static int
fmm_ascmagic(unsigned char *buf, size_t nbytes, char **mime_type)
{
int has_escapes = 0;
unsigned char *s;
char nbuf[HOWMANY + 1]; /* one extra for terminating '\0' */
char *token;
register struct names *p;
int small_nbytes;
char *strtok_state;
unsigned char *tp;
/* these are easy, do them first */
/*
* for troff, look for . + letter + letter or .\"; this must be done to
* disambiguate tar archives' ./file and other trash from real troff
* input.
*/
if (*buf == '.') {
tp = buf + 1;
while (isSPACE(*tp))
++tp; /* skip leading whitespace */
if ((isALNUM(*tp) || *tp == '\\') && (isALNUM(*(tp + 1)) || *tp == '"')) {
strcpy(*mime_type, "application/x-troff");
return 0;
}
}
if ((*buf == 'c' || *buf == 'C') && isSPACE(*(buf + 1))) {
/* Fortran */
strcpy(*mime_type, "text/plain");
return 0;
}
/* look for tokens from names.h - this is expensive!, so we'll limit
* ourselves to only SMALL_HOWMANY bytes */
small_nbytes = (nbytes > SMALL_HOWMANY) ? SMALL_HOWMANY : nbytes;
/* make a copy of the buffer here because strtok() will destroy it */
s = (unsigned char *) memcpy(nbuf, buf, small_nbytes);
s[small_nbytes] = '\0';
has_escapes = (memchr(s, '\033', small_nbytes) != NULL);
while ((token = strtok_r((char *) s, " \t\n\r\f", &strtok_state)) != NULL) {
s = NULL; /* make strtok() keep on tokin' */
for (p = names; p < names + NNAMES; p++) {
if (strEQ(p->name, token)) {
strcpy(*mime_type, types[p->type]);
if (has_escapes)
strcat(*mime_type, " (with escape sequences)");
return 0;
}
}
}
int is_tarball = is_tar(buf, nbytes);
if ( is_tarball == 1 || is_tarball == 2 ) {
/* 1: V7 tar archive */
/* 2: POSIX tar archive */
strcpy(*mime_type, "application/x-tar");
return 0;
}
/* all else fails, but it is ascii... */
strcpy(*mime_type, "text/plain");
return 0;
}
cli_file_t cli_filetype2(int desc, const struct cl_engine *engine)
{
unsigned char smallbuff[MAGIC_BUFFER_SIZE + 1], *decoded, *bigbuff;
int bread, sret;
cli_file_t ret = CL_TYPE_UNKNOWN_DATA;
struct cli_matcher *root;
struct cli_ac_data mdata;
memset(smallbuff, 0, sizeof(smallbuff));
if((bread = read(desc, smallbuff, MAGIC_BUFFER_SIZE)) > 0)
ret = cli_filetype(smallbuff, bread);
if(engine && ret == CL_TYPE_UNKNOWN_TEXT) {
root = engine->root[0];
if(!root)
return ret;
if(cli_ac_initdata(&mdata, root->ac_partsigs, AC_DEFAULT_TRACKLEN))
return ret;
sret = cli_ac_scanbuff(smallbuff, bread, NULL, engine->root[0], &mdata, 1, 0, 0, -1, NULL);
cli_ac_freedata(&mdata);
if(sret >= CL_TYPENO) {
ret = sret;
} else {
if(cli_ac_initdata(&mdata, root->ac_partsigs, AC_DEFAULT_TRACKLEN))
return ret;
decoded = (unsigned char *) cli_utf16toascii((char *) smallbuff, bread);
if(decoded) {
sret = cli_ac_scanbuff(decoded, strlen((char *) decoded), NULL, engine->root[0], &mdata, 1, 0, 0, -1, NULL);
free(decoded);
if(sret == CL_TYPE_HTML)
ret = CL_TYPE_HTML_UTF16;
}
cli_ac_freedata(&mdata);
if((((struct cli_dconf*) engine->dconf)->phishing & PHISHING_CONF_ENTCONV) && ret != CL_TYPE_HTML_UTF16) {
struct entity_conv conv;
