->type = PSTRING;
|
|
l += NPSTRING;
|
|
} else if (strncmp(l, "ldate", NLDATE)==0) {
|
|
m->type = LDATE;
|
|
l += NLDATE;
|
|
} else if (strncmp(l, "beldate", NBELDATE)==0) {
|
|
m->type = BELDATE;
|
|
l += NBELDATE;
|
|
} else if (strncmp(l, "leldate", NLELDATE)==0) {
|
|
m->type = LELDATE;
|
|
l += NLELDATE;
|
|
} else if (strncmp(l, "regex", NREGEX)==0) {
|
|
m->type = REGEX;
|
|
l += sizeof("regex");
|
|
} else {
|
|
magwarn("type %s invalid", l);
|
|
return -1;
|
|
}
|
|
/* New-style anding: "0 byte&0x80 =0x80 dynamically linked" */
|
|
/* New and improved: ~ & | ^ + - * / % -- exciting, isn't it? */
|
|
if (*l == '~') {
|
|
if (STRING != m->type && PSTRING != m->type)
|
|
m->mask_op = OPINVERSE;
|
|
++l;
|
|
}
|
|
switch (*l) {
|
|
case '&':
|
|
m->mask_op |= OPAND;
|
|
++l;
|
|
m->mask = signextend(m, strtoul(l, &l, 0));
|
|
eatsize(&l);
|
|
break;
|
|
case '|':
|
|
m->mask_op |= OPOR;
|
|
++l;
|
|
m->mask = signextend(m, strtoul(l, &l, 0));
|
|
eatsize(&l);
|
|
break;
|
|
case '^':
|
|
m->mask_op |= OPXOR;
|
|
++l;
|
|
m->mask = signextend(m, strtoul(l, &l, 0));
|
|
eatsize(&l);
|
|
break;
|
|
case '+':
|
|
m->mask_op |= OPADD;
|
|
++l;
|
|
m->mask = signextend(m, strtoul(l, &l, 0));
|
|
eatsize(&l);
|
|
break;
|
|
case '-':
|
|
m->mask_op |= OPMINUS;
|
|
++l;
|
|
m->mask = signextend(m, strtoul(l, &l, 0));
|
|
eatsize(&l);
|
|
break;
|
|
case '*':
|
|
m->mask_op |= OPMULTIPLY;
|
|
++l;
|
|
m->mask = signextend(m, strtoul(l, &l, 0));
|
|
eatsize(&l);
|
|
break;
|
|
case '%':
|
|
m->mask_op |= OPMODULO;
|
|
++l;
|
|
m->mask = signextend(m, strtoul(l, &l, 0));
|
|
eatsize(&l);
|
|
break;
|
|
case '/':
|
|
if (STRING != m->type && PSTRING != m->type) {
|
|
m->mask_op |= OPDIVIDE;
|
|
++l;
|
|
m->mask = signextend(m, strtoul(l, &l, 0));
|
|
eatsize(&l);
|
|
} else {
|
|
m->mask = 0L;
|
|
while (!isspace(*++l)) {
|
|
switch (*l) {
|
|
case CHAR_IGNORE_LOWERCASE:
|
|
m->mask |= STRING_IGNORE_LOWERCASE;
|
|
break;
|
|
case CHAR_COMPACT_BLANK:
|
|
m->mask |= STRING_COMPACT_BLANK;
|
|
break;
|
|
case CHAR_COMPACT_OPTIONAL_BLANK:
|
|
m->mask |=
|
|
STRING_COMPACT_OPTIONAL_BLANK;
|
|
break;
|
|
default:
|
|
magwarn("string extension %c invalid",
|
|
*l);
|
|
return -1;
|
|
}
|
|
}
|
|
}
|
|
break;
|
|
}
|
|
/* We used to set mask to all 1's here, instead let's just not do anything
|
|
if mask = 0 (unless you have a better idea) */
|
|
EATAB;
|
|
|
|
switch (*l) {
|
|
case '>':
|
|
case '<':
|
|
/* Old-style anding: "0 byte &0x80 dynamically linked" */
|
|
case '&':
|
|
case '^':
|
|
case '=':
|
|
m->reln = *l;
|
|
++l;
|
|
if (*l == '=') {
|
|
/* HP compat: ignore &= etc. */
|
|
++l;
|
|
}
|
|
break;
|
|
case '!':
|
|
if (m->type != STRING && m->type != PSTRING) {
|
|
m->reln = *l;
|
|
++l;
|
|
break;
|
|
}
|
|
/* FALL THROUGH */
|
|
default:
|
|
if (*l == 'x' && isascii((unsigned char)l[1]) &&
|
|
isspace((unsigned char)l[1])) {
|
|
m->reln = *l;
|
|
++l;
|
|
goto GetDesc; /* Bill The Cat */
|
|
}
|
|
m->reln = '=';
|
|
break;
|
|
}
|
|
EATAB;
|
|
|
|
if (getvalue(m, &l))
|
|
return -1;
|
|
/*
|
|
* TODO finish this macro and start using it!
