mod_mime_magic.c revision 10a4cdd68ef1ca0e54af296fe1d08ac00150c90b
/* ====================================================================
* The Apache Software License, Version 1.1
*
* Copyright (c) 2000 The Apache Software Foundation. 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, 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
* distribution.
*
* 3. The end-user documentation included with the redistribution,
* if any, must include the following acknowledgment:
* "This product includes software developed by the
* Apache Software Foundation (http://www.apache.org/)."
* Alternately, this acknowledgment may appear in the software itself,
* if and wherever such third-party acknowledgments normally appear.
*
* 4. The names "Apache" and "Apache Software Foundation" must
* not be used to endorse or promote products derived from this
* software without prior written permission. For written
* permission, please contact apache@apache.org.
*
* 5. Products derived from this software may not be called "Apache",
* nor may "Apache" appear in their name, without prior written
* permission of the Apache Software Foundation.
*
* THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESSED 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 APACHE SOFTWARE FOUNDATION OR
* ITS 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.
* ====================================================================
*
* This software consists of voluntary contributions made by many
* individuals on behalf of the Apache Software Foundation. For more
* information on the Apache Software Foundation, please see
*
* Portions of this software are based upon public domain software
* originally written at the National Center for Supercomputing Applications,
* University of Illinois, Urbana-Champaign.
*/
/*
* mod_mime_magic: MIME type lookup via file magic numbers
* Copyright (c) 1996-1997 Cisco Systems, Inc.
*
* This software was submitted by Cisco Systems to the Apache Software Foundation in July
* 1997. Future revisions and derivatives of this source code must
* acknowledge Cisco Systems as the original contributor of this module.
* All other licensing and usage conditions are those of the Apache Software Foundation.
*
* Some of this code is derived from the free version of the file command
* originally posted to comp.sources.unix. Copyright info for that program
* is included below as required.
* ---------------------------------------------------------------------------
* - Copyright (c) Ian F. Darwin, 1987. Written by Ian F. Darwin.
*
* This software is not subject to any license of the American Telephone and
* Telegraph Company or of the Regents of the University of California.
*
* Permission is granted to anyone to use this software for any purpose on any
* computer system, and to alter it and redistribute it freely, subject to
* the following restrictions:
*
* 1. The author is not responsible for the consequences of use of this
* software, no matter how awful, even if they arise from flaws in it.
*
* 2. The origin of this software must not be misrepresented, either by
* explicit claim or by omission. Since few users ever read sources, credits
* must appear in the documentation.
*
* 3. Altered versions must be plainly marked as such, and must not be
* misrepresented as being the original software. Since few users ever read
* sources, credits must appear in the documentation.
*
* 4. This notice may not be removed or altered.
* -------------------------------------------------------------------------
*
* For compliance with Mr Darwin's terms: this has been very significantly
* modified from the free "file" command.
* - all-in-one file for compilation convenience when moving from one
* version of Apache to the next.
* - Memory allocation is done through the Apache API's apr_pool_t structure.
* - All functions have had necessary Apache API request or server
* structures passed to them where necessary to call other Apache API
* routines. (i.e. usually for logging, files, or memory allocation in
* itself or a called function.)
* - struct magic has been converted from an array to a single-ended linked
* list because it only grows one record at a time, it's only accessed
* sequentially, and the Apache API has no equivalent of realloc().
* - Functions have been changed to get their parameters from the server
* configuration instead of globals. (It should be reentrant now but has
* not been tested in a threaded environment.)
* - Places where it used to print results to stdout now saves them in a
* list where they're used to set the MIME type in the Apache request
* record.
* - Command-line flags have been removed since they will never be used here.
*
* Ian Kluft <ikluft@cisco.com>
* Engineering Information Framework
* Central Engineering
* Cisco Systems, Inc.
