mod_mime_magic.c

apache简化版

    default:
	/* bogosity, pretend it didn't match */
	matched = 0;
	ap_log_rerror(APLOG_MARK, APLOG_NOERRNO | APLOG_ERR, r,
		    MODNAME ": mcheck: can't happen: invalid relation %d.",
		    m->reln);
	break;
    }

    return matched;
}

/* an optimization over plain strcmp() */
#define    STREQ(a, b)    (*(a) == *(b) && strcmp((a), (b)) == 0)

static int ascmagic(request_rec *r, unsigned char *buf, int nbytes)
{
    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;

    /* 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 == '.') {
	unsigned char *tp = buf + 1;

	while (ap_isspace(*tp))
	    ++tp;		/* skip leading whitespace */
	if ((ap_isalnum(*tp) || *tp == '\\') &&
	     (ap_isalnum(*(tp + 1)) || *tp == '"')) {
	    magic_rsl_puts(r, "application/x-troff");
	    return 1;
	}
    }
    if ((*buf == 'c' || *buf == 'C') && ap_isspace(*(buf + 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 */
    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);
    /* XXX: not multithread safe */
    while ((token = strtok((char *) s, " \t\n\r\f")) != NULL) {
	s = NULL;		/* make strtok() keep on tokin' */
	for (p = names; p < names + NNAMES; p++) {
	    if (STREQ(p->name, token)) {
		magic_rsl_puts(r, types[p->type]);
		if (has_escapes)
		    magic_rsl_puts(r, " (with escape sequences)");
		return 1;
	    }
	}
    }

    switch (is_tar(buf, nbytes)) {
    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[] = {

    {
	"\037\235", 2, {
	    "uncompress", "-c", NULL
	}, 0, "x-compress"
    },
    {
	"\037\213", 2, {
	    "gzip", "-dcq", NULL
	}, 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, {
	    "gzip", "-dcq", NULL
	}, 0, "x-gzip"
    },
};

static int ncompr = sizeof(compr) / sizeof(compr[0]);

static int zmagic(request_rec *r, unsigned char *buf, int nbytes)
{
    unsigned char *newbuf;
    int newsize;
    int i;

    for (i = 0; i < ncompr; i++) {
	if (nbytes < compr[i].maglen)
	    continue;
	if (memcmp(buf, compr[i].magic, compr[i].maglen) == 0)
	    break;
    }

    if (i == ncompr)
	return 0;

    if ((newsize = uncompress(r, i, buf, &newbuf, nbytes)) > 0) {
	tryit(r, newbuf, newsize);

	/* set encoding type in the request record */
	r->content_encoding = compr[i].encoding;
    }
    return 1;
}


struct uncompress_parms {
    request_rec *r;
    int method;
};

static int uncompress_child(void *data, child_info *pinfo)
{
    struct uncompress_parms *parm = data;
#if defined(WIN32)
    int child_pid;
#endif

    if (compr[parm->method].silent) {
	close(STDERR_FILENO);
    }

#if defined(WIN32)
    child_pid = spawnvp(compr[parm->method].argv[0],
			compr[parm->method].argv);
    return (child_pid);
#else
    execvp(compr[parm->method].argv[0], compr[parm->method].argv);
    ap_log_rerror(APLOG_MARK, APLOG_ERR, parm->r,
		MODNAME ": could not execute `%s'.",
		compr[parm->method].argv[0]);
    return -1;
#endif
}


static int uncompress(request_rec *r, int method, const unsigned char *old,
		      unsigned char **newch, int n)
{
    struct uncompress_parms parm;
    BUFF *bin, *bout;
    pool *sub_pool;

    parm.r = r;
    parm.method = method;

    /* 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.
     */
    sub_pool = ap_make_sub_pool(r->pool);

    if (!ap_bspawn_child(sub_pool, uncompress_child, &parm, kill_always,
			 &bin, &bout, NULL)) {
	ap_log_rerror(APLOG_MARK, APLOG_ERR, r,
		    MODNAME ": couldn't spawn uncompress process: %s", r->uri);
	return -1;
    }

    if (ap_bwrite(bin, old, n) != n) {
	ap_destroy_pool(sub_pool);
	ap_log_rerror(APLOG_MARK, APLOG_ERR, r,
		    MODNAME ": write failed.");
	return -1;
    }
    ap_bclose(bin);
    *newch = (unsigned char *) ap_palloc(r->pool, n);
    if ((n = ap_bread(bout, *newch, n)) <= 0) {
	ap_destroy_pool(sub_pool);
	ap_log_rerror(APLOG_MARK, APLOG_ERR, r,
	    MODNAME ": read failed %s", r->filename);
	return -1;
    }
    ap_destroy_pool(sub_pool);
    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 $
 *
 * Comments changed and some code/comments reformatted for file command by Ian
 * Darwin.
 */

#define    isodigit(c)    ( ((c) >= '0') && ((c) <= '7') )