|
|
* #define offsetcheck {if (offset > HOWMANY-1)
|
|
* magwarn("offset too big"); }
|
|
*/
|
|
|
|
/*
|
|
* now get last part - the description
|
|
*/
|
|
GetDesc:
|
|
EATAB;
|
|
if (l[0] == '\b') {
|
|
++l;
|
|
m->nospflag = 1;
|
|
} else if ((l[0] == '\\') && (l[1] == 'b')) {
|
|
++l;
|
|
++l;
|
|
m->nospflag = 1;
|
|
} else
|
|
m->nospflag = 0;
|
|
while ((m->desc[i++] = *l++) != '\0' && i<MAXDESC)
|
|
/* NULLBODY */;
|
|
|
|
#ifndef COMPILE_ONLY
|
|
if (action == CHECK) {
|
|
mdump(m);
|
|
}
|
|
#endif
|
|
++(*nmagicp); /* make room for next */
|
|
return 0;
|
|
}
|
|
|
|
/*
|
|
* Read a numeric value from a pointer, into the value union of a magic
|
|
* pointer, according to the magic type. Update the string pointer to point
|
|
* just after the number read. Return 0 for success, non-zero for failure.
|
|
*/
|
|
static int
|
|
getvalue(struct magic *m, char **p)
|
|
{
|
|
int slen;
|
|
|
|
if (m->type == STRING || m->type == PSTRING || m->type == REGEX) {
|
|
*p = getstr(*p, m->value.s, sizeof(m->value.s), &slen);
|
|
m->vallen = slen;
|
|
} else
|
|
if (m->reln != 'x') {
|
|
m->value.l = signextend(m, strtoul(*p, p, 0));
|
|
eatsize(p);
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
/*
|
|
* Convert a string containing C character escapes. Stop at an unescaped
|
|
* space or tab.
|
|
* Copy the converted version to "p", returning its length in *slen.
|
|
* Return updated scan pointer as function result.
