From: Derick Rethans Date: Fri, 11 Jul 2008 14:10:50 +0000 (+0000) Subject: - Step one for bundling the libmagic library. Some config.m4 issues left. X-Git-Tag: BEFORE_HEAD_NS_CHANGE~1305 X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=90f11b47df3510bacab18e53a9cd9ee482b692c8;p=php - Step one for bundling the libmagic library. Some config.m4 issues left. --- diff --git a/ext/fileinfo/config.m4 b/ext/fileinfo/config.m4 index 686f6e2780..75aff6724e 100644 --- a/ext/fileinfo/config.m4 +++ b/ext/fileinfo/config.m4 @@ -5,48 +5,6 @@ PHP_ARG_WITH(fileinfo, for fileinfo support, [ --with-fileinfo=DIR Include fileinfo support]) if test "$PHP_FILEINFO" != "no"; then - SEARCH_PATH="/usr/local /usr /usr/share/file" - if test -r $PHP_FILEINFO/include/magic.h || test -r $PHP_FILEINFO/magic.h; then - FILEINFO_DIR=$PHP_FILEINFO - else - AC_MSG_CHECKING([for magic files in default path]) - for i in $SEARCH_PATH ; do - if test -r $i/include/magic.h || test -r $i/magic.h; then - FILEINFO_DIR=$i - AC_MSG_RESULT(found in $i) - break - fi - done - fi - - if test -z "$FILEINFO_DIR"; then - AC_MSG_RESULT([not found]) - AC_MSG_ERROR([Please reinstall the libmagic distribution]) - fi - - if test -r "$FILEINFO_DIR/include/magic.h"; then - PHP_ADD_INCLUDE($FILEINFO_DIR/include) - else - PHP_ADD_INCLUDE($FILEINFO_DIR) - fi - - LIBNAME=magic - LIBSYMBOL=magic_open - - PHP_CHECK_FUNC(dl, dlopen) - PHP_CHECK_FUNC(gzgets, z) - PHP_CHECK_FUNC(round, m) - - PHP_CHECK_LIBRARY($LIBNAME,$LIBSYMBOL, - [ - PHP_ADD_LIBRARY_WITH_PATH($LIBNAME, $FILEINFO_DIR/lib, FILEINFO_SHARED_LIBADD) - AC_DEFINE(HAVE_FILEINFOLIB,1,[ ]) - ],[ - AC_MSG_ERROR([wrong magic lib version or lib not found]) - ],[ - -L$FILEINFO_DIR/lib - ]) - MAGIC_MIME_DIRS="/usr/local/share/file /usr/share/file /usr/share/misc/file /etc /usr/share/misc" MAGIC_MIME_FILENAMES="magic magic.mime" @@ -58,10 +16,19 @@ if test "$PHP_FILEINFO" != "no"; then fi done done + + AC_CHECK_FUNCS([utimes]) AC_DEFINE_UNQUOTED(PHP_DEFAULT_MAGIC_FILE,"$PHP_DEFAULT_MAGIC_FILE",[magic file path]) PHP_SUBST(FILEINFO_SHARED_LIBADD) - PHP_NEW_EXTENSION(fileinfo, fileinfo.c, $ext_shared) + PHP_NEW_EXTENSION(fileinfo, fileinfo.c lib/apprentice.c lib/apptype.c lib/ascmagic.c lib/compress.c lib/file.c lib/fsmagic.c lib/funcs.c lib/getopt_long.c lib/is_tar.c lib/magic.c lib/print.c lib/readelf.c lib/softmagic.c, $ext_shared) + +if test "$ext_shared" = "no" || test "$ext_srcdir" != "$abs_srcdir"; then + echo '#include ' > $ext_builddir/libsqlite/src/config.h +else + echo "#include \"$abs_builddir/config.h\"" > $ext_builddir/libsqlite/src/config.h +fi + fi diff --git a/ext/fileinfo/lib/apprentice.c b/ext/fileinfo/lib/apprentice.c new file mode 100644 index 0000000000..438a38e045 --- /dev/null +++ b/ext/fileinfo/lib/apprentice.c @@ -0,0 +1,2131 @@ +/* + * Copyright (c) Ian F. Darwin 1986-1995. + * Software written by Ian F. Darwin and others; + * maintained 1995-present by Christos Zoulas and others. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice immediately at the beginning of the file, without modification, + * this list of conditions, and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR + * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + */ +/* + * apprentice - make one pass through /etc/magic, learning its secrets. + */ + +#include "file.h" +#include "magic.h" +#include "patchlevel.h" +#include +//#ifdef HAVE_UNISTD_H +#include +//#endif +#include +#include +#include +#include +#include +#include +#ifdef QUICK +#include +#endif +#include +#include + +#ifndef lint +FILE_RCSID("@(#)$File: apprentice.c,v 1.132 2008/03/28 18:19:30 christos Exp $") +#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 + +#ifndef MAXPATHLEN +#define MAXPATHLEN 1024 +#endif + +struct magic_entry { + struct magic *mp; + uint32_t cont_count; + uint32_t max_count; +}; + +int file_formats[FILE_NAMES_SIZE]; +const size_t file_nformats = FILE_NAMES_SIZE; +const char *file_names[FILE_NAMES_SIZE]; +const size_t file_nnames = FILE_NAMES_SIZE; + +private int getvalue(struct magic_set *ms, struct magic *, const char **, int); +private int hextoint(int); +private const char *getstr(struct magic_set *, const char *, char *, int, + int *, int); +private int parse(struct magic_set *, struct magic_entry **, uint32_t *, + const char *, size_t, int); +private int parse_mime(struct magic_set *, struct magic_entry **, uint32_t *, + const char *); +private void eatsize(const char **); +private int apprentice_1(struct magic_set *, const char *, int, struct mlist *); +private size_t apprentice_magic_strength(const struct magic *); +private int apprentice_sort(const void *, const void *); +private int apprentice_load(struct magic_set *, struct magic **, uint32_t *, + const char *, int); +private void byteswap(struct magic *, uint32_t); +private void bs1(struct magic *); +private uint16_t swap2(uint16_t); +private uint32_t swap4(uint32_t); +private uint64_t swap8(uint64_t); +private void mkdbname(const char *, char **, int); +private int apprentice_map(struct magic_set *, struct magic **, uint32_t *, + const char *); +private int apprentice_compile(struct magic_set *, struct magic **, uint32_t *, + const char *); +private int check_format_type(const char *, int); +private int check_format(struct magic_set *, struct magic *); +private int get_op(char); + +private size_t maxmagic = 0; +private size_t magicsize = sizeof(struct magic); + +private const char usg_hdr[] = "cont\toffset\ttype\topcode\tmask\tvalue\tdesc"; +private const char mime_marker[] = "!:mime"; +private const size_t mime_marker_len = sizeof(mime_marker) - 1; + +#ifdef COMPILE_ONLY + +int main(int, char *[]); + +int +main(int argc, char *argv[]) +{ + int ret; + struct magic_set *ms; + char *progname; + + if ((progname = strrchr(argv[0], '/')) != NULL) + progname++; + else + progname = argv[0]; + + if (argc != 2) { + (void)fprintf(stderr, "Usage: %s file\n", progname); + return 1; + } + + if ((ms = magic_open(MAGIC_CHECK)) == NULL) { + (void)fprintf(stderr, "%s: %s\n", progname, strerror(errno)); + return 1; + } + ret = magic_compile(ms, argv[1]) == -1 ? 1 : 0; + if (ret == 1) + (void)fprintf(stderr, "%s: %s\n", progname, magic_error(ms)); + magic_close(ms); + return ret; +} +#endif /* COMPILE_ONLY */ + +static const struct type_tbl_s { + const char name[16]; + const size_t len; + const int type; + const int format; +} type_tbl[] = { +# define XX(s) s, (sizeof(s) - 1) +# define XX_NULL "", 0 + { XX("byte"), FILE_BYTE, FILE_FMT_NUM }, + { XX("short"), FILE_SHORT, FILE_FMT_NUM }, + { XX("default"), FILE_DEFAULT, FILE_FMT_STR }, + { XX("long"), FILE_LONG, FILE_FMT_NUM }, + { XX("string"), FILE_STRING, FILE_FMT_STR }, + { XX("date"), FILE_DATE, FILE_FMT_STR }, + { XX("beshort"), FILE_BESHORT, FILE_FMT_NUM }, + { XX("belong"), FILE_BELONG, FILE_FMT_NUM }, + { XX("bedate"), FILE_BEDATE, FILE_FMT_STR }, + { XX("leshort"), FILE_LESHORT, FILE_FMT_NUM }, + { XX("lelong"), FILE_LELONG, FILE_FMT_NUM }, + { XX("ledate"), FILE_LEDATE, FILE_FMT_STR }, + { XX("pstring"), FILE_PSTRING, FILE_FMT_STR }, + { XX("ldate"), FILE_LDATE, FILE_FMT_STR }, + { XX("beldate"), FILE_BELDATE, FILE_FMT_STR }, + { XX("leldate"), FILE_LELDATE, FILE_FMT_STR }, + { XX("regex"), FILE_REGEX, FILE_FMT_STR }, + { XX("bestring16"), FILE_BESTRING16, FILE_FMT_STR }, + { XX("lestring16"), FILE_LESTRING16, FILE_FMT_STR }, + { XX("search"), FILE_SEARCH, FILE_FMT_STR }, + { XX("medate"), FILE_MEDATE, FILE_FMT_STR }, + { XX("meldate"), FILE_MELDATE, FILE_FMT_STR }, + { XX("melong"), FILE_MELONG, FILE_FMT_NUM }, + { XX("quad"), FILE_QUAD, FILE_FMT_QUAD }, + { XX("lequad"), FILE_LEQUAD, FILE_FMT_QUAD }, + { XX("bequad"), FILE_BEQUAD, FILE_FMT_QUAD }, + { XX("qdate"), FILE_QDATE, FILE_FMT_STR }, + { XX("leqdate"), FILE_LEQDATE, FILE_FMT_STR }, + { XX("beqdate"), FILE_BEQDATE, FILE_FMT_STR }, + { XX("qldate"), FILE_QLDATE, FILE_FMT_STR }, + { XX("leqldate"), FILE_LEQLDATE, FILE_FMT_STR }, + { XX("beqldate"), FILE_BEQLDATE, FILE_FMT_STR }, + { XX("float"), FILE_FLOAT, FILE_FMT_FLOAT }, + { XX("befloat"), FILE_BEFLOAT, FILE_FMT_FLOAT }, + { XX("lefloat"), FILE_LEFLOAT, FILE_FMT_FLOAT }, + { XX("double"), FILE_DOUBLE, FILE_FMT_DOUBLE }, + { XX("bedouble"), FILE_BEDOUBLE, FILE_FMT_DOUBLE }, + { XX("ledouble"), FILE_LEDOUBLE, FILE_FMT_DOUBLE }, + { XX_NULL, FILE_INVALID, FILE_FMT_NONE }, +# undef XX +# undef XX_NULL +}; + +private int +get_type(const char *l, const char **t) +{ + const struct type_tbl_s *p; + + for (p = type_tbl; p->len; p++) { + if (strncmp(l, p->name, p->len) == 0) { + if (t) + *t = l + p->len; + break; + } + } + return p->type; +} + +private void +init_file_tables(void) +{ + static int done = 0; + const struct type_tbl_s *p; + + if (done) + return; + done++; + + for (p = type_tbl; p->len; p++) { + assert(p->type < FILE_NAMES_SIZE); + file_names[p->type] = p->name; + file_formats[p->type] = p->format; + } +} + +/* + * Handle one file or directory. + */ +private int +apprentice_1(struct magic_set *ms, const char *fn, int action, + struct mlist *mlist) +{ + struct magic *magic = NULL; + uint32_t nmagic = 0; + struct mlist *ml; + int rv = -1; + int mapped; + + if (magicsize != FILE_MAGICSIZE) { + file_error(ms, 0, "magic element size %lu != %lu", + (unsigned long)sizeof(*magic), + (unsigned long)FILE_MAGICSIZE); + return -1; + } + + if (action == FILE_COMPILE) { + rv = apprentice_load(ms, &magic, &nmagic, fn, action); + if (rv != 0) + return -1; + rv = apprentice_compile(ms, &magic, &nmagic, fn); + free(magic); + return rv; + } + +#ifndef COMPILE_ONLY + if ((rv = apprentice_map(ms, &magic, &nmagic, fn)) == -1) { + if (ms->flags & MAGIC_CHECK) + file_magwarn(ms, "using regular magic file `%s'", fn); + rv = apprentice_load(ms, &magic, &nmagic, fn, action); + if (rv != 0) + return -1; + } + + mapped = rv; + + if (magic == NULL) { + file_delmagic(magic, mapped, nmagic); + return -1; + } + + if ((ml = malloc(sizeof(*ml))) == NULL) { + file_delmagic(magic, mapped, nmagic); + file_oomem(ms, sizeof(*ml)); + return -1; + } + + ml->magic = magic; + ml->nmagic = nmagic; + ml->mapped = mapped; + + mlist->prev->next = ml; + ml->prev = mlist->prev; + ml->next = mlist; + mlist->prev = ml; + + return 0; +#endif /* COMPILE_ONLY */ +} + +protected void +file_delmagic(struct magic *p, int type, size_t entries) +{ + if (p == NULL) + return; + switch (type) { +#ifdef QUICK + case 2: + p--; + (void)munmap((void *)p, sizeof(*p) * (entries + 1)); + break; +#endif + case 1: + p--; + /*FALLTHROUGH*/ + case 0: + free(p); + break; + default: + abort(); + } +} + +/* const char *fn: list of magic files and directories */ +protected struct mlist * +file_apprentice(struct magic_set *ms, const char *fn, int action) +{ + char *p, *mfn; + int file_err, errs = -1; + struct mlist *mlist; + + init_file_tables(); + + if (fn == NULL) + fn = getenv("MAGIC"); + if (fn == NULL) + fn = MAGIC; + + if ((mfn = strdup(fn)) == NULL) { + file_oomem(ms, strlen(fn)); + return NULL; + } + fn = mfn; + + if ((mlist = malloc(sizeof(*mlist))) == NULL) { + free(mfn); + file_oomem(ms, sizeof(*mlist)); + return NULL; + } + mlist->next = mlist->prev = mlist; + + while (fn) { + p = strchr(fn, PATHSEP); + if (p) + *p++ = '\0'; + if (*fn == '\0') + break; + file_err = apprentice_1(ms, fn, action, mlist); + errs = MAX(errs, file_err); + fn = p; + } + if (errs == -1) { + free(mfn); + free(mlist); + mlist = NULL; + file_error(ms, 0, "could not find any magic files!"); + return NULL; + } + free(mfn); + return mlist; +} + +/* + * Get weight of this magic entry, for sorting purposes. + */ +private size_t +apprentice_magic_strength(const struct magic *m) +{ +#define MULT 10 + size_t val = 2 * MULT; /* baseline strength */ + + switch (m->type) { + case FILE_DEFAULT: /* make sure this sorts last */ + return 0; + + case FILE_BYTE: + val += 1 * MULT; + break; + + case FILE_SHORT: + case FILE_LESHORT: + case FILE_BESHORT: + val += 2 * MULT; + break; + + case FILE_LONG: + case FILE_LELONG: + case FILE_BELONG: + case FILE_MELONG: + val += 4 * MULT; + break; + + case FILE_PSTRING: + case FILE_STRING: + val += m->vallen * MULT; + break; + + case FILE_BESTRING16: + case FILE_LESTRING16: + val += m->vallen * MULT / 2; + break; + + case FILE_SEARCH: + case FILE_REGEX: + val += m->vallen * MAX(MULT / m->vallen, 1); + break; + + case FILE_DATE: + case FILE_LEDATE: + case FILE_BEDATE: + case FILE_MEDATE: + case FILE_LDATE: + case FILE_LELDATE: + case FILE_BELDATE: + case FILE_MELDATE: + case FILE_FLOAT: + case FILE_BEFLOAT: + case FILE_LEFLOAT: + val += 4 * MULT; + break; + + case FILE_QUAD: + case FILE_BEQUAD: + case FILE_LEQUAD: + case FILE_QDATE: + case FILE_LEQDATE: + case FILE_BEQDATE: + case FILE_QLDATE: + case FILE_LEQLDATE: + case FILE_BEQLDATE: + case FILE_DOUBLE: + case FILE_BEDOUBLE: + case FILE_LEDOUBLE: + val += 8 * MULT; + break; + + default: + val = 0; + (void)fprintf(stderr, "Bad type %d\n", m->type); + abort(); + } + + switch (m->reln) { + case 'x': /* matches anything penalize */ + case '!': /* matches almost anything penalize */ + val = 0; + break; + + case '=': /* Exact match, prefer */ + val += MULT; + break; + + case '>': + case '<': /* comparison match reduce strength */ + val -= 2 * MULT; + break; + + case '^': + case '&': /* masking bits, we could count them too */ + val -= MULT; + break; + + default: + (void)fprintf(stderr, "Bad relation %c\n", m->reln); + abort(); + } + + if (val == 0) /* ensure we only return 0 for FILE_DEFAULT */ + val = 1; + + return val; +} + +/* + * Sort callback for sorting entries by "strength" (basically length) + */ +private int +apprentice_sort(const void *a, const void *b) +{ + const struct magic_entry *ma = a; + const struct magic_entry *mb = b; + size_t sa = apprentice_magic_strength(ma->mp); + size_t sb = apprentice_magic_strength(mb->mp); + if (sa == sb) + return 0; + else if (sa > sb) + return -1; + else + return 1; +} + +private int +set_test_type(struct magic *mstart, struct magic *m) +{ + switch (m->type) { + case FILE_BYTE: + case FILE_SHORT: + case FILE_LONG: + case FILE_DATE: + case FILE_BESHORT: + case FILE_BELONG: + case FILE_BEDATE: + case FILE_LESHORT: + case FILE_LELONG: + case FILE_LEDATE: + case FILE_LDATE: + case FILE_BELDATE: + case FILE_LELDATE: + case FILE_MEDATE: + case FILE_MELDATE: + case FILE_MELONG: + case FILE_QUAD: + case FILE_LEQUAD: + case FILE_BEQUAD: + case FILE_QDATE: + case FILE_LEQDATE: + case FILE_BEQDATE: + case FILE_QLDATE: + case FILE_LEQLDATE: + case FILE_BEQLDATE: + case FILE_FLOAT: + case FILE_BEFLOAT: + case FILE_LEFLOAT: + case FILE_DOUBLE: + case FILE_BEDOUBLE: + case FILE_LEDOUBLE: + case FILE_STRING: + case FILE_PSTRING: + case FILE_BESTRING16: + case FILE_LESTRING16: + /* binary test, set flag */ + mstart->flag |= BINTEST; + break; + case FILE_REGEX: + case FILE_SEARCH: + /* binary test if pattern is not text */ + if (file_looks_utf8(m->value.s, m->vallen, NULL, NULL) == 0) + mstart->flag |= BINTEST; + break; + case FILE_DEFAULT: + /* can't deduce anything; we shouldn't see this at the + top level anyway */ + break; + case FILE_INVALID: + default: + /* invalid search type, but no need to complain here */ + break; + } +} + +/* + * Load and parse one file. + */ +private void +load_1(struct magic_set *ms, int action, const char *fn, int *errs, + struct magic_entry **marray, uint32_t *marraycount) +{ + char line[BUFSIZ]; + size_t lineno = 0; + FILE *f = fopen(ms->file = fn, "r"); + if (f == NULL) { + if (errno != ENOENT) + file_error(ms, errno, "cannot read magic file `%s'", + fn); + (*errs)++; + } else { + /* read and parse this file */ + for (ms->line = 1; fgets(line, sizeof(line), f) != NULL; ms->line++) { + size_t len; + len = strlen(line); + if (len == 0) /* null line, garbage, etc */ + continue; + if (line[len - 1] == '\n') { + lineno++; + line[len - 1] = '\0'; /* delete newline */ + } + if (line[0] == '\0') /* empty, do not parse */ + continue; + if (line[0] == '#') /* comment, do not parse */ + continue; + if (len > mime_marker_len && + memcmp(line, mime_marker, mime_marker_len) == 0) { + /* MIME type */ + if (parse_mime(ms, marray, marraycount, + line + mime_marker_len) != 0) + (*errs)++; + continue; + } + if (parse(ms, marray, marraycount, line, lineno, action) != 0) + (*errs)++; + } + + (void)fclose(f); + } +} + +/* + * parse a file or directory of files + * const char *fn: name of magic file or directory + */ +private int +apprentice_load(struct magic_set *ms, struct magic **magicp, uint32_t *nmagicp, + const char *fn, int action) +{ + int errs = 0; + struct magic_entry *marray; + uint32_t marraycount, i, mentrycount = 0, starttest; + char subfn[MAXPATHLEN]; + struct stat st; + DIR *dir; + struct dirent *d; + + ms->flags |= MAGIC_CHECK; /* Enable checks for parsed files */ + + maxmagic = MAXMAGIS; + if ((marray = calloc(maxmagic, sizeof(*marray))) == NULL) { + file_oomem(ms, maxmagic * sizeof(*marray)); + return -1; + } + marraycount = 0; + + /* print silly verbose header for USG compat. */ + if (action == FILE_CHECK) + (void)fprintf(stderr, "%s\n", usg_hdr); + + /* load directory or file */ + if (stat(fn, &st) == 0 && S_ISDIR(st.st_mode)) { + dir = opendir(fn); + if (dir) { + while (d = readdir(dir)) { + snprintf(subfn, sizeof(subfn), "%s/%s", + fn, d->d_name); + if (stat(subfn, &st) == 0 && S_ISREG(st.st_mode)) { + load_1(ms, action, subfn, &errs, + &marray, &marraycount); + } + } + closedir(dir); + } else + errs++; + } else + load_1(ms, action, fn, &errs, &marray, &marraycount); + if (errs) + goto out; + + /* Set types of tests */ + for (i = 0; i < marraycount; ) { + if (marray[i].mp->cont_level != 0) { + i++; + continue; + } + + starttest = i; + do { + set_test_type(marray[starttest].mp, marray[i].mp); + if (ms->flags & MAGIC_DEBUG) { + (void)fprintf(stderr, "%s%s%s: %s\n", + marray[i].mp->mimetype, + marray[i].mp->mimetype[0] == '\0' ? "" : "; ", + marray[i].mp->desc[0] ? marray[i].mp->desc : "(no description)", + marray[i].mp->flag & BINTEST ? "binary" : "text"); + if (marray[i].mp->flag & BINTEST) { +#define SYMBOL "text" +#define SYMLEN sizeof(SYMBOL) + char *p = strstr(marray[i].mp->desc, "text"); + if (p && (p == marray[i].mp->desc || isspace(p[-1])) && + (p + SYMLEN - marray[i].mp->desc == MAXstring || + (p[SYMLEN] == '\0' || isspace(p[SYMLEN])))) { + (void)fprintf(stderr, + "*** Possible binary test for text type\n"); + } +#undef SYMBOL +#undef SYMLEN + } + } + } while (++i < marraycount && marray[i].mp->cont_level != 0); + } + + qsort(marray, marraycount, sizeof(*marray), apprentice_sort); + + /* + * Make sure that any level 0 "default" line is last (if one exists). + */ + for (i = 0; i < marraycount; i++) { + if (marray[i].mp->cont_level == 0 && + marray[i].mp->type == FILE_DEFAULT) { + while (++i < marraycount) + if (marray[i].mp->cont_level == 0) + break; + if (i != marraycount) { + ms->line = marray[i].mp->lineno; /* XXX - Ugh! */ + file_magwarn(ms, + "level 0 \"default\" did not sort last"); + } + break; + } + } + + for (i = 0; i < marraycount; i++) + mentrycount += marray[i].cont_count; + + if ((*magicp = malloc(sizeof(**magicp) * mentrycount)) == NULL) { + file_oomem(ms, sizeof(**magicp) * mentrycount); + errs++; + goto out; + } + + mentrycount = 0; + for (i = 0; i < marraycount; i++) { + (void)memcpy(*magicp + mentrycount, marray[i].mp, + marray[i].cont_count * sizeof(**magicp)); + mentrycount += marray[i].cont_count; + } +out: + for (i = 0; i < marraycount; i++) + free(marray[i].mp); + free(marray); + if (errs) { + *magicp = NULL; + *nmagicp = 0; + return errs; + } else { + *nmagicp = mentrycount; + return 0; + } + +} + +/* + * extend the sign bit if the comparison is to be signed + */ +protected uint64_t +file_signextend(struct magic_set *ms, struct magic *m, uint64_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 FILE_BYTE: + v = (char) v; + break; + case FILE_SHORT: + case FILE_BESHORT: + case FILE_LESHORT: + v = (short) v; + break; + case FILE_DATE: + case FILE_BEDATE: + case FILE_LEDATE: + case FILE_MEDATE: + case FILE_LDATE: + case FILE_BELDATE: + case FILE_LELDATE: + case FILE_MELDATE: + case FILE_LONG: + case FILE_BELONG: + case FILE_LELONG: + case FILE_MELONG: + case FILE_FLOAT: + case FILE_BEFLOAT: + case FILE_LEFLOAT: + v = (int32_t) v; + break; + case FILE_QUAD: + case FILE_BEQUAD: + case FILE_LEQUAD: + case FILE_QDATE: + case FILE_QLDATE: + case FILE_BEQDATE: + case FILE_BEQLDATE: + case FILE_LEQDATE: + case FILE_LEQLDATE: + case FILE_DOUBLE: + case FILE_BEDOUBLE: + case FILE_LEDOUBLE: + v = (int64_t) v; + break; + case FILE_STRING: + case FILE_PSTRING: + case FILE_BESTRING16: + case FILE_LESTRING16: + case FILE_REGEX: + case FILE_SEARCH: + case FILE_DEFAULT: + break; + default: + if (ms->flags & MAGIC_CHECK) + file_magwarn(ms, "cannot happen: m->type=%d\n", + m->type); + return ~0U; + } + } + return v; +} + +private int +string_modifier_check(struct magic_set *ms, struct magic *m) +{ + if ((ms->flags & MAGIC_CHECK) == 0) + return 0; + + switch (m->type) { + case FILE_BESTRING16: + case FILE_LESTRING16: + if (m->str_flags != 0) { + file_magwarn(ms, + "no modifiers allowed for 16-bit strings\n"); + return -1; + } + break; + case FILE_STRING: + case FILE_PSTRING: + if ((m->str_flags & REGEX_OFFSET_START) != 0) { + file_magwarn(ms, + "'/%c' only allowed on regex and search\n", + CHAR_REGEX_OFFSET_START); + return -1; + } + break; + case FILE_SEARCH: + if (m->str_range == 0) { + file_magwarn(ms, + "missing range; defaulting to %d\n", + STRING_DEFAULT_RANGE); + m->str_range = STRING_DEFAULT_RANGE; + return -1; + } + break; + case FILE_REGEX: + if ((m->str_flags & STRING_COMPACT_BLANK) != 0) { + file_magwarn(ms, "'/%c' not allowed on regex\n", + CHAR_COMPACT_BLANK); + return -1; + } + if ((m->str_flags & STRING_COMPACT_OPTIONAL_BLANK) != 0) { + file_magwarn(ms, "'/%c' not allowed on regex\n", + CHAR_COMPACT_OPTIONAL_BLANK); + return -1; + } + break; + default: + file_magwarn(ms, "coding error: m->type=%d\n", + m->type); + return -1; + } + return 0; +} + +private int +get_op(char c) +{ + switch (c) { + case '&': + return FILE_OPAND; + case '|': + return FILE_OPOR; + case '^': + return FILE_OPXOR; + case '+': + return FILE_OPADD; + case '-': + return FILE_OPMINUS; + case '*': + return FILE_OPMULTIPLY; + case '/': + return FILE_OPDIVIDE; + case '%': + return FILE_OPMODULO; + default: + return -1; + } +} + +#ifdef ENABLE_CONDITIONALS +private int +get_cond(const char *l, const char **t) +{ + static const struct cond_tbl_s { + char name[8]; + size_t len; + int cond; + } cond_tbl[] = { + { "if", 2, COND_IF }, + { "elif", 4, COND_ELIF }, + { "else", 4, COND_ELSE }, + { "", 0, COND_NONE }, + }; + const struct cond_tbl_s *p; + + for (p = cond_tbl; p->len; p++) { + if (strncmp(l, p->name, p->len) == 0 && + isspace((unsigned char)l[p->len])) { + if (t) + *t = l + p->len; + break; + } + } + return p->cond; +} + +private int +check_cond(struct magic_set *ms, int cond, uint32_t cont_level) +{ + int last_cond; + last_cond = ms->c.li[cont_level].last_cond; + + switch (cond) { + case COND_IF: + if (last_cond != COND_NONE && last_cond != COND_ELIF) { + if (ms->flags & MAGIC_CHECK) + file_magwarn(ms, "syntax error: `if'"); + return -1; + } + last_cond = COND_IF; + break; + + case COND_ELIF: + if (last_cond != COND_IF && last_cond != COND_ELIF) { + if (ms->flags & MAGIC_CHECK) + file_magwarn(ms, "syntax error: `elif'"); + return -1; + } + last_cond = COND_ELIF; + break; + + case COND_ELSE: + if (last_cond != COND_IF && last_cond != COND_ELIF) { + if (ms->flags & MAGIC_CHECK) + file_magwarn(ms, "syntax error: `else'"); + return -1; + } + last_cond = COND_NONE; + break; + + case COND_NONE: + last_cond = COND_NONE; + break; + } + + ms->c.li[cont_level].last_cond = last_cond; + return 0; +} +#endif /* ENABLE_CONDITIONALS */ + +/* + * parse one line from magic file, put into magic[index++] if valid + */ +private int +parse(struct magic_set *ms, struct magic_entry **mentryp, uint32_t *nmentryp, + const char *line, size_t lineno, int action) +{ +#ifdef ENABLE_CONDITIONALS + static uint32_t last_cont_level = 0; +#endif + size_t i; + struct magic_entry *me; + struct magic *m; + const char *l = line; + char *t; + int op; + uint32_t cont_level; + + cont_level = 0; + + while (*l == '>') { + ++l; /* step over */ + cont_level++; + } +#ifdef ENABLE_CONDITIONALS + if (cont_level == 0 || cont_level > last_cont_level) + if (file_check_mem(ms, cont_level) == -1) + return -1; + last_cont_level = cont_level; +#endif + +#define ALLOC_CHUNK (size_t)10 +#define ALLOC_INCR (size_t)200 + + if (cont_level != 0) { + if (*nmentryp == 0) { + file_error(ms, 0, "No current entry for continuation"); + return -1; + } + me = &(*mentryp)[*nmentryp - 1]; + if (me->cont_count == me->max_count) { + struct magic *nm; + size_t cnt = me->max_count + ALLOC_CHUNK; + if ((nm = realloc(me->mp, sizeof(*nm) * cnt)) == NULL) { + file_oomem(ms, sizeof(*nm) * cnt); + return -1; + } + me->mp = m = nm; + me->max_count = cnt; + } + m = &me->mp[me->cont_count++]; + (void)memset(m, 0, sizeof(*m)); + m->cont_level = cont_level; + } else { + if (*nmentryp == maxmagic) { + struct magic_entry *mp; + + maxmagic += ALLOC_INCR; + if ((mp = realloc(*mentryp, sizeof(*mp) * maxmagic)) == + NULL) { + file_oomem(ms, sizeof(*mp) * maxmagic); + return -1; + } + (void)memset(&mp[*nmentryp], 0, sizeof(*mp) * + ALLOC_INCR); + *mentryp = mp; + } + me = &(*mentryp)[*nmentryp]; + if (me->mp == NULL) { + if ((m = malloc(sizeof(*m) * ALLOC_CHUNK)) == NULL) { + file_oomem(ms, sizeof(*m) * ALLOC_CHUNK); + return -1; + } + me->mp = m; + me->max_count = ALLOC_CHUNK; + } else + m = me->mp; + (void)memset(m, 0, sizeof(*m)); + m->cont_level = 0; + me->cont_count = 1; + } + m->lineno = lineno; + + if (*l == '&') { /* m->cont_level == 0 checked below. */ + ++l; /* step over */ + m->flag |= OFFADD; + } + if (*l == '(') { + ++l; /* step over */ + m->flag |= INDIR; + if (m->flag & OFFADD) + m->flag = (m->flag & ~OFFADD) | INDIROFFADD; + + if (*l == '&') { /* m->cont_level == 0 checked below */ + ++l; /* step over */ + m->flag |= OFFADD; + } + } + /* Indirect offsets are not valid at level 0. */ + if (m->cont_level == 0 && (m->flag & (OFFADD | INDIROFFADD))) + if (ms->flags & MAGIC_CHECK) + file_magwarn(ms, "relative offset at level 0"); + + /* get offset, then skip over it */ + m->offset = (uint32_t)strtoul(l, &t, 0); + if (l == t) + if (ms->flags & MAGIC_CHECK) + file_magwarn(ms, "offset `%s' invalid", l); + l = t; + + if (m->flag & INDIR) { + m->in_type = FILE_LONG; + m->in_offset = 0; + /* + * read [.lbs][+-]nnnnn) + */ + if (*l == '.') { + l++; + switch (*l) { + case 'l': + m->in_type = FILE_LELONG; + break; + case 'L': + m->in_type = FILE_BELONG; + break; + case 'm': + m->in_type = FILE_MELONG; + break; + case 'h': + case 's': + m->in_type = FILE_LESHORT; + break; + case 'H': + case 'S': + m->in_type = FILE_BESHORT; + break; + case 'c': + case 'b': + case 'C': + case 'B': + m->in_type = FILE_BYTE; + break; + case 'e': + case 'f': + case 'g': + m->in_type = FILE_LEDOUBLE; + break; + case 'E': + case 'F': + case 'G': + m->in_type = FILE_BEDOUBLE; + break; + default: + if (ms->flags & MAGIC_CHECK) + file_magwarn(ms, + "indirect offset type `%c' invalid", + *l); + break; + } + l++; + } + + m->in_op = 0; + if (*l == '~') { + m->in_op |= FILE_OPINVERSE; + l++; + } + if ((op = get_op(*l)) != -1) { + m->in_op |= op; + l++; + } + if (*l == '(') { + m->in_op |= FILE_OPINDIRECT; + l++; + } + if (isdigit((unsigned char)*l) || *l == '-') { + m->in_offset = (int32_t)strtol(l, &t, 0); + if (l == t) + if (ms->flags & MAGIC_CHECK) + file_magwarn(ms, + "in_offset `%s' invalid", l); + l = t; + } + if (*l++ != ')' || + ((m->in_op & FILE_OPINDIRECT) && *l++ != ')')) + if (ms->flags & MAGIC_CHECK) + file_magwarn(ms, + "missing ')' in indirect offset"); + } + EATAB; + +#ifdef ENABLE_CONDITIONALS + m->cond = get_cond(l, &l); + if (check_cond(ms, m->cond, cont_level) == -1) + return -1; + + EATAB; +#endif + + if (*l == 'u') { + ++l; + m->flag |= UNSIGNED; + } + + m->type = get_type(l, &l); + if (m->type == FILE_INVALID) { + if (ms->flags & MAGIC_CHECK) + file_magwarn(ms, "type `%s' invalid", l); + return -1; + } + + /* New-style anding: "0 byte&0x80 =0x80 dynamically linked" */ + /* New and improved: ~ & | ^ + - * / % -- exciting, isn't it? */ + + m->mask_op = 0; + if (*l == '~') { + if (!IS_STRING(m->type)) + m->mask_op |= FILE_OPINVERSE; + else if (ms->flags & MAGIC_CHECK) + file_magwarn(ms, "'~' invalid for string types"); + ++l; + } + m->str_range = 0; + m->str_flags = 0; + m->num_mask = 0; + if ((op = get_op(*l)) != -1) { + if (!IS_STRING(m->type)) { + uint64_t val; + ++l; + m->mask_op |= op; + val = (uint64_t)strtoull(l, &t, 0); + l = t; + m->num_mask = file_signextend(ms, m, val); + eatsize(&l); + } + else if (op == FILE_OPDIVIDE) { + int have_range = 0; + while (!isspace((unsigned char)*++l)) { + switch (*l) { + case '0': case '1': case '2': + case '3': case '4': case '5': + case '6': case '7': case '8': + case '9': + if (have_range && + (ms->flags & MAGIC_CHECK)) + file_magwarn(ms, + "multiple ranges"); + have_range = 1; + m->str_range = strtoul(l, &t, 0); + if (m->str_range == 0) + file_magwarn(ms, + "zero range"); + l = t - 1; + break; + case CHAR_COMPACT_BLANK: + m->str_flags |= STRING_COMPACT_BLANK; + break; + case CHAR_COMPACT_OPTIONAL_BLANK: + m->str_flags |= + STRING_COMPACT_OPTIONAL_BLANK; + break; + case CHAR_IGNORE_LOWERCASE: + m->str_flags |= STRING_IGNORE_LOWERCASE; + break; + case CHAR_IGNORE_UPPERCASE: + m->str_flags |= STRING_IGNORE_UPPERCASE; + break; + case CHAR_REGEX_OFFSET_START: + m->str_flags |= REGEX_OFFSET_START; + break; + default: + if (ms->flags & MAGIC_CHECK) + file_magwarn(ms, + "string extension `%c' invalid", + *l); + return -1; + } + /* allow multiple '/' for readability */ + if (l[1] == '/' && + !isspace((unsigned char)l[2])) + l++; + } + if (string_modifier_check(ms, m) == -1) + return -1; + } + else { + if (ms->flags & MAGIC_CHECK) + file_magwarn(ms, "invalid string op: %c", *t); + return -1; + } + } + /* + * 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 '!': + m->reln = *l; + ++l; + break; + default: + m->reln = '='; /* the default relation */ + if (*l == 'x' && ((isascii((unsigned char)l[1]) && + isspace((unsigned char)l[1])) || !l[1])) { + m->reln = *l; + ++l; + } + break; + } + /* + * Grab the value part, except for an 'x' reln. + */ + if (m->reln != 'x' && getvalue(ms, m, &l, action)) + 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 + */ + EATAB; + if (l[0] == '\b') { + ++l; + m->flag |= NOSPACE; + } else if ((l[0] == '\\') && (l[1] == 'b')) { + ++l; + ++l; + m->flag |= NOSPACE; + } + for (i = 0; (m->desc[i++] = *l++) != '\0' && i < sizeof(m->desc); ) + continue; + if (i == sizeof(m->desc)) { + m->desc[sizeof(m->desc) - 1] = '\0'; + if (ms->flags & MAGIC_CHECK) + file_magwarn(ms, "description `%s' truncated", m->desc); + } + + /* + * We only do this check while compiling, or if any of the magic + * files were not compiled. + */ + if (ms->flags & MAGIC_CHECK) { + if (check_format(ms, m) == -1) + return -1; + } +#ifndef COMPILE_ONLY + if (action == FILE_CHECK) { + file_mdump(m); + } +#endif + m->mimetype[0] = '\0'; /* initialise MIME type to none */ + if (m->cont_level == 0) + ++(*nmentryp); /* make room for next */ + return 0; +} + +/* + * parse a MIME annotation line from magic file, put into magic[index - 1] + * if valid + */ +private int +parse_mime(struct magic_set *ms, struct magic_entry **mentryp, + uint32_t *nmentryp, const char *line) +{ + size_t i; + const char *l = line; + struct magic *m; + struct magic_entry *me; + + if (*nmentryp == 0) { + file_error(ms, 0, "No current entry for MIME type"); + return -1; + } + + me = &(*mentryp)[*nmentryp - 1]; + m = &me->mp[me->cont_count == 0 ? 