* San Jose, CA, USA
*
* Misc bug fixes May 1997
* Submission to Apache Software Foundation July 1997
*
*/
#include "apr_strings.h"
#include "ap_config.h"
#include "httpd.h"
#include "http_config.h"
#include "http_request.h"
#include "http_core.h"
#include "http_log.h"
#include "http_protocol.h"
#include "util_script.h"
#ifdef HAVE_UNISTD_H
#include <unistd.h>
#endif
#ifdef HAVE_UTIME_H
#include <utime.h>
#endif
/*
* data structures and related constants
*/
#define MODNAME "mod_mime_magic"
#define MIME_MAGIC_DEBUG 0
#define MIME_BINARY_UNKNOWN "application/octet-stream"
#define MIME_TEXT_UNKNOWN "text/plain"
#define MAXMIMESTRING 256
/* HOWMANY must be at least 4096 to make gzip -dcq work */
#define HOWMANY 4096
/* SMALL_HOWMANY limits how much work we do to figure out text files */
#define SMALL_HOWMANY 1024
struct magic {
int lineno; /* line number from magic file */
short flag;
short cont_level; /* level of ">" */
struct {
char type; /* byte short long */
long offset; /* offset from indirection */
} in;
long offset; /* offset to magic number */
unsigned char reln; /* relation (0=eq, '>'=gt, etc) */
char type; /* int, short, long or string. */
char vallen; /* length of string value, if any */
#define BYTE 1
#define SHORT 2
#define LONG 4
#define STRING 5
#define DATE 6
#define BESHORT 7
#define BELONG 8
#define BEDATE 9
#define LESHORT 10
#define LELONG 11
#define LEDATE 12
union VALUETYPE {
unsigned char b;
unsigned short h;
unsigned long l;
char s[MAXstring];
} value; /* either number or string */
unsigned long mask; /* mask before comparison with value */
char nospflag; /* supress space character */
/* NOTE: this string is suspected of overrunning - find it! */
};
/*
* data structures for tar file recognition
* --------------------------------------------------------------------------
* 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.
*
* 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 mode[8];
char uid[8];
char gid[8];
char size[12];
char mtime[12];
char chksum[8];
char linkflag;
char magic[8];
char devmajor[8];
char devminor[8];
} header;
};
/* The magic field is filled with this if uname and gname are valid. */
/*
* file-function prototypes
*/
static int ascmagic(request_rec *, unsigned char *, int);
static int is_tar(unsigned char *, int);
static int softmagic(request_rec *, unsigned char *, int);
static void tryit(request_rec *, unsigned char *, int, int);
static int zmagic(request_rec *, unsigned char *, int);
static int hextoint(int);
static char *getstr(server_rec *, char *, char *, int, int *);
static int match(request_rec *, unsigned char *, int);
struct magic *, int);
static int uncompress(request_rec *, int,
unsigned char **, int);
static long from_oct(int, char *);
/*
* includes for ASCII substring recognition formerly "names.h" in file
* command
*
* Original notes: names and types used by ascmagic in file(1). These tokens are
* here because they can appear anywhere in the first HOWMANY bytes, while
* make HOWMANY too high unless you have a very fast CPU.
*/
/* these types are used to index the apr_table_t 'types': keep em in sync! */
/* HTML inserted in first because this is a web server module now */
#define L_HTML 0 /* HTML */
static char *types[] =
{
"application/binary", /* "can't happen error on names.h/types", */
0
};
static struct names {
char *name;
short type;
} names[] = {
/* These must be sorted by eye for optimal hit rate */
/* Add to this list only after substantial meditation */
{
"<html>", L_HTML
},
{
"<HTML>", L_HTML
},
{
"<head>", L_HTML
},
{
"<HEAD>", L_HTML
},
{
"<title>", L_HTML
},
{
"<TITLE>", L_HTML
},
{
"<h1>", L_HTML
},
{
"<H1>", L_HTML
},
{
"<!--", L_HTML
},
{
"<!DOCTYPE HTML", L_HTML
},
{
"/*", 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
},
{
"real", L_C
},
{
"struct", L_C
},
{
"union", L_C
},
{
"CFLAGS", L_MAKE
},
{
"LDFLAGS", L_MAKE
},
{
"all:", L_MAKE
},
{
".PRECIOUS", L_MAKE
},
/*
* Too many files of text have these words in them. Find another way to
* recognize Fortrash.
*/
#ifdef NOTDEF
{
"subroutine", L_FORT
},
{
"function", L_FORT
},
{
"block", L_FORT
},
{
"common", L_FORT
},
{
"dimension", L_FORT
},
{
"integer", L_FORT
},
{
"data", L_FORT
},
#endif /* NOTDEF */
{
".ascii", L_MACH
},
{
".asciiz", L_MACH
},
{
".byte", L_MACH
},
{
".even", L_MACH
},
{
".globl", L_MACH
},
{
"clr", L_MACH
},
{
"(input,", 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
},
{
NULL, 0
}
};
/*
* Result String List (RSL)
*
* The file(1) command prints its output. Instead, we store the various
* "printed" strings in a list (allocating memory as we go) and concatenate
* them at the end when we finally know how much space they'll need.