|
|
*/
|
|
static char *
|
|
getstr(char *s, char *p, int plen, int *slen)
|
|
{
|
|
char *origs = s, *origp = p;
|
|
char *pmax = p + plen - 1;
|
|
int c;
|
|
int val;
|
|
|
|
while ((c = *s++) != '\0') {
|
|
if (isspace((unsigned char) c))
|
|
break;
|
|
if (p >= pmax) {
|
|
fprintf(stderr, "String too long: %s\n", origs);
|
|
break;
|
|
}
|
|
if(c == '\\') {
|
|
switch(c = *s++) {
|
|
|
|
case '\0':
|
|
goto out;
|
|
|
|
default:
|
|
*p++ = (char) c;
|
|
break;
|
|
|
|
case 'n':
|
|
*p++ = '\n';
|
|
break;
|
|
|
|
case 'r':
|
|
*p++ = '\r';
|
|
break;
|
|
|
|
case 'b':
|
|
*p++ = '\b';
|
|
break;
|
|
|
|
case 't':
|
|
*p++ = '\t';
|
|
break;
|
|
|
|
case 'f':
|
|
*p++ = '\f';
|
|
break;
|
|
|
|
case 'v':
|
|
*p++ = '\v';
|
|
break;
|
|
|
|
/* \ and up to 3 octal digits */
|
|
case '0':
|
|
case '1':
|
|
case '2':
|
|
case '3':
|
|
case '4':
|
|
case '5':
|
|
case '6':
|
|
case '7':
|
|
val = c - '0';
|
|
c = *s++; /* try for 2 */
|
|
if(c >= '0' && c <= '7') {
|
|
val = (val<<3) | (c - '0');
|
|
c = *s++; /* try for 3 */
|
|
if(c >= '0' && c <= '7')
|
|
val = (val<<3) | (c-'0');
|
|
else
|
|
--s;
|
|
}
|
|
else
|
|
--s;
|
|
*p++ = (char)val;
|
|
break;
|
|
|
|
/* \x and up to 2 hex digits */
|
|
case 'x':
|
|
val = 'x'; /* Default if no digits */
|
|
c = hextoint(*s++); /* Get next char */
|
|
if (c >= 0) {
|
|
val = c;
|
|
c = hextoint(*s++);
|
|
if (c >= 0)
|
|
val = (val << 4) + c;
|
|
else
|
|
--s;
|
|
} else
|
|
--s;
|
|
*p++ = (char)val;
|
|
break;
|
|
}
|
|
} else
|
|
*p++ = (char)c;
|
|
}
|
|
out:
|
|
*p = '\0';
|
|
*slen = p - origp;
|
|
return s;
|
|
}
|
|
|
|
|
|
/* Single hex char to int; -1 if not a hex char. */
|
|
static int
|
|
hextoint(int c)
|
|
{
|
|
if (!isascii((unsigned char) c))
|
|
return -1;
|
|
if (isdigit((unsigned char) c))
|
|
return c - '0';
|
|
if ((c >= 'a')&&(c <= 'f'))
|
|
return c + 10 - 'a';
|
|
if (( c>= 'A')&&(c <= 'F'))
|
|
return c + 10 - 'A';
|
|
return -1;
|
|
}
|
|
|
|
|
|
/*
|
|
* Print a string containing C character escapes.
|
|
*/
|
|
void
|
|
showstr(FILE *fp, const char *s, int len)
|
|
{
|
|
char c;
|
|
|
|
for (;;) {
|
|
c = *s++;
|
|
if (len == -1) {
|
|
if (c == '\0')
|
|
break;
|
|
}
|
|
else {
|
|
if (len-- == 0)
|
|
break;
|
|
}
|
|
if(c >= 040 && c <= 0176) /* TODO isprint && !iscntrl */
|
|
(void) fputc(c, fp);
|
|
else {
|
|
(void) fputc('\\', fp);
|
|
switch (c) {
|
|
|
|
case '\n':
|
|
(void) fputc('n', fp);
|
|
break;
|
|
|
|
case '\r':
|
|
(void) fputc('r', fp);
|
|
break;
|
|
|
|
case '\b':
|
|
(void) fputc('b', fp);
|
|
break;
|
|
|
|
case '\t':
|
|
(void) fputc('t', fp);
|
|
break;
|
|
|
|
case '\f':
|
|
(void) fputc('f', fp);
|
|
break;
|
|
|
|
case '\v':
|
|
(void) fputc('v', fp);
|
|
break;
|
|
|
|
default:
|
|
(void) fprintf(fp, "%.3o", c & 0377);
|
|
break;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
/*
|
|
* eatsize(): Eat the size spec from a number [eg. 10UL]
|
|
*/
|
|
static void
|
|
eatsize(char **p)
|
|
{
|
|
char *l = *p;
|
|
|
|
if (LOWCASE(*l) == 'u')
|
|
l++;
|
|
|
|
switch (LOWCASE(*l)) {
|
|
case 'l': /* long */
|
|
case 's': /* short */
|
|
case 'h': /* short */
|
|
case 'b': /* char/byte */
|
|
case 'c': /* char/byte */
|
|
l++;
|
|
/*FALLTHROUGH*/
|
|
default:
|
|
break;
|
|
}
|
|
|
|
*p = l;
|
|
}
|
|
|
|
/*
|
|
* handle a compiled file.