0 : me->cont_count - 1]; + + if (m->mimetype[0] != '\0') { + file_error(ms, 0, "Current entry already has a MIME type: %s\n" + "Description: %s\nNew type: %s", m->mimetype, m->desc, l); + return -1; + } + + EATAB; + for (i = 0; + *l && ((isascii((unsigned char)*l) && isalnum((unsigned char)*l)) + || strchr("-+/.", *l)) && i < sizeof(m->mimetype); + m->mimetype[i++] = *l++) + continue; + if (i == sizeof(m->mimetype)) { + m->desc[sizeof(m->mimetype) - 1] = '\0'; + if (ms->flags & MAGIC_CHECK) + file_magwarn(ms, "MIME type `%s' truncated %zu", + m->mimetype, i); + } else + m->mimetype[i] = '\0'; + + if (i > 0) + return 0; + else + return -1; +} + +private int +check_format_type(const char *ptr, int type) +{ + int quad = 0; + if (*ptr == '\0') { + /* Missing format string; bad */ + return -1; + } + + switch (type) { + case FILE_FMT_QUAD: + quad = 1; + /*FALLTHROUGH*/ + case FILE_FMT_NUM: + if (*ptr == '-') + ptr++; + if (*ptr == '.') + ptr++; + while (isdigit((unsigned char)*ptr)) ptr++; + if (*ptr == '.') + ptr++; + while (isdigit((unsigned char)*ptr)) ptr++; + if (quad) { + if (*ptr++ != 'l') + return -1; + if (*ptr++ != 'l') + return -1; + } + + switch (*ptr++) { + case 'l': + switch (*ptr++) { + case 'i': + case 'd': + case 'u': + case 'x': + case 'X': + return 0; + default: + return -1; + } + + case 'h': + switch (*ptr++) { + case 'h': + switch (*ptr++) { + case 'i': + case 'd': + case 'u': + case 'x': + case 'X': + return 0; + default: + return -1; + } + case 'd': + return 0; + default: + return -1; + } + + case 'i': + case 'c': + case 'd': + case 'u': + case 'x': + case 'X': + return 0; + + default: + return -1; + } + + case FILE_FMT_FLOAT: + case FILE_FMT_DOUBLE: + if (*ptr == '-') + ptr++; + if (*ptr == '.') + ptr++; + while (isdigit((unsigned char)*ptr)) ptr++; + if (*ptr == '.') + ptr++; + while (isdigit((unsigned char)*ptr)) ptr++; + + switch (*ptr++) { + case 'e': + case 'E': + case 'f': + case 'F': + case 'g': + case 'G': + return 0; + + default: + return -1; + } + + + case FILE_FMT_STR: + if (*ptr == '-') + ptr++; + while (isdigit((unsigned char )*ptr)) + ptr++; + if (*ptr == '.') { + ptr++; + while (isdigit((unsigned char )*ptr)) + ptr++; + } + + switch (*ptr++) { + case 's': + return 0; + default: + return -1; + } + + default: + /* internal error */ + abort(); + } + /*NOTREACHED*/ + return -1; +} + +/* + * Check that the optional printf format in description matches + * the type of the magic. + */ +private int +check_format(struct magic_set *ms, struct magic *m) +{ + char *ptr; + + for (ptr = m->desc; *ptr; ptr++) + if (*ptr == '%') + break; + if (*ptr == '\0') { + /* No format string; ok */ + return 1; + } + + assert(file_nformats == file_nnames); + + if (m->type >= file_nformats) { + file_magwarn(ms, "Internal error inconsistency between " + "m->type and format strings"); + return -1; + } + if (file_formats[m->type] == FILE_FMT_NONE) { + file_magwarn(ms, "No format string for `%s' with description " + "`%s'", m->desc, file_names[m->type]); + return -1; + } + + ptr++; + if (check_format_type(ptr, file_formats[m->type]) == -1) { + /* + * TODO: this error message is unhelpful if the format + * string is not one character long + */ + file_magwarn(ms, "Printf format `%c' is not valid for type " + "`%s' in description `%s'", *ptr, + file_names[m->type], m->desc); + return -1; + } + + for (; *ptr; ptr++) { + if (*ptr == '%') { + file_magwarn(ms, + "Too many format strings (should have at most one) " + "for `%s' with description `%s'", + file_names[m->type], m->desc); + return -1; + } + } + 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. + */ +private int +getvalue(struct magic_set *ms, struct magic *m, const char **p, int action) +{ + int slen; + + switch (m->type) { + case FILE_BESTRING16: + case FILE_LESTRING16: + case FILE_STRING: + case FILE_PSTRING: + case FILE_REGEX: + case FILE_SEARCH: + *p = getstr(ms, *p, m->value.s, sizeof(m->value.s), &slen, action); + if (*p == NULL) { + if (ms->flags & MAGIC_CHECK) + file_magwarn(ms, "cannot get string from `%s'", + m->value.s); + return -1; + } + m->vallen = slen; + if (m->type == FILE_PSTRING) + m->vallen++; + return 0; + case FILE_FLOAT: + case FILE_BEFLOAT: + case FILE_LEFLOAT: + if (m->reln != 'x') { + char *ep; +#ifdef HAVE_STRTOF + m->value.f = strtof(*p, &ep); +#else + m->value.f = (float)strtod(*p, &ep); +#endif + *p = ep; + } + return 0; + case FILE_DOUBLE: + case FILE_BEDOUBLE: + case FILE_LEDOUBLE: + if (m->reln != 'x') { + char *ep; + m->value.d = strtod(*p, &ep); + *p = ep; + } + return 0; + default: + if (m->reln != 'x') { + char *ep; + m->value.q = file_signextend(ms, m, + (uint64_t)strtoull(*p, &ep, 0)); + *p = ep; + 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. + */ +private const char * +getstr(struct magic_set *ms, const char *s, char *p, int plen, int *slen, int action) +{ + const char *origs = s; + char *origp = p; + char *pmax = p + plen - 1; + int c; + int val; + + while ((c = *s++) != '\0') { + if (isspace((unsigned char) c)) + break; + if (p >= pmax) { + file_error(ms, 0, "string too long: `%s'", origs); + return NULL; + } + if (c == '\\') { + switch(c = *s++) { + + case '\0': + if (action == FILE_COMPILE) + file_magwarn(ms, "incomplete escape"); + goto out; + + case '\t': + if (action == FILE_COMPILE) { + file_magwarn(ms, + "escaped tab found, use \\t instead"); + action++; + } + /*FALLTHROUGH*/ + default: + if (action == FILE_COMPILE) { + if (isprint((unsigned char)c)) + file_magwarn(ms, + "no need to escape `%c'", c); + else + file_magwarn(ms, + "unknown escape sequence: \\%03o", c); + } + /*FALLTHROUGH*/ + /* space, perhaps force people to use \040? */ + case ' ': +#if 0 + /* + * Other things people escape, but shouldn't need to, + * so we disallow them + */ + case '\'': + case '"': + case '?': +#endif + /* Relations */ + case '>': + case '<': + case '&': + case '^': + case '=': + case '!': + /* and baskslash itself */ + case '\\': + *p++ = (char) c; + break; + + case 'a': + *p++ = '\a'; + break; + + case 'b': + *p++ = '\b'; + break; + + case 'f': + *p++ = '\f'; + break; + + case 'n': + *p++ = '\n'; + break; + + case 'r': + *p++ = '\r'; + break; + + case 't': + *p++ = '\t'; + 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. */ +private 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. + */ +protected void +file_showstr(FILE *fp, const char *s, size_t len) +{ + char c; + + for (;;) { + c = *s++; + if (len == ~0U) { + 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 '\a': + (void) fputc('a', fp); + break; + + case '\b': + (void) fputc('b', fp); + break; + + case '\f': + (void) fputc('f', fp); + break; + + case '\n': + (void) fputc('n', fp); + break; + + case '\r': + (void) fputc('r', fp); + break; + + case '\t': + (void) fputc('t', 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] + */ +private void +eatsize(const char **p) +{ + const 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. + */ +private int +apprentice_map(struct magic_set *ms, struct magic **magicp, uint32_t *nmagicp, + const char *fn) +{ + int fd; + struct stat st; + uint32_t *ptr; + uint32_t version; + int needsbyteswap; + char *dbname = NULL; + void *mm = NULL; + + mkdbname(fn, &dbname, 0); + if (dbname == NULL) + goto error2; + + if ((fd = open(dbname, O_RDONLY|O_BINARY)) == -1) + goto error2; + + if (fstat(fd, &st) == -1) { + file_error(ms, errno, "cannot stat `%s'", dbname); + goto error1; + } + if (st.st_size < 8) { + file_error(ms, 0, "file `%s' is too small", dbname); + goto error1; + } + +#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) { + file_error(ms, errno, "cannot map `%s'", dbname); + goto error1; + } +#define RET 2 +#else + if ((mm = malloc((size_t)st.st_size)) == NULL) { + file_oomem(ms, (size_t)st.st_size); + goto error1; + } + if (read(fd, mm, (size_t)st.st_size) != (size_t)st.st_size) { + file_badread(ms); + goto error1; + } +#define RET 1 +#endif + *magicp = mm; + (void)close(fd); + fd = -1; + ptr = (uint32_t *)(void *)*magicp; + if (*ptr != MAGICNO) { + if (swap4(*ptr) != MAGICNO) { + file_error(ms, 0, "bad magic in `%s'"); + goto error1; + } + needsbyteswap = 1; + } else + needsbyteswap = 0; + if (needsbyteswap) + version = swap4(ptr[1]); + else + version = ptr[1]; + if (version != VERSIONNO) { + file_error(ms, 0, "File %d.%d supports only %d version magic " + "files. `%s' is version %d", FILE_VERSION_MAJOR, patchlevel, + VERSIONNO, dbname, version); + goto error1; + } + *nmagicp = (uint32_t)(st.st_size / sizeof(struct magic)); + if (*nmagicp > 0) + (*nmagicp)--; + (*magicp)++; + if (needsbyteswap) + byteswap(*magicp, *nmagicp); + free(dbname); + return RET; + +error1: + if (fd != -1) + (void)close(fd); + if (mm) { +#ifdef QUICK + (void)munmap((void *)mm, (size_t)st.st_size); +#else + free(mm); +#endif + } else { + *magicp = NULL; + *nmagicp = 0; + } +error2: + free(dbname); + return -1; +} + +private const uint32_t ar[] = { + MAGICNO, VERSIONNO +}; +/* + * handle an mmaped file. + */ +private int +apprentice_compile(struct magic_set *ms, struct magic **magicp, + uint32_t *nmagicp, const char *fn) +{ + int fd; + char *dbname; + int rv = -1; + + mkdbname(fn, &dbname, 1); + + if (dbname == NULL) + goto out; + + if ((fd = open(dbname, O_WRONLY|O_CREAT|O_TRUNC|O_BINARY, 0644)) == -1) { + file_error(ms, errno, "cannot open `%s'", dbname); + goto out; + } + + if (write(fd, ar, sizeof(ar)) != (ssize_t)sizeof(ar)) { + file_error(ms, errno, "error writing `%s'", dbname); + goto out; + } + + if (lseek(fd, (off_t)sizeof(struct magic), SEEK_SET) + != sizeof(struct magic)) { + file_error(ms, errno, "error seeking `%s'", dbname); + goto out; + } + + if (write(fd, *magicp, (sizeof(struct magic) * *nmagicp)) + != (ssize_t)(sizeof(struct magic) * *nmagicp)) { + file_error(ms, errno, "error writing `%s'", dbname); + goto out; + } + + (void)close(fd); + rv = 0; +out: + free(dbname); + return rv; +} + +private const char ext[] = ".mgc"; +/* + * make a dbname + */ +private void +mkdbname(const char *fn, char **buf, int strip) +{ + if (strip) { + const char *p; + if ((p = strrchr(fn, '/')) != NULL) + fn = ++p; + } + + (void)asprintf(buf, "%s%s", fn, ext); + if (*buf && strlen(*buf) > MAXPATHLEN) { + free(*buf); + *buf = NULL; + } +} + +/* + * Byteswap an mmap'ed file if needed + */ +private void +byteswap(struct magic *magic, uint32_t nmagic) +{ + uint32_t i; + for (i = 0; i < nmagic; i++) + bs1(&magic[i]); +} + +/* + * swap a short + */ +private uint16_t +swap2(uint16_t sv) +{ + uint16_t rv; + uint8_t *s = (uint8_t *)(void *)&sv; + uint8_t *d = (uint8_t *)(void *)&rv; + d[0] = s[1]; + d[1] = s[0]; + return rv; +} + +/* + * swap an int + */ +private uint32_t +swap4(uint32_t sv) +{ + uint32_t rv; + uint8_t *s = (uint8_t *)(void *)&sv; + uint8_t *d = (uint8_t *)(void *)&rv; + d[0] = s[3]; + d[1] = s[2]; + d[2] = s[1]; + d[3] = s[0]; + return rv; +} + +/* + * swap a quad + */ +private uint64_t +swap8(uint64_t sv) +{ + uint32_t rv; + uint8_t *s = (uint8_t *)(void *)&sv; + uint8_t *d = (uint8_t *)(void *)&rv; +#if 0 + d[0] = s[3]; + d[1] = s[2]; + d[2] = s[1]; + d[3] = s[0]; + d[4] = s[7]; + d[5] = s[6]; + d[6] = s[5]; + d[7] = s[4]; +#else + d[0] = s[7]; + d[1] = s[6]; + d[2] = s[5]; + d[3] = s[4]; + d[4] = s[3]; + d[5] = s[2]; + d[6] = s[1]; + d[7] = s[0]; +#endif + return rv; +} + +/* + * byteswap a single magic entry + */ +private void +bs1(struct magic *m) +{ + m->cont_level = swap2(m->cont_level); + m->offset = swap4((uint32_t)m->offset); + m->in_offset = swap4((uint32_t)m->in_offset); + m->lineno = swap4((uint32_t)m->lineno); + if (IS_STRING(m->type)) { + m->str_range = swap4(m->str_range); + m->str_flags = swap4(m->str_flags); + } + else { + m->value.q = swap8(m->value.q); + m->num_mask = swap8(m->num_mask); + } +} diff --git a/ext/fileinfo/lib/apptype.c b/ext/fileinfo/lib/apptype.c new file mode 100644 index 0000000000..42cee40e92 --- /dev/null +++ b/ext/fileinfo/lib/apptype.c @@ -0,0 +1,170 @@ +/* + * Adapted from: apptype.c, Written by Eberhard Mattes and put into the + * public domain + * + * Notes: 1. Qualify the filename so that DosQueryAppType does not do extraneous + * searches. + * + * 2. DosQueryAppType will return FAPPTYP_DOS on a file ending with ".com" + * (other than an OS/2 exe or Win exe with this name). Eberhard Mattes + * remarks Tue, 6 Apr 93: Moreover, it reports the type of the (new and very + * bug ridden) Win Emacs as "OS/2 executable". + * + * 3. apptype() uses the filename if given, otherwise a tmp file is created with + * the contents of buf. If buf is not the complete file, apptype can + * incorrectly identify the exe type. The "-z" option of "file" is the reason + * for this ugly code. + */ + +/* + * amai: Darrel Hankerson did the changes described here. + * + * It remains to check the validity of comments (2.) since it's referred to an + * "old" OS/2 version. + * + */ + +#include "file.h" + +#include +#include +#include + + +#ifndef lint +FILE_RCSID("@(#)$File: apptype.c,v 1.7 2007/01/12 17:38:27 christos Exp $") +#endif /* lint */ + +#ifdef __EMX__ +#include +#define INCL_DOSSESMGR +#define INCL_DOSERRORS +#define INCL_DOSFILEMGR +#include +typedef ULONG APPTYPE; + +protected int +file_os2_apptype(struct magic_set *ms, const char *fn, const void *buf, + size_t nb) +{ + APPTYPE rc, type; + char path[_MAX_PATH], drive[_MAX_DRIVE], dir[_MAX_DIR], + fname[_MAX_FNAME], ext[_MAX_EXT]; + char *filename; + FILE *fp; + + if (fn) + filename = strdup(fn); + else if ((filename = tempnam("./", "tmp")) == NULL) { + file_error(ms, errno, "cannot create tempnam"); + return -1; + } + /* qualify the filename to prevent extraneous searches */ + _splitpath(filename, drive, dir, fname, ext); + (void)sprintf(path, "%s%s%s%s", drive, + (*dir == '\0') ? "./" : dir, + fname, + (*ext == '\0') ? "." : ext); + + if (fn == NULL) { + if ((fp = fopen(path, "wb")) == NULL) { + file_error(ms, errno, "cannot open tmp file `%s'", path); + return -1; + } + if (fwrite(buf, 1, nb, fp) != nb) { + file_error(ms, errno, "cannot write tmp file `%s'", + path); + return -1; + } + (void)fclose(fp); + } + rc = DosQueryAppType(path, &type); + + if (fn == NULL) { + unlink(path); + free(filename); + } +#if 0 + if (rc == ERROR_INVALID_EXE_SIGNATURE) + printf("%s: not an executable file\n", fname); + else if (rc == ERROR_FILE_NOT_FOUND) + printf("%s: not found\n", fname); + else if (rc == ERROR_ACCESS_DENIED) + printf("%s: access denied\n", fname); + else if (rc != 0) + printf("%s: error code = %lu\n", fname, rc); + else +#else + + /* + * for our purpose here it's sufficient to just ignore the error and + * return w/o success (=0) + */ + + if (rc) + return (0); + +#endif + + if (type & FAPPTYP_32BIT) + if (file_printf(ms, "32-bit ") == -1) + return -1; + if (type & FAPPTYP_PHYSDRV) { + if (file_printf(ms, "physical device driver") == -1) + return -1; + } else if (type & FAPPTYP_VIRTDRV) { + if (file_printf(ms, "virtual device driver") == -1) + return -1; + } else if (type & FAPPTYP_DLL) { + if (type & FAPPTYP_PROTDLL) + if (file_printf(ms, "protected ") == -1) + return -1; + if (file_printf(ms, "DLL") == -1) + return -1; + } else if (type & (FAPPTYP_WINDOWSREAL | FAPPTYP_WINDOWSPROT)) { + if (file_printf(ms, "Windows executable") == -1) + return -1; + } else if (type & FAPPTYP_DOS) { + /* + * The API routine is partially broken on filenames ending + * ".com". + */ + if (stricmp(ext, ".com") == 0) + if (strncmp((const char *)buf, "MZ", 2)) + return (0); + if (file_printf(ms, "DOS executable") == -1) + return -1; + /* ---------------------------------------- */ + /* Might learn more from the magic(4) entry */ + if (file_printf(ms, ", magic(4)-> ") == -1) + return -1; + return (0); + /* ---------------------------------------- */ + } else if (type & FAPPTYP_BOUND) { + if (file_printf(ms, "bound executable") == -1) + return -1; + } else if ((type & 7) == FAPPTYP_WINDOWAPI) { + if (file_printf(ms, "PM executable") == -1) + return -1; + } else if (file_printf(ms, "OS/2 executable") == -1) + return -1; + + switch (type & (FAPPTYP_NOTWINDOWCOMPAT | + FAPPTYP_WINDOWCOMPAT | + FAPPTYP_WINDOWAPI)) { + case FAPPTYP_NOTWINDOWCOMPAT: + if (file_printf(ms, " [NOTWINDOWCOMPAT]") == -1) + return -1; + break; + case FAPPTYP_WINDOWCOMPAT: + if (file_printf(ms, " [WINDOWCOMPAT]") == -1) + return -1; + break; + case FAPPTYP_WINDOWAPI: + if (file_printf(ms, " [WINDOWAPI]") == -1) + return -1; + break; + } + return 1; +} +#endif diff --git a/ext/fileinfo/lib/ascmagic.c b/ext/fileinfo/lib/ascmagic.c new file mode 100644 index 0000000000..12380ce1de --- /dev/null +++ b/ext/fileinfo/lib/ascmagic.c @@ -0,0 +1,792 @@ +/* + * Copyright (c) Ian F. Darwin 1986-1995. + * Software written by Ian F. Darwin and others; + * maintained 1995-present by Christos Zoulas and others. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice immediately at the beginning of the file, without modification, + * this list of conditions, and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR + * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + */ +/* + * ASCII magic -- file types that we know based on keywords + * that can appear anywhere in the file. + * + * Extensively modified by Eric Fischer in July, 2000, + * to handle character codes other than ASCII on a unified basis. + * + * Joerg Wunsch wrote the original support for 8-bit + * international characters, now subsumed into this file. + */ + +#include "file.h" +#include "magic.h" +#include +#include +#include +#include +#include +#ifdef HAVE_UNISTD_H +#include +#endif +#include "names.h" + +#ifndef lint +FILE_RCSID("@(#)$File: ascmagic.c,v 1.62 2008/03/01 22:21:48 rrt Exp $") +#endif /* lint */ + +#define MAXLINELEN 300 /* longest sane line length */ +#define ISSPC(x) ((x) == ' ' || (x) == '\t' || (x) == '\r' || (x) == '\n' \ + || (x) == 0x85 || (x) == '\f') + +private int looks_ascii(const unsigned char *, size_t, unichar *, size_t *); +private int looks_utf8_with_BOM(const unsigned char *, size_t, unichar *, + size_t *); +protected int file_looks_utf8(const unsigned char *, size_t, unichar *, size_t *); +private int looks_ucs16(const unsigned char *, size_t, unichar *, size_t *); +private int looks_latin1(const unsigned char *, size_t, unichar *, size_t *); +private int looks_extended(const unsigned char *, size_t, unichar *, size_t *); +private void from_ebcdic(const unsigned char *, size_t, unsigned char *); +private int ascmatch(const unsigned char *, const unichar *, size_t); +private unsigned char *encode_utf8(unsigned char *, size_t, unichar *, size_t); + + +protected int +file_ascmagic(struct magic_set *ms, const unsigned char *buf, size_t nbytes) +{ + size_t i; + unsigned char *nbuf = NULL, *utf8_buf = NULL, *utf8_end; + unichar *ubuf = NULL; + size_t ulen, mlen; + const struct names *p; + int rv = -1; + int mime = ms->flags & MAGIC_MIME; + + const char *code = NULL; + const char *code_mime = NULL; + const char *type = NULL; + const char *subtype = NULL; + const char *subtype_mime = NULL; + + int has_escapes = 0; + int has_backspace = 0; + int seen_cr = 0; + + int n_crlf = 0; + int n_lf = 0; + int n_cr = 0; + int n_nel = 0; + + size_t last_line_end = (size_t)-1; + int has_long_lines = 0; + + /* + * Undo the NUL-termination kindly provided by process() + * but leave at least one byte to look at + */ + while (nbytes > 1 && buf[nbytes - 1] == '\0') + nbytes--; + + if ((nbuf = calloc(1, (nbytes + 1) * sizeof(nbuf[0]))) == NULL) + goto done; + if ((ubuf = calloc(1, (nbytes + 1) * sizeof(ubuf[0]))) == NULL) + goto done; + + /* + * Then try to determine whether it's any character code we can + * identify. Each of these tests, if it succeeds, will leave + * the text converted into one-unichar-per-character Unicode in + * ubuf, and the number of characters converted in ulen. + */ + if (looks_ascii(buf, nbytes, ubuf, &ulen)) { + code = "ASCII"; + code_mime = "us-ascii"; + type = "text"; + } else if (looks_utf8_with_BOM(buf, nbytes, ubuf, &ulen) > 0) { + code = "UTF-8 Unicode (with BOM)"; + code_mime = "utf-8"; + type = "text"; + } else if (file_looks_utf8(buf, nbytes, ubuf, &ulen) > 1) { + code = "UTF-8 Unicode"; + code_mime = "utf-8"; + type = "text"; + } else if ((i = looks_ucs16(buf, nbytes, ubuf, &ulen)) != 0) { + if (i == 1) + code = "Little-endian UTF-16 Unicode"; + else + code = "Big-endian UTF-16 Unicode"; + + type = "character data"; + code_mime = "utf-16"; /* is this defined? */ + } else if (looks_latin1(buf, nbytes, ubuf, &ulen)) { + code = "ISO-8859"; + type = "text"; + code_mime = "iso-8859-1"; + } else if (looks_extended(buf, nbytes, ubuf, &ulen)) { + code = "Non-ISO extended-ASCII"; + type = "text"; + code_mime = "unknown"; + } else { + from_ebcdic(buf, nbytes, nbuf); + + if (looks_ascii(nbuf, nbytes, ubuf, &ulen)) { + code = "EBCDIC"; + type = "character data"; + code_mime = "ebcdic"; + } else if (looks_latin1(nbuf, nbytes, ubuf, &ulen)) { + code = "International EBCDIC"; + type = "character data"; + code_mime = "ebcdic"; + } else { + rv = 0; + goto done; /* doesn't look like text at all */ + } + } + + if (nbytes <= 1) { + rv = 0; + goto done; + } + + /* Convert ubuf to UTF-8 and try text soft magic */ + /* If original was ASCII or UTF-8, could use nbuf instead of + re-converting. */ + /* malloc size is a conservative overestimate; could be + re-converting improved, or at least realloced after + re-converting conversion. */ + mlen = ulen * 6; + if ((utf8_buf = malloc(mlen)) == NULL) { + file_oomem(ms, mlen); + goto done; + } + if ((utf8_end = encode_utf8(utf8_buf, mlen, ubuf, ulen)) == NULL) + goto done; + if (file_softmagic(ms, utf8_buf, utf8_end - utf8_buf, TEXTTEST) != 0) { + rv = 1; + goto done; + } + + /* look for tokens from names.h - this is expensive! */ + if ((ms->flags & MAGIC_NO_CHECK_TOKENS) != 0) + goto subtype_identified; + + i = 0; + while (i < ulen) { + size_t end; + + /* skip past any leading space */ + while (i < ulen && ISSPC(ubuf[i])) + i++; + if (i >= ulen) + break; + + /* find the next whitespace */ + for (end = i + 1; end < nbytes; end++) + if (ISSPC(ubuf[end])) + break; + + /* compare the word thus isolated against the token list */ + for (p = names; p < names + NNAMES; p++) { + if (ascmatch((const unsigned char *)p->name, ubuf + i, + end - i)) { + subtype = types[p->type].human; + subtype_mime = types[p->type].mime; + goto subtype_identified; + } + } + + i = end; + } + +subtype_identified: + + /* Now try to discover other details about the file. */ + for (i = 0; i < ulen; i++) { + if (ubuf[i] == '\n') { + if (seen_cr) + n_crlf++; + else + n_lf++; + last_line_end = i; + } else if (seen_cr) + n_cr++; + + seen_cr = (ubuf[i] == '\r'); + if (seen_cr) + last_line_end = i; + + if (ubuf[i] == 0x85) { /* X3.64/ECMA-43 "next line" character */ + n_nel++; + last_line_end = i; + } + + /* If this line is _longer_ than MAXLINELEN, remember it. */ + if (i > last_line_end + MAXLINELEN) + has_long_lines = 1; + + if (ubuf[i] == '\033') + has_escapes = 1; + if (ubuf[i] == '\b') + has_backspace = 1; + } + + /* Beware, if the data has been truncated, the final CR could have + been followed by a LF. If we have HOWMANY bytes, it indicates + that the data might have been truncated, probably even before + this function was called. */ + if (seen_cr && nbytes < HOWMANY) + n_cr++; + + if (mime) { + if (mime & MAGIC_MIME_TYPE) { + if (subtype_mime) { + if (file_printf(ms, subtype_mime) == -1) + goto done; + } else { + if (file_printf(ms, "text/plain") == -1) + goto done; + } + } + + if ((mime == 0 || mime == MAGIC_MIME) && code_mime) { + if ((mime & MAGIC_MIME_TYPE) && + file_printf(ms, " charset=") == -1) + goto done; + if (file_printf(ms, code_mime) == -1) + goto done; + } + + if (mime == MAGIC_MIME_ENCODING) + file_printf(ms, "binary"); + } else { + if (file_printf(ms, code) == -1) + goto done; + + if (subtype) { + if (file_printf(ms, " ") == -1) + goto done; + if (file_printf(ms, subtype) == -1) + goto done; + } + + if (file_printf(ms, " ") == -1) + goto done; + if (file_printf(ms, type) == -1) + goto done; + + if (has_long_lines) + if (file_printf(ms, ", with very long lines") == -1) + goto done; + + /* + * Only report line terminators if we find one other than LF, + * or if we find none at all. + */ + if ((n_crlf == 0 && n_cr == 0 && n_nel == 0 && n_lf == 0) || + (n_crlf != 0 || n_cr != 0 || n_nel != 0)) { + if (file_printf(ms, ", with") == -1) + goto done; + + if (n_crlf == 0 && n_cr == 0 && n_nel == 0 && n_lf == 0) { + if (file_printf(ms, " no") == -1) + goto done; + } else { + if (n_crlf) { + if (file_printf(ms, " CRLF") == -1) + goto done; + if (n_cr || n_lf || n_nel) + if (file_printf(ms, ",") == -1) + goto done; + } + if (n_cr) { + if (file_printf(ms, " CR") == -1) + goto done; + if (n_lf || n_nel) + if (file_printf(ms, ",") == -1) + goto done; + } + if (n_lf) { + if (file_printf(ms, " LF") == -1) + goto done; + if (n_nel) + if (file_printf(ms, ",") == -1) + goto done; + } + if (n_nel) + if (file_printf(ms, " NEL") == -1) + goto done; + } + + if (file_printf(ms, " line terminators") == -1) + goto done; + } + + if (has_escapes) + if (file_printf(ms, ", with escape sequences") == -1) + goto done; + if (has_backspace) + if (file_printf(ms, ", with overstriking") == -1) + goto done; + } + rv = 1; +done: + if (nbuf) + free(nbuf); + if (ubuf) + free(ubuf); + if (utf8_buf) + free(utf8_buf); + + return rv; +} + +private int +ascmatch(const unsigned char *s, const unichar *us, size_t ulen) +{ + size_t i; + + for (i = 0; i < ulen; i++) { + if (s[i] != us[i]) + return 0; + } + + if (s[i]) + return 0; + else + return 1; +} + +/* + * This table reflects a particular philosophy about what constitutes + * "text," and there is room for disagreement about it. + * + * Version 3.31 of the file command considered a file to be ASCII if + * each of its characters was approved by either the isascii() or + * isalpha() function. On most systems, this would mean that any + * file consisting only of characters in the range 0x00 ... 0x7F + * would be called ASCII text, but many systems might reasonably + * consider some characters outside this range to be alphabetic, + * so the file command would call such characters ASCII. It might + * have been more accurate to call this "considered textual on the + * local system" than "ASCII." + * + * It considered a file to be "International language text" if each + * of its characters was either an ASCII printing character (according + * to the real ASCII standard, not the above test), a character in + * the range 0x80 ... 0xFF, or one of the following control characters: + * backspace, tab, line feed, vertical tab, form feed, carriage return, + * escape. No attempt was made to determine the language in which files + * of this type were written. + * + * + * The table below considers a file to be ASCII if all of its characters + * are either ASCII printing characters (again, according to the X3.4 + * standard, not isascii()) or any of the following controls: bell, + * backspace, tab, line feed, form feed, carriage return, esc, nextline. + * + * I include bell because some programs (particularly shell scripts) + * use it literally, even though it is rare in normal text. I exclude + * vertical tab because it never seems to be used in real text. I also + * include, with hesitation, the X3.64/ECMA-43 control nextline (0x85), + * because that's what the dd EBCDIC->ASCII table maps the EBCDIC newline + * character to. It might be more appropriate to include it in the 8859 + * set instead of the ASCII set, but it's got to be included in *something* + * we recognize or EBCDIC files aren't going to be considered textual. + * Some old Unix source files use SO/SI (^N/^O) to shift between Greek + * and Latin characters, so these should possibly be allowed. But they + * make a real mess on VT100-style displays if they're not paired properly, + * so we are probably better off not calling them text. + * + * A file is considered to be ISO-8859 text if its characters are all + * either ASCII, according to the above definition, or printing characters + * from the ISO-8859 8-bit extension, characters 0xA0 ... 