*/
typedef struct magic_rsl_s {
char *str; /* string, possibly a fragment */
} magic_rsl;
/*
* Apache module configuration structures
*/
/* per-server info */
typedef struct {
const char *magicfile; /* where magic be found */
/* per-request info */
typedef struct {
unsigned suf_recursion; /* recursion depth in suffix check */
/*
* configuration functions - called by Apache API routines
*/
{
/* allocate the config - use pcalloc because it needs to be zeroed */
return apr_pcalloc(p, sizeof(magic_server_config_rec));
}
{
apr_palloc(p, sizeof(magic_server_config_rec));
return new;
}
{
if (!conf) {
return MODNAME ": server structure not allocated";
}
return NULL;
}
/*
* configuration file commands - exported to Apache API
*/
static const command_rec mime_magic_cmds[] =
{
"Path to MIME Magic file (in file(1) format)"),
{NULL}
};
/*
* RSL (result string list) processing routines
*
* These collect strings that would have been printed in fragments by file(1)
* into a list of magic_rsl structures with the strings. When complete,
* they're concatenated together to become the MIME content and encoding
* types.
*
* return value conventions for these functions: functions which return int:
* failure = -1, other = result functions which return pointers: failure = 0,
* other = result
*/
/* allocate a per-request structure and put it in the request record */
{
sizeof(magic_req_rec));
return req_dat;
}
/* add a string to the result string list for this request */
/* it is the responsibility of the caller to allocate "str" */
{
/* make sure we have a list to put it in */
if (!req_dat) {
MODNAME ": request config should not be NULL");
if (!(req_dat = magic_set_config(r))) {
/* failure */
return -1;
}
}
/* allocate the list entry */
/* fill it */
/* append to the list */
}
else {
}
/* success */
return 0;
}
/* RSL hook for puts-type functions */
{
return magic_rsl_add(r, str);
}
/* RSL hook for printf-type functions */
{
char buf[MAXMIMESTRING];
/* assemble the string into the buffer */
/* add the buffer to the list */
}
/* RSL hook for putchar-type functions */
static int magic_rsl_putchar(request_rec *r, char c)
{
char str[2];
/* high overhead for 1 char - just hope they don't do this much */
str[0] = c;
return magic_rsl_add(r, str);
}
/* allocate and copy a contiguous string from a result string list */
{
char *result; /* return value */
cur_pos, /* current position within fragment */
res_pos; /* position in result string */
/* allocate the result string */
/* loop through and collect the string */
res_pos = 0;
/* loop to the first fragment */
if (cur_frag < start_frag)
continue;
/* loop through and collect chars */
cur_pos++) {
if (cur_frag >= start_frag
break;
}
}
}
}
/* clean up and return */
#if MIME_MAGIC_DEBUG
#endif
return result;
}
/* states for the state-machine algorithm in magic_rsl_to_request() */
typedef enum {
} rsl_states;
/* process the RSL and set the MIME info in the request record */
static int magic_rsl_to_request(request_rec *r)
{
cur_pos, /* current position within fragment */
type_frag, /* content type starting point: fragment */
type_pos, /* content type starting point: position */
type_len, /* content type length */
encoding_frag, /* content encoding starting point: fragment */
encoding_pos, /* content encoding starting point: position */
encoding_len; /* content encoding length */
/* check if we have a result */
/* empty - no match, we defer to other Apache modules */
return DECLINED;
}
/* start searching for the type and encoding */
/* loop through the characters in the fragment */
/* process whitespace actions for each state */
if (state == rsl_leading_space) {
/* eat whitespace in this state */
continue;
}
/* whitespace: type has no slash! */
return DECLINED;
}
else if (state == rsl_subtype) {
/* whitespace: end of MIME type */
state++;
continue;
}
else if (state == rsl_separator) {
/* eat whitespace in this state */
continue;
}
else if (state == rsl_encoding) {
/* whitespace: end of MIME encoding */
/* we're done */
break;
}
else {
/* should not be possible */
/* abandon malfunctioning module */
return DECLINED;
}
/* NOTREACHED */
}
/* copy the char and go to rsl_subtype state */
type_len++;
state++;
}
else {
/* process non-space actions for each state */
if (state == rsl_leading_space) {
/* non-space: begin MIME type */
state++;
type_len = 1;
continue;
}
state == rsl_subtype) {
/* non-space: adds to type */
type_len++;
continue;
}
else if (state == rsl_separator) {
/* non-space: begin MIME encoding */
state++;
encoding_len = 1;
continue;
}
else if (state == rsl_encoding) {
/* non-space: adds to encoding */
encoding_len++;
continue;
}
else {
/* should not be possible */
/* abandon malfunctioning module */
return DECLINED;
}
/* NOTREACHED */
}
/* NOTREACHED */
}
}
/* if we ended prior to state rsl_subtype, we had incomplete info */
state != rsl_encoding) {
/* defer to other modules */
return DECLINED;
}
/* save the info in the request record */
state == rsl_encoding) {
char *tmp;
/* XXX: this could be done at config time I'm sure... but I'm
* confused by all this magic_rsl stuff. -djg */
r->content_type = tmp;
}
if (state == rsl_encoding) {
char *tmp;
/* XXX: this could be done at config time I'm sure... but I'm
* confused by all this magic_rsl stuff. -djg */
r->content_encoding = tmp;
}
/* detect memory allocation errors */
if (!r->content_type ||
return HTTP_INTERNAL_SERVER_ERROR;
}
/* success! */
return OK;
}
/*
* magic_process - process input file r Apache API request record
* (formerly called "process" in file command, prefix added for clarity) Opens
* the file and reads a fixed-size buffer to begin processing the contents.