|
|
*/
|
|
static int
|
|
apprentice_map(struct magic **magicp, uint32_t *nmagicp, const char *fn,
|
|
int action)
|
|
{
|
|
int fd;
|
|
struct stat st;
|
|
uint32_t *ptr;
|
|
uint32_t version;
|
|
int needsbyteswap;
|
|
char *dbname = mkdbname(fn);
|
|
void *mm;
|
|
|
|
if (dbname == NULL)
|
|
return -1;
|
|
|
|
if ((fd = open(dbname, O_RDONLY)) == -1)
|
|
return -1;
|
|
|
|
if (fstat(fd, &st) == -1) {
|
|
(void)fprintf(stderr, "%s: Cannot stat `%s' (%s)\n",
|
|
progname, dbname, strerror(errno));
|
|
goto error;
|
|
}
|
|
|
|
#ifdef QUICK
|
|
if ((mm = mmap(0, (size_t)st.st_size, PROT_READ|PROT_WRITE,
|
|
MAP_PRIVATE|MAP_FILE, fd, (off_t)0)) == MAP_FAILED) {
|
|
(void)fprintf(stderr, "%s: Cannot map `%s' (%s)\n",
|
|
progname, dbname, strerror(errno));
|
|
goto error;
|
|
}
|
|
#else
|
|
if ((mm = malloc((size_t)st.st_size)) == NULL) {
|
|
(void) fprintf(stderr, "%s: Out of memory (%s).\n", progname,
|
|
strerror(errno));
|
|
goto error;
|
|
}
|
|
if (read(fd, mm, (size_t)st.st_size) != (size_t)st.st_size) {
|
|
(void) fprintf(stderr, "%s: Read failed (%s).\n", progname,
|
|
strerror(errno));
|
|
goto error;
|
|
}
|
|
#endif
|
|
*magicp = mm;
|
|
(void)close(fd);
|
|
fd = -1;
|
|
ptr = (uint32_t *) *magicp;
|
|
if (*ptr != MAGICNO) {
|
|
if (swap4(*ptr) != MAGICNO) {
|
|
(void)fprintf(stderr, "%s: Bad magic in `%s'\n",
|
|
progname, dbname);
|
|
goto error;
|
|
}
|
|
needsbyteswap = 1;
|
|
} else
|
|
needsbyteswap = 0;
|
|
if (needsbyteswap)
|
|
version = swap4(ptr[1]);
|
|
else
|
|
version = ptr[1];
|
|
if (version != VERSIONNO) {
|
|
(void)fprintf(stderr,
|
|
"%s: version mismatch (%d != %d) in `%s'\n",
|
|
progname, version, VERSIONNO, dbname);
|
|
goto error;
|
|
}
|
|
*nmagicp = (st.st_size / sizeof(struct magic)) - 1;
|
|
(*magicp)++;
|
|
if (needsbyteswap)
|
|
byteswap(*magicp, *nmagicp);
|
|
return 0;
|
|
|
|
error:
|
|
if (fd != -1)
|
|
(void)close(fd);
|
|
if (mm) {
|
|
#ifdef QUICK
|
|
(void)munmap(mm, (size_t)st.st_size);
|
|
#else
|
|
free(mm);
|
|
#endif
|
|
} else {
|
|
*magicp = NULL;
|
|
*nmagicp = 0;
|
|
}
|
|
return -1;
|
|
}
|
|
|
|
/*
|
|
* handle an mmaped file.