const size_t conv_size = 2*bread < 256 ? 256 : 2*bread;
if(init_entity_converter(&conv,UNKNOWN,conv_size) == 0) {
int end = 0;
m_area_t area;
area.buffer = (unsigned char *) smallbuff;
area.length = bread;
area.offset = 0;
while(!end) {
if(cli_ac_initdata(&mdata, root->ac_partsigs, AC_DEFAULT_TRACKLEN))
return ret;
decoded = encoding_norm_readline(&conv, NULL, &area, bread);
if(decoded) {
sret = cli_ac_scanbuff(decoded, strlen((const char *) decoded), NULL, engine->root[0], &mdata, 1, 0, 0, -1, NULL);
free(decoded);
if(sret == CL_TYPE_HTML) {
ret = CL_TYPE_HTML;
end = 1;
}
} else
end = 1;
cli_ac_freedata(&mdata);
}
entity_norm_done(&conv);
} else {
cli_warnmsg("cli_filetype2: Error initializing entity converter\n");
}
}
}
}
if(ret == CL_TYPE_UNKNOWN_DATA || ret == CL_TYPE_UNKNOWN_TEXT) {
if(!(bigbuff = (unsigned char *) cli_calloc(37638 + 1, sizeof(unsigned char))))
return ret;
lseek(desc, 0, SEEK_SET);
if((bread = read(desc, bigbuff, 37638)) > 0) {
bigbuff[bread] = 0;
switch(is_tar(bigbuff, bread)) {
case 1:
ret = CL_TYPE_OLD_TAR;
cli_dbgmsg("Recognized old fashioned tar file\n");
break;
case 2:
ret = CL_TYPE_POSIX_TAR;
cli_dbgmsg("Recognized POSIX tar file\n");
break;
}
}
if(ret == CL_TYPE_UNKNOWN_DATA || ret == CL_TYPE_UNKNOWN_TEXT) {
if(!memcmp(bigbuff + 32769, "CD001" , 5) || !memcmp(bigbuff + 37633, "CD001" , 5)) {
cli_dbgmsg("Recognized ISO 9660 CD-ROM data\n");
ret = CL_TYPE_DATA;
} else if(!memcmp(bigbuff + 32776, "CDROM" , 5)) {
cli_dbgmsg("Recognized High Sierra CD-ROM data\n");
ret = CL_TYPE_DATA;
}
}
free(bigbuff);
}
return ret;
}
static gboolean
extract_archive (GFile *destination,
GFile *archive_file,
guint strip_components,
GError **error)
{
ArchiveType type;
g_autofree char *archive_path = NULL;
archive_path = g_file_get_path (archive_file);
g_debug ("Uncompress %s\n", archive_path);
type = get_type (archive_file);
if (is_tar (type))
{
g_autofree char *strip_components_str = g_strdup_printf ("--strip-components=%u", strip_components);
/* tar_decompress_flag can return NULL, so put it last */
if (!tar (destination, error, "xf", archive_path, "--no-same-owner",
strip_components_str, tar_decompress_flag (type), NULL))
return FALSE;
}
else if (type == ZIP)
{
g_autoptr(GFile) zip_dest = NULL;
zip_dest = create_uncompress_directory (destination, strip_components, error);
if (zip_dest == NULL)
return FALSE;
if (!unzip (zip_dest, error, archive_path, NULL))
return FALSE;
if (strip_components > 0 &&
!strip_components_into (destination, zip_dest, strip_components, error))
return FALSE;
}
else if (type == RPM)
{
g_autoptr(GFile) rpm_dest = NULL;
rpm_dest = create_uncompress_directory (destination, strip_components, error);
if (rpm_dest == NULL)
return FALSE;
if (!unrpm (rpm_dest, archive_path, error))
return FALSE;
if (strip_components > 0 &&
!strip_components_into (destination, rpm_dest, strip_components, error))
return FALSE;
}
else
{
g_set_error (error, G_IO_ERROR, G_IO_ERROR_FAILED, "Unknown archive format of '%s'", archive_path);
return FALSE;
}
return TRUE;
}