*/
static int magic_process(request_rec *r)
{
int nbytes = 0; /* number of bytes read from a datafile */
int result;
/*
* first try judging the file based on its filesystem status
*/
case DONE:
magic_rsl_putchar(r, '\n');
return OK;
case OK:
break;
default:
/* fatal error, bail out */
return result;
}
/* We can't open it, but we were able to stat it. */
/* let some other handler decide what the problem is */
return DECLINED;
}
/*
* try looking at the first HOWMANY bytes
*/
return HTTP_INTERNAL_SERVER_ERROR;
}
if (nbytes == 0)
else {
}
(void) magic_rsl_putchar(r, '\n');
return OK;
}
{
/*
* Try compression stuff
*/
if (checkzmagic == 1) {
return;
}
/*
*/
return;
/*
* try known keywords, check for ascii-ness too.
*/
return;
/*
* abandon hope, all ye who remain here
*/
}
#define EATAB {while (apr_isspace((unsigned char) *l)) ++l;}
/*
* apprentice - load configuration from the magic file r
* API request record
*/
{
apr_file_t *f = NULL;
int errs = 0;
int lineno;
#if MIME_MAGIC_DEBUG
int rule = 0;
#endif
if (result != APR_SUCCESS) {
return -1;
}
/* set up the magic list (empty) */
/* parse it */
int ws_offset;
/* delete newline */
if (line[0]) {
}
/* skip leading whitespace */
ws_offset = 0;
ws_offset++;
}
/* skip blank lines */
continue;
}
/* comment, do not parse */
continue;
#if MIME_MAGIC_DEBUG
/* if we get here, we're going to use it so count it */
rule++;
#endif
/* parse it */
++errs;
}
(void) apr_close(f);
#if MIME_MAGIC_DEBUG
MODNAME ": apprentice conf=%x file=%s m=%s m->next=%s last=%s",
conf,
MODNAME ": apprentice read %d lines, %d rules, %d errors",
#endif
#if MIME_MAGIC_DEBUG
prevm = 0;
MODNAME ": apprentice test");
apr_isprint(((unsigned long) m) & 255)) {
MODNAME ": apprentice: POINTER CLOBBERED! "
"m=\"%c%c%c%c\" line=%d",
(((unsigned long) m) >> 24) & 255,
(((unsigned long) m) >> 16) & 255,
(((unsigned long) m) >> 8) & 255,
((unsigned long) m) & 255,
break;
}
prevm = m;
}
#endif
return (errs ? -1 : 0);
}
/*
* extend the sign bit if the comparison is to be signed
*/
{
switch (m->type) {
/*
* Do not remove the casts below. They are vital. When later
* compared with the data, the sign extension must have happened.
*/
case BYTE:
v = (char) v;
break;
case SHORT:
case BESHORT:
case LESHORT:
v = (short) v;
break;
case DATE:
case BEDATE:
case LEDATE:
case LONG:
case BELONG:
case LELONG:
v = (long) v;
break;
case STRING:
break;
default:
return -1;
}
return v;
}
/*
* parse one line from magic file, put into magic[index++] if valid
*/
{
struct magic *m;
char *t, *s;
/* allocate magic structure entry */
/* append to linked list */
}
else {
}
/* set values in magic structure */
m->flag = 0;
m->cont_level = 0;
while (*l == '>') {
++l; /* step over */
m->cont_level++;
}
if (m->cont_level != 0 && *l == '(') {
++l; /* step over */
}
/* get offset, then skip over it */
if (l == t) {
MODNAME ": offset %s invalid", l);
}
l = t;
/*
* read [.lbs][+-]nnnnn)
*/
if (*l == '.') {
switch (*++l) {
case 'l':
break;
case 's':
break;
case 'b':
break;
default:
MODNAME ": indirect offset type %c invalid", *l);
break;
}
l++;
}
s = l;
if (*l == '+' || *l == '-')
l++;
if (apr_isdigit((unsigned char) *l)) {
if (*s == '-')
}
else
t = l;
if (*t++ != ')') {
MODNAME ": missing ')' in indirect offset");
}
l = t;
}
while (apr_isdigit((unsigned char) *l))
++l;
#define NBYTE 4
#define NSHORT 5
#define NLONG 4
#define NSTRING 6
#define NDATE 4
#define NBESHORT 7
#define NBELONG 6
#define NBEDATE 6