|
|
*/
|
|
static int
|
|
apprentice_compile(struct magic **magicp, uint32_t *nmagicp, const char *fn,
|
|
int action)
|
|
{
|
|
int fd;
|
|
char *dbname = mkdbname(fn);
|
|
static const uint32_t ar[] = {
|
|
MAGICNO, VERSIONNO
|
|
};
|
|
|
|
if (dbname == NULL)
|
|
return -1;
|
|
|
|
if ((fd = open(dbname, O_WRONLY|O_CREAT|O_TRUNC, 0644)) == -1) {
|
|
(void)fprintf(stderr, "%s: Cannot open `%s' (%s)\n",
|
|
progname, dbname, strerror(errno));
|
|
return -1;
|
|
}
|
|
|
|
if (write(fd, ar, sizeof(ar)) != sizeof(ar)) {
|
|
(void)fprintf(stderr, "%s: error writing `%s' (%s)\n",
|
|
progname, dbname, strerror(errno));
|
|
return -1;
|
|
}
|
|
|
|
if (lseek(fd, sizeof(struct magic), SEEK_SET) != sizeof(struct magic)) {
|
|
(void)fprintf(stderr, "%s: error seeking `%s' (%s)\n",
|
|
progname, dbname, strerror(errno));
|
|
return -1;
|
|
}
|
|
|
|
if (write(fd, *magicp, sizeof(struct magic) * *nmagicp)
|
|
!= sizeof(struct magic) * *nmagicp) {
|
|
(void)fprintf(stderr, "%s: error writing `%s' (%s)\n",
|
|
progname, dbname, strerror(errno));
|
|
return -1;
|
|
}
|
|
|
|
(void)close(fd);
|
|
return 0;
|
|
}
|
|
|
|
/*
|
|
* make a dbname
|
|
*/
|
|
char *
|
|
mkdbname(const char *fn)
|
|
{
|
|
static const char ext[] = ".mgc";
|
|
static char *buf = NULL;
|
|
size_t len = strlen(fn) + sizeof(ext) + 1;
|
|
if (buf == NULL)
|
|
buf = malloc(len);
|
|
else
|
|
buf = realloc(buf, len);
|
|
if (buf == NULL) {
|
|
(void) fprintf(stderr, "%s: Out of memory (%s).\n", progname,
|
|
strerror(errno));
|
|
return NULL;
|
|
}
|
|
(void)strcpy(buf, fn);
|
|
(void)strcat(buf, ext);
|
|
return buf;
|
|
}
|
|
|
|
/*
|
|
* Byteswap an mmap'ed file if needed
|
|
*/
|
|
static void
|
|
byteswap(struct magic *magic, uint32_t nmagic)
|
|
{
|
|
uint32_t i;
|
|
for (i = 0; i < nmagic; i++)
|
|
bs1(&magic[i]);
|
|
}
|
|
|
|
/*
|
|
* swap a short
|
|
*/
|
|
static uint16_t
|
|
swap2(uint16_t sv)
|
|
{
|
|
uint16_t rv;
|
|
uint8_t *s = (uint8_t *) &sv;
|
|
uint8_t *d = (uint8_t *) &rv;
|
|
d[0] = s[1];
|
|
d[1] = s[0];
|
|
return rv;
|
|
}
|
|
|
|
/*
|
|
* swap an int
|
|
*/
|
|
static uint32_t
|
|
swap4(uint32_t sv)
|
|
{
|
|
uint32_t rv;
|
|
uint8_t *s = (uint8_t *) &sv;
|
|
uint8_t *d = (uint8_t *) &rv;
|
|
d[0] = s[3];
|
|
d[1] = s[2];
|
|
d[2] = s[1];
|
|
d[3] = s[0];
|
|
return rv;
|
|
}
|
|
|
|
/*
|
|
* byteswap a single magic entry
|
|
*/
|
|
static
|
|
void bs1(struct magic *m)
|
|
{
|
|
m->cont_level = swap2(m->cont_level);
|
|
m->offset = swap4(m->offset);
|
|
m->in_offset = swap4(m->in_offset);
|
|
if (m->type != STRING)
|
|
m->value.l = swap4(m->value.l);
|
|
m->mask = swap4(m->mask);
|
|
}
|