0xFF. + * + * Finally, a file is considered to be international text from some other + * character code if its characters are all either ISO-8859 (according to + * the above definition) or characters in the range 0x80 ... 0x9F, which + * ISO-8859 considers to be control characters but the IBM PC and Macintosh + * consider to be printing characters. + */ + +#define F 0 /* character never appears in text */ +#define T 1 /* character appears in plain ASCII text */ +#define I 2 /* character appears in ISO-8859 text */ +#define X 3 /* character appears in non-ISO extended ASCII (Mac, IBM PC) */ + +private char text_chars[256] = { + /* BEL BS HT LF FF CR */ + F, F, F, F, F, F, F, T, T, T, T, F, T, T, F, F, /* 0x0X */ + /* ESC */ + F, F, F, F, F, F, F, F, F, F, F, T, F, F, F, F, /* 0x1X */ + T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x2X */ + T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x3X */ + T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x4X */ + T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x5X */ + T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x6X */ + T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, F, /* 0x7X */ + /* NEL */ + X, X, X, X, X, T, X, X, X, X, X, X, X, X, X, X, /* 0x8X */ + X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, /* 0x9X */ + I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xaX */ + I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xbX */ + I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xcX */ + I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xdX */ + I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xeX */ + I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I /* 0xfX */ +}; + +private int +looks_ascii(const unsigned char *buf, size_t nbytes, unichar *ubuf, + size_t *ulen) +{ + size_t i; + + *ulen = 0; + + for (i = 0; i < nbytes; i++) { + int t = text_chars[buf[i]]; + + if (t != T) + return 0; + + ubuf[(*ulen)++] = buf[i]; + } + + return 1; +} + +private int +looks_latin1(const unsigned char *buf, size_t nbytes, unichar *ubuf, size_t *ulen) +{ + size_t i; + + *ulen = 0; + + for (i = 0; i < nbytes; i++) { + int t = text_chars[buf[i]]; + + if (t != T && t != I) + return 0; + + ubuf[(*ulen)++] = buf[i]; + } + + return 1; +} + +private int +looks_extended(const unsigned char *buf, size_t nbytes, unichar *ubuf, + size_t *ulen) +{ + size_t i; + + *ulen = 0; + + for (i = 0; i < nbytes; i++) { + int t = text_chars[buf[i]]; + + if (t != T && t != I && t != X) + return 0; + + ubuf[(*ulen)++] = buf[i]; + } + + return 1; +} + +/* + * Encode Unicode string as UTF-8, returning pointer to character + * after end of string, or NULL if an invalid character is found. + */ +private unsigned char * +encode_utf8(unsigned char *buf, size_t len, unichar *ubuf, size_t ulen) +{ + size_t i; + unsigned char *end = buf + len; + + for (i = 0; i < ulen; i++) { + if (ubuf[i] <= 0x7f) { + if (end - buf < 1) + return NULL; + *buf++ = (unsigned char)ubuf[i]; + } else if (ubuf[i] <= 0x7ff) { + if (end - buf < 2) + return NULL; + *buf++ = (unsigned char)((ubuf[i] >> 6) + 0xc0); + *buf++ = (unsigned char)((ubuf[i] & 0x3f) + 0x80); + } else if (ubuf[i] <= 0xffff) { + if (end - buf < 3) + return NULL; + *buf++ = (unsigned char)((ubuf[i] >> 12) + 0xe0); + *buf++ = (unsigned char)(((ubuf[i] >> 6) & 0x3f) + 0x80); + *buf++ = (unsigned char)((ubuf[i] & 0x3f) + 0x80); + } else if (ubuf[i] <= 0x1fffff) { + if (end - buf < 4) + return NULL; + *buf++ = (unsigned char)((ubuf[i] >> 18) + 0xf0); + *buf++ = (unsigned char)(((ubuf[i] >> 12) & 0x3f) + 0x80); + *buf++ = (unsigned char)(((ubuf[i] >> 6) & 0x3f) + 0x80); + *buf++ = (unsigned char)((ubuf[i] & 0x3f) + 0x80); + } else if (ubuf[i] <= 0x3ffffff) { + if (end - buf < 5) + return NULL; + *buf++ = (unsigned char)((ubuf[i] >> 24) + 0xf8); + *buf++ = (unsigned char)(((ubuf[i] >> 18) & 0x3f) + 0x80); + *buf++ = (unsigned char)(((ubuf[i] >> 12) & 0x3f) + 0x80); + *buf++ = (unsigned char)(((ubuf[i] >> 6) & 0x3f) + 0x80); + *buf++ = (unsigned char)((ubuf[i] & 0x3f) + 0x80); + } else if (ubuf[i] <= 0x7fffffff) { + if (end - buf < 6) + return NULL; + *buf++ = (unsigned char)((ubuf[i] >> 30) + 0xfc); + *buf++ = (unsigned char)(((ubuf[i] >> 24) & 0x3f) + 0x80); + *buf++ = (unsigned char)(((ubuf[i] >> 18) & 0x3f) + 0x80); + *buf++ = (unsigned char)(((ubuf[i] >> 12) & 0x3f) + 0x80); + *buf++ = (unsigned char)(((ubuf[i] >> 6) & 0x3f) + 0x80); + *buf++ = (unsigned char)((ubuf[i] & 0x3f) + 0x80); + } else /* Invalid character */ + return NULL; + } + + return buf; +} + +/* + * Decide whether some text looks like UTF-8. Returns: + * + * -1: invalid UTF-8 + * 0: uses odd control characters, so doesn't look like text + * 1: 7-bit text + * 2: definitely UTF-8 text (valid high-bit set bytes) + * + * If ubuf is non-NULL on entry, text is decoded into ubuf, *ulen; + * ubuf must be big enough! + */ +protected int +file_looks_utf8(const unsigned char *buf, size_t nbytes, unichar *ubuf, size_t *ulen) +{ + size_t i; + int n; + unichar c; + int gotone = 0, ctrl = 0; + + if (ubuf) + *ulen = 0; + + for (i = 0; i < nbytes; i++) { + if ((buf[i] & 0x80) == 0) { /* 0xxxxxxx is plain ASCII */ + /* + * Even if the whole file is valid UTF-8 sequences, + * still reject it if it uses weird control characters. + */ + + if (text_chars[buf[i]] != T) + ctrl = 1; + + if (ubuf) + ubuf[(*ulen)++] = buf[i]; + } else if ((buf[i] & 0x40) == 0) { /* 10xxxxxx never 1st byte */ + return -1; + } else { /* 11xxxxxx begins UTF-8 */ + int following; + + if ((buf[i] & 0x20) == 0) { /* 110xxxxx */ + c = buf[i] & 0x1f; + following = 1; + } else if ((buf[i] & 0x10) == 0) { /* 1110xxxx */ + c = buf[i] & 0x0f; + following = 2; + } else if ((buf[i] & 0x08) == 0) { /* 11110xxx */ + c = buf[i] & 0x07; + following = 3; + } else if ((buf[i] & 0x04) == 0) { /* 111110xx */ + c = buf[i] & 0x03; + following = 4; + } else if ((buf[i] & 0x02) == 0) { /* 1111110x */ + c = buf[i] & 0x01; + following = 5; + } else + return -1; + + for (n = 0; n < following; n++) { + i++; + if (i >= nbytes) + goto done; + + if ((buf[i] & 0x80) == 0 || (buf[i] & 0x40)) + return -1; + + c = (c << 6) + (buf[i] & 0x3f); + } + + if (ubuf) + ubuf[(*ulen)++] = c; + gotone = 1; + } + } +done: + return ctrl ? 0 : (gotone ? 2 : 1); +} + +/* + * Decide whether some text looks like UTF-8 with BOM. If there is no + * BOM, return -1; otherwise return the result of looks_utf8 on the + * rest of the text. + */ +private int +looks_utf8_with_BOM(const unsigned char *buf, size_t nbytes, unichar *ubuf, + size_t *ulen) +{ + if (nbytes > 3 && buf[0] == 0xef && buf[1] == 0xbb && buf[2] == 0xbf) + return file_looks_utf8(buf + 3, nbytes - 3, ubuf, ulen); + else + return -1; +} + +private int +looks_ucs16(const unsigned char *buf, size_t nbytes, unichar *ubuf, + size_t *ulen) +{ + int bigend; + size_t i; + + if (nbytes < 2) + return 0; + + if (buf[0] == 0xff && buf[1] == 0xfe) + bigend = 0; + else if (buf[0] == 0xfe && buf[1] == 0xff) + bigend = 1; + else + return 0; + + *ulen = 0; + + for (i = 2; i + 1 < nbytes; i += 2) { + /* XXX fix to properly handle chars > 65536 */ + + if (bigend) + ubuf[(*ulen)++] = buf[i + 1] + 256 * buf[i]; + else + ubuf[(*ulen)++] = buf[i] + 256 * buf[i + 1]; + + if (ubuf[*ulen - 1] == 0xfffe) + return 0; + if (ubuf[*ulen - 1] < 128 && + text_chars[(size_t)ubuf[*ulen - 1]] != T) + return 0; + } + + return 1 + bigend; +} + +#undef F +#undef T +#undef I +#undef X + +/* + * This table maps each EBCDIC character to an (8-bit extended) ASCII + * character, as specified in the rationale for the dd(1) command in + * draft 11.2 (September, 1991) of the POSIX P1003.2 standard. + * + * Unfortunately it does not seem to correspond exactly to any of the + * five variants of EBCDIC documented in IBM's _Enterprise Systems + * Architecture/390: Principles of Operation_, SA22-7201-06, Seventh + * Edition, July, 1999, pp. I-1 - I-4. + * + * Fortunately, though, all versions of EBCDIC, including this one, agree + * on most of the printing characters that also appear in (7-bit) ASCII. + * Of these, only '|', '!', '~', '^', '[', and ']' are in question at all. + * + * Fortunately too, there is general agreement that codes 0x00 through + * 0x3F represent control characters, 0x41 a nonbreaking space, and the + * remainder printing characters. + * + * This is sufficient to allow us to identify EBCDIC text and to distinguish + * between old-style and internationalized examples of text. + */ + +private unsigned char ebcdic_to_ascii[] = { + 0, 1, 2, 3, 156, 9, 134, 127, 151, 141, 142, 11, 12, 13, 14, 15, + 16, 17, 18, 19, 157, 133, 8, 135, 24, 25, 146, 143, 28, 29, 30, 31, +128, 129, 130, 131, 132, 10, 23, 27, 136, 137, 138, 139, 140, 5, 6, 7, +144, 145, 22, 147, 148, 149, 150, 4, 152, 153, 154, 155, 20, 21, 158, 26, +' ', 160, 161, 162, 163, 164, 165, 166, 167, 168, 213, '.', '<', '(', '+', '|', +'&', 169, 170, 171, 172, 173, 174, 175, 176, 177, '!', '$', '*', ')', ';', '~', +'-', '/', 178, 179, 180, 181, 182, 183, 184, 185, 203, ',', '%', '_', '>', '?', +186, 187, 188, 189, 190, 191, 192, 193, 194, '`', ':', '#', '@', '\'','=', '"', +195, 'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 196, 197, 198, 199, 200, 201, +202, 'j', 'k', 'l', 'm', 'n', 'o', 'p', 'q', 'r', '^', 204, 205, 206, 207, 208, +209, 229, 's', 't', 'u', 'v', 'w', 'x', 'y', 'z', 210, 211, 212, '[', 214, 215, +216, 217, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 228, ']', 230, 231, +'{', 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 232, 233, 234, 235, 236, 237, +'}', 'J', 'K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R', 238, 239, 240, 241, 242, 243, +'\\',159, 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z', 244, 245, 246, 247, 248, 249, +'0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 250, 251, 252, 253, 254, 255 +}; + +#ifdef notdef +/* + * The following EBCDIC-to-ASCII table may relate more closely to reality, + * or at least to modern reality. It comes from + * + * http://ftp.s390.ibm.com/products/oe/bpxqp9.html + * + * and maps the characters of EBCDIC code page 1047 (the code used for + * Unix-derived software on IBM's 390 systems) to the corresponding + * characters from ISO 8859-1. + * + * If this table is used instead of the above one, some of the special + * cases for the NEL character can be taken out of the code. + */ + +private unsigned char ebcdic_1047_to_8859[] = { +0x00,0x01,0x02,0x03,0x9C,0x09,0x86,0x7F,0x97,0x8D,0x8E,0x0B,0x0C,0x0D,0x0E,0x0F, +0x10,0x11,0x12,0x13,0x9D,0x0A,0x08,0x87,0x18,0x19,0x92,0x8F,0x1C,0x1D,0x1E,0x1F, +0x80,0x81,0x82,0x83,0x84,0x85,0x17,0x1B,0x88,0x89,0x8A,0x8B,0x8C,0x05,0x06,0x07, +0x90,0x91,0x16,0x93,0x94,0x95,0x96,0x04,0x98,0x99,0x9A,0x9B,0x14,0x15,0x9E,0x1A, +0x20,0xA0,0xE2,0xE4,0xE0,0xE1,0xE3,0xE5,0xE7,0xF1,0xA2,0x2E,0x3C,0x28,0x2B,0x7C, +0x26,0xE9,0xEA,0xEB,0xE8,0xED,0xEE,0xEF,0xEC,0xDF,0x21,0x24,0x2A,0x29,0x3B,0x5E, +0x2D,0x2F,0xC2,0xC4,0xC0,0xC1,0xC3,0xC5,0xC7,0xD1,0xA6,0x2C,0x25,0x5F,0x3E,0x3F, +0xF8,0xC9,0xCA,0xCB,0xC8,0xCD,0xCE,0xCF,0xCC,0x60,0x3A,0x23,0x40,0x27,0x3D,0x22, +0xD8,0x61,0x62,0x63,0x64,0x65,0x66,0x67,0x68,0x69,0xAB,0xBB,0xF0,0xFD,0xFE,0xB1, +0xB0,0x6A,0x6B,0x6C,0x6D,0x6E,0x6F,0x70,0x71,0x72,0xAA,0xBA,0xE6,0xB8,0xC6,0xA4, +0xB5,0x7E,0x73,0x74,0x75,0x76,0x77,0x78,0x79,0x7A,0xA1,0xBF,0xD0,0x5B,0xDE,0xAE, +0xAC,0xA3,0xA5,0xB7,0xA9,0xA7,0xB6,0xBC,0xBD,0xBE,0xDD,0xA8,0xAF,0x5D,0xB4,0xD7, +0x7B,0x41,0x42,0x43,0x44,0x45,0x46,0x47,0x48,0x49,0xAD,0xF4,0xF6,0xF2,0xF3,0xF5, +0x7D,0x4A,0x4B,0x4C,0x4D,0x4E,0x4F,0x50,0x51,0x52,0xB9,0xFB,0xFC,0xF9,0xFA,0xFF, +0x5C,0xF7,0x53,0x54,0x55,0x56,0x57,0x58,0x59,0x5A,0xB2,0xD4,0xD6,0xD2,0xD3,0xD5, +0x30,0x31,0x32,0x33,0x34,0x35,0x36,0x37,0x38,0x39,0xB3,0xDB,0xDC,0xD9,0xDA,0x9F +}; +#endif + +/* + * Copy buf[0 ... nbytes-1] into out[], translating EBCDIC to ASCII. + */ +private void +from_ebcdic(const unsigned char *buf, size_t nbytes, unsigned char *out) +{ + size_t i; + + for (i = 0; i < nbytes; i++) { + out[i] = ebcdic_to_ascii[buf[i]]; + } +} diff --git a/ext/fileinfo/lib/compress.c b/ext/fileinfo/lib/compress.c new file mode 100644 index 0000000000..9c1228ce3f --- /dev/null +++ b/ext/fileinfo/lib/compress.c @@ -0,0 +1,492 @@ +/* + * Copyright (c) Ian F. Darwin 1986-1995. + * Software written by Ian F. Darwin and others; + * maintained 1995-present by Christos Zoulas and others. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice immediately at the beginning of the file, without modification, + * this list of conditions, and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR + * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + */ +/* + * compress routines: + * zmagic() - returns 0 if not recognized, uncompresses and prints + * information if recognized + * uncompress(method, old, n, newch) - uncompress old into new, + * using method, return sizeof new + */ +#include "file.h" +#include "magic.h" +#include +#include +#ifdef HAVE_UNISTD_H +#include +#endif +#include +#include +#include +#include +#ifdef HAVE_SYS_WAIT_H +#include +#endif +#if defined(HAVE_SYS_TIME_H) +#include +#endif +#if defined(HAVE_ZLIB_H) && defined(HAVE_LIBZ) +#define BUILTIN_DECOMPRESS +#include +#endif + + +#ifndef lint +FILE_RCSID("@(#)$File: compress.c,v 1.56 2008/02/07 00:58:52 christos Exp $") +#endif + +private const struct { + const char magic[8]; + size_t maglen; + const char *argv[3]; + int silent; +} compr[] = { + { "\037\235", 2, { "gzip", "-cdq", NULL }, 1 }, /* compressed */ + /* Uncompress can get stuck; so use gzip first if we have it + * Idea from Damien Clark, thanks! */ + { "\037\235", 2, { "uncompress", "-c", NULL }, 1 }, /* compressed */ + { "\037\213", 2, { "gzip", "-cdq", NULL }, 1 }, /* gzipped */ + { "\037\236", 2, { "gzip", "-cdq", NULL }, 1 }, /* frozen */ + { "\037\240", 2, { "gzip", "-cdq", NULL }, 1 }, /* SCO LZH */ + /* the standard pack utilities do not accept standard input */ + { "\037\036", 2, { "gzip", "-cdq", NULL }, 0 }, /* packed */ + { "PK\3\4", 4, { "gzip", "-cdq", NULL }, 1 }, /* pkzipped, */ + /* ...only first file examined */ + { "BZh", 3, { "bzip2", "-cd", NULL }, 1 }, /* bzip2-ed */ +}; + +private size_t ncompr = sizeof(compr) / sizeof(compr[0]); + +#define NODATA ((size_t)~0) + + +private ssize_t swrite(int, const void *, size_t); +private size_t uncompressbuf(struct magic_set *, int, size_t, + const unsigned char *, unsigned char **, size_t); +#ifdef BUILTIN_DECOMPRESS +private size_t uncompressgzipped(struct magic_set *, const unsigned char *, + unsigned char **, size_t); +#endif + +protected int +file_zmagic(struct magic_set *ms, int fd, const char *name, + const unsigned char *buf, size_t nbytes) +{ + unsigned char *newbuf = NULL; + size_t i, nsz; + int rv = 0; + int mime = ms->flags & MAGIC_MIME; + + if ((ms->flags & MAGIC_COMPRESS) == 0) + return 0; + + for (i = 0; i < ncompr; i++) { + if (nbytes < compr[i].maglen) + continue; + if (memcmp(buf, compr[i].magic, compr[i].maglen) == 0 && + (nsz = uncompressbuf(ms, fd, i, buf, &newbuf, + nbytes)) != NODATA) { + ms->flags &= ~MAGIC_COMPRESS; + rv = -1; + if (file_buffer(ms, -1, name, newbuf, nsz) == -1) + goto error; + + if (mime == MAGIC_MIME || mime == 0) { + if (file_printf(ms, mime ? + " compressed-encoding=" : " (") == -1) + goto error; + } + + if ((mime == 0 || mime & MAGIC_MIME_ENCODING) && + file_buffer(ms, -1, NULL, buf, nbytes) == -1) + goto error; + + if (!mime && file_printf(ms, ")") == -1) + goto error; + rv = 1; + break; + } + } +error: + if (newbuf) + free(newbuf); + ms->flags |= MAGIC_COMPRESS; + return rv; +} + +/* + * `safe' write for sockets and pipes. + */ +private ssize_t +swrite(int fd, const void *buf, size_t n) +{ + int rv; + size_t rn = n; + + do + switch (rv = write(fd, buf, n)) { + case -1: + if (errno == EINTR) + continue; + return -1; + default: + n -= rv; + buf = ((const char *)buf) + rv; + break; + } + while (n > 0); + return rn; +} + + +/* + * `safe' read for sockets and pipes. + */ +protected ssize_t +sread(int fd, void *buf, size_t n, int canbepipe) +{ + int rv, cnt; +#ifdef FIONREAD + int t = 0; +#endif + size_t rn = n; + + if (fd == STDIN_FILENO) + goto nocheck; + +#ifdef FIONREAD + if ((canbepipe && (ioctl(fd, FIONREAD, &t) == -1)) || (t == 0)) { +#ifdef FD_ZERO + for (cnt = 0;; cnt++) { + fd_set check; + struct timeval tout = {0, 100 * 1000}; + int selrv; + + FD_ZERO(&check); + FD_SET(fd, &check); + + /* + * Avoid soft deadlock: do not read if there + * is nothing to read from sockets and pipes. + */ + selrv = select(fd + 1, &check, NULL, NULL, &tout); + if (selrv == -1) { + if (errno == EINTR || errno == EAGAIN) + continue; + } else if (selrv == 0 && cnt >= 5) { + return 0; + } else + break; + } +#endif + (void)ioctl(fd, FIONREAD, &t); + } + + if (t > 0 && (size_t)t < n) { + n = t; + rn = n; + } +#endif + +nocheck: + do + switch ((rv = read(fd, buf, n))) { + case -1: + if (errno == EINTR) + continue; + return -1; + case 0: + return rn - n; + default: + n -= rv; + buf = ((char *)buf) + rv; + break; + } + while (n > 0); + return rn; +} + +protected int +file_pipe2file(struct magic_set *ms, int fd, const void *startbuf, + size_t nbytes) +{ + char buf[4096]; + int r, tfd; + + (void)strcpy(buf, "/tmp/file.XXXXXX"); +#ifndef HAVE_MKSTEMP + { + char *ptr = mktemp(buf); + tfd = open(ptr, O_RDWR|O_TRUNC|O_EXCL|O_CREAT, 0600); + r = errno; + (void)unlink(ptr); + errno = r; + } +#else + tfd = mkstemp(buf); + r = errno; + (void)unlink(buf); + errno = r; +#endif + if (tfd == -1) { + file_error(ms, errno, + "cannot create temporary file for pipe copy"); + return -1; + } + + if (swrite(tfd, startbuf, nbytes) != (ssize_t)nbytes) + r = 1; + else { + while ((r = sread(fd, buf, sizeof(buf), 1)) > 0) + if (swrite(tfd, buf, (size_t)r) != r) + break; + } + + switch (r) { + case -1: + file_error(ms, errno, "error copying from pipe to temp file"); + return -1; + case 0: + break; + default: + file_error(ms, errno, "error while writing to temp file"); + return -1; + } + + /* + * We duplicate the file descriptor, because fclose on a + * tmpfile will delete the file, but any open descriptors + * can still access the phantom inode. + */ + if ((fd = dup2(tfd, fd)) == -1) { + file_error(ms, errno, "could not dup descriptor for temp file"); + return -1; + } + (void)close(tfd); + if (lseek(fd, (off_t)0, SEEK_SET) == (off_t)-1) { + file_badseek(ms); + return -1; + } + return fd; +} + +#ifdef BUILTIN_DECOMPRESS + +#define FHCRC (1 << 1) +#define FEXTRA (1 << 2) +#define FNAME (1 << 3) +#define FCOMMENT (1 << 4) + +private size_t +uncompressgzipped(struct magic_set *ms, const unsigned char *old, + unsigned char **newch, size_t n) +{ + unsigned char flg = old[3]; + size_t data_start = 10; + z_stream z; + int rc; + + if (flg & FEXTRA) { + if (data_start+1 >= n) + return 0; + data_start += 2 + old[data_start] + old[data_start + 1] * 256; + } + if (flg & FNAME) { + while(data_start < n && old[data_start]) + data_start++; + data_start++; + } + if(flg & FCOMMENT) { + while(data_start < n && old[data_start]) + data_start++; + data_start++; + } + if(flg & FHCRC) + data_start += 2; + + if (data_start >= n) + return 0; + if ((*newch = (unsigned char *)malloc(HOWMANY + 1)) == NULL) { + return 0; + } + + /* XXX: const castaway, via strchr */ + z.next_in = (Bytef *)strchr((const char *)old + data_start, + old[data_start]); + z.avail_in = n - data_start; + z.next_out = *newch; + z.avail_out = HOWMANY; + z.zalloc = Z_NULL; + z.zfree = Z_NULL; + z.opaque = Z_NULL; + + rc = inflateInit2(&z, -15); + if (rc != Z_OK) { + file_error(ms, 0, "zlib: %s", z.msg); + return 0; + } + + rc = inflate(&z, Z_SYNC_FLUSH); + if (rc != Z_OK && rc != Z_STREAM_END) { + file_error(ms, 0, "zlib: %s", z.msg); + return 0; + } + + n = (size_t)z.total_out; + (void)inflateEnd(&z); + + /* let's keep the nul-terminate tradition */ + (*newch)[n] = '\0'; + + return n; +} +#endif + +private size_t +uncompressbuf(struct magic_set *ms, int fd, size_t method, + const unsigned char *old, unsigned char **newch, size_t n) +{ + int fdin[2], fdout[2]; + int r; + +#ifdef BUILTIN_DECOMPRESS + /* FIXME: This doesn't cope with bzip2 */ + if (method == 2) + return uncompressgzipped(ms, old, newch, n); +#endif + (void)fflush(stdout); + (void)fflush(stderr); + + if ((fd != -1 && pipe(fdin) == -1) || pipe(fdout) == -1) { + file_error(ms, errno, "cannot create pipe"); + return NODATA; + } + switch (fork()) { + case 0: /* child */ + (void) close(0); + if (fd != -1) { + (void) dup(fd); + (void) lseek(0, (off_t)0, SEEK_SET); + } else { + (void) dup(fdin[0]); + (void) close(fdin[0]); + (void) close(fdin[1]); + } + + (void) close(1); + (void) dup(fdout[1]); + (void) close(fdout[0]); + (void) close(fdout[1]); +#ifndef DEBUG + if (compr[method].silent) + (void)close(2); +#endif + + (void)execvp(compr[method].argv[0], + (char *const *)(intptr_t)compr[method].argv); +#ifdef DEBUG + (void)fprintf(stderr, "exec `%s' failed (%s)\n", + compr[method].argv[0], strerror(errno)); +#endif + exit(1); + /*NOTREACHED*/ + case -1: + file_error(ms, errno, "could not fork"); + return NODATA; + + default: /* parent */ + (void) close(fdout[1]); + if (fd == -1) { + (void) close(fdin[0]); + /* + * fork again, to avoid blocking because both + * pipes filled + */ + switch (fork()) { + case 0: /* child */ + (void)close(fdout[0]); + if (swrite(fdin[1], old, n) != (ssize_t)n) { +#ifdef DEBUG + (void)fprintf(stderr, + "Write failed (%s)\n", + strerror(errno)); +#endif + exit(1); + } + exit(0); + /*NOTREACHED*/ + + case -1: +#ifdef DEBUG + (void)fprintf(stderr, "Fork failed (%s)\n", + strerror(errno)); +#endif + exit(1); + /*NOTREACHED*/ + + default: /* parent */ + break; + } + (void) close(fdin[1]); + fdin[1] = -1; + } + + if ((*newch = (unsigned char *) malloc(HOWMANY + 1)) == NULL) { +#ifdef DEBUG + (void)fprintf(stderr, "Malloc failed (%s)\n", + strerror(errno)); +#endif + n = 0; + goto err; + } + if ((r = sread(fdout[0], *newch, HOWMANY, 0)) <= 0) { +#ifdef DEBUG + (void)fprintf(stderr, "Read failed (%s)\n", + strerror(errno)); +#endif + free(*newch); + n = 0; + newch[0] = '\0'; + goto err; + } else { + n = r; + } + /* NUL terminate, as every buffer is handled here. */ + (*newch)[n] = '\0'; +err: + if (fdin[1] != -1) + (void) close(fdin[1]); + (void) close(fdout[0]); +#ifdef WNOHANG + while (waitpid(-1, NULL, WNOHANG) != -1) + continue; +#else + (void)wait(NULL); +#endif + return n; + } +} diff --git a/ext/fileinfo/lib/elfclass.h b/ext/fileinfo/lib/elfclass.h new file mode 100644 index 0000000000..7f6d8adce6 --- /dev/null +++ b/ext/fileinfo/lib/elfclass.h @@ -0,0 +1,67 @@ +/* + * Copyright (c) Christos Zoulas 2008. + * All Rights Reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice immediately at the beginning of the file, without modification, + * this list of conditions, and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR + * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + */ + if (nbytes <= sizeof(elfhdr)) + return 0; + + u.l = 1; + (void)memcpy(&elfhdr, buf, sizeof elfhdr); + swap = (u.c[sizeof(int32_t) - 1] + 1) != elfhdr.e_ident[EI_DATA]; + + type = elf_getu16(swap, elfhdr.e_type); + switch (type) { +#ifdef ELFCORE + case ET_CORE: + if (dophn_core(ms, class, swap, fd, + (off_t)elf_getu(swap, elfhdr.e_phoff), + elf_getu16(swap, elfhdr.e_phnum), + (size_t)elf_getu16(swap, elfhdr.e_phentsize), + fsize, &flags) == -1) + return -1; + break; +#endif + case ET_EXEC: + case ET_DYN: + if (dophn_exec(ms, class, swap, fd, + (off_t)elf_getu(swap, elfhdr.e_phoff), + elf_getu16(swap, elfhdr.e_phnum), + (size_t)elf_getu16(swap, elfhdr.e_phentsize), + fsize, &flags) == -1) + return -1; + /*FALLTHROUGH*/ + case ET_REL: + if (doshn(ms, class, swap, fd, + (off_t)elf_getu(swap, elfhdr.e_shoff), + elf_getu16(swap, elfhdr.e_shnum), + (size_t)elf_getu16(swap, elfhdr.e_shentsize), + &flags) == -1) + return -1; + break; + + default: + break; + } + return 1; diff --git a/ext/fileinfo/lib/file.c b/ext/fileinfo/lib/file.c new file mode 100644 index 0000000000..89d163284b --- /dev/null +++ b/ext/fileinfo/lib/file.c @@ -0,0 +1,481 @@ +/* + * Copyright (c) Ian F. Darwin 1986-1995. + * Software written by Ian F. Darwin and others; + * maintained 1995-present by Christos Zoulas and others. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice immediately at the beginning of the file, without modification, + * this list of conditions, and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR + * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + */ +/* + * file - find type of a file or files - main program. + */ + +#include "file.h" +#include "magic.h" + +#include +#include +#include +#include +#include +#include /* for MAXPATHLEN */ +#include +#ifdef RESTORE_TIME +# if (__COHERENT__ >= 0x420) +# include +# else +# ifdef USE_UTIMES +# include +# else +# include +# endif +# endif +#endif +#ifdef HAVE_UNISTD_H +#include /* for read() */ +#endif +#ifdef HAVE_LOCALE_H +#include +#endif +#ifdef HAVE_WCHAR_H +#include +#endif + +#include +#ifndef HAVE_GETOPT_LONG +int getopt_long(int argc, char * const *argv, const char *optstring, const struct option *longopts, int *longindex); +#endif + +#include /* for byte swapping */ + +#include "patchlevel.h" + +#ifndef lint +FILE_RCSID("@(#)$File: file.c,v 1.119 2008/02/07 00:58:52 christos Exp $") +#endif /* lint */ + + +#ifdef S_IFLNK +#define SYMLINKFLAG "Lh" +#else +#define SYMLINKFLAG "" +#endif + +# define USAGE "Usage: %s [-bcik" SYMLINKFLAG "nNrsvz0] [-e test] [-f namefile] [-F separator] [-m magicfiles] file...\n %s -C -m magicfiles\n" + +#ifndef MAXPATHLEN +#define MAXPATHLEN 512 +#endif + +private int /* Global command-line options */ + bflag = 0, /* brief output format */ + nopad = 0, /* Don't pad output */ + nobuffer = 0, /* Do not buffer stdout */ + nulsep = 0; /* Append '\0' to the separator */ + +private const char *magicfile = 0; /* where the magic is */ +private const char *default_magicfile = MAGIC; +private const char *separator = ":"; /* Default field separator */ + +private char *progname; /* used throughout */ + +private struct magic_set *magic; + +private void unwrap(char *); +private void usage(void); +private void help(void); + +int main(int, char *[]); +private void process(const char *, int); +private void load(const char *, int); + + +/* + * main - parse arguments and handle options + */ +int +main(int argc, char *argv[]) +{ + int c; + size_t i; + int action = 0, didsomefiles = 0, errflg = 0; + int flags = 0; + char *home, *usermagic; + struct stat sb; + static const char hmagic[] = "/.magic"; +#define OPTSTRING "bcCde:f:F:hikLm:nNprsvz0" + int longindex; + static const struct option long_options[] = + { +#define OPT(shortname, longname, opt, doc) \ + {longname, opt, NULL, shortname}, +#define OPT_LONGONLY(longname, opt, doc) \ + {longname, opt, NULL, 0}, +#include "file_opts.h" +#undef OPT +#undef OPT_LONGONLY + {0, 0, NULL, 0} +}; + + static const struct { + const char *name; + int value; + } nv[] = { + { "apptype", MAGIC_NO_CHECK_APPTYPE }, + { "ascii", MAGIC_NO_CHECK_ASCII }, + { "compress", MAGIC_NO_CHECK_COMPRESS }, + { "elf", MAGIC_NO_CHECK_ELF }, + { "soft", MAGIC_NO_CHECK_SOFT }, + { "tar", MAGIC_NO_CHECK_TAR }, + { "tokens", MAGIC_NO_CHECK_TOKENS }, + }; + + /* makes islower etc work for other langs */ + (void)setlocale(LC_CTYPE, ""); + +#ifdef __EMX__ + /* sh-like wildcard expansion! Shouldn't hurt at least ... */ + _wildcard(&argc, &argv); +#endif + + if ((progname = strrchr(argv[0], '/')) != NULL) + progname++; + else + progname = argv[0]; + + magicfile = default_magicfile; + if ((usermagic = getenv("MAGIC")) != NULL) + magicfile = usermagic; + else + if ((home = getenv("HOME")) != NULL) { + if ((usermagic = malloc(strlen(home) + + sizeof(hmagic))) != NULL) { + (void)strcpy(usermagic, home); + (void)strcat(usermagic, hmagic); + if (stat(usermagic, &sb)<0) + free(usermagic); + else + magicfile = usermagic; + } + } + +#ifdef S_IFLNK + flags |= getenv("POSIXLY_CORRECT") ? MAGIC_SYMLINK : 0; +#endif + while ((c = getopt_long(argc, argv, OPTSTRING, long_options, + &longindex)) != -1) + switch (c) { + case 0 : + switch (longindex) { + case 0: + help(); + break; + case 10: + flags |= MAGIC_MIME_TYPE; + break; + case 11: + flags |= MAGIC_MIME_ENCODING; + break; + } + break; + case '0': + nulsep = 1; + break; + case 'b': + bflag++; + break; + case 'c': + action = FILE_CHECK; + break; + case 'C': + action = FILE_COMPILE; + break; + case 'd': + flags |= MAGIC_DEBUG|MAGIC_CHECK; + break; + case 'e': + for (i = 0; i < sizeof(nv) / sizeof(nv[0]); i++) + if (strcmp(nv[i].name, optarg) == 0) + break; + + if (i == sizeof(nv) / sizeof(nv[0])) + errflg++; + else + flags |= nv[i].value; + break; + + case 'f': + if(action) + usage(); + load(magicfile, flags); + unwrap(optarg); + ++didsomefiles; + break; + case 'F': + separator = optarg; + break; + case 'i': + flags |= MAGIC_MIME; + break; + case 'k': + flags |= MAGIC_CONTINUE; + break; + case 'm': + magicfile = optarg; + break; + case 'n': + ++nobuffer; + break; + case 'N': + ++nopad; + break; +#if defined(HAVE_UTIME) || defined(HAVE_UTIMES) + case 'p': + flags |= MAGIC_PRESERVE_ATIME; + break; +#endif + case 'r': + flags |= MAGIC_RAW; + break; + case 's': + flags |= MAGIC_DEVICES; + break; + case 'v': + (void)fprintf(stderr, "%s-%d.