#define NLESHORT 7
#define NLELONG 6
#define NLEDATE 6
if (*l == 'u') {
++l;
}
/* get type, skip it */
l += NBYTE;
}
l += NSHORT;
}
l += NLONG;
}
l += NSTRING;
}
l += NDATE;
}
l += NBESHORT;
}
l += NBELONG;
}
l += NBEDATE;
}
l += NLESHORT;
}
l += NLELONG;
}
l += NLEDATE;
}
else {
MODNAME ": type %s invalid", l);
return -1;
}
/* New-style anding: "0 byte&0x80 =0x80 dynamically linked" */
if (*l == '&') {
++l;
}
else
m->mask = ~0L;
switch (*l) {
case '>':
case '<':
/* Old-style anding: "0 byte &0x80 dynamically linked" */
case '&':
case '^':
case '=':
m->reln = *l;
++l;
break;
case '!':
m->reln = *l;
++l;
break;
}
/* FALL THROUGH */
default:
m->reln = *l;
++l;
goto GetDesc; /* Bill The Cat */
}
m->reln = '=';
break;
}
return -1;
/*
* now get last part - the description
*/
if (l[0] == '\b') {
++l;
m->nospflag = 1;
}
else if ((l[0] == '\\') && (l[1] == 'b')) {
++l;
++l;
m->nospflag = 1;
}
else
m->nospflag = 0;
#if MIME_MAGIC_DEBUG
MODNAME ": parse line=%d m=%x next=%x cont=%d desc=%s",
#endif /* MIME_MAGIC_DEBUG */
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.
*/
{
int slen;
}
else if (m->reln != 'x')
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.
*/
{
register int c;
register int val;
while ((c = *s++) != '\0') {
if (apr_isspace((unsigned char) c))
break;
if (p >= pmax) {
break;
}
if (c == '\\') {
switch (c = *s++) {
case '\0':
goto out;
default:
*p++ = (char) c;
break;
case 'n':
*p++ = '\n';
break;
case 'r':
*p++ = '\r';
break;
case 'b':
*p++ = '\b';
break;
case 't':
*p++ = '\t';
break;
case 'f':
*p++ = '\f';
break;
case 'v':
*p++ = '\v';
break;
/* \ and up to 3 octal digits */
case '0':
case '1':
case '2':
case '3':
case '4':
case '5':
case '6':
case '7':
val = c - '0';
c = *s++; /* try for 2 */
if (c >= '0' && c <= '7') {
c = *s++; /* try for 3 */
if (c >= '0' && c <= '7')
else
--s;
}
else
--s;
*p++ = (char) val;
break;
/* \x and up to 3 hex digits */
case 'x':
c = hextoint(*s++); /* Get next char */
if (c >= 0) {
val = c;
c = hextoint(*s++);
if (c >= 0) {
c = hextoint(*s++);
if (c >= 0) {
}
else
--s;
}
else
--s;
}
else
--s;
*p++ = (char) val;
break;
}
}
else
*p++ = (char) c;
}
out:
*p = '\0';
return s;
}
/* Single hex char to int; -1 if not a hex char. */
static int hextoint(int c)
{
if (apr_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;
}
/*
* return DONE to indicate it's been handled
* return OK to indicate it's a regular file still needing handling
* other returns indicate a failure of some sort
*/
{
case APR_DIR:
return DONE;
case APR_CHR:
/*
* (void) magic_rsl_printf(r,"character special (%d/%d)",
* major(sb->st_rdev), minor(sb->st_rdev));
*/
(void) magic_rsl_puts(r, MIME_BINARY_UNKNOWN);
return DONE;
case APR_BLK:
/*
* (void) magic_rsl_printf(r,"block special (%d/%d)",
* major(sb->st_rdev), minor(sb->st_rdev));
*/
(void) magic_rsl_puts(r, MIME_BINARY_UNKNOWN);
return DONE;
/* TODO add code to handle V7 MUX and Blit MUX files */
case APR_PIPE:
/*
* magic_rsl_puts(r,"fifo (named pipe)");
*/
(void) magic_rsl_puts(r, MIME_BINARY_UNKNOWN);
return DONE;
case APR_LNK:
/* We used stat(), the only possible reason for this is that the
* symlink is broken.
*/
return HTTP_INTERNAL_SERVER_ERROR;
case APR_SOCK:
return DONE;
case APR_REG:
break;
default:
return HTTP_INTERNAL_SERVER_ERROR;
}
/*
* regular file, check next possibility
*/
return DONE;
}
return OK;
}
/*
* apprentice.c). Passed the name and FILE * of one file to be typed.