/*
 * 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.
 */

static int is_tar(unsigned char *buf, int nbytes)
{
    register union record *header = (union record *) buf;
    register int i;
    register long sum, recsum;
    register 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 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. */
    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 (0 == strcmp(header->header.magic, TMAGIC))
	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).
 */
static long from_oct(int digs, char *where)
{
    register long value;

    while (ap_isspace(*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 && !ap_isspace(*where))
	return -1;		/* Ended on non-space/nul */

    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;
    request_rec *sub;

#if MIME_MAGIC_DEBUG
    ap_log_rerror(APLOG_MARK, APLOG_NOERRNO | APLOG_DEBUG, r,
		MODNAME ": revision_suffix checking %s", r->filename);
#endif /* MIME_MAGIC_DEBUG */

    /* check for recognized revision suffix */
    suffix_pos = strlen(r->filename) - 1;
    if (!ap_isdigit(r->filename[suffix_pos])) {
	return 0;
    }
    while (suffix_pos >= 0 && ap_isdigit(r->filename[suffix_pos]))
	suffix_pos--;
    if (suffix_pos < 0 || r->filename[suffix_pos] != '@') {
	return 0;
    }

    /* perform sub-request for the file name without the suffix */
    result = 0;
    sub_filename = ap_pstrndup(r->pool, r->filename, suffix_pos);
#if MIME_MAGIC_DEBUG
    ap_log_rerror(APLOG_MARK, APLOG_NOERRNO | APLOG_DEBUG, r,
		MODNAME ": subrequest lookup for %s", sub_filename);
#endif /* MIME_MAGIC_DEBUG */
    sub = ap_sub_req_lookup_file(sub_filename, r);

    /* extract content type/encoding/language from sub-request */
    if (sub->content_type) {
	r->content_type = ap_pstrdup(r->pool, sub->content_type);
#if MIME_MAGIC_DEBUG
	ap_log_rerror(APLOG_MARK, APLOG_NOERRNO | APLOG_DEBUG, r,
		    MODNAME ": subrequest %s got %s",
		    sub_filename, r->content_type);
#endif /* MIME_MAGIC_DEBUG */
	if (sub->content_encoding)
	    r->content_encoding =
		ap_pstrdup(r->pool, sub->content_encoding);
	if (sub->content_language)
	    r->content_language =
		ap_pstrdup(r->pool, sub->content_language);
	result = 1;
    }

    /* clean up */
    ap_destroy_sub_req(sub);

    return result;
}

/*
 * initialize the module
 */

static void magic_init(server_rec *main_server, pool *p)
{
    int result;
    magic_server_config_rec *conf;
    magic_server_config_rec *main_conf;
    server_rec *s;
#if MIME_MAGIC_DEBUG
    struct magic *m, *prevm;
#endif /* MIME_MAGIC_DEBUG */

    main_conf = ap_get_module_config(main_server->module_config, &mime_magic_module);
    for (s = main_server; s; s = s->next) {
	conf = ap_get_module_config(s->module_config, &mime_magic_module);
	if (conf->magicfile == NULL && s != main_server) {
	    /* inherits from the parent */
	    *conf = *main_conf;
	}
	else if (conf->magicfile) {
	    result = apprentice(s, p);
	    if (result == -1)
		return;
#if MIME_MAGIC_DEBUG
	    prevm = 0;
	    ap_log_error(APLOG_MARK, APLOG_NOERRNO | APLOG_DEBUG, s,
			MODNAME ": magic_init 1 test");
	    for (m = conf->magic; m; m = m->next) {
		if (ap_isprint((((unsigned long) m) >> 24) & 255) &&
		    ap_isprint((((unsigned long) m) >> 16) & 255) &&
		    ap_isprint((((unsigned long) m) >> 8) & 255) &&
		    ap_isprint(((unsigned long) m) & 255)) {
		    ap_log_error(APLOG_MARK, APLOG_NOERRNO | APLOG_DEBUG, s,
				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,
				prevm ? prevm->lineno : -1);
		    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;
    magic_server_config_rec *conf;

    /* the file has to exist */
    if (r->finfo.st_mode == 0 || !r->filename) {
	return DECLINED;
    }

    /* was someone else already here? */
    if (r->content_type) {
	return DECLINED;
    }

    conf = ap_get_module_config(r->server->module_config, &mime_magic_module);
    if (!conf || !conf->magic) {
	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 */
	if ((result = magic_process(r)) != OK) {
	    return result;
	}
    }

    /* if we have any results, put them in the request structure */
    return magic_rsl_to_request(r);
}

/*
 * Apache API module interface
 */

module mime_magic_module =
{
    STANDARD_MODULE_STUFF,
    magic_init,			/* initializer */
    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 table */
    NULL,			/* handlers */
    NULL,			/* filename translation */
    NULL,			/* check_user_id */
    NULL,			/* check auth */
    NULL,			/* check access */
    magic_find_ct,		/* type_checker */
    NULL,			/* fixups */
    NULL,			/* logger */
    NULL,			/* header parser */
    NULL,			/* child_init */
    NULL,			/* child_exit */
    NULL			/* post read-request */
};