%.2d\n", progname, + FILE_VERSION_MAJOR, patchlevel); + (void)fprintf(stderr, "magic file from %s\n", + magicfile); + return 1; + case 'z': + flags |= MAGIC_COMPRESS; + break; +#ifdef S_IFLNK + case 'L': + flags |= MAGIC_SYMLINK; + break; + case 'h': + flags &= ~MAGIC_SYMLINK; + break; +#endif + case '?': + default: + errflg++; + break; + } + + if (errflg) { + usage(); + } + + switch(action) { + case FILE_CHECK: + case FILE_COMPILE: + magic = magic_open(flags|MAGIC_CHECK); + if (magic == NULL) { + (void)fprintf(stderr, "%s: %s\n", progname, + strerror(errno)); + return 1; + } + c = action == FILE_CHECK ? magic_check(magic, magicfile) : + magic_compile(magic, magicfile); + if (c == -1) { + (void)fprintf(stderr, "%s: %s\n", progname, + magic_error(magic)); + return -1; + } + return 0; + default: + load(magicfile, flags); + break; + } + + if (optind == argc) { + if (!didsomefiles) { + usage(); + } + } + else { + size_t j, wid, nw; + for (wid = 0, j = (size_t)optind; j < (size_t)argc; j++) { + nw = file_mbswidth(argv[j]); + if (nw > wid) + wid = nw; + } + /* + * If bflag is only set twice, set it depending on + * number of files [this is undocumented, and subject to change] + */ + if (bflag == 2) { + bflag = optind >= argc - 1; + } + for (; optind < argc; optind++) + process(argv[optind], wid); + } + + c = magic->haderr ? 1 : 0; + magic_close(magic); + return c; +} + + +private void +/*ARGSUSED*/ +load(const char *m, int flags) +{ + if (magic || m == NULL) + return; + magic = magic_open(flags); + if (magic == NULL) { + (void)fprintf(stderr, "%s: %s\n", progname, strerror(errno)); + exit(1); + } + if (magic_load(magic, magicfile) == -1) { + (void)fprintf(stderr, "%s: %s\n", + progname, magic_error(magic)); + exit(1); + } +} + +/* + * unwrap -- read a file of filenames, do each one. + */ +private void +unwrap(char *fn) +{ + char buf[MAXPATHLEN]; + FILE *f; + int wid = 0, cwid; + + if (strcmp("-", fn) == 0) { + f = stdin; + wid = 1; + } else { + if ((f = fopen(fn, "r")) == NULL) { + (void)fprintf(stderr, "%s: Cannot open `%s' (%s).\n", + progname, fn, strerror(errno)); + exit(1); + } + + while (fgets(buf, MAXPATHLEN, f) != NULL) { + buf[strcspn(buf, "\n")] = '\0'; + cwid = file_mbswidth(buf); + if (cwid > wid) + wid = cwid; + } + + rewind(f); + } + + while (fgets(buf, sizeof(buf), f) != NULL) { + buf[strcspn(buf, "\n")] = '\0'; + process(buf, wid); + if(nobuffer) + (void)fflush(stdout); + } + + (void)fclose(f); +} + +/* + * Called for each input file on the command line (or in a list of files) + */ +private void +process(const char *inname, int wid) +{ + const char *type; + int std_in = strcmp(inname, "-") == 0; + + if (wid > 0 && !bflag) { + (void)printf("%s", std_in ? "/dev/stdin" : inname); + if (nulsep) + (void)putc('\0', stdout); + else + (void)printf("%s", separator); + (void)printf("%*s ", + (int) (nopad ? 0 : (wid - file_mbswidth(inname))), ""); + } + + type = magic_file(magic, std_in ? NULL : inname); + if (type == NULL) + (void)printf("ERROR: %s\n", magic_error(magic)); + else + (void)printf("%s\n", type); +} + +size_t +file_mbswidth(const char *s) +{ +#if defined(HAVE_WCHAR_H) && defined(HAVE_MBRTOWC) && defined(HAVE_WCWIDTH) + size_t bytesconsumed, old_n, n, width = 0; + mbstate_t state; + wchar_t nextchar; + (void)memset(&state, 0, sizeof(mbstate_t)); + old_n = n = strlen(s); + + while (n > 0) { + bytesconsumed = mbrtowc(&nextchar, s, n, &state); + if (bytesconsumed == (size_t)(-1) || + bytesconsumed == (size_t)(-2)) { + /* Something went wrong, return something reasonable */ + return old_n; + } + if (s[0] == '\n') { + /* + * do what strlen() would do, so that caller + * is always right + */ + width++; + } else + width += wcwidth(nextchar); + + s += bytesconsumed, n -= bytesconsumed; + } + return width; +#else + return strlen(s); +#endif +} + +private void +usage(void) +{ + (void)fprintf(stderr, USAGE, progname, progname); + (void)fputs("Try `file --help' for more information.\n", stderr); + exit(1); +} + +private void +help(void) +{ + (void)fputs( +"Usage: file [OPTION...] [FILE...]\n" +"Determine type of FILEs.\n" +"\n", stderr); +#define OPT(shortname, longname, opt, doc) \ + fprintf(stderr, " -%c, --" longname doc, shortname); +#define OPT_LONGONLY(longname, opt, doc) \ + fprintf(stderr, " --" longname doc); +#include "file_opts.h" +#undef OPT +#undef OPT_LONGONLY + exit(0); +} diff --git a/ext/fileinfo/lib/file.h b/ext/fileinfo/lib/file.h new file mode 100644 index 0000000000..828189a877 --- /dev/null +++ b/ext/fileinfo/lib/file.h @@ -0,0 +1,392 @@ +/* + * Copyright (c) Ian F. Darwin 1986-1995. + * Software written by Ian F. Darwin and others; + * maintained 1995-present by Christos Zoulas and others. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice immediately at the beginning of the file, without modification, + * this list of conditions, and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR + * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + */ +/* + * file.h - definitions for file(1) program + * @(#)$File: file.h,v 1.103 2008/03/01 22:21:49 rrt Exp $ + */ + +#ifndef __file_h__ +#define __file_h__ + +//#ifdef HAVE_CONFIG_H +#include "config.h" +//#endif */ + +#include /* Include that here, to make sure __P gets defined */ +#include +#include /* For open and flags */ +#ifdef HAVE_STDINT_H +#include +#endif +#ifdef HAVE_INTTYPES_H +#include +#endif +#include +#include +/* Do this here and now, because struct stat gets re-defined on solaris */ +#include +#include + +#define ENABLE_CONDITIONALS + +#ifndef MAGIC +#define MAGIC "/etc/magic" +#endif + +#ifdef __EMX__ +#define PATHSEP ';' +#else +#define PATHSEP ':' +#endif + +#define private static +#ifndef protected +#define protected +#endif +#define public + +#ifndef __GNUC_PREREQ__ +#ifdef __GNUC__ +#define __GNUC_PREREQ__(x, y) \ + ((__GNUC__ == (x) && __GNUC_MINOR__ >= (y)) || \ + (__GNUC__ > (x))) +#else +#define __GNUC_PREREQ__(x, y) 0 +#endif +#endif + +#ifndef MIN +#define MIN(a,b) (((a) < (b)) ? (a) : (b)) +#endif + +#ifndef MAX +#define MAX(a,b) (((a) > (b)) ? (a) : (b)) +#endif + +#ifndef HOWMANY +# define HOWMANY (256 * 1024) /* how much of the file to look at */ +#endif +#define MAXMAGIS 8192 /* max entries in any one magic file + or directory */ +#define MAXDESC 64 /* max leng of text description/MIME type */ +#define MAXstring 32 /* max leng of "string" types */ + +#define MAGICNO 0xF11E041C +#define VERSIONNO 5 +#define FILE_MAGICSIZE (32 * 6) + +#define FILE_LOAD 0 +#define FILE_CHECK 1 +#define FILE_COMPILE 2 + +struct magic { + /* Word 1 */ + uint16_t cont_level; /* level of ">" */ + uint8_t flag; +#define INDIR 0x01 /* if '(...)' appears */ +#define OFFADD 0x02 /* if '>&' or '>...(&' appears */ +#define INDIROFFADD 0x04 /* if '>&(' appears */ +#define UNSIGNED 0x08 /* comparison is unsigned */ +#define NOSPACE 0x10 /* suppress space character before output */ +#define BINTEST 0x20 /* test is for a binary type (set only + for top-level tests) */ +#define TEXTTEST 0 /* for passing to file_softmagic */ + + uint8_t dummy1; + + /* Word 2 */ + uint8_t reln; /* relation (0=eq, '>'=gt, etc) */ + uint8_t vallen; /* length of string value, if any */ + uint8_t type; /* comparison type (FILE_*) */ + uint8_t in_type; /* type of indirection */ +#define FILE_INVALID 0 +#define FILE_BYTE 1 +#define FILE_SHORT 2 +#define FILE_DEFAULT 3 +#define FILE_LONG 4 +#define FILE_STRING 5 +#define FILE_DATE 6 +#define FILE_BESHORT 7 +#define FILE_BELONG 8 +#define FILE_BEDATE 9 +#define FILE_LESHORT 10 +#define FILE_LELONG 11 +#define FILE_LEDATE 12 +#define FILE_PSTRING 13 +#define FILE_LDATE 14 +#define FILE_BELDATE 15 +#define FILE_LELDATE 16 +#define FILE_REGEX 17 +#define FILE_BESTRING16 18 +#define FILE_LESTRING16 19 +#define FILE_SEARCH 20 +#define FILE_MEDATE 21 +#define FILE_MELDATE 22 +#define FILE_MELONG 23 +#define FILE_QUAD 24 +#define FILE_LEQUAD 25 +#define FILE_BEQUAD 26 +#define FILE_QDATE 27 +#define FILE_LEQDATE 28 +#define FILE_BEQDATE 29 +#define FILE_QLDATE 30 +#define FILE_LEQLDATE 31 +#define FILE_BEQLDATE 32 +#define FILE_FLOAT 33 +#define FILE_BEFLOAT 34 +#define FILE_LEFLOAT 35 +#define FILE_DOUBLE 36 +#define FILE_BEDOUBLE 37 +#define FILE_LEDOUBLE 38 +#define FILE_NAMES_SIZE 39/* size of array to contain all names */ + +#define IS_STRING(t) \ + ((t) == FILE_STRING || \ + (t) == FILE_PSTRING || \ + (t) == FILE_BESTRING16 || \ + (t) == FILE_LESTRING16 || \ + (t) == FILE_REGEX || \ + (t) == FILE_SEARCH || \ + (t) == FILE_DEFAULT) + +#define FILE_FMT_NONE 0 +#define FILE_FMT_NUM 1 /* "cduxXi" */ +#define FILE_FMT_STR 2 /* "s" */ +#define FILE_FMT_QUAD 3 /* "ll" */ +#define FILE_FMT_FLOAT 4 /* "eEfFgG" */ +#define FILE_FMT_DOUBLE 5 /* "eEfFgG" */ + + /* Word 3 */ + uint8_t in_op; /* operator for indirection */ + uint8_t mask_op; /* operator for mask */ +#ifdef ENABLE_CONDITIONALS + uint8_t cond; /* conditional type */ + uint8_t dummy2; +#else + uint8_t dummy2; + uint8_t dummy3; +#endif + +#define FILE_OPS "&|^+-*/%" +#define FILE_OPAND 0 +#define FILE_OPOR 1 +#define FILE_OPXOR 2 +#define FILE_OPADD 3 +#define FILE_OPMINUS 4 +#define FILE_OPMULTIPLY 5 +#define FILE_OPDIVIDE 6 +#define FILE_OPMODULO 7 +#define FILE_OPS_MASK 0x07 /* mask for above ops */ +#define FILE_UNUSED_1 0x08 +#define FILE_UNUSED_2 0x10 +#define FILE_UNUSED_3 0x20 +#define FILE_OPINVERSE 0x40 +#define FILE_OPINDIRECT 0x80 + +#ifdef ENABLE_CONDITIONALS +#define COND_NONE 0 +#define COND_IF 1 +#define COND_ELIF 2 +#define COND_ELSE 3 +#endif /* ENABLE_CONDITIONALS */ + + /* Word 4 */ + uint32_t offset; /* offset to magic number */ + /* Word 5 */ + int32_t in_offset; /* offset from indirection */ + /* Word 6 */ + uint32_t lineno; /* line number in magic file */ + /* Word 7,8 */ + union { + uint64_t _mask; /* for use with numeric and date types */ + struct { + uint32_t _count; /* repeat/line count */ + uint32_t _flags; /* modifier flags */ + } _s; /* for use with string types */ + } _u; +#define num_mask _u._mask +#define str_range _u._s._count +#define str_flags _u._s._flags + + /* Words 9-16 */ + union VALUETYPE { + uint8_t b; + uint16_t h; + uint32_t l; + uint64_t q; + uint8_t hs[2]; /* 2 bytes of a fixed-endian "short" */ + uint8_t hl[4]; /* 4 bytes of a fixed-endian "long" */ + uint8_t hq[8]; /* 8 bytes of a fixed-endian "quad" */ + char s[MAXstring]; /* the search string or regex pattern */ + float f; + double d; + } value; /* either number or string */ + /* Words 17..31 */ + char desc[MAXDESC]; /* description */ + /* Words 32..47 */ + char mimetype[MAXDESC]; /* MIME type */ +}; + +#define BIT(A) (1 << (A)) +#define STRING_COMPACT_BLANK BIT(0) +#define STRING_COMPACT_OPTIONAL_BLANK BIT(1) +#define STRING_IGNORE_LOWERCASE BIT(2) +#define STRING_IGNORE_UPPERCASE BIT(3) +#define REGEX_OFFSET_START BIT(4) +#define CHAR_COMPACT_BLANK 'B' +#define CHAR_COMPACT_OPTIONAL_BLANK 'b' +#define CHAR_IGNORE_LOWERCASE 'c' +#define CHAR_IGNORE_UPPERCASE 'C' +#define CHAR_REGEX_OFFSET_START 's' +#define STRING_IGNORE_CASE (STRING_IGNORE_LOWERCASE|STRING_IGNORE_UPPERCASE) +#define STRING_DEFAULT_RANGE 100 + + +/* list of magic entries */ +struct mlist { + struct magic *magic; /* array of magic entries */ + uint32_t nmagic; /* number of entries in array */ + int mapped; /* allocation type: 0 => apprentice_file + * 1 => apprentice_map + malloc + * 2 => apprentice_map + mmap */ + struct mlist *next, *prev; +}; + +struct magic_set { + struct mlist *mlist; + struct cont { + size_t len; + struct level_info { + int32_t off; + int got_match; +#ifdef ENABLE_CONDITIONALS + int last_match; + int last_cond; /* used for error checking by parse() */ +#endif + } *li; + } c; + struct out { + char *buf; /* Accumulation buffer */ + char *pbuf; /* Printable buffer */ + } o; + uint32_t offset; + int error; + int flags; + int haderr; + const char *file; + size_t line; /* current magic line number */ + + /* data for searches */ + struct { + const char *s; /* start of search in original source */ + size_t s_len; /* length of search region */ + size_t offset; /* starting offset in source: XXX - should this be off_t? */ + size_t rm_len; /* match length */ + } search; + + /* FIXME: Make the string dynamically allocated so that e.g. + strings matched in files can be longer than MAXstring */ + union VALUETYPE ms_value; /* either number or string */ +}; + +/* Type for Unicode characters */ +typedef unsigned long unichar; + +struct stat; +protected const char *file_fmttime(uint32_t, int); +protected int file_buffer(struct magic_set *, int, const char *, const void *, + size_t); +protected int file_fsmagic(struct magic_set *, const char *, struct stat *); +protected int file_pipe2file(struct magic_set *, int, const void *, size_t); +protected int file_printf(struct magic_set *, const char *, ...); +protected int file_reset(struct magic_set *); +protected int file_tryelf(struct magic_set *, int, const unsigned char *, + size_t); +protected int file_zmagic(struct magic_set *, int, const char *, + const unsigned char *, size_t); +protected int file_ascmagic(struct magic_set *, const unsigned char *, size_t); +protected int file_is_tar(struct magic_set *, const unsigned char *, size_t); +protected int file_softmagic(struct magic_set *, const unsigned char *, size_t, int); +protected struct mlist *file_apprentice(struct magic_set *, const char *, int); +protected uint64_t file_signextend(struct magic_set *, struct magic *, + uint64_t); +protected void file_delmagic(struct magic *, int type, size_t entries); +protected void file_badread(struct magic_set *); +protected void file_badseek(struct magic_set *); +protected void file_oomem(struct magic_set *, size_t); +protected void file_error(struct magic_set *, int, const char *, ...); +protected void file_magerror(struct magic_set *, const char *, ...); +protected void file_magwarn(struct magic_set *, const char *, ...); +protected void file_mdump(struct magic *); +protected void file_showstr(FILE *, const char *, size_t); +protected size_t file_mbswidth(const char *); +protected const char *file_getbuffer(struct magic_set *); +protected ssize_t sread(int, void *, size_t, int); +protected int file_check_mem(struct magic_set *, unsigned int); +protected int file_looks_utf8(const unsigned char *, size_t, unichar *, size_t *); + +#ifndef COMPILE_ONLY +extern const char *file_names[]; +extern const size_t file_nnames; +#endif + +#ifndef HAVE_STRERROR +extern int sys_nerr; +extern char *sys_errlist[]; +#define strerror(e) \ + (((e) >= 0 && (e) < sys_nerr) ? sys_errlist[(e)] : "Unknown error") +#endif + +#ifndef HAVE_STRTOUL +#define strtoul(a, b, c) strtol(a, b, c) +#endif + +#ifndef HAVE_VASPRINTF +int vasprintf(char **ptr, const char *format_string, va_list vargs); +#endif +#ifndef HAVE_ASPRINTF +int asprintf(char **ptr, const char *format_string, ...); +#endif + +#if defined(HAVE_MMAP) && defined(HAVE_SYS_MMAN_H) && !defined(QUICK) +#define QUICK +#endif + +#ifndef O_BINARY +#define O_BINARY 0 +#endif + +#ifdef __GNUC__ +static const char *rcsid(const char *) __attribute__((__used__)); +#endif +#define FILE_RCSID(id) \ +static const char *rcsid(const char *p) { \ + return rcsid(p = id); \ +} + +#endif /* __file_h__ */ diff --git a/ext/fileinfo/lib/file_opts.h b/ext/fileinfo/lib/file_opts.h new file mode 100644 index 0000000000..46bc08ae4b --- /dev/null +++ b/ext/fileinfo/lib/file_opts.h @@ -0,0 +1,48 @@ +/* + * Table of command-line options + * + * The first column specifies the short name, if any, or 0 if none. + * The second column specifies the long name. + * The third column specifies whether it takes a parameter. + * The fourth column is the documentation. + * + * N.B. The long options' order must correspond to the code in file.c, + * and OPTSTRING must be kept up-to-date with the short options. + * Pay particular attention to the numbers of long-only options in the + * switch statement! + */ + +OPT_LONGONLY("help", 0, " display this help and exit\n") +OPT('v', "version", 0, " output version information and exit\n") +OPT('m', "magic-file", 1, " LIST use LIST as a colon-separated list of magic\n" + " number files\n") +OPT('z', "uncompress", 0, " try to look inside compressed files\n") +OPT('b', "brief", 0, " do not prepend filenames to output lines\n") +OPT('c', "checking-printout", 0, " print the parsed form of the magic file, use in\n" + " conjunction with -m to debug a new magic file\n" + " before installing it\n") +OPT('e', "exclude", 1, " TEST exclude TEST from the list of test to be\n" + " performed for file. Valid tests are:\n" + " ascii, apptype, compress, elf, soft, tar, tokens, troff\n") +OPT('f', "files-from", 1, " FILE read the filenames to be examined from FILE\n") +OPT('F', "separator", 1, " STRING use string as separator instead of `:'\n") +OPT('i', "mime", 0, " output MIME type strings (--mime-type and\n" + " --mime-encoding)\n") +OPT_LONGONLY("mime-type", 0, " output the MIME type\n") +OPT_LONGONLY("mime-encoding", 0, " output the MIME encoding\n") +OPT('k', "keep-going", 0, " don't stop at the first match\n") +#ifdef S_IFLNK +OPT('L', "dereference", 0, " follow symlinks (default)\n") +OPT('h', "no-dereference", 0, " don't follow symlinks\n") +#endif +OPT('n', "no-buffer", 0, " do not buffer output\n") +OPT('N', "no-pad", 0, " do not pad output\n") +OPT('0', "print0", 0, " terminate filenames with ASCII NUL\n") +#if defined(HAVE_UTIME) || defined(HAVE_UTIMES) +OPT('p', "preserve-date", 0, " preserve access times on files\n") +#endif +OPT('r', "raw", 0, " don't translate unprintable chars to \\ooo\n") +OPT('s', "special-files", 0, " treat special (block/char devices) files as\n" + " ordinary ones\n") +OPT('C', "compile", 0, " compile file specified by -m\n") +OPT('d', "debug", 0, " print debugging messages\n") diff --git a/ext/fileinfo/lib/fsmagic.c b/ext/fileinfo/lib/fsmagic.c new file mode 100644 index 0000000000..0e51da96f0 --- /dev/null +++ b/ext/fileinfo/lib/fsmagic.c @@ -0,0 +1,312 @@ +/* + * Copyright (c) Ian F. Darwin 1986-1995. + * Software written by Ian F. Darwin and others; + * maintained 1995-present by Christos Zoulas and others. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice immediately at the beginning of the file, without modification, + * this list of conditions, and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR + * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + */ +/* + * fsmagic - magic based on filesystem info - directory, special files, etc. + */ + +#include "file.h" +#include "magic.h" +#include +#ifdef HAVE_UNISTD_H +#include +#endif +#include +#include +/* Since major is a function on SVR4, we cannot use `ifndef major'. */ +#ifdef MAJOR_IN_MKDEV +# include +# define HAVE_MAJOR +#endif +#ifdef MAJOR_IN_SYSMACROS +# include +# define HAVE_MAJOR +#endif +#ifdef major /* Might be defined in sys/types.h. */ +# define HAVE_MAJOR +#endif + +#ifndef HAVE_MAJOR +# define major(dev) (((dev) >> 8) & 0xff) +# define minor(dev) ((dev) & 0xff) +#endif +#undef HAVE_MAJOR + +#ifndef lint +FILE_RCSID("@(#)$File: fsmagic.c,v 1.50 2008/02/12 17:22:54 rrt Exp $") +#endif /* lint */ + +private int +bad_link(struct magic_set *ms, int err, char *buf) +{ + char *errfmt; + if (err == ELOOP) + errfmt = "symbolic link in a loop"; + else + errfmt = "broken symbolic link to `%s'"; + if (ms->flags & MAGIC_ERROR) { + file_error(ms, err, errfmt, buf); + return -1; + } + if (file_printf(ms, errfmt, buf) == -1) + return -1; + return 1; +} + +protected int +file_fsmagic(struct magic_set *ms, const char *fn, struct stat *sb) +{ + int ret = 0; + int mime = ms->flags & MAGIC_MIME; +#ifdef S_IFLNK + char buf[BUFSIZ+4]; + int nch; + struct stat tstatbuf; +#endif + + if (fn == NULL) + return 0; + + /* + * Fstat is cheaper but fails for files you don't have read perms on. + * On 4.2BSD and similar systems, use lstat() to identify symlinks. + */ +#ifdef S_IFLNK + if ((ms->flags & MAGIC_SYMLINK) == 0) + ret = lstat(fn, sb); + else +#endif + ret = stat(fn, sb); /* don't merge into if; see "ret =" above */ + + if (ret) { + if (ms->flags & MAGIC_ERROR) { + file_error(ms, errno, "cannot stat `%s'", fn); + return -1; + } + if (file_printf(ms, "cannot open `%s' (%s)", + fn, strerror(errno)) == -1) + return -1; + return 1; + } + + if (mime) { + if ((sb->st_mode & S_IFMT) != S_IFREG) { + if ((mime & MAGIC_MIME_TYPE) && + file_printf(ms, "application/x-not-regular-file") + == -1) + return -1; + return 1; + } + } + else { +#ifdef S_ISUID + if (sb->st_mode & S_ISUID) + if (file_printf(ms, "setuid ") == -1) + return -1; +#endif +#ifdef S_ISGID + if (sb->st_mode & S_ISGID) + if (file_printf(ms, "setgid ") == -1) + return -1; +#endif +#ifdef S_ISVTX + if (sb->st_mode & S_ISVTX) + if (file_printf(ms, "sticky ") == -1) + return -1; +#endif + } + + switch (sb->st_mode & S_IFMT) { + case S_IFDIR: + if (file_printf(ms, "directory") == -1) + return -1; + return 1; +#ifdef S_IFCHR + case S_IFCHR: + /* + * If -s has been specified, treat character special files + * like ordinary files. Otherwise, just report that they + * are block special files and go on to the next file. + */ + if ((ms->flags & MAGIC_DEVICES) != 0) + break; +#ifdef HAVE_STAT_ST_RDEV +# ifdef dv_unit + if (file_printf(ms, "character special (%d/%d/%d)", + major(sb->st_rdev), dv_unit(sb->st_rdev), + dv_subunit(sb->st_rdev)) == -1) + return -1; +# else + if (file_printf(ms, "character special (%ld/%ld)", + (long) major(sb->st_rdev), (long) minor(sb->st_rdev)) == -1) + return -1; +# endif +#else + if (file_printf(ms, "character special") == -1) + return -1; +#endif + return 1; +#endif +#ifdef S_IFBLK + case S_IFBLK: + /* + * If -s has been specified, treat block special files + * like ordinary files. Otherwise, just report that they + * are block special files and go on to the next file. + */ + if ((ms->flags & MAGIC_DEVICES) != 0) + break; +#ifdef HAVE_STAT_ST_RDEV +# ifdef dv_unit + if (file_printf(ms, "block special (%d/%d/%d)", + major(sb->st_rdev), dv_unit(sb->st_rdev), + dv_subunit(sb->st_rdev)) == -1) + return -1; +# else + if (file_printf(ms, "block special (%ld/%ld)", + (long)major(sb->st_rdev), (long)minor(sb->st_rdev)) == -1) + return -1; +# endif +#else + if (file_printf(ms, "block special") == -1) + return -1; +#endif + return 1; +#endif + /* TODO add code to handle V7 MUX and Blit MUX files */ +#ifdef S_IFIFO + case S_IFIFO: + if((ms->flags & MAGIC_DEVICES) != 0) + break; + if (file_printf(ms, "fifo (named pipe)") == -1) + return -1; + return 1; +#endif +#ifdef S_IFDOOR + case S_IFDOOR: + if (file_printf(ms, "door") == -1) + return -1; + return 1; +#endif +#ifdef S_IFLNK + case S_IFLNK: + if ((nch = readlink(fn, buf, BUFSIZ-1)) <= 0) { + if (ms->flags & MAGIC_ERROR) { + file_error(ms, errno, "unreadable symlink `%s'", + fn); + return -1; + } + if (file_printf(ms, + "unreadable symlink `%s' (%s)", fn, + strerror(errno)) == -1) + return -1; + return 1; + } + buf[nch] = '\0'; /* readlink(2) does not do this */ + + /* If broken symlink, say so and quit early. */ + if (*buf == '/') { + if (stat(buf, &tstatbuf) < 0) + return bad_link(ms, errno, buf); + } else { + char *tmp; + char buf2[BUFSIZ+BUFSIZ+4]; + + if ((tmp = strrchr(fn, '/')) == NULL) { + tmp = buf; /* in current directory anyway */ + } else { + if (tmp - fn + 1 > BUFSIZ) { + if (ms->flags & MAGIC_ERROR) { + file_error(ms, 0, + "path too long: `%s'", buf); + return -1; + } + if (file_printf(ms, + "path too long: `%s'", fn) == -1) + return -1; + return 1; + } + (void)strcpy(buf2, fn); /* take dir part */ + buf2[tmp - fn + 1] = '\0'; + (void)strcat(buf2, buf); /* plus (rel) link */ + tmp = buf2; + } + if (stat(tmp, &tstatbuf) < 0) + return bad_link(ms, errno, buf); + } + + /* Otherwise, handle it. */ + if ((ms->flags & MAGIC_SYMLINK) != 0) { + const char *p; + ms->flags &= MAGIC_SYMLINK; + p = magic_file(ms, buf); + ms->flags |= MAGIC_SYMLINK; + return p != NULL ? 1 : -1; + } else { /* just print what it points to */ + if (file_printf(ms, "symbolic link to `%s'", + buf) == -1) + return -1; + } + return 1; +#endif +#ifdef S_IFSOCK +#ifndef __COHERENT__ + case S_IFSOCK: + if (file_printf(ms, "socket") == -1) + return -1; + return 1; +#endif +#endif + case S_IFREG: + break; + default: + file_error(ms, 0, "invalid mode 0%o", sb->st_mode); + return -1; + /*NOTREACHED*/ + } + + /* + * regular file, check next possibility + * + * If stat() tells us the file has zero length, report here that + * the file is empty, so we can skip all the work of opening and + * reading the file. + * But if the -s option has been given, we skip this optimization, + * since on some systems, stat() reports zero size for raw disk + * partitions. (If the block special device really has zero length, + * the fact that it is empty will be detected and reported correctly + * when we read the file.) + */ + if ((ms->flags & MAGIC_DEVICES) == 0 && sb->st_size == 0) { + if ((!mime || (mime & MAGIC_MIME_TYPE)) && + file_printf(ms, mime ? "application/x-empty" : + "empty") == -1) + return -1; + return 1; + } + return 0; +} diff --git a/ext/fileinfo/lib/funcs.c b/ext/fileinfo/lib/funcs.c new file mode 100644 index 0000000000..d9e84fa6da --- /dev/null +++ b/ext/fileinfo/lib/funcs.c @@ -0,0 +1,331 @@ +/* + * Copyright (c) Christos Zoulas 2003. + * All Rights Reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice immediately at the beginning of the file, without modification, + * this list of conditions, and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR + * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + */ +#include "file.h" +#include "magic.h" +#include +#include +#include +#include +#if defined(HAVE_WCHAR_H) +#include +#endif +#if defined(HAVE_WCTYPE_H) +#include +#endif + +#ifndef lint +FILE_RCSID("@(#)$File: funcs.c,v 1.39 2008/03/01 22:21:49 rrt Exp $") +#endif /* lint */ + +/* + * Like printf, only we append to a buffer. + */ +protected int +file_printf(struct magic_set *ms, const char *fmt, ...) +{ + va_list ap; + size_t size; + int len; + char *buf, *newstr; + + va_start(ap, fmt); + len = vasprintf(&buf, fmt, ap); + if (len < 0) + goto out; + va_end(ap); + + if (ms->o.buf != NULL) { + len = asprintf(&newstr, "%s%s", ms->o.buf, buf); + free(buf); + if (len < 0) + goto out; + free(ms->o.buf); + buf = newstr; + } + ms->o.buf = buf; + return 0; +out: + file_error(ms, errno, "vasprintf failed"); + return -1; +} + +/* + * error - print best error message possible + */ +/*VARARGS*/ +private void +file_error_core(struct magic_set *ms, int error, const char *f, va_list va, + uint32_t lineno) +{ + /* Only the first error is ok */ + if (ms->haderr) + return; + if (lineno != 0) { + free(ms->o.buf); + ms->o.buf = NULL; + file_printf(ms, "line %u: ", lineno); + } + file_printf(ms, f, va); + if (error > 0) + file_printf(ms, " (%s)", strerror(error)); + ms->haderr++; + ms->error = error; +} + +/*VARARGS*/ +protected void +file_error(struct magic_set *ms, int error, const char *f, ...) +{ + va_list va; + va_start(va, f); + file_error_core(ms, error, f, va, 0); + va_end(va); +} + +/* + * Print an error with magic line number. + */ +/*VARARGS*/ +protected void +file_magerror(struct magic_set *ms, const char *f, ...) +{ + va_list va; + va_start(va, f); + file_error_core(ms, 0, f, va, ms->line); + va_end(va); +} + +protected void +file_oomem(struct magic_set *ms, size_t len) +{ + file_error(ms, errno, "cannot allocate %zu bytes", len); +} + +protected void +file_badseek(struct magic_set *ms) +{ + file_error(ms, errno, "error seeking"); +} + +protected void +file_badread(struct magic_set *ms) +{ + file_error(ms, errno, "error reading"); +} + +#ifndef COMPILE_ONLY +protected int +file_buffer(struct magic_set *ms, int fd, const char *inname, const void *buf, + size_t nb) +{ + int m; + int mime = ms->flags & MAGIC_MIME; + + if (nb == 0) { + if ((!mime || (mime & MAGIC_MIME_TYPE)) && + file_printf(ms, mime ? "application/x-empty" : + "empty") == -1) + return -1; + return 1; + } else if (nb == 1) { + if ((!mime || (mime & MAGIC_MIME_TYPE)) && + file_printf(ms, mime ? "application/octet-stream" : + "very short file (no magic)") == -1) + return -1; + return 1; + } + +#ifdef __EMX__ + if ((ms->flags & MAGIC_NO_CHECK_APPTYPE) == 0 && inname) { + switch (file_os2_apptype(ms, inname, buf, nb)) { + case -1: + return -1; + case 0: + break; + default: + return 1; + } + } +#endif + + /* try compression stuff */ + if ((ms->flags & MAGIC_NO_CHECK_COMPRESS) != 0 || + (m = file_zmagic(ms, fd, inname, buf, nb)) == 0) { + /* Check if we have a tar file */ + if ((ms->flags & MAGIC_NO_CHECK_TAR) != 0 || + (m = file_is_tar(ms, buf, nb)) == 0) { + /* try tests in /etc/magic (or surrogate magic file) */ + if ((ms->flags & MAGIC_NO_CHECK_SOFT) != 0 || + (m = file_softmagic(ms, buf, nb, BINTEST)) == 0) { + /* try known keywords, check whether it is ASCII */ + if ((ms->flags & MAGIC_NO_CHECK_ASCII) != 0 || + (m = file_ascmagic(ms, buf, nb)) == 0) { + /* abandon hope, all ye who remain here */ + if ((!mime || (mime & MAGIC_MIME_TYPE)) && + file_printf(ms, mime ? "application/octet-stream" : + "data") == -1) + return -1; + m = 1; + } + } + } + } +#ifdef BUILTIN_ELF + if ((ms->flags & MAGIC_NO_CHECK_ELF) == 0 && m == 1 && + nb > 5 && fd != -1) { + /* + * We matched something in the file, so this *might* + * be an ELF file, and the file is at least 5 bytes + * long, so if it's an ELF file it has at least one + * byte past the ELF magic number - try extracting + * information from the ELF headers that cannot easily + * be extracted with rules in the magic file. + */ + (void)file_tryelf(ms, fd, buf, nb); + } +#endif + return m; +} +#endif + +protected int +file_reset(struct magic_set *ms) +{ + if (ms->mlist == NULL) { + file_error(ms, 0, "no magic files loaded"); + return -1; + } + ms->o.buf = NULL; + ms->haderr = 0; + ms->error = -1; + return 0; +} + +#define OCTALIFY(n, o) \ + /*LINTED*/ \ + (void)(*(n)++ = '\\', \ + *(n)++ = (((uint32_t)*(o) >> 6) & 3) + '0', \ + *(n)++ = (((uint32_t)*(o) >> 3) & 7) + '0', \ + *(n)++ = (((uint32_t)*(o) >> 0) & 7) + '0', \ + (o)++) + +protected const char * +file_getbuffer(struct magic_set *ms) +{ + char *pbuf, *op, *np; + size_t psize, len; + + if (ms->haderr) + return NULL; + + if (ms->flags & MAGIC_RAW) + return ms->o.buf; + + /* * 4 is for octal representation, + 1 is for NUL */ + len = strlen(ms->o.buf); + if (len > (SIZE_MAX - 1) / 4) { + file_oomem(ms, len); + return NULL; + } + psize = len * 4 + 1; + if ((pbuf = realloc(ms->o.pbuf, psize)) == NULL) { + file_oomem(ms, psize); + return NULL; + } + ms->o.pbuf = pbuf; + +#if defined(HAVE_WCHAR_H) && defined(HAVE_MBRTOWC) && defined(HAVE_WCWIDTH) + { + mbstate_t state; + wchar_t nextchar; + int mb_conv = 1; + size_t bytesconsumed; + char *eop; + (void)memset(&state, 0, sizeof(mbstate_t)); + + np = ms->o.pbuf; + op = ms->o.buf; + eop = op + len; + + while (op < eop) { + bytesconsumed = mbrtowc(&nextchar, op, + (size_t)(eop - op), &state); + if (bytesconsumed == (size_t)(-1) || + bytesconsumed == (size_t)(-2)) { + mb_conv = 0; + break; + } + + if (iswprint(nextchar)) { + (void)memcpy(np, op, bytesconsumed); + op += bytesconsumed; + np += bytesconsumed; + } else { + while (bytesconsumed-- > 0) + OCTALIFY(np, op); + } + } + *np = '\0'; + + /* Parsing succeeded as a multi-byte sequence */ + if (mb_conv != 0) + return ms->o.pbuf; + } +#endif + + for (np = ms->o.pbuf, op = ms->o.buf; *op; op++) { + if (isprint((unsigned char)*op)) { + *np++ = *op; + } else { + OCTALIFY(np, op); + } + } + *np = '\0'; + return ms->o.pbuf; +} + +protected int +file_check_mem(struct magic_set *ms, unsigned int level) +{ + size_t len; + + if (level >= ms->c.len) { + len = (ms->c.len += 20) * sizeof(*ms->c.li); + ms->c.li = (ms->c.li == NULL) ? malloc(len) : + realloc(ms->c.li, len); + if (ms->c.li == NULL) { + file_oomem(ms, len); + return -1; + } + } + ms->c.li[level].got_match = 0; +#ifdef ENABLE_CONDITIONALS + ms->c.li[level].last_match = 0; + ms->c.li[level].last_cond = COND_NONE; +#endif /* ENABLE_CONDITIONALS */ + return 0; +} diff --git a/ext/fileinfo/lib/getopt_long.c b/ext/fileinfo/lib/getopt_long.c new file mode 100644 index 0000000000..e6efd0f584 --- /dev/null +++ b/ext/fileinfo/lib/getopt_long.c @@ -0,0 +1,482 @@ +/* $NetBSD: getopt_long.c,v 1.21.4.1 2008/01/09 01:34:14 matt Exp $ */ + +/*- + * Copyright (c) 2000 The NetBSD Foundation, Inc. + * All rights reserved. + * + * This code is derived from software contributed to The NetBSD Foundation + * by Dieter Baron and Thomas Klausner. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS + * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED + * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR + * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS + * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR + * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF + * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS + * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN + * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) + * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE + * POSSIBILITY OF SUCH DAMAGE. + */ + +#include +#include +#include +#include +#include +#include + +#define REPLACE_GETOPT + +#define _DIAGASSERT assert + +#ifdef REPLACE_GETOPT +#ifdef __weak_alias +__weak_alias(getopt,_getopt) +#endif +int opterr = 1; /* if error message should be printed */ +int optind = 1; /* index into parent argv vector */ +int optopt = '?'; /* character checked for validity */ +int optreset; /* reset getopt */ +char *optarg; /* argument associated with option */ +#elif HAVE_NBTOOL_CONFIG_H && !HAVE_DECL_OPTRESET +static int optreset; +#endif + +#ifdef __weak_alias +__weak_alias(getopt_long,_getopt_long) +#endif + +#define IGNORE_FIRST (*options == '-' || *options == '+') +#define PRINT_ERROR ((opterr) && ((*options != ':') \ + || (IGNORE_FIRST && options[1] != ':'))) +#define IS_POSIXLY_CORRECT (getenv("POSIXLY_CORRECT") != NULL) +#define PERMUTE (!IS_POSIXLY_CORRECT && !IGNORE_FIRST) +/* XXX: GNU ignores PC if *options == '-' */ +#define IN_ORDER (!IS_POSIXLY_CORRECT && *options == '-') + +/* return values */ +#define BADCH (int)'?' +#define BADARG ((IGNORE_FIRST && options[1] == ':') \ + || (*options == ':') ? (int)':' : (int)'?') +#define INORDER (int)1 + +#define EMSG "" + +static int getopt_internal __P((int, char **, const char *)); +static int gcd __P((int, int)); +static void permute_args __P((int, int, int, char **)); + +static const char *place = EMSG; /* option letter processing */ + +/* XXX: set optreset to 1 rather than these two */ +static int nonopt_start = -1; /* first non option argument (for permute) */ +static int nonopt_end = -1; /* first option after non options (for permute) */ + +/* Error messages */ +static const char recargchar[] = "option requires an argument -- %c"; +static const char recargstring[] = "option requires an argument -- %s"; +static const char ambig[] = "ambiguous option -- %.*s"; +static const char noarg[] = "option doesn't take an argument -- %.*s"; +static const char illoptchar[] = "unknown option -- %c"; +static const char illoptstring[] = "unknown option -- %s"; + + +/* + * Compute the greatest common divisor of a and b. + */ +static int +gcd(a, b) + int a; + int b; +{ + int c; + + c = a % b; + while (c != 0) { + a = b; + b = c; + c = a % b; + } + + return b; +} + +/* + * Exchange the block from nonopt_start to nonopt_end with the block + * from nonopt_end to opt_end (keeping the same order of arguments + * in each block). + */ +static void +permute_args(panonopt_start, panonopt_end, opt_end, nargv) + int panonopt_start; + int panonopt_end; + int opt_end; + char **nargv; +{ + int cstart, cyclelen, i, j, ncycle, nnonopts, nopts, pos; + char *swap; + + _DIAGASSERT(nargv != NULL); + + /* + * compute lengths of blocks and number and size of cycles + */ + nnonopts = panonopt_end - panonopt_start; + nopts = opt_end - panonopt_end; + ncycle = gcd(nnonopts, nopts); + cyclelen = (opt_end - panonopt_start) / ncycle; + + for (i = 0; i < ncycle; i++) { + cstart = panonopt_end+i; + pos = cstart; + for (j = 0; j < cyclelen; j++) { + if (pos >= panonopt_end) + pos -= nnonopts; + else + pos += nopts; + swap = nargv[pos]; + nargv[pos] = nargv[cstart]; + nargv[cstart] = swap; + } + } +} + +/* + * getopt_internal -- + * Parse argc/argv argument vector. Called by user level routines. + * Returns -2 if -- is found (can be long option or end of options marker). + */ +static int +getopt_internal(nargc, nargv, options) + int nargc; + char **nargv; + const char *options; +{ + char *oli; /* option letter list index */ + int optchar; + + _DIAGASSERT(nargv != NULL); + _DIAGASSERT(options != NULL); + + optarg = NULL; + + /* + * XXX Some programs (like rsyncd) expect to be able to + * XXX re-initialize optind to 0 and have getopt_long(3) + * XXX properly function again. Work around this braindamage. + */ + if (optind == 0) + optind = 1; + + if (optreset) + nonopt_start = nonopt_end = -1; +start: + if (optreset || !*place) { /* update scanning pointer */ + optreset = 0; + if (optind >= nargc) { /* end of argument vector */ + place = EMSG; + if (nonopt_end != -1) { + /* do permutation, if we have to */ + permute_args(nonopt_start, nonopt_end, + optind, nargv); + optind -= nonopt_end - nonopt_start; + } + else if (nonopt_start != -1) { + /* + * If we skipped non-options, set optind + * to the first of them. + */ + optind = nonopt_start; + } + nonopt_start = nonopt_end = -1; + return -1; + } + if ((*(place = nargv[optind]) != '-') + || (place[1] == '\0')) { /* found non-option */ + place = EMSG; + if (IN_ORDER) { + /* + * GNU extension: + * return non-option as argument to option 1 + */ + optarg = nargv[optind++]; + return INORDER; + } + if (!PERMUTE) { + /* + * if no permutation wanted, stop parsing + * at first non-option + */ + return -1; + } + /* do permutation */ + if (nonopt_start == -1) + nonopt_start = optind; + else if (nonopt_end != -1) { + permute_args(nonopt_start, nonopt_end, + optind, nargv); + nonopt_start = optind - + (nonopt_end - nonopt_start); + nonopt_end = -1; + } + optind++; + /* process next argument */ + goto start; + } + if (nonopt_start != -1 && nonopt_end == -1) + nonopt_end = optind; + if (place[1] && *++place == '-') { /* found "--" */ + place++; + return -2; + } + } + if ((optchar = (int)*place++) == (int)':' || + (oli = strchr(options + (IGNORE_FIRST ? 1 : 0), optchar)) == NULL) { + /* option letter unknown or ':' */ + if (!*place) + ++optind; + if (PRINT_ERROR) + warnx(illoptchar, optchar); + optopt = optchar; + return BADCH; + } + if (optchar == 'W' && oli[1] == ';') { /* -W long-option */ + /* XXX: what if no long options provided (called by getopt)? */ + if (*place) + return -2; + + if (++optind >= nargc) { /* no arg */ + place = EMSG; + if (PRINT_ERROR) + warnx(recargchar, optchar); + optopt = optchar; + return BADARG; + } else /* white space */ + place = nargv[optind]; + /* + * Handle -W arg the same as --arg (which causes getopt to + * stop parsing). + */ + return -2; + } + if (*++oli != ':') { /* doesn't take argument */ + if (!*place) + ++optind; + } else { /* takes (optional) argument */ + optarg = NULL; + if (*place) /* no white space */ + optarg = (char *)place; + /* XXX: disable test for :: if PC? (GNU doesn't) */ + else if (oli[1] != ':') { /* arg not optional */ + if (++optind >= nargc) { /* no arg */ + place = EMSG; + if (PRINT_ERROR) + warnx(recargchar, optchar); + optopt = optchar; + return BADARG; + } else + optarg = nargv[optind]; + } + place = EMSG; + ++optind; + } + /* dump back option letter */ + return optchar; +} + +#ifdef REPLACE_GETOPT +/* + * getopt -- + * Parse argc/argv argument vector. + * + * [eventually this will replace the real getopt] + */ +int +getopt(nargc, nargv, options) + int nargc; + char * const *nargv; + const char *options; +{ + int retval; + + _DIAGASSERT(nargv != NULL); + _DIAGASSERT(options != NULL); + + retval = getopt_internal(nargc, (char **)nargv, options); + if (retval == -2) { + ++optind; + /* + * We found an option (--), so if we skipped non-options, + * we have to permute. + */ + if (nonopt_end != -1) { + permute_args(nonopt_start, nonopt_end, optind, + (char **)nargv); + optind -= nonopt_end - nonopt_start; + } + nonopt_start = nonopt_end = -1; + retval = -1; + } + return retval; +} +#endif + +/* + * getopt_long -- + * Parse argc/argv argument vector. + */ +int +getopt_long(nargc, nargv, options, long_options, idx) + int nargc; + char * const *nargv; + const char *options; + const struct option *long_options; + int *idx; +{ + int retval; + +#define IDENTICAL_INTERPRETATION(_x, _y) \ + (long_options[(_x)].has_arg == long_options[(_y)].has_arg && \ + long_options[(_x)].flag == long_options[(_y)].flag && \ + long_options[(_x)].val == long_options[(_y)].val) + + _DIAGASSERT(nargv != NULL); + _DIAGASSERT(options != NULL); + _DIAGASSERT(long_options != NULL); + /* idx may be NULL */ + + retval = getopt_internal(nargc, (char **)nargv, options); + if (retval == -2) { + char *current_argv, *has_equal; + size_t current_argv_len; + int i, ambiguous, match; + + current_argv = (char *)place; + match = -1; + ambiguous = 0; + + optind++; + place = EMSG; + + if (*current_argv == '\0') { /* found "--" */ + /* + * We found an option (--), so if we skipped + * non-options, we have to permute. + */ + if (nonopt_end != -1) { + permute_args(nonopt_start, nonopt_end, + optind, (char **)nargv); + optind -= nonopt_end - nonopt_start; + } + nonopt_start = nonopt_end = -1; + return -1; + } + if ((has_equal = strchr(current_argv, '=')) != NULL) { + /* argument found (--option=arg) */ + current_argv_len = has_equal - current_argv; + has_equal++; + } else + current_argv_len = strlen(current_argv); + + for (i = 0; long_options[i].name; i++) { + /* find matching long option */ + if (strncmp(current_argv, long_options[i].name, + current_argv_len)) + continue; + + if (strlen(long_options[i].name) == + (unsigned)current_argv_len) { + /* exact match */ + match = i; + ambiguous = 0; + break; + } + if (match == -1) /* partial match */ + match = i; + else if (!IDENTICAL_INTERPRETATION(i, match)) + ambiguous = 1; + } + if (ambiguous) { + /* ambiguous abbreviation */ + if (PRINT_ERROR) + warnx(ambig, (int)current_argv_len, + current_argv); + optopt = 0; + return BADCH; + } + if (match != -1) { /* option found */ + if (long_options[match].has_arg == no_argument + && has_equal) { + if (PRINT_ERROR) + warnx(noarg, (int)current_argv_len, + current_argv); + /* + * XXX: GNU sets optopt to val regardless of + * flag + */ + if (long_options[match].flag == NULL) + optopt = long_options[match].val; + else + optopt = 0; + return BADARG; + } + if (long_options[match].has_arg == required_argument || + long_options[match].has_arg == optional_argument) { + if (has_equal) + optarg = has_equal; + else if (long_options[match].has_arg == + required_argument) { + /* + * optional argument doesn't use + * next nargv + */ + optarg = nargv[optind++]; + } + } + if ((long_options[match].has_arg == required_argument) + && (optarg == NULL)) { + /* + * Missing argument; leading ':' + * indicates no error should be generated + */ + if (PRINT_ERROR) + warnx(recargstring, current_argv); + /* + * XXX: GNU sets optopt to val regardless + * of flag + */ + if (long_options[match].flag == NULL) + optopt = long_options[match].val; + else + optopt = 0; + --optind; + return BADARG; + } + } else { /* unknown option */ + if (PRINT_ERROR) + warnx(illoptstring, current_argv); + optopt = 0; + return BADCH; + } + if (long_options[match].flag) { + *long_options[match].flag = long_options[match].val; + retval = 0; + } else + retval = long_options[match].val; + if (idx) + *idx = match; + } + return retval; +#undef IDENTICAL_INTERPRETATION +} diff --git a/ext/fileinfo/lib/is_tar.c b/ext/fileinfo/lib/is_tar.c new file mode 100644 index 0000000000..a9311117ed --- /dev/null +++ b/ext/fileinfo/lib/is_tar.c @@ -0,0 +1,156 @@ +/* + * Copyright (c) Ian F. Darwin 1986-1995. + * Software written by Ian F. Darwin and others; + * maintained 1995-present by Christos Zoulas and others. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice immediately at the beginning of the file, without modification, + * this list of conditions, and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR + * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + */ +/* + * is_tar() -- figure out whether file is a tar archive. + * + * Stolen (by the author!) from the public domain tar program: + * Public Domain version written 26 Aug 1985 John Gilmore (ihnp4!hoptoad!gnu). + * + * @(#)list.c 1.18 9/23/86 Public Domain - gnu + * + * Comments changed and some code/comments reformatted + * for file command by Ian Darwin. + */ + +#include "file.h" +#include "magic.h" +#include +#include +#include +#include "tar.h" + +#ifndef lint +FILE_RCSID("@(#)$File: is_tar.c,v 1.31 2008/02/04 20:51:17 christos Exp $") +#endif + +#define isodigit(c) ( ((c) >= '0') && ((c) <= '7') ) + +private int is_tar(const unsigned char *, size_t); +private int from_oct(int, const char *); /* Decode octal number */ + +static const char tartype[][32] = { + "tar archive", + "POSIX tar archive", + "POSIX tar archive (GNU)", +}; + +protected int +file_is_tar(struct magic_set *ms, const unsigned char *buf, size_t nbytes) +{ + /* + * Do the tar test first, because if the first file in the tar + * archive starts with a dot, we can confuse it with an nroff file. + */ + int tar = is_tar(buf, nbytes); + int mime = ms->flags & MAGIC_MIME; + + if (tar < 1 || tar > 3) + return 0; + + if (mime == MAGIC_MIME_ENCODING) + return 0; + + if (file_printf(ms, mime ? "application/x-tar" : + tartype[tar - 1]) == -1) + return -1; + return 1; +} + +/* + * Return + * 0 if the checksum is bad (i.e., probably not a tar archive), + * 1 for old UNIX tar file, + * 2 for Unix Std (POSIX) tar file, + * 3 for GNU tar file. + */ +private int +is_tar(const unsigned char *buf, size_t nbytes) +{ + const union record *header = (const union record *)(const void *)buf; + int i; + int sum, recsum; + const char *p; + + if (nbytes < sizeof(union record)) + return 0; + + recsum = from_oct(8, header->header.chksum); + + sum = 0; + p = header->charptr; + for (i = sizeof(union record); --i >= 0;) { + /* + * We cannot use unsigned char here because of old compilers, + * e.g. V7. + */ + sum += 0xFF & *p++; + } + + /* Adjust checksum to count the "chksum" field as blanks. */ + for (i = sizeof(header->header.chksum); --i >= 0;) + sum -= 0xFF & header->header.chksum[i]; + sum += ' '* sizeof header->header.chksum; + + if (sum != recsum) + return 0; /* Not a tar archive */ + + if (strcmp(header->header.magic, GNUTMAGIC) == 0) + return 3; /* GNU Unix Standard tar archive */ + if (strcmp(header->header.magic, TMAGIC) == 0) + return 2; /* Unix Standard tar archive */ + + return 1; /* Old fashioned tar archive */ +} + + +/* + * Quick and dirty octal conversion. + * + * Result is -1 if the field is invalid (all blank, or nonoctal). + */ +private int +from_oct(int digs, const char *where) +{ + int value; + + while (isspace((unsigned char)*where)) { /* Skip spaces */ + where++; + if (--digs <= 0) + return -1; /* All blank field */ + } + value = 0; + while (digs > 0 && isodigit(*where)) { /* Scan til nonoctal */ + value = (value << 3) | (*where++ - '0'); + --digs; + } + + if (digs > 0 && *where && !isspace((unsigned char)*where)) + return -1; /* Ended on non-space/nul */ + + return value; +} diff --git a/ext/fileinfo/lib/magic.c b/ext/fileinfo/lib/magic.c new file mode 100644 index 0000000000..12c38f260d --- /dev/null +++ b/ext/fileinfo/lib/magic.c @@ -0,0 +1,397 @@ +/* + * Copyright (c) Christos Zoulas 2003. + * All Rights Reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice immediately at the beginning of the file, without modification, + * this list of conditions, and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR + * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + */ + +#include "file.h" +#include "magic.h" + +#include +#include +#include +#include +#include +#include /* for MAXPATHLEN */ +#include +#ifdef QUICK +#include +#endif +#include /* for PIPE_BUF */ + +#if defined(HAVE_UTIMES) +# include +#elif defined(HAVE_UTIME) +# if defined(HAVE_SYS_UTIME_H) +# include +# elif defined(HAVE_UTIME_H) +# include +# endif +#endif + +#ifdef HAVE_UNISTD_H +#include /* for read() */ +#endif + +#ifdef HAVE_LOCALE_H +#include +#endif + +#include /* for byte swapping */ + +#include "patchlevel.h" + +#ifndef lint +FILE_RCSID("@(#)$File: magic.c,v 1.50 2008/02/19 00:58:59 rrt Exp $") +#endif /* lint */ + +#ifndef PIPE_BUF +/* Get the PIPE_BUF from pathconf */ +#ifdef _PC_PIPE_BUF +#define PIPE_BUF pathconf(".", _PC_PIPE_BUF) +#else +#define PIPE_BUF 512 +#endif +#endif + +#ifdef __EMX__ +private char *apptypeName = NULL; +protected int file_os2_apptype(struct magic_set *ms, const char *fn, + const void *buf, size_t nb); +#endif /* __EMX__ */ + +private void free_mlist(struct mlist *); +private void close_and_restore(const struct magic_set *, const char *, int, + const struct stat *); +private int info_from_stat(struct magic_set *, mode_t); +#ifndef COMPILE_ONLY +private const char *file_or_fd(struct magic_set *, const char *, int); +#endif + +#ifndef STDIN_FILENO +#define STDIN_FILENO 0 +#endif + +public struct magic_set * +magic_open(int flags) +{ + struct magic_set *ms; + + if ((ms = calloc((size_t)1, sizeof(struct magic_set))) == NULL) + return NULL; + + if (magic_setflags(ms, flags) == -1) { + errno = EINVAL; + goto free; + } + + ms->o.buf = ms->o.pbuf = NULL; + + ms->c.li = malloc((ms->c.len = 10) * sizeof(*ms->c.li)); + if (ms->c.li == NULL) + goto free; + + ms->haderr = 0; + ms->error = -1; + ms->mlist = NULL; + ms->file = "unknown"; + ms->line = 0; + return ms; +free: + free(ms); + return NULL; +} + +private void +free_mlist(struct mlist *mlist) +{ + struct mlist *ml; + + if (mlist == NULL) + return; + + for (ml = mlist->next; ml != mlist;) { + struct mlist *next = ml->next; + struct magic *mg = ml->magic; + file_delmagic(mg, ml->mapped, ml->nmagic); + free(ml); + ml = next; + } + free(ml); +} + +private int +info_from_stat(struct magic_set *ms, mode_t md) +{ + /* We cannot open it, but we were able to stat it. */ + if (md & 0222) + if (file_printf(ms, "writable, ") == -1) + return -1; + if (md & 0111) + if (file_printf(ms, "executable, ") == -1) + return -1; + if (S_ISREG(md)) + if (file_printf(ms, "regular file, ") == -1) + return -1; + if (file_printf(ms, "no read permission") == -1) + return -1; + return 0; +} + +public void +magic_close(struct magic_set *ms) +{ + free_mlist(ms->mlist); + free(ms->o.pbuf); + free(ms->o.buf); + free(ms->c.li); + free(ms); +} + +/* + * load a magic file + */ +public int +magic_load(struct magic_set *ms, const char *magicfile) +{ + struct mlist *ml = file_apprentice(ms, magicfile, FILE_LOAD); + if (ml) { + free_mlist(ms->mlist); + ms->mlist = ml; + return 0; + } + return -1; +} + +public int +magic_compile(struct magic_set *ms, const char *magicfile) +{ + struct mlist *ml = file_apprentice(ms, magicfile, FILE_COMPILE); + free_mlist(ml); + return ml ? 0 : -1; +} + +public int +magic_check(struct magic_set *ms, const char *magicfile) +{ + struct mlist *ml = file_apprentice(ms, magicfile, FILE_CHECK); + free_mlist(ml); + return ml ? 0 : -1; +} + +private void +close_and_restore(const struct magic_set *ms, const char *name, int fd, + const struct stat *sb) +{ + if (fd == STDIN_FILENO) + return; + (void) close(fd); + + if ((ms->flags & MAGIC_PRESERVE_ATIME) != 0) { + /* + * Try to restore access, modification times if read it. + * This is really *bad* because it will modify the status + * time of the file... And of course this will affect + * backup programs + */ +#ifdef HAVE_UTIMES + struct timeval utsbuf[2]; + (void)memset(utsbuf, 0, sizeof(utsbuf)); + utsbuf[0].tv_sec = sb->st_atime; + utsbuf[1].tv_sec = sb->st_mtime; + + (void) utimes(name, utsbuf); /* don't care if loses */ +#elif defined(HAVE_UTIME_H) || defined(HAVE_SYS_UTIME_H) + struct utimbuf utbuf; + + (void)memset(utbuf, 0, sizeof(utbuf)); + utbuf.actime = sb->st_atime; + utbuf.modtime = sb->st_mtime; + (void) utime(name, &utbuf); /* don't care if loses */ +#endif + } +} + +#ifndef COMPILE_ONLY + +/* + * find type of descriptor + */ +public const char * +magic_descriptor(struct magic_set *ms, int fd) +{ + return file_or_fd(ms, NULL, fd); +} + +/* + * find type of named file + */ +public const char * +magic_file(struct magic_set *ms, const char *inname) +{ + return file_or_fd(ms, inname, STDIN_FILENO); +} + +private const char * +file_or_fd(struct magic_set *ms, const char *inname, int fd) +{ + int rv = -1; + unsigned char *buf; + struct stat sb; + ssize_t nbytes = 0; /* number of bytes read from a datafile */ + int ispipe = 0; + + /* + * one extra for terminating '\0', and + * some overlapping space for matches near EOF + */ +#define SLOP (1 + sizeof(union VALUETYPE)) + if ((buf = malloc(HOWMANY + SLOP)) == NULL) + return NULL; + + if (file_reset(ms) == -1) + goto done; + + switch (file_fsmagic(ms, inname, &sb)) { + case -1: /* error */ + goto done; + case 0: /* nothing found */ + break; + default: /* matched it and printed type */ + rv = 0; + goto done; + } + + if (inname == NULL) { + if (fstat(fd, &sb) == 0 && S_ISFIFO(sb.st_mode)) + ispipe = 1; + } else { + int flags = O_RDONLY|O_BINARY; + + if (stat(inname, &sb) == 0 && S_ISFIFO(sb.st_mode)) { + flags |= O_NONBLOCK; + ispipe = 1; + } + + errno = 0; + if ((fd = open(inname, flags)) < 0) { +#ifdef __CYGWIN__ + /* FIXME: Do this with EXEEXT from autotools */ + char *tmp = alloca(strlen(inname) + 5); + (void)strcat(strcpy(tmp, inname), ".exe"); + if ((fd = open(tmp, flags)) < 0) { +#endif + fprintf(stderr, "couldn't open file\n"); + if (info_from_stat(ms, sb.st_mode) == -1) + goto done; + rv = 0; + goto done; +#ifdef __CYGWIN__ + } +#endif + } +#ifdef O_NONBLOCK + if ((flags = fcntl(fd, F_GETFL)) != -1) { + flags &= ~O_NONBLOCK; + (void)fcntl(fd, F_SETFL, flags); + } +#endif + } + + /* + * try looking at the first HOWMANY bytes + */ + if (ispipe) { + ssize_t r = 0; + + while ((r = sread(fd, (void *)&buf[nbytes], + (size_t)(HOWMANY - nbytes), 1)) > 0) { + nbytes += r; + if (r < PIPE_BUF) break; + } + + if (nbytes == 0) { + /* We can not read it, but we were able to stat it. */ + if (info_from_stat(ms, sb.st_mode) == -1) + goto done; + rv = 0; + goto done; + } + + } else { + if ((nbytes = read(fd, (char *)buf, HOWMANY)) == -1) { + file_error(ms, errno, "cannot read `%s'", inname); + goto done; + } + } + + (void)memset(buf + nbytes, 0, SLOP); /* NUL terminate */ + if (file_buffer(ms, fd, inname, buf, (size_t)nbytes) == -1) + goto done; + rv = 0; +done: + free(buf); + close_and_restore(ms, inname, fd, &sb); + return rv == 0 ? file_getbuffer(ms) : NULL; +} + + +public const char * +magic_buffer(struct magic_set *ms, const void *buf, size_t nb) +{ + if (file_reset(ms) == -1) + return NULL; + /* + * The main work is done here! + * We have the file name and/or the data buffer to be identified. + */ + if (file_buffer(ms, -1, NULL, buf, nb) == -1) { + return NULL; + } + return file_getbuffer(ms); +} +#endif + +public const char * +magic_error(struct magic_set *ms) +{ + return ms->haderr ? ms->o.buf : NULL; +} + +public int +magic_errno(struct magic_set *ms) +{ + return ms->haderr ? ms->error : 0; +} + +public int +magic_setflags(struct magic_set *ms, int flags) +{ +#if !defined(HAVE_UTIME) && !defined(HAVE_UTIMES) + if (flags & MAGIC_PRESERVE_ATIME) + return -1; +#endif + ms->flags = flags; + return 0; +} diff --git a/ext/fileinfo/lib/magic.h b/ext/fileinfo/lib/magic.h new file mode 100644 index 0000000000..ecdd53c1f9 --- /dev/null +++ b/ext/fileinfo/lib/magic.h @@ -0,0 +1,82 @@ +/* + * Copyright (c) Christos Zoulas 2003. + * All Rights Reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice immediately at the beginning of the file, without modification, + * this list of conditions, and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR + * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + */ +#ifndef _MAGIC_H +#define _MAGIC_H + +#include + +#define MAGIC_NONE 0x000000 /* No flags */ +#define MAGIC_DEBUG 0x000001 /* Turn on debugging */ +#define MAGIC_SYMLINK 0x000002 /* Follow symlinks */ +#define MAGIC_COMPRESS 0x000004 /* Check inside compressed files */ +#define MAGIC_DEVICES 0x000008 /* Look at the contents of devices */ +#define MAGIC_MIME_TYPE 0x000010 /* Return only the MIME type */ +#define MAGIC_CONTINUE 0x000020 /* Return all matches */ +#define MAGIC_CHECK 0x000040 /* Print warnings to stderr */ +#define MAGIC_PRESERVE_ATIME 0x000080 /* Restore access time on exit */ +#define MAGIC_RAW 0x000100 /* Don't translate unprint chars */ +#define MAGIC_ERROR 0x000200 /* Handle ENOENT etc as real errors */ +#define MAGIC_MIME_ENCODING 0x000400 /* Return only the MIME encoding */ +#define MAGIC_MIME (MAGIC_MIME_TYPE|MAGIC_MIME_ENCODING) +#define MAGIC_NO_CHECK_COMPRESS 0x001000 /* Don't check for compressed files */ +#define MAGIC_NO_CHECK_TAR 0x002000 /* Don't check for tar files */ +#define MAGIC_NO_CHECK_SOFT 0x004000 /* Don't check magic entries */ +#define MAGIC_NO_CHECK_APPTYPE 0x008000 /* Don't check application type */ +#define MAGIC_NO_CHECK_ELF 0x010000 /* Don't check for elf details */ +#define MAGIC_NO_CHECK_ASCII 0x020000 /* Don't check for ascii files */ +#define MAGIC_NO_CHECK_TOKENS 0x100000 /* Don't check ascii/tokens */ + +/* Defined for backwards compatibility; do nothing */ +#define MAGIC_NO_CHECK_FORTRAN 0x000000 /* Don't check ascii/fortran */ +#define MAGIC_NO_CHECK_TROFF 0x000000 /* Don't check ascii/troff */ + + +#ifdef __cplusplus +extern "C" { +#endif + +typedef struct magic_set *magic_t; +magic_t magic_open(int); +void magic_close(magic_t); + +const char *magic_file(magic_t, const char *); +const char *magic_descriptor(magic_t, int); +const char *magic_buffer(magic_t, const void *, size_t); + +const char *magic_error(magic_t); +int magic_setflags(magic_t, int); + +int magic_load(magic_t, const char *); +int magic_compile(magic_t, const char *); +int magic_check(magic_t, const char *); +int magic_errno(magic_t); + +#ifdef __cplusplus +}; +#endif + +#endif /* _MAGIC_H */ diff --git a/ext/fileinfo/lib/names.h b/ext/fileinfo/lib/names.h new file mode 100644 index 0000000000..2682edcc3a --- /dev/null +++ b/ext/fileinfo/lib/names.h @@ -0,0 +1,173 @@ +/* + * Copyright (c) Ian F. Darwin 1986-1995. + * Software written by Ian F. Darwin and others; + * maintained 1995-present by Christos Zoulas and others. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice immediately at the beginning of the file, without modification, + * this list of conditions, and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR + * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + */ +/* + * Names.h - names and types used by ascmagic in file(1). + * These tokens are here because they can appear anywhere in + * the first HOWMANY bytes, while tokens in MAGIC must + * appear at fixed offsets into the file. Don't make HOWMANY + * too high unless you have a very fast CPU. + * + * $File: names.h,v 1.32 2008/02/11 00:19:29 rrt Exp $ + */ + +/* + modified by Chris Lowth - 9 April 2000 + to add mime type strings to the types table. +*/ + +/* these types are used to index the table 'types': keep em in sync! */ +#define L_C 0 /* first and foremost on UNIX */ +#define L_CC 1 /* Bjarne's postincrement */ +#define L_MAKE 2 /* Makefiles */ +#define L_PLI 3 /* PL/1 */ +#define L_MACH 4 /* some kinda assembler */ +#define L_ENG 5 /* English */ +#define L_PAS 6 /* Pascal */ +#define L_MAIL 7 /* Electronic mail */ +#define L_NEWS 8 /* Usenet Netnews */ +#define L_JAVA 9 /* Java code */ +#define L_HTML 10 /* HTML */ +#define L_BCPL 11 /* BCPL */ +#define L_M4 12 /* M4 */ +#define L_PO 13 /* PO */ + +static const struct { + char human[48]; + char mime[16]; +} types[] = { + { "C program", "text/x-c", }, + { "C++ program", "text/x-c++" }, + { "make commands", "text/x-makefile" }, + { "PL/1 program", "text/x-pl1" }, + { "assembler program", "text/x-asm" }, + { "English", "text/plain" }, + { "Pascal program", "text/x-pascal" }, + { "mail", "text/x-mail" }, + { "news", "text/x-news" }, + { "Java program", "text/x-java" }, + { "HTML document", "text/html", }, + { "BCPL program", "text/x-bcpl" }, + { "M4 macro language pre-processor", "text/x-m4" }, + { "PO (gettext message catalogue)", "text/x-po" }, + { "cannot happen error on names.h/types", "error/x-error" } +}; + +/* + * XXX - how should we distinguish Java from C++? + * The trick used in a Debian snapshot, of having "extends" or "implements" + * as tags for Java, doesn't work very well, given that those keywords + * are often preceded by "class", which flags it as C++. + * + * Perhaps we need to be able to say + * + * If "class" then + * + * if "extends" or "implements" then + * Java + * else + * C++ + * endif + * + * Or should we use other keywords, such as "package" or "import"? + * Unfortunately, Ada95 uses "package", and Modula-3 uses "import", + * although I infer from the language spec at + * + * http://www.research.digital.com/SRC/m3defn/html/m3.html + * + * that Modula-3 uses "IMPORT" rather than "import", i.e. it must be + * in all caps. + * + * So, for now, we go with "import". We must put it before the C++ + * stuff, so that we don't misidentify Java as C++. Not using "package" + * means we won't identify stuff that defines a package but imports + * nothing; hopefully, very little Java code imports nothing (one of the + * reasons for doing OO programming is to import as much as possible + * and write only what you need to, right?). + * + * Unfortunately, "import" may cause us to misidentify English text + * as Java, as it comes after "the" and "The". Perhaps we need a fancier + * heuristic to identify Java? + */ +static const struct names { + char name[14]; + short type; +} names[] = { + /* These must be sorted by eye for optimal hit rate */ + /* Add to this list only after substantial meditation */ + {"msgid", L_PO}, + {"dnl", L_M4}, + {"import", L_JAVA}, + {"\"libhdr\"", L_BCPL}, + {"\"LIBHDR\"", L_BCPL}, + {"//", L_CC}, + {"template", L_CC}, + {"virtual", L_CC}, + {"class", L_CC}, + {"public:", L_CC}, + {"private:", L_CC}, + {"/*", L_C}, /* must precede "The", "the", etc. */ + {"#include", L_C}, + {"char", L_C}, + {"The", L_ENG}, + {"the", L_ENG}, + {"double", L_C}, + {"extern", L_C}, + {"float", L_C}, + {"struct", L_C}, + {"union", L_C}, + {"CFLAGS", L_MAKE}, + {"LDFLAGS", L_MAKE}, + {"all:", L_MAKE}, + {".PRECIOUS", L_MAKE}, + {".ascii", L_MACH}, + {".asciiz", L_MACH}, + {".byte", L_MACH}, + {".even", L_MACH}, + {".globl", L_MACH}, + {".text", L_MACH}, + {"clr", L_MACH}, + {"(input,", L_PAS}, + {"program", L_PAS}, + {"record", L_PAS}, + {"dcl", L_PLI}, + {"Received:", L_MAIL}, + {">From", L_MAIL}, + {"Return-Path:",L_MAIL}, + {"Cc:", L_MAIL}, + {"Newsgroups:", L_NEWS}, + {"Path:", L_NEWS}, + {"Organization:",L_NEWS}, + {"href=", L_HTML}, + {"HREF=", L_HTML}, + {" magic_buffer + * - keep only the first error + * - manual page: new sentence, new line + * - fix typo in api function (magic_buf -> magic_buffer) + * + * Revision 1.44 2003/03/23 22:23:31 christos + * finish librarification. + * + * Revision 1.43 2003/03/23 21:16:26 christos + * update copyrights. + * + * Revision 1.42 2003/03/23 04:06:05 christos + * Library re-organization + * + * Revision 1.41 2003/02/27 20:53:45 christos + * - fix memory allocation problem (Jeff Johnson) + * - fix stack overflow corruption (David Endler) + * - fixes from NetBSD source (Antti Kantee) + * - magic fixes + * + * Revision 1.40 2003/02/08 18:33:53 christos + * - detect inttypes.h too (Dave Love ) + * - eliminate unsigned char warnings (Petter Reinholdtsen ) + * - better elf PT_NOTE handling (Nalin Dahyabhai ) + * - add options to format the output differently + * - much more magic. + * + * Revision 1.39 2002/07/03 18:57:52 christos + * - ansify/c99ize + * - more magic + * - better COMPILE_ONLY support. + * - new magic files. + * - fix solaris compilation problems. + * + * Revision 1.38 2002/05/16 18:45:56 christos + * - pt_note elf additions from NetBSD + * - EMX os specific changes (Alexander Mai) + * - stdint.h detection, acconfig.h fixes (Maciej W. Rozycki, Franz Korntner) + * - regex file additions (Kim Cromie) + * - getopt_long support and misc cleanups (Michael Piefel) + * - many magic fixes and additions + * + * Revision 1.37 2001/09/03 14:44:22 christos + * daylight/tm_isdst detection + * magic fixes + * don't eat the whole file if it has only nulls + * + * Revision 1.36 2001/07/22 21:04:15 christos + * - magic fixes + * - add new operators, pascal strings, UTC date printing, $HOME/.magic + * [from "Tom N Harris" ] + * + * Revision 1.35 2001/04/24 14:40:25 christos + * - rename magic file sgi to mips and fix it + * - add support for building magic.mgc + * - portability fixes for mmap() + * - try gzip before uncompress, because uncompress sometimes hangs + * - be more conservative about pipe reads and writes + * - many magic fixes + * + * Revision 1.34 2001/03/12 05:05:57 christos + * - new compiled magic format + * - lots of magic additions + * + * Revision 1.33 2000/11/13 00:30:50 christos + * - wordperfect magic fix: freebsd pr 9388 + * - more msdos fixes from freebsd pr's 20131 and 20812 + * - sas and spss magic [Bruce Foster] + * - mkinstalldirs [John Fremlin] + * - sgi opengl fixes [Michael Pruett] + * - netbsd magic fixes [Ignatios Souvatzis] + * - audio additions [Michael Pruett] + * - fix problem with non ansi RCSID [Andreas Ley] + * - oggs magic [Felix von Leitner] + * - gmon magic [Eugen Dedu] + * - TNEF magic [Joomy] + * - netpbm magic and misc other image stuff [Bryan Henderson] + * + * Revision 1.32 2000/08/05 18:24:18 christos + * Correct indianness detection in elf (Charles Hannum) + * FreeBSD elf core support (Guy Harris) + * Use gzip in systems that don't have uncompress (Anthon van der Neut) + * Internationalization/EBCDIC support (Eric Fisher) + * Many many magic changes + * + * Revision 1.31 2000/05/14 17:58:36 christos + * - new magic for claris files + * - new magic for mathematica and maple files + * - new magic for msvc files + * - new -k flag to keep going matching all possible entries + * - add the word executable on #! magic files, and fix the usage of + * the word script + * - lots of other magic fixes + * - fix typo test -> text + * + * Revision 1.30 2000/04/11 02:41:17 christos + * - add support for mime output (-i) + * - make sure we free memory in case realloc fails + * - magic fixes + * + * Revision 1.29 1999/11/28 20:02:29 christos + * new string/[Bcb] magic from anthon, and adjustments to the magic files to + * use it. + * + * Revision 1.28 1999/10/31 22:11:48 christos + * - add "char" type for compatibility with HP/UX + * - recognize HP/UX syntax &=n etc. + * - include errno.h for CYGWIN + * - conditionalize the S_IS* macros + * - revert the SHT_DYNSYM test that broke the linux stripped binaries test + * - lots of Magdir changes + * + * Revision 1.27 1999/02/14 17:21:41 christos + * Automake support and misc cleanups from Rainer Orth + * Enable reading character and block special files from Dale R. Worley + * + * Revision 1.26 1998/09/12 13:19:39 christos + * - add support for bi-endian indirect offsets (Richard Verhoeven) + * - add recognition for bcpl (Joseph Myers) + * - remove non magic files from Magdir to avoid difficulties building + * on os2 where files are case independent + * - magic fixes. + * + * Revision 1.25 1998/06/27 14:04:04 christos + * OLF patch Guy Harris + * Recognize java/html (debian linux) + * Const poisoning (debian linux) + * More magic! + * + * Revision 1.24 1998/02/15 23:20:38 christos + * Autoconf patch: Felix von Leitner + * More magic fixes + * Elf64 fixes + * + * Revision 1.23 1997/11/05 16:03:37 christos + * - correct elf prps offset for SunOS-2.5.1 [guy@netapp.com] + * - handle 64 bit time_t's correctly [ewt@redhat.com] + * - new mime style magic [clarosse@netvista.net] + * - new TI calculator magic [rmcguire@freenet.columbus.oh.us] + * - new figlet fonts [obrien@freebsd.org] + * - new cisco magic, and elf fixes [jhawk@bbnplanet.com] + * - -b flag addition, and x86 filesystem magic [vax@linkhead.paranoia.com] + * - s/Mpeg/MPEG, header and elf typo fixes [guy@netapp.com] + * - Windows/NT registry files, audio code [guy@netapp.com] + * - libGrx graphics lib fonts [guy@netapp.com] + * - PNG fixes [guy@netapp.com] + * - more m$ document magic [guy@netapp.com] + * - PPD files [guy@netapp.com] + * - archive magic cleanup [guy@netapp.com] + * - linux kernel magic cleanup [guy@netapp.com] + * - lecter magic [guy@netapp.com] + * - vgetty magic [guy@netapp.com] + * - sniffer additions [guy@netapp.com] + * + * Revision 1.22 1997/01/15 17:23:24 christos + * - add support for elf core files: find the program name under SVR4 [Ken Pizzini] + * - print strings only up to the first carriage return [various] + * - freebsd international ascii support [J Wunsch] + * - magic fixes and additions [Guy Harris] + * - 64 bit fixes [Larry Schwimmer] + * - support for both utime and utimes, but don't restore file access times + * by default [various] + * - \xXX only takes 2 hex digits, not 3. + * - re-implement support for core files [Guy Harris] + * + * Revision 1.21 1996/10/05 18:15:29 christos + * Segregate elf stuff and conditionally enable it with -DBUILTIN_ELF + * More magic fixes + * + * Revision 1.20 1996/06/22 22:15:52 christos + * - support relative offsets of the form >& + * - fix bug with truncating magic strings that contain \n + * - file -f - did not read from stdin as documented + * - support elf file parsing using our own elf support. + * - as always magdir fixes and additions. + * + * Revision 1.19 1995/10/27 23:14:46 christos + * Ability to parse colon separated list of magic files + * New LEGAL.NOTICE + * Various magic file changes + * + * Revision 1.18 1995/05/20 22:09:21 christos + * Passed incorrect argument to eatsize(). + * Use %ld and %lx where appropriate. + * Remove unused variables + * ELF support for both big and little endian + * Fixes for small files again. + * + * Revision 1.17 1995/04/28 17:29:13 christos + * - Incorrect nroff detection fix from der Mouse + * - Lost and incorrect magic entries. + * - Added ELF stripped binary detection [in C; ugh] + * - Look for $MAGIC to find the magic file. + * - Eat trailing size specifications from numbers i.e. ignore 10L + * - More fixes for very short files + * + * Revision 1.16 1995/03/25 22:06:45 christos + * - use strtoul() where it exists. + * - fix sign-extend bug + * - try to detect tar archives before nroff files, otherwise + * tar files where the first file starts with a . will not work + * + * Revision 1.15 1995/01/21 21:03:35 christos + * Added CSECTION for the file man page + * Added version flag -v + * Fixed bug with -f input flag (from iorio@violet.berkeley.edu) + * Lots of magic fixes and reorganization... + * + * Revision 1.14 1994/05/03 17:58:23 christos + * changes from mycroft@gnu.ai.mit.edu (Charles Hannum) for unsigned + * + * Revision 1.13 1994/01/21 01:27:01 christos + * Fixed null termination bug from Don Seeley at BSDI in ascmagic.c + * + * Revision 1.12 1993/10/27 20:59:05 christos + * Changed -z flag to understand gzip format too. + * Moved builtin compression detection to a table, and move + * the compress magic entry out of the source. + * Made printing of numbers unsigned, and added the mask to it. + * Changed the buffer size to 8k, because gzip will refuse to + * unzip just a few bytes. + * + * Revision 1.11 1993/09/24 18:49:06 christos + * Fixed small bug in softmagic.c introduced by + * copying the data to be examined out of the input + * buffer. Changed the Makefile to use sed to create + * the correct man pages. + * + * Revision 1.10 1993/09/23 21:56:23 christos + * Passed purify. Fixed indirections. Fixed byte order printing. + * Fixed segmentation faults caused by referencing past the end + * of the magic buffer. Fixed bus errors caused by referencing + * unaligned shorts or longs. + * + * Revision 1.9 1993/03/24 14:23:40 ian + * Batch of minor changes from several contributors. + * + * Revision 1.8 93/02/19 15:01:26 ian + * Numerous changes from Guy Harris too numerous to mention but including + * byte-order independance, fixing "old-style masking", etc. etc. A bugfix + * for broken symlinks from martin@@d255s004.zfe.siemens.de. + * + * Revision 1.7 93/01/05 14:57:27 ian + * Couple of nits picked by Christos (again, thanks). + * + * Revision 1.6 93/01/05 13:51:09 ian + * Lotsa work on the Magic directory. + * + * Revision 1.5 92/09/14 14:54:51 ian + * Fix a tiny null-pointer bug in previous fix for tar archive + uncompress. + * + */ diff --git a/ext/fileinfo/lib/print.c b/ext/fileinfo/lib/print.c new file mode 100644 index 0000000000..c41e71eac5 --- /dev/null +++ b/ext/fileinfo/lib/print.c @@ -0,0 +1,236 @@ +/* + * Copyright (c) Ian F. Darwin 1986-1995. + * Software written by Ian F. Darwin and others; + * maintained 1995-present by Christos Zoulas and others. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice immediately at the beginning of the file, without modification, + * this list of conditions, and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR + * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + */ +/* + * print.c - debugging printout routines + */ + +#include "file.h" +#include +#include +#include +#include +#include +#ifdef HAVE_UNISTD_H +#include +#endif +#include + +#ifndef lint +FILE_RCSID("@(#)$File: print.c,v 1.63 2008/02/17 19:28:54 rrt Exp $") +#endif /* lint */ + +#define SZOF(a) (sizeof(a) / sizeof(a[0])) + +#ifndef COMPILE_ONLY +protected void +file_mdump(struct magic *m) +{ + private const char optyp[] = { FILE_OPS }; + + (void) fprintf(stderr, "[%u", m->lineno); + (void) fprintf(stderr, ">>>>>>>> %u" + 8 - (m->cont_level & 7), + m->offset); + + if (m->flag & INDIR) { + (void) fprintf(stderr, "(%s,", + /* Note: type is unsigned */ + (m->in_type < file_nnames) ? + file_names[m->in_type] : "*bad*"); + if (m->in_op & FILE_OPINVERSE) + (void) fputc('~', stderr); + (void) fprintf(stderr, "%c%u),", + ((m->in_op & FILE_OPS_MASK) < SZOF(optyp)) ? + optyp[m->in_op & FILE_OPS_MASK] : '?', + m->in_offset); + } + (void) fprintf(stderr, " %s%s", (m->flag & UNSIGNED) ? "u" : "", + /* Note: type is unsigned */ + (m->type < file_nnames) ? file_names[m->type] : "*bad*"); + if (m->mask_op & FILE_OPINVERSE) + (void) fputc('~', stderr); + + if (IS_STRING(m->type)) { + if (m->str_flags) { + (void) fputc('/', stderr); + if (m->str_flags & STRING_COMPACT_BLANK) + (void) fputc(CHAR_COMPACT_BLANK, stderr); + if (m->str_flags & STRING_COMPACT_OPTIONAL_BLANK) + (void) fputc(CHAR_COMPACT_OPTIONAL_BLANK, + stderr); + if (m->str_flags & STRING_IGNORE_LOWERCASE) + (void) fputc(CHAR_IGNORE_LOWERCASE, stderr); + if (m->str_flags & STRING_IGNORE_UPPERCASE) + (void) fputc(CHAR_IGNORE_UPPERCASE, stderr); + if (m->str_flags & REGEX_OFFSET_START) + (void) fputc(CHAR_REGEX_OFFSET_START, stderr); + } + if (m->str_range) + (void) fprintf(stderr, "/%u", m->str_range); + } + else { + if ((m->mask_op & FILE_OPS_MASK) < SZOF(optyp)) + (void) fputc(optyp[m->mask_op & FILE_OPS_MASK], stderr); + else + (void) fputc('?', stderr); + + if (m->num_mask) { + (void) fprintf(stderr, "%.8llx", + (unsigned long long)m->num_mask); + } + } + (void) fprintf(stderr, ",%c", m->reln); + + if (m->reln != 'x') { + switch (m->type) { + case FILE_BYTE: + case FILE_SHORT: + case FILE_LONG: + case FILE_LESHORT: + case FILE_LELONG: + case FILE_MELONG: + case FILE_BESHORT: + case FILE_BELONG: + (void) fprintf(stderr, "%d", m->value.l); + break; + case FILE_BEQUAD: + case FILE_LEQUAD: + case FILE_QUAD: + (void) fprintf(stderr, "%lld", + (unsigned long long)m->value.q); + break; + case FILE_PSTRING: + case FILE_STRING: + case FILE_REGEX: + case FILE_BESTRING16: + case FILE_LESTRING16: + case FILE_SEARCH: + file_showstr(stderr, m->value.s, (size_t)m->vallen); + break; + case FILE_DATE: + case FILE_LEDATE: + case FILE_BEDATE: + case FILE_MEDATE: + (void)fprintf(stderr, "%s,", + file_fmttime(m->value.l, 1)); + break; + case FILE_LDATE: + case FILE_LELDATE: + case FILE_BELDATE: + case FILE_MELDATE: + (void)fprintf(stderr, "%s,", + file_fmttime(m->value.l, 0)); + break; + case FILE_QDATE: + case FILE_LEQDATE: + case FILE_BEQDATE: + (void)fprintf(stderr, "%s,", + file_fmttime((uint32_t)m->value.q, 1)); + break; + case FILE_QLDATE: + case FILE_LEQLDATE: + case FILE_BEQLDATE: + (void)fprintf(stderr, "%s,", + file_fmttime((uint32_t)m->value.q, 0)); + break; + case FILE_FLOAT: + case FILE_BEFLOAT: + case FILE_LEFLOAT: + (void) fprintf(stderr, "%G", m->value.f); + break; + case FILE_DOUBLE: + case FILE_BEDOUBLE: + case FILE_LEDOUBLE: + (void) fprintf(stderr, "%G", m->value.d); + break; + case FILE_DEFAULT: + /* XXX - do anything here? */ + break; + default: + (void) fputs("*bad*", stderr); + break; + } + } + (void) fprintf(stderr, ",\"%s\"]\n", m->desc); +} +#endif + +/*VARARGS*/ +protected void +file_magwarn(struct magic_set *ms, const char *f, ...) +{ + va_list va; + + /* cuz we use stdout for most, stderr here */ + (void) fflush(stdout); + + if (ms->file) + (void) fprintf(stderr, "%s, %lu: ", ms->file, + (unsigned long)ms->line); + (void) fprintf(stderr, "Warning: "); + va_start(va, f); + (void) vfprintf(stderr, f, va); + va_end(va); + (void) fputc('\n', stderr); +} + +protected const char * +file_fmttime(uint32_t v, int local) +{ + char *pp; + time_t t = (time_t)v; + struct tm *tm; + + if (local) { + pp = ctime(&t); + } else { +#ifndef HAVE_DAYLIGHT + private int daylight = 0; +#ifdef HAVE_TM_ISDST + private time_t now = (time_t)0; + + if (now == (time_t)0) { + struct tm *tm1; + (void)time(&now); + tm1 = localtime(&now); + if (tm1 == NULL) + return "*Invalid time*"; + daylight = tm1->tm_isdst; + } +#endif /* HAVE_TM_ISDST */ +#endif /* HAVE_DAYLIGHT */ + if (daylight) + t += 3600; + tm = gmtime(&t); + if (tm == NULL) + return "*Invalid time*"; + pp = asctime(tm); + } + + pp[strcspn(pp, "\n")] = '\0'; + return pp; +} diff --git a/ext/fileinfo/lib/readelf.c b/ext/fileinfo/lib/readelf.c new file mode 100644 index 0000000000..8e0a8fe32d --- /dev/null +++ b/ext/fileinfo/lib/readelf.c @@ -0,0 +1,1016 @@ +/* + * Copyright (c) Christos Zoulas 2003. + * All Rights Reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice immediately at the beginning of the file, without modification, + * this list of conditions, and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR + * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + */ +#include "file.h" + +#ifdef BUILTIN_ELF +#include +#include +#include +#ifdef HAVE_UNISTD_H +#include +#endif + +#include "readelf.h" +#include "magic.h" + +#ifndef lint +FILE_RCSID("@(#)$File: readelf.c,v 1.73 2008/03/27 22:00:28 christos Exp $") +#endif + +#ifdef ELFCORE +private int dophn_core(struct magic_set *, int, int, int, off_t, int, size_t, + off_t, int *); +#endif +private int dophn_exec(struct magic_set *, int, int, int, off_t, int, size_t, + off_t, int *); +private int doshn(struct magic_set *, int, int, int, off_t, int, size_t, int *); +private size_t donote(struct magic_set *, unsigned char *, size_t, size_t, int, + int, size_t, int *); + +#define ELF_ALIGN(a) ((((a) + align - 1) / align) * align) + +#define isquote(c) (strchr("'\"`", (c)) != NULL) + +private uint16_t getu16(int, uint16_t); +private uint32_t getu32(int, uint32_t); +private uint64_t getu64(int, uint64_t); + +private uint16_t +getu16(int swap, uint16_t value) +{ + union { + uint16_t ui; + char c[2]; + } retval, tmpval; + + if (swap) { + tmpval.ui = value; + + retval.c[0] = tmpval.c[1]; + retval.c[1] = tmpval.c[0]; + + return retval.ui; + } else + return value; +} + +private uint32_t +getu32(int swap, uint32_t value) +{ + union { + uint32_t ui; + char c[4]; + } retval, tmpval; + + if (swap) { + tmpval.ui = value; + + retval.c[0] = tmpval.c[3]; + retval.c[1] = tmpval.c[2]; + retval.c[2] = tmpval.c[1]; + retval.c[3] = tmpval.c[0]; + + return retval.ui; + } else + return value; +} + +private uint64_t +getu64(int swap, uint64_t value) +{ + union { + uint64_t ui; + char c[8]; + } retval, tmpval; + + if (swap) { + tmpval.ui = value; + + retval.c[0] = tmpval.c[7]; + retval.c[1] = tmpval.c[6]; + retval.c[2] = tmpval.c[5]; + retval.c[3] = tmpval.c[4]; + retval.c[4] = tmpval.c[3]; + retval.c[5] = tmpval.c[2]; + retval.c[6] = tmpval.c[1]; + retval.c[7] = tmpval.c[0]; + + return retval.ui; + } else + return value; +} + +#define elf_getu16(swap, value) getu16(swap, value) +#define elf_getu32(swap, value) getu32(swap, value) +#ifdef USE_ARRAY_FOR_64BIT_TYPES +# define elf_getu64(swap, array) \ + ((swap ? ((uint64_t)elf_getu32(swap, array[0])) << 32 : elf_getu32(swap, array[0])) + \ + (swap ? elf_getu32(swap, array[1]) : ((uint64_t)elf_getu32(swap, array[1]) << 32))) +#else +# define elf_getu64(swap, value) getu64(swap, value) +#endif + +#define xsh_addr (class == ELFCLASS32 \ + ? (void *) &sh32 \ + : (void *) &sh64) +#define xsh_sizeof (class == ELFCLASS32 \ + ? sizeof sh32 \ + : sizeof sh64) +#define xsh_size (class == ELFCLASS32 \ + ? elf_getu32(swap, sh32.sh_size) \ + : elf_getu64(swap, sh64.sh_size)) +#define xsh_offset (class == ELFCLASS32 \ + ? elf_getu32(swap, sh32.sh_offset) \ + : elf_getu64(swap, sh64.sh_offset)) +#define xsh_type (class == ELFCLASS32 \ + ? elf_getu32(swap, sh32.sh_type) \ + : elf_getu32(swap, sh64.sh_type)) +#define xph_addr (class == ELFCLASS32 \ + ? (void *) &ph32 \ + : (void *) &ph64) +#define xph_sizeof (class == ELFCLASS32 \ + ? sizeof ph32 \ + : sizeof ph64) +#define xph_type (class == ELFCLASS32 \ + ? elf_getu32(swap, ph32.p_type) \ + : elf_getu32(swap, ph64.p_type)) +#define xph_offset (off_t)(class == ELFCLASS32 \ + ? elf_getu32(swap, ph32.p_offset) \ + : elf_getu64(swap, ph64.p_offset)) +#define xph_align (size_t)((class == ELFCLASS32 \ + ? (off_t) (ph32.p_align ? \ + elf_getu32(swap, ph32.p_align) : 4) \ + : (off_t) (ph64.p_align ? \ + elf_getu64(swap, ph64.p_align) : 4))) +#define xph_filesz (size_t)((class == ELFCLASS32 \ + ? elf_getu32(swap, ph32.p_filesz) \ + : elf_getu64(swap, ph64.p_filesz))) +#define xnh_addr (class == ELFCLASS32 \ + ? (void *) &nh32 \ + : (void *) &nh64) +#define xph_memsz (size_t)((class == ELFCLASS32 \ + ? elf_getu32(swap, ph32.p_memsz) \ + : elf_getu64(swap, ph64.p_memsz))) +#define xnh_sizeof (class == ELFCLASS32 \ + ? sizeof nh32 \ + : sizeof nh64) +#define xnh_type (class == ELFCLASS32 \ + ? elf_getu32(swap, nh32.n_type) \ + : elf_getu32(swap, nh64.n_type)) +#define xnh_namesz (class == ELFCLASS32 \ + ? elf_getu32(swap, nh32.n_namesz) \ + : elf_getu32(swap, nh64.n_namesz)) +#define xnh_descsz (class == ELFCLASS32 \ + ? elf_getu32(swap, nh32.n_descsz) \ + : elf_getu32(swap, nh64.n_descsz)) +#define prpsoffsets(i) (class == ELFCLASS32 \ + ? prpsoffsets32[i] \ + : prpsoffsets64[i]) + +#ifdef ELFCORE +/* + * Try larger offsets first to avoid false matches + * from earlier data that happen to look like strings. + */ +static const size_t prpsoffsets32[] = { +#ifdef USE_NT_PSINFO + 104, /* SunOS 5.x (command line) */ + 88, /* SunOS 5.x (short name) */ +#endif /* USE_NT_PSINFO */ + + 100, /* SunOS 5.x (command line) */ + 84, /* SunOS 5.x (short name) */ + + 44, /* Linux (command line) */ + 28, /* Linux 2.0.36 (short name) */ + + 8, /* FreeBSD */ +}; + +static const size_t prpsoffsets64[] = { +#ifdef USE_NT_PSINFO + 152, /* SunOS 5.x (command line) */ + 136, /* SunOS 5.x (short name) */ +#endif /* USE_NT_PSINFO */ + + 136, /* SunOS 5.x, 64-bit (command line) */ + 120, /* SunOS 5.x, 64-bit (short name) */ + + 56, /* Linux (command line) */ + 40, /* Linux (tested on core from 2.4.x, short name) */ + + 16, /* FreeBSD, 64-bit */ +}; + +#define NOFFSETS32 (sizeof prpsoffsets32 / sizeof prpsoffsets32[0]) +#define NOFFSETS64 (sizeof prpsoffsets64 / sizeof prpsoffsets64[0]) + +#define NOFFSETS (class == ELFCLASS32 ? NOFFSETS32 : NOFFSETS64) + +/* + * Look through the program headers of an executable image, searching + * for a PT_NOTE section of type NT_PRPSINFO, with a name "CORE" or + * "FreeBSD"; if one is found, try looking in various places in its + * contents for a 16-character string containing only printable + * characters - if found, that string should be the name of the program + * that dropped core. Note: right after that 16-character string is, + * at least in SunOS 5.x (and possibly other SVR4-flavored systems) and + * Linux, a longer string (80 characters, in 5.x, probably other + * SVR4-flavored systems, and Linux) containing the start of the + * command line for that program. + * + * SunOS 5.x core files contain two PT_NOTE sections, with the types + * NT_PRPSINFO (old) and NT_PSINFO (new). These structs contain the + * same info about the command name and command line, so it probably + * isn't worthwhile to look for NT_PSINFO, but the offsets are provided + * above (see USE_NT_PSINFO), in case we ever decide to do so. The + * NT_PRPSINFO and NT_PSINFO sections are always in order and adjacent; + * the SunOS 5.x file command relies on this (and prefers the latter). + * + * The signal number probably appears in a section of type NT_PRSTATUS, + * but that's also rather OS-dependent, in ways that are harder to + * dissect with heuristics, so I'm not bothering with the signal number. + * (I suppose the signal number could be of interest in situations where + * you don't have the binary of the program that dropped core; if you + * *do* have that binary, the debugger will probably tell you what + * signal it was.) + */ + +#define OS_STYLE_SVR4 0 +#define OS_STYLE_FREEBSD 1 +#define OS_STYLE_NETBSD 2 + +private const char os_style_names[][8] = { + "SVR4", + "FreeBSD", + "NetBSD", +}; + +#define FLAGS_DID_CORE 1 +#define FLAGS_DID_NOTE 2 +#define FLAGS_DID_CORE_STYLE 4 + +private int +dophn_core(struct magic_set *ms, int class, int swap, int fd, off_t off, + int num, size_t size, off_t fsize, int *flags) +{ + Elf32_Phdr ph32; + Elf64_Phdr ph64; + size_t offset; + unsigned char nbuf[BUFSIZ]; + ssize_t bufsize; + off_t savedoffset; + struct stat st; + + if (fstat(fd, &st) < 0) { + file_badread(ms); + return -1; + } + + if (size != xph_sizeof) { + if (file_printf(ms, ", corrupted program header size") == -1) + return -1; + return 0; + } + + /* + * Loop through all the program headers. + */ + for ( ; num; num--) { + if ((savedoffset = lseek(fd, off, SEEK_SET)) == (off_t)-1) { + file_badseek(ms); + return -1; + } + if (read(fd, xph_addr, xph_sizeof) == -1) { + file_badread(ms); + return -1; + } + if (xph_offset > fsize) { + if (lseek(fd, savedoffset, SEEK_SET) == (off_t)-1) { + file_badseek(ms); + return -1; + } + continue; + } + + off += size; + if (xph_type != PT_NOTE) + continue; + + /* + * This is a PT_NOTE section; loop through all the notes + * in the section. + */ + if (lseek(fd, xph_offset, SEEK_SET) == (off_t)-1) { + file_badseek(ms); + return -1; + } + bufsize = read(fd, nbuf, + ((xph_filesz < sizeof(nbuf)) ? xph_filesz : sizeof(nbuf))); + if (bufsize == -1) { + file_badread(ms); + return -1; + } + offset = 0; + for (;;) { + if (offset >= (size_t)bufsize) + break; + offset = donote(ms, nbuf, offset, (size_t)bufsize, + class, swap, 4, flags); + if (offset == 0) + break; + + } + } + return 0; +} +#endif + +private size_t +donote(struct magic_set *ms, unsigned char *nbuf, size_t offset, size_t size, + int class, int swap, size_t align, int *flags) +{ + Elf32_Nhdr nh32; + Elf64_Nhdr nh64; + size_t noff, doff; +#ifdef ELFCORE + int os_style = -1; +#endif + uint32_t namesz, descsz; + + (void)memcpy(xnh_addr, &nbuf[offset], xnh_sizeof); + offset += xnh_sizeof; + + namesz = xnh_namesz; + descsz = xnh_descsz; + if ((namesz == 0) && (descsz == 0)) { + /* + * We're out of note headers. + */ + return (offset >= size) ? offset : size; + } + + if (namesz & 0x80000000) { + (void)file_printf(ms, ", bad note name size 0x%lx", + (unsigned long)namesz); + return offset; + } + + if (descsz & 0x80000000) { + (void)file_printf(ms, ", bad note description size 0x%lx", + (unsigned long)descsz); + return offset; + } + + + noff = offset; + doff = ELF_ALIGN(offset + namesz); + + if (offset + namesz > size) { + /* + * We're past the end of the buffer. + */ + return doff; + } + + offset = ELF_ALIGN(doff + descsz); + if (doff + descsz > size) { + /* + * We're past the end of the buffer. + */ + return (offset >= size) ? offset : size; + } + + if (*flags & FLAGS_DID_NOTE) + goto core; + + if (namesz == 4 && strcmp((char *)&nbuf[noff], "GNU") == 0 && + xnh_type == NT_GNU_VERSION && descsz == 16) { + uint32_t desc[4]; + (void)memcpy(desc, &nbuf[doff], sizeof(desc)); + + if (file_printf(ms, ", for GNU/") == -1) + return size; + switch (elf_getu32(swap, desc[0])) { + case GNU_OS_LINUX: + if (file_printf(ms, "Linux") == -1) + return size; + break; + case GNU_OS_HURD: + if (file_printf(ms, "Hurd") == -1) + return size; + break; + case GNU_OS_SOLARIS: + if (file_printf(ms, "Solaris") == -1) + return size; + break; + case GNU_OS_KFREEBSD: + if (file_printf(ms, "kFreeBSD") == -1) + return size; + break; + case GNU_OS_KNETBSD: + if (file_printf(ms, "kNetBSD") == -1) + return size; + break; + default: + if (file_printf(ms, "") == -1) + return size; + } + if (file_printf(ms, " %d.