*/
/* ARGSUSED1 *//* nbytes passed for regularity, maybe need later */
{
return 1;
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.
*/
{
#if MIME_MAGIC_DEBUG
int rule_counter = 0;
#endif
int cont_level = 0;
int need_separator = 0;
union VALUETYPE p;
struct magic *m;
#if MIME_MAGIC_DEBUG
MODNAME ": match conf=%x file=%s m=%s m->next=%s last=%s",
conf,
#endif
#if MIME_MAGIC_DEBUG
apr_isprint(((unsigned long) m) & 255)) {
MODNAME ": match: POINTER CLOBBERED! "
"m=\"%c%c%c%c\"",
(((unsigned long) m) >> 24) & 255,
(((unsigned long) m) >> 16) & 255,
(((unsigned long) m) >> 8) & 255,
((unsigned long) m) & 255);
break;
}
}
#endif
#if MIME_MAGIC_DEBUG
rule_counter++;
#endif
/* check if main entry matches */
!mcheck(r, &p, m)) {
/*
* main entry didn't match, flush its continuations
*/
continue;
}
#if MIME_MAGIC_DEBUG
rule_counter++;
MODNAME ": line=%d mc=%x mc->next=%x cont=%d desc=%s",
#endif
/*
* this trick allows us to keep *m in sync when the continue
* advances the pointer
*/
m = m_cont;
}
continue;
}
/* if we get here, the main entry rule was a match */
/* this will be the last run through the loop */
#if MIME_MAGIC_DEBUG
MODNAME ": rule matched, line=%d type=%d %s",
#endif
/* print the match */
mprint(r, &p, m);
/*
* If we printed something, we'll need to print a blank before we
* print something else.
*/
if (m->desc[0])
need_separator = 1;
/* and any continuations that match */
cont_level++;
/*
* while (m && m->next && m->next->cont_level != 0 && ( m = m->next
* ))
*/
m = m->next;
while (m && (m->cont_level != 0)) {
#if MIME_MAGIC_DEBUG
MODNAME ": match line=%d cont=%d type=%d %s",
#endif
if (cont_level >= m->cont_level) {
if (cont_level > m->cont_level) {
/*
* We're at the end of the level "cont_level"
* continuations.
*/
cont_level = m->cont_level;
}
mcheck(r, &p, m)) {
/*
* 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->nospflag == 0)
&& (m->desc[0] != '\0')
) {
(void) magic_rsl_putchar(r, ' ');
need_separator = 0;
}
mprint(r, &p, m);
if (m->desc[0])
need_separator = 1;
/*
* If we see any continuations at a higher level, process
* them.
*/
cont_level++;
}
}
/* move to next continuation record */
m = m->next;
}
#if MIME_MAGIC_DEBUG
#endif
return 1; /* all through */
}
#if MIME_MAGIC_DEBUG
#endif
return 0; /* no match at all */
}
{
unsigned long v;
switch (m->type) {
case BYTE:
v = p->b;
break;
case SHORT:
case BESHORT:
case LESHORT:
v = p->h;
break;
case LONG:
case BELONG:
case LELONG:
v = p->l;
break;
case STRING:
if (m->reln == '=') {
}
else {
(void) magic_rsl_printf(r, m->desc, p->s);
}
return;
case DATE:
case BEDATE:
case LEDATE:
/* XXX: not multithread safe */
*rt = '\0';
return;
default:
MODNAME ": invalid m->type (%d) in mprint().",
m->type);
return;
}
(void) magic_rsl_printf(r, m->desc, (unsigned long) v);
}
/*
* Convert the byte order of the data we are looking at
*/
{
char *rt;
switch (m->type) {
case BYTE:
case SHORT:
case LONG:
case DATE:
return 1;
case STRING:
/* Null terminate and eat the return */
p->s[sizeof(p->s) - 1] = '\0';
*rt = '\0';
return 1;
case BESHORT:
return 1;
case BELONG:
case BEDATE:
p->l = (long)
return 1;
case LESHORT:
return 1;
case LELONG:
case LEDATE:
p->l = (long)
return 1;
default:
return 0;
}
}
{
return 0;
if (!mconvert(r, p, m))
return 0;
case BYTE:
break;
case SHORT:
break;
case LONG:
break;
}
return 0;
if (!mconvert(r, p, m))
return 0;
}
return 1;
}
{
register unsigned long l = m->value.l;
register unsigned long v;
int matched;
MODNAME ": BOINK");
return 1;
}
switch (m->type) {
case BYTE:
v = p->b;
break;
case SHORT:
case BESHORT:
case LESHORT:
v = p->h;
break;
case LONG:
case BELONG:
case LELONG:
case DATE:
case BEDATE:
case LEDATE:
v = p->l;
break;
case STRING:
l = 0;
/*
* What we want here is: v = strncmp(m->value.s, p->s, m->vallen);
* but ignoring any nulls. bcmp doesn't give -/+/0 and isn't
* universally available anyway.
