mime_magic.c

php-4.4.7学习linux时下载的源代码

		if (cur_frag < start_frag)
			continue;

		/* loop through and collect chars */
		for (cur_pos = (cur_frag == start_frag) ? start_pos : 0;
			 frag->str[cur_pos];
			 cur_pos++) {
			if (cur_frag >= start_frag
				&& cur_pos >= start_pos
				&& res_pos <= len) {
				result[res_pos++] = frag->str[cur_pos];
				if (res_pos > len) {
					break;
				}
			}
		}
    }

    /* clean up and return */
    result[res_pos] = 0;

    return result;
}

/*
 * 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(char *filename TSRMLS_DC)
{
	php_stream *stream;
    unsigned char buf[HOWMANY + 1];	/* one extra for terminating '\0' */
    int nbytes = 0;		/* number of bytes read from a datafile */
    int result;

    /*
     * first try judging the file based on its filesystem status
     */
    switch ((result = fsmagic(filename TSRMLS_CC))) {
    case MIME_MAGIC_DONE:
		magic_rsl_putchar('\n');
		return MIME_MAGIC_OK;
    case MIME_MAGIC_OK:
		break;
    default:
		/* fatal error, bail out */
		return result;
    }

    stream = php_stream_open_wrapper(filename, "rb", IGNORE_PATH | ENFORCE_SAFE_MODE | REPORT_ERRORS, NULL);

    if (stream == NULL) {
		/* We can't open it, but we were able to stat it. */
		php_error(E_WARNING,
					  MODNAME ": can't read `%s'", filename);
		/* let some other handler decide what the problem is */
		return MIME_MAGIC_DECLINED;
    }

    /*
     * try looking at the first HOWMANY bytes
     */
    if ((nbytes = php_stream_read(stream, (char *) buf, sizeof(buf) - 1)) == -1) {
		php_error(E_WARNING,
					  MODNAME ": read failed: %s", filename);
		return MIME_MAGIC_ERROR;
    }

    if (nbytes == 0)
		magic_rsl_puts(MIME_TEXT_UNKNOWN);
    else {
		buf[nbytes++] = '\0';	/* null-terminate it */
		tryit(buf, nbytes, 1); 
    }

    (void) php_stream_close(stream);
    (void) magic_rsl_putchar('\n');

    return MIME_MAGIC_OK;
}

static void tryit(unsigned char *buf, int nb, int checkzmagic)
{
    /*
     * Try compression stuff
     */
#if HAVE_ZLIB && !defined(COMPILE_DL_ZLIB)
	if (checkzmagic == 1) {  
		if (zmagic(buf, nb) == 1)
			return;
	}
#endif

    /*
     * try tests in /etc/magic (or surrogate magic file)
     */
    if (softmagic(buf, nb) == 1)
		return;

    /*
     * try known keywords, check for ascii-ness too.
     */
    if (ascmagic(buf, nb) == 1)
		return;

    /*
     * abandon hope, all ye who remain here
     */
    magic_rsl_puts(MIME_BINARY_UNKNOWN);
}


/*
 * return MIME_MAGIC_DONE to indicate it's been handled
 * return MIME_MAGIC_OK to indicate it's a regular file still needing handling
 * other returns indicate a failure of some sort
 */
static int fsmagic(char *filename TSRMLS_DC)
{
	php_stream_statbuf stat_ssb;