%d.%d", elf_getu32(swap, desc[1]), + elf_getu32(swap, desc[2]), elf_getu32(swap, desc[3])) == -1) + return size; + *flags |= FLAGS_DID_NOTE; + return size; + } + + if (namesz == 7 && strcmp((char *)&nbuf[noff], "NetBSD") == 0 && + xnh_type == NT_NETBSD_VERSION && descsz == 4) { + uint32_t desc; + (void)memcpy(&desc, &nbuf[doff], sizeof(desc)); + desc = elf_getu32(swap, desc); + + if (file_printf(ms, ", for NetBSD") == -1) + return size; + /* + * The version number used to be stuck as 199905, and was thus + * basically content-free. Newer versions of NetBSD have fixed + * this and now use the encoding of __NetBSD_Version__: + * + * MMmmrrpp00 + * + * M = major version + * m = minor version + * r = release ["",A-Z,Z[A-Z] but numeric] + * p = patchlevel + */ + if (desc > 100000000U) { + uint32_t ver_patch = (desc / 100) % 100; + uint32_t ver_rel = (desc / 10000) % 100; + uint32_t ver_min = (desc / 1000000) % 100; + uint32_t ver_maj = desc / 100000000; + + if (file_printf(ms, " %u.%u", ver_maj, ver_min) == -1) + return size; + if (ver_rel == 0 && ver_patch != 0) { + if (file_printf(ms, ".%u", ver_patch) == -1) + return size; + } else if (ver_rel != 0) { + while (ver_rel > 26) { + if (file_printf(ms, "Z") == -1) + return size; + ver_rel -= 26; + } + if (file_printf(ms, "%c", 'A' + ver_rel - 1) + == -1) + return size; + } + } + *flags |= FLAGS_DID_NOTE; + return size; + } + + if (namesz == 8 && strcmp((char *)&nbuf[noff], "FreeBSD") == 0 && + xnh_type == NT_FREEBSD_VERSION && descsz == 4) { + uint32_t desc; + (void)memcpy(&desc, &nbuf[doff], sizeof(desc)); + desc = elf_getu32(swap, desc); + if (file_printf(ms, ", for FreeBSD") == -1) + return size; + + /* + * Contents is __FreeBSD_version, whose relation to OS + * versions is defined by a huge table in the Porter's + * Handbook. This is the general scheme: + * + * Releases: + * Mmp000 (before 4.10) + * Mmi0p0 (before 5.0) + * Mmm0p0 + * + * Development branches: + * Mmpxxx (before 4.6) + * Mmp1xx (before 4.10) + * Mmi1xx (before 5.0) + * M000xx (pre-M.0) + * Mmm1xx + * + * M = major version + * m = minor version + * i = minor version increment (491000 -> 4.10) + * p = patchlevel + * x = revision + * + * The first release of FreeBSD to use ELF by default + * was version 3.0. + */ + if (desc == 460002) { + if (file_printf(ms, " 4.6.2") == -1) + return size; + } else if (desc < 460100) { + if (file_printf(ms, " %d.%d", desc / 100000, + desc / 10000 % 10) == -1) + return size; + if (desc / 1000 % 10 > 0) + if (file_printf(ms, ".%d", desc / 1000 % 10) + == -1) + return size; + if ((desc % 1000 > 0) || (desc % 100000 == 0)) + if (file_printf(ms, " (%d)", desc) == -1) + return size; + } else if (desc < 500000) { + if (file_printf(ms, " %d.%d", desc / 100000, + desc / 10000 % 10 + desc / 1000 % 10) == -1) + return size; + if (desc / 100 % 10 > 0) { + if (file_printf(ms, " (%d)", desc) == -1) + return size; + } else if (desc / 10 % 10 > 0) { + if (file_printf(ms, ".%d", desc / 10 % 10) + == -1) + return size; + } + } else { + if (file_printf(ms, " %d.%d", desc / 100000, + desc / 1000 % 100) == -1) + return size; + if ((desc / 100 % 10 > 0) || + (desc % 100000 / 100 == 0)) { + if (file_printf(ms, " (%d)", desc) == -1) + return size; + } else if (desc / 10 % 10 > 0) { + if (file_printf(ms, ".%d", desc / 10 % 10) + == -1) + return size; + } + } + *flags |= FLAGS_DID_NOTE; + return size; + } + + if (namesz == 8 && strcmp((char *)&nbuf[noff], "OpenBSD") == 0 && + xnh_type == NT_OPENBSD_VERSION && descsz == 4) { + if (file_printf(ms, ", for OpenBSD") == -1) + return size; + /* Content of note is always 0 */ + *flags |= FLAGS_DID_NOTE; + return size; + } + + if (namesz == 10 && strcmp((char *)&nbuf[noff], "DragonFly") == 0 && + xnh_type == NT_DRAGONFLY_VERSION && descsz == 4) { + uint32_t desc; + if (file_printf(ms, ", for DragonFly") == -1) + return size; + (void)memcpy(&desc, &nbuf[doff], sizeof(desc)); + desc = elf_getu32(swap, desc); + if (file_printf(ms, " %d.%d.%d", desc / 100000, + desc / 10000 % 10, desc % 10000) == -1) + return size; + *flags |= FLAGS_DID_NOTE; + return size; + } + +core: + /* + * Sigh. The 2.0.36 kernel in Debian 2.1, at + * least, doesn't correctly implement name + * sections, in core dumps, as specified by + * the "Program Linking" section of "UNIX(R) System + * V Release 4 Programmer's Guide: ANSI C and + * Programming Support Tools", because my copy + * clearly says "The first 'namesz' bytes in 'name' + * contain a *null-terminated* [emphasis mine] + * character representation of the entry's owner + * or originator", but the 2.0.36 kernel code + * doesn't include the terminating null in the + * name.... + */ + if ((namesz == 4 && strncmp((char *)&nbuf[noff], "CORE", 4) == 0) || + (namesz == 5 && strcmp((char *)&nbuf[noff], "CORE") == 0)) { + os_style = OS_STYLE_SVR4; + } + + if ((namesz == 8 && strcmp((char *)&nbuf[noff], "FreeBSD") == 0)) { + os_style = OS_STYLE_FREEBSD; + } + + if ((namesz >= 11 && strncmp((char *)&nbuf[noff], "NetBSD-CORE", 11) + == 0)) { + os_style = OS_STYLE_NETBSD; + } + +#ifdef ELFCORE + if ((*flags & FLAGS_DID_CORE) != 0) + return size; + + if (os_style != -1 && (*flags & FLAGS_DID_CORE_STYLE) == 0) { + if (file_printf(ms, ", %s-style", os_style_names[os_style]) + == -1) + return size; + *flags |= FLAGS_DID_CORE_STYLE; + } + + switch (os_style) { + case OS_STYLE_NETBSD: + if (xnh_type == NT_NETBSD_CORE_PROCINFO) { + uint32_t signo; + /* + * Extract the program name. It is at + * offset 0x7c, and is up to 32-bytes, + * including the terminating NUL. + */ + if (file_printf(ms, ", from '%.31s'", + &nbuf[doff + 0x7c]) == -1) + return size; + + /* + * Extract the signal number. It is at + * offset 0x08. + */ + (void)memcpy(&signo, &nbuf[doff + 0x08], + sizeof(signo)); + if (file_printf(ms, " (signal %u)", + elf_getu32(swap, signo)) == -1) + return size; + *flags |= FLAGS_DID_CORE; + return size; + } + break; + + default: + if (xnh_type == NT_PRPSINFO) { + size_t i, j; + unsigned char c; + /* + * Extract the program name. We assume + * it to be 16 characters (that's what it + * is in SunOS 5.x and Linux). + * + * Unfortunately, it's at a different offset + * in various OSes, so try multiple offsets. + * If the characters aren't all printable, + * reject it. + */ + for (i = 0; i < NOFFSETS; i++) { + unsigned char *cname, *cp; + size_t reloffset = prpsoffsets(i); + size_t noffset = doff + reloffset; + for (j = 0; j < 16; j++, noffset++, + reloffset++) { + /* + * Make sure we're not past + * the end of the buffer; if + * we are, just give up. + */ + if (noffset >= size) + goto tryanother; + + /* + * Make sure we're not past + * the end of the contents; + * if we are, this obviously + * isn't the right offset. + */ + if (reloffset >= descsz) + goto tryanother; + + c = nbuf[noffset]; + if (c == '\0') { + /* + * A '\0' at the + * beginning is + * obviously wrong. + * Any other '\0' + * means we're done. + */ + if (j == 0) + goto tryanother; + else + break; + } else { + /* + * A nonprintable + * character is also + * wrong. + */ + if (!isprint(c) || isquote(c)) + goto tryanother; + } + } + /* + * Well, that worked. + */ + cname = (unsigned char *) + &nbuf[doff + prpsoffsets(i)]; + for (cp = cname; *cp && isprint(*cp); cp++) + continue; + /* + * Linux apparently appends a space at the end + * of the command line: remove it. + */ + while (cp > cname && isspace(cp[-1])) + cp--; + if (file_printf(ms, ", from '%.*s'", + (int)(cp - cname), cname) == -1) + return size; + *flags |= FLAGS_DID_CORE; + return size; + + tryanother: + ; + } + } + break; + } +#endif + return offset; +} + +private int +doshn(struct magic_set *ms, int class, int swap, int fd, off_t off, int num, + size_t size, int *flags) +{ + Elf32_Shdr sh32; + Elf64_Shdr sh64; + int stripped = 1; + void *nbuf; + off_t noff; + + if (size != xsh_sizeof) { + if (file_printf(ms, ", corrupted section header size") == -1) + return -1; + return 0; + } + + if (lseek(fd, off, SEEK_SET) == (off_t)-1) { + file_badseek(ms); + return -1; + } + + for ( ; num; num--) { + if (read(fd, xsh_addr, xsh_sizeof) == -1) { + file_badread(ms); + return -1; + } + switch (xsh_type) { + case SHT_SYMTAB: +#if 0 + case SHT_DYNSYM: +#endif + stripped = 0; + break; + case SHT_NOTE: + if ((off = lseek(fd, (off_t)0, SEEK_CUR)) == + (off_t)-1) { + file_badread(ms); + return -1; + } + if ((nbuf = malloc((size_t)xsh_size)) == NULL) { + file_error(ms, errno, "Cannot allocate memory" + " for note"); + return -1; + } + if ((noff = lseek(fd, (off_t)xsh_offset, SEEK_SET)) == + (off_t)-1) { + file_badread(ms); + free(nbuf); + return -1; + } + if (read(fd, nbuf, (size_t)xsh_size) != + (ssize_t)xsh_size) { + free(nbuf); + file_badread(ms); + return -1; + } + + noff = 0; + for (;;) { + if (noff >= (size_t)xsh_size) + break; + noff = donote(ms, nbuf, (size_t)noff, + (size_t)xsh_size, class, swap, 4, + flags); + if (noff == 0) + break; + } + if ((lseek(fd, off, SEEK_SET)) == (off_t)-1) { + free(nbuf); + file_badread(ms); + return -1; + } + free(nbuf); + break; + } + } + if (file_printf(ms, ", %sstripped", stripped ? "" : "not ") == -1) + return -1; + return 0; +} + +/* + * Look through the program headers of an executable image, searching + * for a PT_INTERP section; if one is found, it's dynamically linked, + * otherwise it's statically linked. + */ +private int +dophn_exec(struct magic_set *ms, int class, int swap, int fd, off_t off, + int num, size_t size, off_t fsize, int *flags) +{ + Elf32_Phdr ph32; + Elf64_Phdr ph64; + const char *linking_style = "statically"; + const char *shared_libraries = ""; + unsigned char nbuf[BUFSIZ]; + int bufsize; + size_t offset, align; + off_t savedoffset = (off_t)-1; + struct stat st; + + if (fstat(fd, &st) < 0) { + file_badread(ms); + return -1; + } + + if (size != xph_sizeof) { + if (file_printf(ms, ", corrupted program header size") == -1) + return -1; + return 0; + } + + if (lseek(fd, off, SEEK_SET) == (off_t)-1) { + file_badseek(ms); + return -1; + } + + for ( ; num; num--) { + if (read(fd, xph_addr, xph_sizeof) == -1) { + file_badread(ms); + return -1; + } + if (xph_offset > st.st_size && savedoffset != (off_t)-1) { + if (lseek(fd, savedoffset, SEEK_SET) == (off_t)-1) { + file_badseek(ms); + return -1; + } + continue; + } + + if ((savedoffset = lseek(fd, (off_t)0, SEEK_CUR)) == (off_t)-1) { + file_badseek(ms); + return -1; + } + + if (xph_offset > fsize) { + if (lseek(fd, savedoffset, SEEK_SET) == (off_t)-1) { + file_badseek(ms); + return -1; + } + continue; + } + + switch (xph_type) { + case PT_DYNAMIC: + linking_style = "dynamically"; + break; + case PT_INTERP: + shared_libraries = " (uses shared libs)"; + break; + case PT_NOTE: + if ((align = xph_align) & 0x80000000) { + if (file_printf(ms, + ", invalid note alignment 0x%lx", + (unsigned long)align) == -1) + return -1; + align = 4; + } + /* + * This is a PT_NOTE section; loop through all the notes + * in the section. + */ + if (lseek(fd, xph_offset, SEEK_SET) + == (off_t)-1) { + file_badseek(ms); + return -1; + } + bufsize = read(fd, nbuf, ((xph_filesz < sizeof(nbuf)) ? + xph_filesz : sizeof(nbuf))); + if (bufsize == -1) { + file_badread(ms); + return -1; + } + offset = 0; + for (;;) { + if (offset >= (size_t)bufsize) + break; + offset = donote(ms, nbuf, offset, + (size_t)bufsize, class, swap, align, + flags); + if (offset == 0) + break; + } + if (lseek(fd, savedoffset, SEEK_SET) == (off_t)-1) { + file_badseek(ms); + return -1; + } + break; + default: + break; + } + } + if (file_printf(ms, ", %s linked%s", linking_style, shared_libraries) + == -1) + return -1; + return 0; +} + + +protected int +file_tryelf(struct magic_set *ms, int fd, const unsigned char *buf, + size_t nbytes) +{ + union { + int32_t l; + char c[sizeof (int32_t)]; + } u; + int class; + int swap; + struct stat st; + off_t fsize; + int flags = 0; + Elf32_Ehdr elf32hdr; + Elf64_Ehdr elf64hdr; + uint16_t type; + + if (ms->flags & MAGIC_MIME) + return 0; + /* + * ELF executables have multiple section headers in arbitrary + * file locations and thus file(1) cannot determine it from easily. + * Instead we traverse thru all section headers until a symbol table + * one is found or else the binary is stripped. + * Return immediately if it's not ELF (so we avoid pipe2file unless needed). + */ + if (buf[EI_MAG0] != ELFMAG0 + || (buf[EI_MAG1] != ELFMAG1 && buf[EI_MAG1] != OLFMAG1) + || buf[EI_MAG2] != ELFMAG2 || buf[EI_MAG3] != ELFMAG3) + return 0; + + /* + * If we cannot seek, it must be a pipe, socket or fifo. + */ + if((lseek(fd, (off_t)0, SEEK_SET) == (off_t)-1) && (errno == ESPIPE)) + fd = file_pipe2file(ms, fd, buf, nbytes); + + if (fstat(fd, &st) == -1) { + file_badread(ms); + return -1; + } + fsize = st.st_size; + + class = buf[EI_CLASS]; + + switch (class) { + case ELFCLASS32: +#undef elf_getu +#define elf_getu(a, b) elf_getu32(a, b) +#undef elfhdr +#define elfhdr elf32hdr +#include "elfclass.h" + case ELFCLASS64: +#undef elf_getu +#define elf_getu(a, b) elf_getu64(a, b) +#undef elfhdr +#define elfhdr elf64hdr +#include "elfclass.h" + default: + if (file_printf(ms, ", unknown class %d", class) == -1) + return -1; + break; + } + return 0; +} +#endif diff --git a/ext/fileinfo/lib/readelf.h b/ext/fileinfo/lib/readelf.h new file mode 100644 index 0000000000..2848f6dede --- /dev/null +++ b/ext/fileinfo/lib/readelf.h @@ -0,0 +1,236 @@ +/* + * Copyright (c) Christos Zoulas 2003. + * All Rights Reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice immediately at the beginning of the file, without modification, + * this list of conditions, and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR + * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + */ +/* + * @(#)Id: readelf.h,v 1.9 2002/05/16 18:45:56 christos Exp + * + * Provide elf data structures for non-elf machines, allowing file + * non-elf hosts to determine if an elf binary is stripped. + * Note: cobbled from the linux header file, with modifications + */ +#ifndef __fake_elf_h__ +#define __fake_elf_h__ + +#if HAVE_STDINT_H +#include +#endif + +typedef uint32_t Elf32_Addr; +typedef uint32_t Elf32_Off; +typedef uint16_t Elf32_Half; +typedef uint32_t Elf32_Word; +typedef uint8_t Elf32_Char; + +#if SIZEOF_LONG_LONG != 8 +#define USE_ARRAY_FOR_64BIT_TYPES +typedef uint32_t Elf64_Addr[2]; +typedef uint32_t Elf64_Off[2]; +typedef uint32_t Elf64_Xword[2]; +#else +#undef USE_ARRAY_FOR_64BIT_TYPES +typedef uint64_t Elf64_Addr; +typedef uint64_t Elf64_Off; +typedef uint64_t Elf64_Xword; +#endif +typedef uint16_t Elf64_Half; +typedef uint32_t Elf64_Word; +typedef uint8_t Elf64_Char; + +#define EI_NIDENT 16 + +typedef struct { + Elf32_Char e_ident[EI_NIDENT]; + Elf32_Half e_type; + Elf32_Half e_machine; + Elf32_Word e_version; + Elf32_Addr e_entry; /* Entry point */ + Elf32_Off e_phoff; + Elf32_Off e_shoff; + Elf32_Word e_flags; + Elf32_Half e_ehsize; + Elf32_Half e_phentsize; + Elf32_Half e_phnum; + Elf32_Half e_shentsize; + Elf32_Half e_shnum; + Elf32_Half e_shstrndx; +} Elf32_Ehdr; + +typedef struct { + Elf64_Char e_ident[EI_NIDENT]; + Elf64_Half e_type; + Elf64_Half e_machine; + Elf64_Word e_version; + Elf64_Addr e_entry; /* Entry point */ + Elf64_Off e_phoff; + Elf64_Off e_shoff; + Elf64_Word e_flags; + Elf64_Half e_ehsize; + Elf64_Half e_phentsize; + Elf64_Half e_phnum; + Elf64_Half e_shentsize; + Elf64_Half e_shnum; + Elf64_Half e_shstrndx; +} Elf64_Ehdr; + +/* e_type */ +#define ET_REL 1 +#define ET_EXEC 2 +#define ET_DYN 3 +#define ET_CORE 4 + +/* sh_type */ +#define SHT_SYMTAB 2 +#define SHT_NOTE 7 +#define SHT_DYNSYM 11 + +/* elf type */ +#define ELFDATANONE 0 /* e_ident[EI_DATA] */ +#define ELFDATA2LSB 1 +#define ELFDATA2MSB 2 + +/* elf class */ +#define ELFCLASSNONE 0 +#define ELFCLASS32 1 +#define ELFCLASS64 2 + +/* magic number */ +#define EI_MAG0 0 /* e_ident[] indexes */ +#define EI_MAG1 1 +#define EI_MAG2 2 +#define EI_MAG3 3 +#define EI_CLASS 4 +#define EI_DATA 5 +#define EI_VERSION 6 +#define EI_PAD 7 + +#define ELFMAG0 0x7f /* EI_MAG */ +#define ELFMAG1 'E' +#define ELFMAG2 'L' +#define ELFMAG3 'F' +#define ELFMAG "\177ELF" + +#define OLFMAG1 'O' +#define OLFMAG "\177OLF" + +typedef struct { + Elf32_Word p_type; + Elf32_Off p_offset; + Elf32_Addr p_vaddr; + Elf32_Addr p_paddr; + Elf32_Word p_filesz; + Elf32_Word p_memsz; + Elf32_Word p_flags; + Elf32_Word p_align; +} Elf32_Phdr; + +typedef struct { + Elf64_Word p_type; + Elf64_Word p_flags; + Elf64_Off p_offset; + Elf64_Addr p_vaddr; + Elf64_Addr p_paddr; + Elf64_Xword p_filesz; + Elf64_Xword p_memsz; + Elf64_Xword p_align; +} Elf64_Phdr; + +#define PT_NULL 0 /* p_type */ +#define PT_LOAD 1 +#define PT_DYNAMIC 2 +#define PT_INTERP 3 +#define PT_NOTE 4 +#define PT_SHLIB 5 +#define PT_PHDR 6 +#define PT_NUM 7 + +typedef struct { + Elf32_Word sh_name; + Elf32_Word sh_type; + Elf32_Word sh_flags; + Elf32_Addr sh_addr; + Elf32_Off sh_offset; + Elf32_Word sh_size; + Elf32_Word sh_link; + Elf32_Word sh_info; + Elf32_Word sh_addralign; + Elf32_Word sh_entsize; +} Elf32_Shdr; + +typedef struct { + Elf64_Word sh_name; + Elf64_Word sh_type; + Elf64_Off sh_flags; + Elf64_Addr sh_addr; + Elf64_Off sh_offset; + Elf64_Off sh_size; + Elf64_Word sh_link; + Elf64_Word sh_info; + Elf64_Off sh_addralign; + Elf64_Off sh_entsize; +} Elf64_Shdr; + +#define NT_NETBSD_CORE_PROCINFO 1 + +/* Note header in a PT_NOTE section */ +typedef struct elf_note { + Elf32_Word n_namesz; /* Name size */ + Elf32_Word n_descsz; /* Content size */ + Elf32_Word n_type; /* Content type */ +} Elf32_Nhdr; + +typedef struct { + Elf64_Word n_namesz; + Elf64_Word n_descsz; + Elf64_Word n_type; +} Elf64_Nhdr; + +/* Notes used in ET_CORE */ +#define NT_PRSTATUS 1 +#define NT_PRFPREG 2 +#define NT_PRPSINFO 3 +#define NT_PRXREG 4 +#define NT_TASKSTRUCT 4 +#define NT_PLATFORM 5 +#define NT_AUXV 6 + +/* Note types used in executables */ +/* NetBSD executables (name = "NetBSD") */ +#define NT_NETBSD_VERSION 1 +#define NT_NETBSD_EMULATION 2 +#define NT_FREEBSD_VERSION 1 +#define NT_OPENBSD_VERSION 1 +#define NT_DRAGONFLY_VERSION 1 +/* GNU executables (name = "GNU") */ +#define NT_GNU_VERSION 1 + +/* GNU OS tags */ +#define GNU_OS_LINUX 0 +#define GNU_OS_HURD 1 +#define GNU_OS_SOLARIS 2 +#define GNU_OS_KFREEBSD 3 +#define GNU_OS_KNETBSD 4 + +#endif diff --git a/ext/fileinfo/lib/softmagic.c b/ext/fileinfo/lib/softmagic.c new file mode 100644 index 0000000000..bca75cc92e --- /dev/null +++ b/ext/fileinfo/lib/softmagic.c @@ -0,0 +1,1825 @@ +/* + * Copyright (c) Ian F. Darwin 1986-1995. + * Software written by Ian F. Darwin and others; + * maintained 1995-present by Christos Zoulas and others. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice immediately at the beginning of the file, without modification, + * this list of conditions, and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR + * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + */ +/* + * softmagic - interpret variable magic from MAGIC + */ + +#include "file.h" +#include "magic.h" +#include +#include +#include +#include + + +#ifndef lint +FILE_RCSID("@(#)$File: softmagic.c,v 1.117 2008/03/01 22:21:49 rrt Exp $") +#endif /* lint */ + +private int match(struct magic_set *, struct magic *, uint32_t, + const unsigned char *, size_t, int); +private int mget(struct magic_set *, const unsigned char *, + struct magic *, size_t, unsigned int); +private int magiccheck(struct magic_set *, struct magic *); +private int32_t mprint(struct magic_set *, struct magic *); +private void mdebug(uint32_t, const char *, size_t); +private int mcopy(struct magic_set *, union VALUETYPE *, int, int, + const unsigned char *, uint32_t, size_t, size_t); +private int mconvert(struct magic_set *, struct magic *); +private int print_sep(struct magic_set *, int); +private void cvt_8(union VALUETYPE *, const struct magic *); +private void cvt_16(union VALUETYPE *, const struct magic *); +private void cvt_32(union VALUETYPE *, const struct magic *); +private void cvt_64(union VALUETYPE *, const struct magic *); + +/* + * Macro to give description string according to whether we want plain + * text or MIME type + */ +#define MAGIC_DESC ((ms->flags & MAGIC_MIME) ? m->mimetype : m->desc) + +/* + * softmagic - lookup one file in parsed, in-memory copy of database + * Passed the name and FILE * of one file to be typed. + */ +/*ARGSUSED1*/ /* nbytes passed for regularity, maybe need later */ +protected int +file_softmagic(struct magic_set *ms, const unsigned char *buf, size_t nbytes, int mode) +{ + struct mlist *ml; + int rv; + for (ml = ms->mlist->next; ml != ms->mlist; ml = ml->next) + if ((rv = match(ms, ml->magic, ml->nmagic, buf, nbytes, mode)) != 0) + return rv; + + return 0; +} + +/* + * Go through the whole list, stopping if you find a match. Process all + * the continuations of that match before returning. + * + * We support multi-level continuations: + * + * At any time when processing a successful top-level match, there is a + * current continuation level; it represents the level of the last + * successfully matched continuation. + * + * Continuations above that level are skipped as, if we see one, it + * means that the continuation that controls them - i.e, the + * lower-level continuation preceding them - failed to match. + * + * Continuations below that level are processed as, if we see one, + * it means we've finished processing or skipping higher-level + * continuations under the control of a successful or unsuccessful + * lower-level continuation, and are now seeing the next lower-level + * continuation and should process it. The current continuation + * level reverts to the level of the one we're seeing. + * + * Continuations at the current level are processed as, if we see + * one, there's no lower-level continuation that may have failed. + * + * If a continuation matches, we bump the current continuation level + * so that higher-level continuations are processed. + */ +private int +match(struct magic_set *ms, struct magic *magic, uint32_t nmagic, + const unsigned char *s, size_t nbytes, int mode) +{ + uint32_t magindex = 0; + unsigned int cont_level = 0; + int need_separator = 0; + int returnval = 0; /* if a match is found it is set to 1*/ + int firstline = 1; /* a flag to print X\n X\n- X */ + int printed_something = 0; + + if (file_check_mem(ms, cont_level) == -1) + return -1; + + for (magindex = 0; magindex < nmagic; magindex++) { + int flush; + struct magic *m = &magic[magindex]; + + if ((m->flag & BINTEST) != mode) { + /* Skip sub-tests */ + while (magic[magindex + 1].cont_level != 0 && + ++magindex < nmagic) + continue; + continue; /* Skip to next top-level test*/ + } + + ms->offset = m->offset; + ms->line = m->lineno; + + /* if main entry matches, print it... */ + flush = !mget(ms, s, m, nbytes, cont_level); + if (flush) { + if (m->reln == '!') + flush = 0; + } else { + switch (magiccheck(ms, m)) { + case -1: + return -1; + case 0: + flush++; + break; + default: + break; + } + } + if (flush) { + /* + * main entry didn't match, + * flush its continuations + */ + while (magindex < nmagic - 1 && + magic[magindex + 1].cont_level != 0) + magindex++; + continue; + } + + /* + * If we are going to print something, we'll need to print + * a blank before we print something else. + */ + if (*MAGIC_DESC) { + need_separator = 1; + printed_something = 1; + if (print_sep(ms, firstline) == -1) + return -1; + } + + if ((ms->c.li[cont_level].off = mprint(ms, m)) == -1) + return -1; + + /* and any continuations that match */ + if (file_check_mem(ms, ++cont_level) == -1) + return -1; + + while (magic[magindex+1].cont_level != 0 && + ++magindex < nmagic) { + m = &magic[magindex]; + ms->line = m->lineno; /* for messages */ + + if (cont_level < m->cont_level) + continue; + if (cont_level > m->cont_level) { + /* + * We're at the end of the level + * "cont_level" continuations. + */ + cont_level = m->cont_level; + } + ms->offset = m->offset; + if (m->flag & OFFADD) { + ms->offset += + ms->c.li[cont_level - 1].off; + } + +#ifdef ENABLE_CONDITIONALS + if (m->cond == COND_ELSE || + m->cond == COND_ELIF) { + if (ms->c.li[cont_level].last_match == 1) + continue; + } +#endif + flush = !mget(ms, s, m, nbytes, cont_level); + if (flush && m->reln != '!') + continue; + + switch (flush ? 1 : magiccheck(ms, m)) { + case -1: + return -1; + case 0: +#ifdef ENABLE_CONDITIONALS + ms->c.li[cont_level].last_match = 0; +#endif + break; + default: +#ifdef ENABLE_CONDITIONALS + ms->c.li[cont_level].last_match = 1; +#endif + if (m->type != FILE_DEFAULT) + ms->c.li[cont_level].got_match = 1; + else if (ms->c.li[cont_level].got_match) { + ms->c.li[cont_level].got_match = 0; + break; + } + /* + * If we are going to print something, + * make sure that we have a separator first. + */ + if (*MAGIC_DESC) { + printed_something = 1; + if (print_sep(ms, firstline) == -1) + return -1; + } + /* + * This continuation matched. Print + * its message, with a blank before it + * if the previous item printed and + * this item isn't empty. + */ + /* space if previous printed */ + if (need_separator + && ((m->flag & NOSPACE) == 0) + && *MAGIC_DESC) { + if (file_printf(ms, " ") == -1) + return -1; + need_separator = 0; + } + if ((ms->c.li[cont_level].off = mprint(ms, m)) == -1) + return -1; + if (*MAGIC_DESC) + need_separator = 1; + + /* + * If we see any continuations + * at a higher level, + * process them. + */ + if (file_check_mem(ms, ++cont_level) == -1) + return -1; + break; + } + } + if (printed_something) { + firstline = 0; + returnval = 1; + } + if ((ms->flags & MAGIC_CONTINUE) == 0 && printed_something) { + return 1; /* don't keep searching */ + } + } + return returnval; /* This is hit if -k is set or there is no match */ +} + +private int +check_fmt(struct magic_set *ms, struct magic *m) +{ + regex_t rx; + int rc; + + if (strchr(MAGIC_DESC, '%') == NULL) + return 0; + + rc = regcomp(&rx, "%[-0-9\\.]*s", REG_EXTENDED|REG_NOSUB); + if (rc) { + char errmsg[512]; + (void)regerror(rc, &rx, errmsg, sizeof(errmsg)); + file_magerror(ms, "regex error %d, (%s)", rc, errmsg); + return -1; + } else { + rc = regexec(&rx, MAGIC_DESC, 0, 0, 0); + regfree(&rx); + return !rc; + } +} + +#ifndef HAVE_STRNDUP +char * strndup(const char *, size_t); + +char * +strndup(const char *str, size_t n) +{ + size_t len; + char *copy; + + len = strlen(str); + if (len > n) + len = n; + if (!