*/
v = 0;
{
register unsigned char *a = (unsigned char *) m->value.s;
register unsigned char *b = (unsigned char *) p->s;
while (--len >= 0)
if ((v = *b++ - *a++) != 0)
break;
}
break;
default:
/* bogosity, pretend that it just wasn't a match */
return 0;
}
switch (m->reln) {
case 'x':
#if MIME_MAGIC_DEBUG
"%lu == *any* = 1", v);
#endif
matched = 1;
break;
case '!':
matched = v != l;
#if MIME_MAGIC_DEBUG
"%lu != %lu = %d", v, l, matched);
#endif
break;
case '=':
matched = v == l;
#if MIME_MAGIC_DEBUG
"%lu == %lu = %d", v, l, matched);
#endif
break;
case '>':
matched = v > l;
#if MIME_MAGIC_DEBUG
"%lu > %lu = %d", v, l, matched);
#endif
}
else {
matched = (long) v > (long) l;
#if MIME_MAGIC_DEBUG
"%ld > %ld = %d", v, l, matched);
#endif
}
break;
case '<':
matched = v < l;
#if MIME_MAGIC_DEBUG
"%lu < %lu = %d", v, l, matched);
#endif
}
else {
matched = (long) v < (long) l;
#if MIME_MAGIC_DEBUG
"%ld < %ld = %d", v, l, matched);
#endif
}
break;
case '&':
matched = (v & l) == l;
#if MIME_MAGIC_DEBUG
"((%lx & %lx) == %lx) = %d", v, l, l, matched);
#endif
break;
case '^':
matched = (v & l) != l;
#if MIME_MAGIC_DEBUG
"((%lx & %lx) != %lx) = %d", v, l, l, matched);
#endif
break;
default:
/* bogosity, pretend it didn't match */
matched = 0;
MODNAME ": mcheck: can't happen: invalid relation %d.",
m->reln);
break;
}
return matched;
}
/* an optimization over plain strcmp() */
{
int has_escapes = 0;
unsigned char *s;
char *token;
register struct names *p;
int small_nbytes;
/* 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 == '.') {
while (apr_isspace(*tp))
++tp; /* skip leading whitespace */
magic_rsl_puts(r, "application/x-troff");
return 1;
}
}
/* Fortran */
magic_rsl_puts(r, "text/plain");
return 1;
}
/* look for tokens from names.h - this is expensive!, so we'll limit
* ourselves to only SMALL_HOWMANY bytes */
/* make a copy of the buffer here because strtok() will destroy it */
s[small_nbytes] = '\0';
/* XXX: not multithread safe */
s = NULL; /* make strtok() keep on tokin' */
if (has_escapes)
magic_rsl_puts(r, " (with escape sequences)");
return 1;
}
}
}
case 1:
/* V7 tar archive */
magic_rsl_puts(r, "application/x-tar");
return 1;
case 2:
/* POSIX tar archive */
magic_rsl_puts(r, "application/x-tar");
return 1;
}
/* all else fails, but it is ascii... */
if (has_escapes) {
/* text with escape sequences */
/* we leave this open for further differentiation later */
magic_rsl_puts(r, "text/plain");
}
else {
/* plain text */
magic_rsl_puts(r, "text/plain");
}
return 1;
}
/*
* compress routines: zmagic() - returns 0 if not recognized, uncompresses
* and prints information if recognized uncompress(s, method, old, n, newch)
* - uncompress old into new, using method, return sizeof new
*/
static struct {
char *magic;
int maglen;
char *argv[3];
int silent;
char *encoding; /* MUST be lowercase */
} compr[] = {
/* we use gzip here rather than uncompress because we have to pass
* it a full filename -- and uncompress only considers filenames
* ending with .Z
*/
{
"\037\235", 2, {
}, 0, "x-compress"
},
{
"\037\213", 2, {
}, 1, "x-gzip"
},
/*
* XXX pcat does not work, cause I don't know how to make it read stdin,
* so we use gzip
*/
{
"\037\036", 2, {
}, 0, "x-gzip"
},
};
{
unsigned char *newbuf;
int newsize;
int i;
for (i = 0; i < ncompr; i++) {
continue;
break;
}
if (i == ncompr)
return 0;
/* set encoding type in the request record */
}
return 1;
}
struct uncompress_parms {
request_rec *r;
int method;
};
{
int rc = 1;
char *new_argv[4];
char **env;
request_rec *r = parm->r;
/* Something bad happened, tell the world. */
"couldn't setup child process: %s", r->filename);
}
else {
}
if (rc != APR_SUCCESS) {
/* Bad things happened. Everyone should have cleaned up. */
MODNAME ": could not execute `%s'.",
}
else {
/* Fill in BUFF structure for parents pipe to child's stdout */
if (!iol)
return APR_EBADF;
if (script_in) {
}
}
}
return (rc);
}
unsigned char **newch, int n)
{
struct uncompress_parms parm;
parm.r = r;
/* We make a sub_pool so that we can collect our child early, otherwise
* there are cases (i.e. generating directory indicies with mod_autoindex)
* where we would end up with LOTS of zombies.