	if(!php_stream_stat_path(filename, &stat_ssb)) {
		return MIME_MAGIC_OK;
	}

    switch (stat_ssb.sb.st_mode & S_IFMT) {
    case S_IFDIR:
		magic_rsl_puts(DIR_MAGIC_TYPE);
		return MIME_MAGIC_DONE;
    case S_IFCHR:
		/*
		 * (void) magic_rsl_printf(r,"character special (%d/%d)",
		 * major(sb->st_rdev), minor(sb->st_rdev));
		 */
		(void) magic_rsl_puts(MIME_BINARY_UNKNOWN);
		return MIME_MAGIC_DONE;
#ifdef S_IFBLK
    case S_IFBLK:
		/*
		 * (void) magic_rsl_printf(r,"block special (%d/%d)",
		 * major(sb->st_rdev), minor(sb->st_rdev));
		 */
		(void) magic_rsl_puts(MIME_BINARY_UNKNOWN);
		return MIME_MAGIC_DONE;
		/* TODO add code to handle V7 MUX and Blit MUX files */
#endif
#ifdef    S_IFIFO
    case S_IFIFO:
		/*
		 * magic_rsl_puts(r,"fifo (named pipe)");
		 */
		(void) magic_rsl_puts(MIME_BINARY_UNKNOWN);
		return MIME_MAGIC_DONE;
#endif
#ifdef    S_IFLNK
    case S_IFLNK:
		/* We used stat(), the only possible reason for this is that the
		 * symlink is broken.
		 */
		php_error(E_WARNING,
					  MODNAME ": broken symlink (%s)", filename);
		return MIME_MAGIC_ERROR;
#endif
#ifdef    S_IFSOCK
#ifndef __COHERENT__
    case S_IFSOCK:
		magic_rsl_puts(MIME_BINARY_UNKNOWN);
		return MIME_MAGIC_DONE;
#endif
#endif
    case S_IFREG:
		break;
	case 0:
		break;
    default:
		php_error(E_WARNING,
					  MODNAME ": invalid mode 0%o.", (unsigned int)stat_ssb.sb.st_mode);
		return MIME_MAGIC_ERROR;
    }

    /*
     * regular file, check next possibility
     */
    if (stat_ssb.sb.st_size == 0) {
		magic_rsl_puts(MIME_TEXT_UNKNOWN);
		return MIME_MAGIC_DONE;
    }
    return MIME_MAGIC_OK;
}

/*
 * softmagic - lookup one file in database (already read from /etc/magic by
 * apprentice.c). Passed the name and FILE * of one file to be typed.
 */
/* ARGSUSED1 *//* nbytes passed for regularity, maybe need later */
static int softmagic(unsigned char *buf, int nbytes)
{
    if (match(buf, nbytes))
		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.
 */
static int match(unsigned char *s, int nbytes)
{
    int cont_level = 0;
    int need_separator = 0;
    union VALUETYPE p;
    magic_server_config_rec *conf = &mime_global;
    struct magic *m;

    for (m = conf->magic; m; m = m->next) {
		/* check if main entry matches */
		if (!mget(&p, s, m, nbytes) ||
			!mcheck(&p, m)) {
			struct magic *m_cont;

			/*
			 * main entry didn't match, flush its continuations
			 */
			if (!m->next || (m->next->cont_level == 0)) {
				continue;
			}

			m_cont = m->next;
			while (m_cont && (m_cont->cont_level != 0)) {
				/*
				 * this trick allows us to keep *m in sync when the continue
				 * advances the pointer
				 */
				m = m_cont;
				m_cont = m_cont->next;
			}
			continue;
		}

		/* if we get here, the main entry rule was a match */
		/* this will be the last run through the loop */

		/* print the match */
		mprint(&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 (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;
				}
				if (mget(&p, s, m, nbytes) &&
					mcheck(&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(' ');
						need_separator = 0;
					}
					mprint(&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;
		}
		return 1;		/* all through */
    }
    return 0;			/* no match at all */
}

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

static int ascmagic(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 (isspace(*tp))
			++tp;		/* skip leading whitespace */
		if ((isalnum(*tp) || *tp == '\\') &&
			(isalnum(*(tp + 1)) || *tp == '"')) {
			magic_rsl_puts("application/x-troff");
			return 1;
		}
    }
    if ((*buf == 'c' || *buf == 'C') && isspace(*(buf + 1))) {
		/* Fortran */
		magic_rsl_puts("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(types[p->type]);
				if (has_escapes)
					magic_rsl_puts(" (with escape sequences)");
				return 1;
			}
		}
    }

    switch (is_tar(buf, nbytes)) {
    case 1:
		/* V7 tar archive */
		magic_rsl_puts("application/x-tar");
		return 1;
    case 2:
		/* POSIX tar archive */
		magic_rsl_puts("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("text/plain");
    }
    else {
		/* plain text */
		magic_rsl_puts("text/plain");
    }
    return 1;
}

/*
 * 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).
 */