(copy = malloc(len + 1))) + return (NULL); + (void) memcpy(copy, str, len + 1); + copy[len] = '\0'; + return (copy); +} +#endif /* HAVE_STRNDUP */ + +private int32_t +mprint(struct magic_set *ms, struct magic *m) +{ + uint64_t v; + float vf; + double vd; + int64_t t = 0; + char *buf; + union VALUETYPE *p = &ms->ms_value; + + switch (m->type) { + case FILE_BYTE: + v = file_signextend(ms, m, (uint64_t)p->b); + switch (check_fmt(ms, m)) { + case -1: + return -1; + case 1: + if (asprintf(&buf, "%c", (unsigned char)v) < 0) + return -1; + if (file_printf(ms, MAGIC_DESC, buf) == -1) + return -1; + break; + default: + if (file_printf(ms, MAGIC_DESC, (unsigned char) v) == -1) + return -1; + break; + } + t = ms->offset + sizeof(char); + break; + + case FILE_SHORT: + case FILE_BESHORT: + case FILE_LESHORT: + v = file_signextend(ms, m, (uint64_t)p->h); + switch (check_fmt(ms, m)) { + case -1: + return -1; + case 1: + if (asprintf(&buf, "%hu", (unsigned short)v) < 0) + return -1; + if (file_printf(ms, MAGIC_DESC, buf) == -1) + return -1; + break; + default: + if (file_printf(ms, MAGIC_DESC, (unsigned short) v) == -1) + return -1; + break; + } + t = ms->offset + sizeof(short); + break; + + case FILE_LONG: + case FILE_BELONG: + case FILE_LELONG: + case FILE_MELONG: + v = file_signextend(ms, m, (uint64_t)p->l); + switch (check_fmt(ms, m)) { + case -1: + return -1; + case 1: + if (asprintf(&buf, "%u", (uint32_t)v) < 0) + return -1; + if (file_printf(ms, MAGIC_DESC, buf) == -1) + return -1; + break; + default: + if (file_printf(ms, MAGIC_DESC, (uint32_t) v) == -1) + return -1; + break; + } + t = ms->offset + sizeof(int32_t); + break; + + case FILE_QUAD: + case FILE_BEQUAD: + case FILE_LEQUAD: + v = file_signextend(ms, m, p->q); + if (file_printf(ms, MAGIC_DESC, (uint64_t) v) == -1) + return -1; + t = ms->offset + sizeof(int64_t); + break; + + case FILE_STRING: + case FILE_PSTRING: + case FILE_BESTRING16: + case FILE_LESTRING16: + if (m->reln == '=' || m->reln == '!') { + if (file_printf(ms, MAGIC_DESC, m->value.s) == -1) + return -1; + t = ms->offset + m->vallen; + } + else { + if (*m->value.s == '\0') + p->s[strcspn(p->s, "\n")] = '\0'; + if (file_printf(ms, MAGIC_DESC, p->s) == -1) + return -1; + t = ms->offset + strlen(p->s); + if (m->type == FILE_PSTRING) + t++; + } + break; + + case FILE_DATE: + case FILE_BEDATE: + case FILE_LEDATE: + case FILE_MEDATE: + if (file_printf(ms, MAGIC_DESC, file_fmttime(p->l, 1)) == -1) + return -1; + t = ms->offset + sizeof(time_t); + break; + + case FILE_LDATE: + case FILE_BELDATE: + case FILE_LELDATE: + case FILE_MELDATE: + if (file_printf(ms, MAGIC_DESC, file_fmttime(p->l, 0)) == -1) + return -1; + t = ms->offset + sizeof(time_t); + break; + + case FILE_QDATE: + case FILE_BEQDATE: + case FILE_LEQDATE: + if (file_printf(ms, MAGIC_DESC, file_fmttime((uint32_t)p->q, 1)) + == -1) + return -1; + t = ms->offset + sizeof(uint64_t); + break; + + case FILE_QLDATE: + case FILE_BEQLDATE: + case FILE_LEQLDATE: + if (file_printf(ms, MAGIC_DESC, file_fmttime((uint32_t)p->q, 0)) + == -1) + return -1; + t = ms->offset + sizeof(uint64_t); + break; + + case FILE_FLOAT: + case FILE_BEFLOAT: + case FILE_LEFLOAT: + vf = p->f; + switch (check_fmt(ms, m)) { + case -1: + return -1; + case 1: + if (asprintf(&buf, "%g", vf) < 0) + return -1; + if (file_printf(ms, MAGIC_DESC, buf) == -1) + return -1; + break; + default: + if (file_printf(ms, MAGIC_DESC, vf) == -1) + return -1; + break; + } + t = ms->offset + sizeof(float); + break; + + case FILE_DOUBLE: + case FILE_BEDOUBLE: + case FILE_LEDOUBLE: + vd = p->d; + switch (check_fmt(ms, m)) { + case -1: + return -1; + case 1: + if (asprintf(&buf, "%g", vd) < 0) + return -1; + if (file_printf(ms, MAGIC_DESC, buf) == -1) + return -1; + break; + default: + if (file_printf(ms, MAGIC_DESC, vd) == -1) + return -1; + break; + } + t = ms->offset + sizeof(double); + break; + + case FILE_REGEX: { + char *cp; + int rval; + + cp = strndup((const char *)ms->search.s, ms->search.rm_len); + if (cp == NULL) { + file_oomem(ms, ms->search.rm_len); + return -1; + } + rval = file_printf(ms, MAGIC_DESC, cp); + free(cp); + + if (rval == -1) + return -1; + + if ((m->str_flags & REGEX_OFFSET_START)) + t = ms->search.offset; + else + t = ms->search.offset + ms->search.rm_len; + break; + } + + case FILE_SEARCH: + if (file_printf(ms, MAGIC_DESC, m->value.s) == -1) + return -1; + if ((m->str_flags & REGEX_OFFSET_START)) + t = ms->search.offset; + else + t = ms->search.offset + m->vallen; + break; + + case FILE_DEFAULT: + if (file_printf(ms, MAGIC_DESC, m->value.s) == -1) + return -1; + t = ms->offset; + break; + + default: + file_magerror(ms, "invalid m->type (%d) in mprint()", m->type); + return -1; + } + return(t); +} + + +#define DO_CVT(fld, cast) \ + if (m->num_mask) \ + switch (m->mask_op & FILE_OPS_MASK) { \ + case FILE_OPAND: \ + p->fld &= cast m->num_mask; \ + break; \ + case FILE_OPOR: \ + p->fld |= cast m->num_mask; \ + break; \ + case FILE_OPXOR: \ + p->fld ^= cast m->num_mask; \ + break; \ + case FILE_OPADD: \ + p->fld += cast m->num_mask; \ + break; \ + case FILE_OPMINUS: \ + p->fld -= cast m->num_mask; \ + break; \ + case FILE_OPMULTIPLY: \ + p->fld *= cast m->num_mask; \ + break; \ + case FILE_OPDIVIDE: \ + p->fld /= cast m->num_mask; \ + break; \ + case FILE_OPMODULO: \ + p->fld %= cast m->num_mask; \ + break; \ + } \ + if (m->mask_op & FILE_OPINVERSE) \ + p->fld = ~p->fld \ + +private void +cvt_8(union VALUETYPE *p, const struct magic *m) +{ + DO_CVT(b, (uint8_t)); +} + +private void +cvt_16(union VALUETYPE *p, const struct magic *m) +{ + DO_CVT(h, (uint16_t)); +} + +private void +cvt_32(union VALUETYPE *p, const struct magic *m) +{ + DO_CVT(l, (uint32_t)); +} + +private void +cvt_64(union VALUETYPE *p, const struct magic *m) +{ + DO_CVT(q, (uint64_t)); +} + +#define DO_CVT2(fld, cast) \ + if (m->num_mask) \ + switch (m->mask_op & FILE_OPS_MASK) { \ + case FILE_OPADD: \ + p->fld += cast m->num_mask; \ + break; \ + case FILE_OPMINUS: \ + p->fld -= cast m->num_mask; \ + break; \ + case FILE_OPMULTIPLY: \ + p->fld *= cast m->num_mask; \ + break; \ + case FILE_OPDIVIDE: \ + p->fld /= cast m->num_mask; \ + break; \ + } \ + +private void +cvt_float(union VALUETYPE *p, const struct magic *m) +{ + DO_CVT2(f, (float)); +} + +private void +cvt_double(union VALUETYPE *p, const struct magic *m) +{ + DO_CVT2(d, (double)); +} + +/* + * Convert the byte order of the data we are looking at + * While we're here, let's apply the mask operation + * (unless you have a better idea) + */ +private int +mconvert(struct magic_set *ms, struct magic *m) +{ + union VALUETYPE *p = &ms->ms_value; + + switch (m->type) { + case FILE_BYTE: + cvt_8(p, m); + return 1; + case FILE_SHORT: + cvt_16(p, m); + return 1; + case FILE_LONG: + case FILE_DATE: + case FILE_LDATE: + cvt_32(p, m); + return 1; + case FILE_QUAD: + case FILE_QDATE: + case FILE_QLDATE: + cvt_64(p, m); + return 1; + case FILE_STRING: + case FILE_BESTRING16: + case FILE_LESTRING16: { + size_t len; + + /* Null terminate and eat *trailing* return */ + p->s[sizeof(p->s) - 1] = '\0'; + len = strlen(p->s); + if (len-- && p->s[len] == '\n') + p->s[len] = '\0'; + return 1; + } + case FILE_PSTRING: { + char *ptr1 = p->s, *ptr2 = ptr1 + 1; + size_t len = *p->s; + if (len >= sizeof(p->s)) + len = sizeof(p->s) - 1; + while (len--) + *ptr1++ = *ptr2++; + *ptr1 = '\0'; + len = strlen(p->s); + if (len-- && p->s[len] == '\n') + p->s[len] = '\0'; + return 1; + } + case FILE_BESHORT: + p->h = (short)((p->hs[0]<<8)|(p->hs[1])); + cvt_16(p, m); + return 1; + case FILE_BELONG: + case FILE_BEDATE: + case FILE_BELDATE: + p->l = (int32_t) + ((p->hl[0]<<24)|(p->hl[1]<<16)|(p->hl[2]<<8)|(p->hl[3])); + cvt_32(p, m); + return 1; + case FILE_BEQUAD: + case FILE_BEQDATE: + case FILE_BEQLDATE: + p->q = (uint64_t) + (((uint64_t)p->hq[0]<<56)|((uint64_t)p->hq[1]<<48)| + ((uint64_t)p->hq[2]<<40)|((uint64_t)p->hq[3]<<32)| + ((uint64_t)p->hq[4]<<24)|((uint64_t)p->hq[5]<<16)| + ((uint64_t)p->hq[6]<<8)|((uint64_t)p->hq[7])); + cvt_64(p, m); + return 1; + case FILE_LESHORT: + p->h = (short)((p->hs[1]<<8)|(p->hs[0])); + cvt_16(p, m); + return 1; + case FILE_LELONG: + case FILE_LEDATE: + case FILE_LELDATE: + p->l = (int32_t) + ((p->hl[3]<<24)|(p->hl[2]<<16)|(p->hl[1]<<8)|(p->hl[0])); + cvt_32(p, m); + return 1; + case FILE_LEQUAD: + case FILE_LEQDATE: + case FILE_LEQLDATE: + p->q = (uint64_t) + (((uint64_t)p->hq[7]<<56)|((uint64_t)p->hq[6]<<48)| + ((uint64_t)p->hq[5]<<40)|((uint64_t)p->hq[4]<<32)| + ((uint64_t)p->hq[3]<<24)|((uint64_t)p->hq[2]<<16)| + ((uint64_t)p->hq[1]<<8)|((uint64_t)p->hq[0])); + cvt_64(p, m); + return 1; + case FILE_MELONG: + case FILE_MEDATE: + case FILE_MELDATE: + p->l = (int32_t) + ((p->hl[1]<<24)|(p->hl[0]<<16)|(p->hl[3]<<8)|(p->hl[2])); + cvt_32(p, m); + return 1; + case FILE_FLOAT: + cvt_float(p, m); + return 1; + case FILE_BEFLOAT: + p->l = ((uint32_t)p->hl[0]<<24)|((uint32_t)p->hl[1]<<16)| + ((uint32_t)p->hl[2]<<8) |((uint32_t)p->hl[3]); + cvt_float(p, m); + return 1; + case FILE_LEFLOAT: + p->l = ((uint32_t)p->hl[3]<<24)|((uint32_t)p->hl[2]<<16)| + ((uint32_t)p->hl[1]<<8) |((uint32_t)p->hl[0]); + cvt_float(p, m); + return 1; + case FILE_DOUBLE: + cvt_double(p, m); + return 1; + case FILE_BEDOUBLE: + p->q = ((uint64_t)p->hq[0]<<56)|((uint64_t)p->hq[1]<<48)| + ((uint64_t)p->hq[2]<<40)|((uint64_t)p->hq[3]<<32)| + ((uint64_t)p->hq[4]<<24)|((uint64_t)p->hq[5]<<16)| + ((uint64_t)p->hq[6]<<8) |((uint64_t)p->hq[7]); + cvt_double(p, m); + return 1; + case FILE_LEDOUBLE: + p->q = ((uint64_t)p->hq[7]<<56)|((uint64_t)p->hq[6]<<48)| + ((uint64_t)p->hq[5]<<40)|((uint64_t)p->hq[4]<<32)| + ((uint64_t)p->hq[3]<<24)|((uint64_t)p->hq[2]<<16)| + ((uint64_t)p->hq[1]<<8) |((uint64_t)p->hq[0]); + cvt_double(p, m); + return 1; + case FILE_REGEX: + case FILE_SEARCH: + case FILE_DEFAULT: + return 1; + default: + file_magerror(ms, "invalid type %d in mconvert()", m->type); + return 0; + } +} + + +private void +mdebug(uint32_t offset, const char *str, size_t len) +{ + (void) fprintf(stderr, "mget @%d: ", offset); + file_showstr(stderr, str, len); + (void) fputc('\n', stderr); + (void) fputc('\n', stderr); +} + +private int +mcopy(struct magic_set *ms, union VALUETYPE *p, int type, int indir, + const unsigned char *s, uint32_t offset, size_t nbytes, size_t linecnt) +{ + /* + * Note: FILE_SEARCH and FILE_REGEX do not actually copy + * anything, but setup pointers into the source + */ + if (indir == 0) { + switch (type) { + case FILE_SEARCH: + ms->search.s = (const char *)s + offset; + ms->search.s_len = nbytes - offset; + ms->search.offset = offset; + return 0; + + case FILE_REGEX: { + const char *b; + const char *c; + const char *last; /* end of search region */ + const char *buf; /* start of search region */ + size_t lines; + + if (s == NULL) { + ms->search.s_len = 0; + ms->search.s = NULL; + return 0; + } + buf = (const char *)s + offset; + last = (const char *)s + nbytes; + /* mget() guarantees buf <= last */ + for (lines = linecnt, b = buf; + lines && ((b = strchr(c = b, '\n')) || (b = strchr(c, '\r'))); + lines--, b++) { + last = b; + if (b[0] == '\r' && b[1] == '\n') + b++; + } + if (lines) + last = (const char *)s + nbytes; + + ms->search.s = buf; + ms->search.s_len = last - buf; + ms->search.offset = offset; + ms->search.rm_len = 0; + return 0; + } + case FILE_BESTRING16: + case FILE_LESTRING16: { + const unsigned char *src = s + offset; + const unsigned char *esrc = s + nbytes; + char *dst = p->s; + char *edst = &p->s[sizeof(p->s) - 1]; + + if (type == FILE_BESTRING16) + src++; + + /* check for pointer overflow */ + if (src < s) { + file_magerror(ms, "invalid offset %zu in mcopy()", + offset); + return -1; + } + for (/*EMPTY*/; src < esrc; src += 2, dst++) { + if (dst < edst) + *dst = *src; + else + break; + if (*dst == '\0') { + if (type == FILE_BESTRING16 ? + *(src - 1) != '\0' : + *(src + 1) != '\0') + *dst = ' '; + } + } + *edst = '\0'; + return 0; + } + case FILE_STRING: /* XXX - these two should not need */ + case FILE_PSTRING: /* to copy anything, but do anyway. */ + default: + break; + } + } + + if (offset >= nbytes) { + (void)memset(p, '\0', sizeof(*p)); + return 0; + } + if (nbytes - offset < sizeof(*p)) + nbytes = nbytes - offset; + else + nbytes = sizeof(*p); + + (void)memcpy(p, s + offset, nbytes); + + /* + * the usefulness of padding with zeroes eludes me, it + * might even cause problems + */ + if (nbytes < sizeof(*p)) + (void)memset(((char *)(void *)p) + nbytes, '\0', + sizeof(*p) - nbytes); + return 0; +} + +private int +mget(struct magic_set *ms, const unsigned char *s, + struct magic *m, size_t nbytes, unsigned int cont_level) +{ + uint32_t offset = ms->offset; + uint32_t count = m->str_range; + union VALUETYPE *p = &ms->ms_value; + + if (mcopy(ms, p, m->type, m->flag & INDIR, s, offset, nbytes, count) == -1) + return -1; + + if ((ms->flags & MAGIC_DEBUG) != 0) { + mdebug(offset, (char *)(void *)p, sizeof(union VALUETYPE)); + file_mdump(m); + } + + if (m->flag & INDIR) { + int off = m->in_offset; + if (m->in_op & FILE_OPINDIRECT) { + const union VALUETYPE *q = + ((const void *)(s + offset + off)); + switch (m->in_type) { + case FILE_BYTE: + off = q->b; + break; + case FILE_SHORT: + off = q->h; + break; + case FILE_BESHORT: + off = (short)((q->hs[0]<<8)|(q->hs[1])); + break; + case FILE_LESHORT: + off = (short)((q->hs[1]<<8)|(q->hs[0])); + break; + case FILE_LONG: + off = q->l; + break; + case FILE_BELONG: + off = (int32_t)((q->hl[0]<<24)|(q->hl[1]<<16)| + (q->hl[2]<<8)|(q->hl[3])); + break; + case FILE_LELONG: + off = (int32_t)((q->hl[3]<<24)|(q->hl[2]<<16)| + (q->hl[1]<<8)|(q->hl[0])); + break; + case FILE_MELONG: + off = (int32_t)((q->hl[1]<<24)|(q->hl[0]<<16)| + (q->hl[3]<<8)|(q->hl[2])); + break; + } + } + switch (m->in_type) { + case FILE_BYTE: + if (nbytes < (offset + 1)) + return 0; + if (off) { + switch (m->in_op & FILE_OPS_MASK) { + case FILE_OPAND: + offset = p->b & off; + break; + case FILE_OPOR: + offset = p->b | off; + break; + case FILE_OPXOR: + offset = p->b ^ off; + break; + case FILE_OPADD: + offset = p->b + off; + break; + case FILE_OPMINUS: + offset = p->b - off; + break; + case FILE_OPMULTIPLY: + offset = p->b * off; + break; + case FILE_OPDIVIDE: + offset = p->b / off; + break; + case FILE_OPMODULO: + offset = p->b % off; + break; + } + } else + offset = p->b; + if (m->in_op & FILE_OPINVERSE) + offset = ~offset; + break; + case FILE_BESHORT: + if (nbytes < (offset + 2)) + return 0; + if (off) { + switch (m->in_op & FILE_OPS_MASK) { + case FILE_OPAND: + offset = (short)((p->hs[0]<<8)| + (p->hs[1])) & + off; + break; + case FILE_OPOR: + offset = (short)((p->hs[0]<<8)| + (p->hs[1])) | + off; + break; + case FILE_OPXOR: + offset = (short)((p->hs[0]<<8)| + (p->hs[1])) ^ + off; + break; + case FILE_OPADD: + offset = (short)((p->hs[0]<<8)| + (p->hs[1])) + + off; + break; + case FILE_OPMINUS: + offset = (short)((p->hs[0]<<8)| + (p->hs[1])) - + off; + break; + case FILE_OPMULTIPLY: + offset = (short)((p->hs[0]<<8)| + (p->hs[1])) * + off; + break; + case FILE_OPDIVIDE: + offset = (short)((p->hs[0]<<8)| + (p->hs[1])) / + off; + break; + case FILE_OPMODULO: + offset = (short)((p->hs[0]<<8)| + (p->hs[1])) % + off; + break; + } + } else + offset = (short)((p->hs[0]<<8)| + (p->hs[1])); + if (m->in_op & FILE_OPINVERSE) + offset = ~offset; + break; + case FILE_LESHORT: + if (nbytes < (offset + 2)) + return 0; + if (off) { + switch (m->in_op & FILE_OPS_MASK) { + case FILE_OPAND: + offset = (short)((p->hs[1]<<8)| + (p->hs[0])) & + off; + break; + case FILE_OPOR: + offset = (short)((p->hs[1]<<8)| + (p->hs[0])) | + off; + break; + case FILE_OPXOR: + offset = (short)((p->hs[1]<<8)| + (p->hs[0])) ^ + off; + break; + case FILE_OPADD: + offset = (short)((p->hs[1]<<8)| + (p->hs[0])) + + off; + break; + case FILE_OPMINUS: + offset = (short)((p->hs[1]<<8)| + (p->hs[0])) - + off; + break; + case FILE_OPMULTIPLY: + offset = (short)((p->hs[1]<<8)| + (p->hs[0])) * + off; + break; + case FILE_OPDIVIDE: + offset = (short)((p->hs[1]<<8)| + (p->hs[0])) / + off; + break; + case FILE_OPMODULO: + offset = (short)((p->hs[1]<<8)| + (p->hs[0])) % + off; + break; + } + } else + offset = (short)((p->hs[1]<<8)| + (p->hs[0])); + if (m->in_op & FILE_OPINVERSE) + offset = ~offset; + break; + case FILE_SHORT: + if (nbytes < (offset + 2)) + return 0; + if (off) { + switch (m->in_op & FILE_OPS_MASK) { + case FILE_OPAND: + offset = p->h & off; + break; + case FILE_OPOR: + offset = p->h | off; + break; + case FILE_OPXOR: + offset = p->h ^ off; + break; + case FILE_OPADD: + offset = p->h + off; + break; + case FILE_OPMINUS: + offset = p->h - off; + break; + case FILE_OPMULTIPLY: + offset = p->h * off; + break; + case FILE_OPDIVIDE: + offset = p->h / off; + break; + case FILE_OPMODULO: + offset = p->h % off; + break; + } + } + else + offset = p->h; + if (m->in_op & FILE_OPINVERSE) + offset = ~offset; + break; + case FILE_BELONG: + if (nbytes < (offset + 4)) + return 0; + if (off) { + switch (m->in_op & FILE_OPS_MASK) { + case FILE_OPAND: + offset = (int32_t)((p->hl[0]<<24)| + (p->hl[1]<<16)| + (p->hl[2]<<8)| + (p->hl[3])) & + off; + break; + case FILE_OPOR: + offset = (int32_t)((p->hl[0]<<24)| + (p->hl[1]<<16)| + (p->hl[2]<<8)| + (p->hl[3])) | + off; + break; + case FILE_OPXOR: + offset = (int32_t)((p->hl[0]<<24)| + (p->hl[1]<<16)| + (p->hl[2]<<8)| + (p->hl[3])) ^ + off; + break; + case FILE_OPADD: + offset = (int32_t)((p->hl[0]<<24)| + (p->hl[1]<<16)| + (p->hl[2]<<8)| + (p->hl[3])) + + off; + break; + case FILE_OPMINUS: + offset = (int32_t)((p->hl[0]<<24)| + (p->hl[1]<<16)| + (p->hl[2]<<8)| + (p->hl[3])) - + off; + break; + case FILE_OPMULTIPLY: + offset = (int32_t)((p->hl[0]<<24)| + (p->hl[1]<<16)| + (p->hl[2]<<8)| + (p->hl[3])) * + off; + break; + case FILE_OPDIVIDE: + offset = (int32_t)((p->hl[0]<<24)| + (p->hl[1]<<16)| + (p->hl[2]<<8)| + (p->hl[3])) / + off; + break; + case FILE_OPMODULO: + offset = (int32_t)((p->hl[0]<<24)| + (p->hl[1]<<16)| + (p->hl[2]<<8)| + (p->hl[3])) % + off; + break; + } + } else + offset = (int32_t)((p->hl[0]<<24)| + (p->hl[1]<<16)| + (p->hl[2]<<8)| + (p->hl[3])); + if (m->in_op & FILE_OPINVERSE) + offset = ~offset; + break; + case FILE_LELONG: + if (nbytes < (offset + 4)) + return 0; + if (off) { + switch (m->in_op & FILE_OPS_MASK) { + case FILE_OPAND: + offset = (int32_t)((p->hl[3]<<24)| + (p->hl[2]<<16)| + (p->hl[1]<<8)| + (p->hl[0])) & + off; + break; + case FILE_OPOR: + offset = (int32_t)((p->hl[3]<<24)| + (p->hl[2]<<16)| + (p->hl[1]<<8)| + (p->hl[0])) | + off; + break; + case FILE_OPXOR: + offset = (int32_t)((p->hl[3]<<24)| + (p->hl[2]<<16)| + (p->hl[1]<<8)| + (p->hl[0])) ^ + off; + break; + case FILE_OPADD: + offset = (int32_t)((p->hl[3]<<24)| + (p->hl[2]<<16)| + (p->hl[1]<<8)| + (p->hl[0])) + + off; + break; + case FILE_OPMINUS: + offset = (int32_t)((p->hl[3]<<24)| + (p->hl[2]<<16)| + (p->hl[1]<<8)| + (p->hl[0])) - + off; + break; + case FILE_OPMULTIPLY: + offset = (int32_t)((p->hl[3]<<24)| + (p->hl[2]<<16)| + (p->hl[1]<<8)| + (p->hl[0])) * + off; + break; + case FILE_OPDIVIDE: + offset = (int32_t)((p->hl[3]<<24)| + (p->hl[2]<<16)| + (p->hl[1]<<8)| + (p->hl[0])) / + off; + break; + case FILE_OPMODULO: + offset = (int32_t)((p->hl[3]<<24)| + (p->hl[2]<<16)| + (p->hl[1]<<8)| + (p->hl[0])) % + off; + break; + } + } else + offset = (int32_t)((p->hl[3]<<24)| + (p->hl[2]<<16)| + (p->hl[1]<<8)| + (p->hl[0])); + if (m->in_op & FILE_OPINVERSE) + offset = ~offset; + break; + case FILE_MELONG: + if (nbytes < (offset + 4)) + return 0; + if (off) { + switch (m->in_op & FILE_OPS_MASK) { + case FILE_OPAND: + offset = (int32_t)((p->hl[1]<<24)| + (p->hl[0]<<16)| + (p->hl[3]<<8)| + (p->hl[2])) & + off; + break; + case FILE_OPOR: + offset = (int32_t)((p->hl[1]<<24)| + (p->hl[0]<<16)| + (p->hl[3]<<8)| + (p->hl[2])) | + off; + break; + case FILE_OPXOR: + offset = (int32_t)((p->hl[1]<<24)| + (p->hl[0]<<16)| + (p->hl[3]<<8)| + (p->hl[2])) ^ + off; + break; + case FILE_OPADD: + offset = (int32_t)((p->hl[1]<<24)| + (p->hl[0]<<16)| + (p->hl[3]<<8)| + (p->hl[2])) + + off; + break; + case FILE_OPMINUS: + offset = (int32_t)((p->hl[1]<<24)| + (p->hl[0]<<16)| + (p->hl[3]<<8)| + (p->hl[2])) - + off; + break; + case FILE_OPMULTIPLY: + offset = (int32_t)((p->hl[1]<<24)| + (p->hl[0]<<16)| + (p->hl[3]<<8)| + (p->hl[2])) * + off; + break; + case FILE_OPDIVIDE: + offset = (int32_t)((p->hl[1]<<24)| + (p->hl[0]<<16)| + (p->hl[3]<<8)| + (p->hl[2])) / + off; + break; + case FILE_OPMODULO: + offset = (int32_t)((p->hl[1]<<24)| + (p->hl[0]<<16)| + (p->hl[3]<<8)| + (p->hl[2])) % + off; + break; + } + } else + offset = (int32_t)((p->hl[1]<<24)| + (p->hl[0]<<16)| + (p->hl[3]<<8)| + (p->hl[2])); + if (m->in_op & FILE_OPINVERSE) + offset = ~offset; + break; + case FILE_LONG: + if (nbytes < (offset + 4)) + return 0; + if (off) { + switch (m->in_op & FILE_OPS_MASK) { + case FILE_OPAND: + offset = p->l & off; + break; + case FILE_OPOR: + offset = p->l | off; + break; + case FILE_OPXOR: + offset = p->l ^ off; + break; + case FILE_OPADD: + offset = p->l + off; + break; + case FILE_OPMINUS: + offset = p->l - off; + break; + case FILE_OPMULTIPLY: + offset = p->l * off; + break; + case FILE_OPDIVIDE: + offset = p->l / off; + break; + case FILE_OPMODULO: + offset = p->l % off; + break; + } + } else + offset = p->l; + if (m->in_op & FILE_OPINVERSE) + offset = ~offset; + break; + } + + if (m->flag & INDIROFFADD) + offset += ms->c.li[cont_level-1].off; + if (mcopy(ms, p, m->type, 0, s, offset, nbytes, count) == -1) + return -1; + ms->offset = offset; + + if ((ms->flags & MAGIC_DEBUG) != 0) { + mdebug(offset, (char *)(void *)p, + sizeof(union VALUETYPE)); + file_mdump(m); + } + } + + /* Verify we have enough data to match magic type */ + switch (m->type) { + case FILE_BYTE: + if (nbytes < (offset + 1)) /* should alway be true */ + return 0; + break; + + case FILE_SHORT: + case FILE_BESHORT: + case FILE_LESHORT: + if (nbytes < (offset + 2)) + return 0; + break; + + case FILE_LONG: + case FILE_BELONG: + case FILE_LELONG: + case FILE_MELONG: + case FILE_DATE: + case FILE_BEDATE: + case FILE_LEDATE: + case FILE_MEDATE: + case FILE_LDATE: + case FILE_BELDATE: + case FILE_LELDATE: + case FILE_MELDATE: + case FILE_FLOAT: + case FILE_BEFLOAT: + case FILE_LEFLOAT: + if (nbytes < (offset + 4)) + return 0; + break; + + case FILE_DOUBLE: + case FILE_BEDOUBLE: + case FILE_LEDOUBLE: + if (nbytes < (offset + 8)) + return 0; + break; + + case FILE_STRING: + case FILE_PSTRING: + case FILE_SEARCH: + if (nbytes < (offset + m->vallen)) + return 0; + break; + + case FILE_REGEX: + if (nbytes < offset) + return 0; + break; + + case FILE_DEFAULT: /* nothing to check */ + default: + break; + } + if (!mconvert(ms, m)) + return 0; + return 1; +} + +private uint64_t +file_strncmp(const char *s1, const char *s2, size_t len, uint32_t flags) +{ + /* + * Convert the source args to unsigned here so that (1) the + * compare will be unsigned as it is in strncmp() and (2) so + * the ctype functions will work correctly without extra + * casting. + */ + const unsigned char *a = (const unsigned char *)s1; + const unsigned char *b = (const unsigned char *)s2; + uint64_t v; + + /* + * What we want here is v = strncmp(s1, s2, len), + * but ignoring any nulls. + */ + v = 0; + if (0L == flags) { /* normal string: do it fast */ + while (len-- > 0) + if ((v = *b++ - *a++) != '\0') + break; + } + else { /* combine the others */ + while (len-- > 0) { + if ((flags & STRING_IGNORE_LOWERCASE) && + islower(*a)) { + if ((v = tolower(*b++) - *a++) != '\0') + break; + } + else if ((flags & STRING_IGNORE_UPPERCASE) && + isupper(*a)) { + if ((v = toupper(*b++) - *a++) != '\0') + break; + } + else if ((flags & STRING_COMPACT_BLANK) && + isspace(*a)) { + a++; + if (isspace(*b++)) { + while (isspace(*b)) + b++; + } + else { + v = 1; + break; + } + } + else if ((flags & STRING_COMPACT_OPTIONAL_BLANK) && + isspace(*a)) { + a++; + while (isspace(*b)) + b++; + } + else { + if ((v = *b++ - *a++) != '\0') + break; + } + } + } + return v; +} + +private uint64_t +file_strncmp16(const char *a, const char *b, size_t len, uint32_t flags) +{ + /* + * XXX - The 16-bit string compare probably needs to be done + * differently, especially if the flags are to be supported. + * At the moment, I am unsure. + */ + flags = 0; + return file_strncmp(a, b, len, flags); +} + +private int +magiccheck(struct magic_set *ms, struct magic *m) +{ + uint64_t l = m->value.q; + uint64_t v; + float fl, fv; + double dl, dv; + int matched; + union VALUETYPE *p = &ms->ms_value; + + switch (m->type) { + case FILE_BYTE: + v = p->b; + break; + + case FILE_SHORT: + case FILE_BESHORT: + case FILE_LESHORT: + v = p->h; + break; + + case FILE_LONG: + case FILE_BELONG: + case FILE_LELONG: + case FILE_MELONG: + case FILE_DATE: + case FILE_BEDATE: + case FILE_LEDATE: + case FILE_MEDATE: + case FILE_LDATE: + case FILE_BELDATE: + case FILE_LELDATE: + case FILE_MELDATE: + v = p->l; + break; + + case FILE_QUAD: + case FILE_LEQUAD: + case FILE_BEQUAD: + case FILE_QDATE: + case FILE_BEQDATE: + case FILE_LEQDATE: + case FILE_QLDATE: + case FILE_BEQLDATE: + case FILE_LEQLDATE: + v = p->q; + break; + + case FILE_FLOAT: + case FILE_BEFLOAT: + case FILE_LEFLOAT: + fl = m->value.f; + fv = p->f; + switch (m->reln) { + case 'x': + matched = 1; + break; + + case '!': + matched = fv != fl; + break; + + case '=': + matched = fv == fl; + break; + + case '>': + matched = fv > fl; + break; + + case '<': + matched = fv < fl; + break; + + default: + matched = 0; + file_magerror(ms, "cannot happen with float: invalid relation `%c'", m->reln); + return -1; + } + return matched; + + case FILE_DOUBLE: + case FILE_BEDOUBLE: + case FILE_LEDOUBLE: + dl = m->value.d; + dv = p->d; + switch (m->reln) { + case 'x': + matched = 1; + break; + + case '!': + matched = dv != dl; + break; + + case '=': + matched = dv == dl; + break; + + case '>': + matched = dv > dl; + break; + + case '<': + matched = dv < dl; + break; + + default: + matched = 0; + file_magerror(ms, "cannot happen with double: invalid relation `%c'", m->reln); + return -1; + } + return matched; + + case FILE_DEFAULT: + l = 0; + v = 0; + break; + + case FILE_STRING: + case FILE_PSTRING: + l = 0; + v = file_strncmp(m->value.s, p->s, (size_t)m->vallen, m->str_flags); + break; + + case FILE_BESTRING16: + case FILE_LESTRING16: + l = 0; + v = file_strncmp16(m->value.s, p->s, (size_t)m->vallen, m->str_flags); + break; + + case FILE_SEARCH: { /* search ms->search.s for the string m->value.s */ + size_t slen; + size_t idx; + + if (ms->search.s == NULL) + return 0; + + slen = MIN(m->vallen, sizeof(m->value.s)); + l = 0; + v = 0; + + for (idx = 0; m->str_range == 0 || idx < m->str_range; idx++) { + if (slen + idx > ms->search.s_len) + break; + + v = file_strncmp(m->value.s, ms->search.s + idx, slen, m->str_flags); + if (v == 0) { /* found match */ + ms->search.offset += idx; + break; + } + } + break; + } + case FILE_REGEX: { + int rc; + regex_t rx; + char errmsg[512]; + + if (ms->search.s == NULL) + return 0; + + l = 0; + rc = regcomp(&rx, m->value.s, + REG_EXTENDED|REG_NEWLINE| + ((m->str_flags & STRING_IGNORE_CASE) ? REG_ICASE : 0)); + if (rc) { + (void)regerror(rc, &rx, errmsg, sizeof(errmsg)); + file_magerror(ms, "regex error %d, (%s)", + rc, errmsg); + v = (uint64_t)-1; + } + else { + regmatch_t pmatch[1]; +#ifndef REG_STARTEND +#define REG_STARTEND 0 + size_t l = ms->search.s_len - 1; + char c = ms->search.s[l]; + ((char *)(intptr_t)ms->search.s)[l] = '\0'; +#else + pmatch[0].rm_so = 0; + pmatch[0].rm_eo = ms->search.s_len; +#endif + rc = regexec(&rx, (const char *)ms->search.s, + 1, pmatch, REG_STARTEND); +#if REG_STARTEND == 0 + ((char *)(intptr_t)ms->search.s)[l] = c; +#endif + switch (rc) { + case 0: + ms->search.s += (int)pmatch[0].rm_so; + ms->search.offset += (size_t)pmatch[0].rm_so; + ms->search.rm_len = + (size_t)(pmatch[0].rm_eo - pmatch[0].rm_so); + v = 0; + break; + + case REG_NOMATCH: + v = 1; + break; + + default: + (void)regerror(rc, &rx, errmsg, sizeof(errmsg)); + file_magerror(ms, "regexec error %d, (%s)", + rc, errmsg); + v = (uint64_t)-1; + break; + } + regfree(&rx); + } + if (v == (uint64_t)-1) + return -1; + break; + } + default: + file_magerror(ms, "invalid type %d in magiccheck()", m->type); + return -1; + } + + v = file_signextend(ms, m, v); + + switch (m->reln) { + case 'x': + if ((ms->flags & MAGIC_DEBUG) != 0) + (void) fprintf(stderr, "%llu == *any* = 1\n", + (unsigned long long)v); + matched = 1; + break; + + case '!': + matched = v != l; + if ((ms->flags & MAGIC_DEBUG) != 0) + (void) fprintf(stderr, "%llu != %llu = %d\n", + (unsigned long long)v, (unsigned long long)l, + matched); + break; + + case '=': + matched = v == l; + if ((ms->flags & MAGIC_DEBUG) != 0) + (void) fprintf(stderr, "%llu == %llu = %d\n", + (unsigned long long)v, (unsigned long long)l, + matched); + break; + + case '>': + if (m->flag & UNSIGNED) { + matched = v > l; + if ((ms->flags & MAGIC_DEBUG) != 0) + (void) fprintf(stderr, "%llu > %llu = %d\n", + (unsigned long long)v, + (unsigned long long)l, matched); + } + else { + matched = (int64_t) v > (int64_t) l; + if ((ms->flags & MAGIC_DEBUG) != 0) + (void) fprintf(stderr, "%lld > %lld = %d\n", + (long long)v, (long long)l, matched); + } + break; + + case '<': + if (m->flag & UNSIGNED) { + matched = v < l; + if ((ms->flags & MAGIC_DEBUG) != 0) + (void) fprintf(stderr, "%llu < %llu = %d\n", + (unsigned long long)v, + (unsigned long long)l, matched); + } + else { + matched = (int64_t) v < (int64_t) l; + if ((ms->flags & MAGIC_DEBUG) != 0) + (void) fprintf(stderr, "%lld < %lld = %d\n", + (long long)v, (long long)l, matched); + } + break; + + case '&': + matched = (v & l) == l; + if ((ms->flags & MAGIC_DEBUG) != 0) + (void) fprintf(stderr, "((%llx & %llx) == %llx) = %d\n", + (unsigned long long)v, (unsigned long long)l, + (unsigned long long)l, matched); + break; + + case '^': + matched = (v & l) != l; + if ((ms->flags & MAGIC_DEBUG) != 0) + (void) fprintf(stderr, "((%llx & %llx) != %llx) = %d\n", + (unsigned long long)v, (unsigned long long)l, + (unsigned long long)l, matched); + break; + + default: + matched = 0; + file_magerror(ms, "cannot happen: invalid relation `%c'", + m->reln); + return -1; + } + + return matched; +} + +private int +print_sep(struct magic_set *ms, int firstline) +{ + if (firstline) + return 0; + /* + * we found another match + * put a newline and '-' to do some simple formatting + */ + return file_printf(ms, "\n- "); +} diff --git a/ext/fileinfo/lib/tar.h b/ext/fileinfo/lib/tar.h new file mode 100644 index 0000000000..fa2390ab7d --- /dev/null +++ b/ext/fileinfo/lib/tar.h @@ -0,0 +1,73 @@ +/* + * Copyright (c) Ian F. Darwin 1986-1995. + * Software written by Ian F. Darwin and others; + * maintained 1995-present by Christos Zoulas and others. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice immediately at the beginning of the file, without modification, + * this list of conditions, and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR + * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + */ +/* + * Header file for public domain tar (tape archive) program. + * + * @(#)tar.h 1.20 86/10/29 Public Domain. + * + * Created 25 August 1985 by John Gilmore, ihnp4!hoptoad!gnu. + * + * $File: tar.h,v 1.12 2008/02/07 00:58:52 christos Exp $ # checkin only + */ + +/* + * Header block on tape. + * + * I'm going to use traditional DP naming conventions here. + * A "block" is a big chunk of stuff that we do I/O on. + * A "record" is a piece of info that we care about. + * Typically many "record"s fit into a "block". + */ +#define RECORDSIZE 512 +#define NAMSIZ 100 +#define TUNMLEN 32 +#define TGNMLEN 32 + +union record { + char charptr[RECORDSIZE]; + struct header { + char name[NAMSIZ]; + char mode[8]; + char uid[8]; + char gid[8]; + char size[12]; + char mtime[12]; + char chksum[8]; + char linkflag; + char linkname[NAMSIZ]; + char magic[8]; + char uname[TUNMLEN]; + char gname[TGNMLEN]; + char devmajor[8]; + char devminor[8]; + } header; +}; + +/* The magic field is filled with this if uname and gname are valid. */ +#define TMAGIC "ustar" /* 5 chars and a null */ +#define GNUTMAGIC "ustar " /* 7 chars and a null */