add apple creator/type

19 files changed, 106 insertions, 33 deletions

diff --git a/ChangeLog b/ChangeLog
index 1cd8d338..42a5c132 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,7 @@
+2008-10-18 16:45  Christos Zoulas <christos@astron.com>
+	
+	* Added APPLE file creator/type
+
 2008-10-12 10:20  Christos Zoulas <christos@astron.com>
 
 	* Added CDF parsing
diff --git a/doc/file.man b/doc/file.man
index 4b046868..f033770b 100644
--- a/doc/file.man
+++ b/doc/file.man
@@ -1,4 +1,4 @@
-.\" $File: file.man,v 1.75 2008/10/09 17:25:01 christos Exp $
+.\" $File: file.man,v 1.76 2008/10/18 20:47:47 christos Exp $
 .Dd October 9, 2008
 .Dt FILE __CSECTION__
 .Os
@@ -182,6 +182,8 @@ Check for
 application type (only on EMX).
 .It ascii
 Check for various types of ascii files.
+.It cdf
+Don't look for Compound Document Files.
 .It compress
 Don't look for, or inside compressed files.
 .It elf
diff --git a/doc/magic.man b/doc/magic.man
index 314a0147..893199a4 100644
--- a/doc/magic.man
+++ b/doc/magic.man
@@ -1,4 +1,4 @@
-.\" $File: magic.man,v 1.57 2008/08/30 09:50:20 christos Exp $
+.\" $File: magic.man,v 1.58 2008/10/18 20:47:47 christos Exp $
 .Dd August 30, 2008
 .Dt MAGIC __FSECTION__
 .Os
@@ -290,6 +290,11 @@ added before it: multiple matches are normally separated by a single
 space.
 .El
 .Pp
+An APPLE 4+4 character APPLE creator and type can be specified as:
+.Bd -literal -offset indent 
+!:apple	CREATYPE
+.Ed
+.Pp
 A MIME type is given on a separate line, which must be the next
 non-blank or comment line after the magic line that identifies the
 file type, and has the following format:
diff --git a/magic/Magdir/compress b/magic/Magdir/compress
index e2e4e032..016d0932 100644
--- a/magic/Magdir/compress
+++ b/magic/Magdir/compress
@@ -11,6 +11,7 @@
 # standard unix compress
 0	string		\037\235	compress'd data
 !:mime	application/x-compress
+!:apple	LZIVZIVU
 >2	byte&0x80	>0		block compressed
 >2	byte&0x1f	x		%d bits
 
diff --git a/magic/Magdir/images b/magic/Magdir/images
index cccc70be..22474538 100644
--- a/magic/Magdir/images
+++ b/magic/Magdir/images
@@ -110,6 +110,7 @@
 # GIF
 0	string		GIF8		GIF image data
 !:mime	image/gif
+!:apple	8BIMGIFf
 >4	string		7a		\b, version 8%s,
 >4	string		9a		\b, version 8%s,
 >6	leshort		>0		%hd x
diff --git a/magic/Magdir/jpeg b/magic/Magdir/jpeg
index d728de6e..4470be4a 100644
--- a/magic/Magdir/jpeg
+++ b/magic/Magdir/jpeg
@@ -10,6 +10,7 @@
 #
 0	beshort		0xffd8		JPEG image data
 !:mime	image/jpeg
+!:apple	8BIMJPEG
 !:strength +1
 >6	string		JFIF		\b, JFIF standard
 # The following added by Erik Rossen <rossen@freesurf.ch> 1999-09-06
diff --git a/magic/Magdir/macintosh b/magic/Magdir/macintosh
index 77187a39..ca665ded 100644
--- a/magic/Magdir/macintosh
+++ b/magic/Magdir/macintosh
@@ -11,6 +11,8 @@
 # Stuffit archives are the de facto standard of compression for Macintosh
 # files obtained from most archives. (franklsm@tuns.ca)
 0	string		SIT!			StuffIt Archive (data)
+!:mime	application/x-stuffit
+!:apple	SIT!SIT!
 >2	string		x			: %s
 0	string		SITD			StuffIt Deluxe (data)
 >2	string		x			: %s
@@ -20,6 +22,7 @@
 # Newer StuffIt archives (grant@netbsd.org)
 0	string		StuffIt			StuffIt Archive
 !:mime	application/x-stuffit
+!:apple	SIT!SIT!
 #>162	string		>0			: %s
 