*/
return -1;
return -1;
}
if (n == 0) {
return -1;
}
return n;
}
/*
* is_tar() -- figure out whether file is a tar archive.
*
* Stolen (by author of file utility) 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 $Id: mod_mime_magic.c,v 1.7
* 1997/06/24 00:41:02 ikluft Exp ikluft $
*
* Darwin.
*/
/*
* 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.
*/
{
register int i;
register char *p;
return 0;
sum = 0;
for (i = sizeof(union record); --i >= 0;) {
/*
* We can't use unsigned char here because of old compilers, e.g. V7.
*/
sum += 0xFF & *p++;
}
/* Adjust checksum to count the "chksum" field as blanks. */
return 0; /* Not a tar archive */
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).
*/
{
register long value;
where++;
if (--digs <= 0)
return -1; /* All blank field */
}
value = 0;
--digs;
}
return value;
}
/*
* Check for file-revision suffix
*
* This is for an obscure document control system used on an intranet.
* The web representation of each file's revision has an @1, @2, etc
* appended with the revision number. This needs to be stripped off to
* find the file suffix, which can be recognized by sending the name back
* through a sub-request. The base file name (without the @num suffix)
* must exist because its type will be used as the result.
*/
static int revision_suffix(request_rec *r)
{
int suffix_pos, result;
char *sub_filename;
#if MIME_MAGIC_DEBUG
#endif /* MIME_MAGIC_DEBUG */
/* check for recognized revision suffix */
return 0;
}
suffix_pos--;
return 0;
}
/* perform sub-request for the file name without the suffix */
result = 0;
#if MIME_MAGIC_DEBUG
#endif /* MIME_MAGIC_DEBUG */
if (sub->content_type) {
#if MIME_MAGIC_DEBUG
MODNAME ": subrequest %s got %s",
sub_filename, r->content_type);
#endif /* MIME_MAGIC_DEBUG */
if (sub->content_encoding)
r->content_encoding =
if (sub->content_language)
r->content_language =
result = 1;
}
/* clean up */
return result;
}
/*
* initialize the module
*/
{
int result;
server_rec *s;
#if MIME_MAGIC_DEBUG
#endif /* MIME_MAGIC_DEBUG */
for (s = main_server; s; s = s->next) {
/* inherits from the parent */
}
result = apprentice(s, p);
if (result == -1)
return;
#if MIME_MAGIC_DEBUG
prevm = 0;
MODNAME ": magic_init 1 test");
apr_isprint(((unsigned long) m) & 255)) {
MODNAME ": magic_init 1: POINTER CLOBBERED! "
"m=\"%c%c%c%c\" line=%d",
(((unsigned long) m) >> 24) & 255,
(((unsigned long) m) >> 16) & 255,
(((unsigned long) m) >> 8) & 255,
((unsigned long) m) & 255,
break;
}
prevm = m;
}
#endif
}
}
}
/*
* Find the Content-Type from any resource this module has available
*/
static int magic_find_ct(request_rec *r)
{
int result;
/* the file has to exist */
return DECLINED;
}
/* was someone else already here? */
if (r->content_type) {
return DECLINED;
}
return DECLINED;
}
/* initialize per-request info */
if (!magic_set_config(r)) {
return HTTP_INTERNAL_SERVER_ERROR;
}
/* try excluding file-revision suffixes */
if (revision_suffix(r) != 1) {
/* process it based on the file contents */
return result;
}
}
/* if we have any results, put them in the request structure */
return magic_rsl_to_request(r);
}
static void register_hooks(void)
{
/* mod_mime_magic should be run after mod_mime, if at all. */
}
/*
* Apache API module interface
*/
{
NULL, /* dir config creator */
NULL, /* dir merger --- default is to override */
create_magic_server_config, /* server config */
merge_magic_server_config, /* merge server config */
mime_magic_cmds, /* command apr_table_t */
NULL, /* handlers */
register_hooks /* register hooks */
};