 # Macintosh Applications and Installation binaries (franklsm@tuns.ca)
diff --git a/magic/Magdir/printer b/magic/Magdir/printer
index 15a17582..e25d03e9 100644
--- a/magic/Magdir/printer
+++ b/magic/Magdir/printer
@@ -6,6 +6,7 @@
 # PostScript, updated by Daniel Quinlan (quinlan@yggdrasil.com)
 0	string		%!		PostScript document text
 !:mime	application/postscript
+!:apple	ASPSTEXT
 >2	string		PS-Adobe-	conforming
 >>11	string		>\0		DSC level %.3s
 >>>15	string		EPS		\b, type %s
@@ -16,6 +17,7 @@
 # Some PCs have the annoying habit of adding a ^D as a document separator
 0	string		\004%!		PostScript document text
 !:mime	application/postscript
+!:apple	ASPSTEXT
 >3	string		PS-Adobe-	conforming
 >>12	string		>\0		DSC level %.3s
 >>>16	string		EPS		\b, type %s
diff --git a/src/apprentice.c b/src/apprentice.c
index 244906aa..7ecfd8bd 100644
--- a/src/apprentice.c
+++ b/src/apprentice.c
@@ -49,7 +49,7 @@
 #include <dirent.h>
 
 #ifndef	lint
-FILE_RCSID("@(#)$File: apprentice.c,v 1.141 2008/08/31 07:58:00 christos Exp $")
+FILE_RCSID("@(#)$File: apprentice.c,v 1.142 2008/10/18 20:47:48 christos Exp $")
 #endif	/* lint */
 
 #define	EATAB {while (isascii((unsigned char) *l) && \
@@ -115,8 +115,8 @@ private int check_format_type(const char *, int);
 private int check_format(struct magic_set *, struct magic *);
 private int get_op(char);
 private int parse_mime(struct magic_set *, struct magic_entry *, const char *);
-private int parse_strength(struct magic_set *, struct magic_entry *,
-    const char *);
+private int parse_strength(struct magic_set *, struct magic_entry *, const char *);
+private int parse_apple(struct magic_set *, struct magic_entry *, const char *);
 
 
 private size_t maxmagic = 0;
@@ -131,6 +131,7 @@ private struct {
 } bang[] = {
 #define	DECLARE_FIELD(name) { # name, sizeof(# name) - 1, parse_ ## name }
 	DECLARE_FIELD(mime),
+	DECLARE_FIELD(apple),
 	DECLARE_FIELD(strength),
 #undef	DECLARE_FIELD
 	{ NULL, 0, NULL }
@@ -1474,6 +1475,38 @@ out:
 }
 
 /*
+ * Parse an Apple CREATOR/TYPE annotation from magic file and put it into magic[index - 1]
+ */
+private int
+parse_apple(struct magic_set *ms, struct magic_entry *me, const char *line)
+{
+	size_t i;
+	const char *l = line;
+	struct magic *m = &me->mp[me->cont_count == 0 ? 0 : me->cont_count - 1];
+
+	if (m->apple[0] != '\0') {
+		file_magwarn(ms, "Current entry already has a APPLE type `%.8s',"
+		    " new type `%s'", m->mimetype, l);
+		return -1;
+	}	
+
+	EATAB;
+	for (i = 0; *l && ((isascii((unsigned char)*l) && isalnum((unsigned char)*l))
+	     || strchr("-+/.", *l)) && i < sizeof(m->apple); m->apple[i++] = *l++)
+		continue;
+	if (i == sizeof(m->apple) && *l) {
+		if (ms->flags & MAGIC_CHECK)
+			file_magwarn(ms, "APPLE type `%s' truncated %zu",
+			    line, i);
+	}
+
+	if (i > 0)
+		return 0;
+	else
+		return -1;
+}
+
+/*
  * parse a MIME annotation line from magic file, put into magic[index - 1]
  * if valid
  */
@@ -1491,10 +1524,8 @@ parse_mime(struct magic_set *ms, struct magic_entry *me, const char *line)
 	}	
 
 	EATAB;
-	for (i = 0;
-	     *l && ((isascii((unsigned char)*l) && isalnum((unsigned char)*l))
-	     || strchr("-+/.", *l)) && i < sizeof(m->mimetype);
-	     m->mimetype[i++] = *l++)
+	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';
diff --git a/src/ascmagic.c b/src/ascmagic.c
index 841dea00..c40942f9 100644
--- a/src/ascmagic.c
+++ b/src/ascmagic.c
@@ -49,7 +49,7 @@
 #include "names.h"
 
 #ifndef	lint
-FILE_RCSID("@(#)$File: ascmagic.c,v 1.66 2008/10/16 16:31:16 christos Exp $")
+FILE_RCSID("@(#)$File: ascmagic.c,v 1.67 2008/10/18 20:47:48 christos Exp $")
 #endif	/* lint */
 
 #define MAXLINELEN 300	/* longest sane line length */
@@ -97,6 +97,9 @@ file_ascmagic(struct magic_set *ms, const unsigned char *buf, size_t nbytes)
 	size_t last_line_end = (size_t)-1;
 	int has_long_lines = 0;
 
+	if (ms->flags & MAGIC_APPLE)
+		return 0;
+
 	/*
 	 * Undo the NUL-termination kindly provided by process()
 	 * but leave at least one byte to look at
diff --git a/src/file.c b/src/file.c
index effa06a0..ab3e6c59 100644
--- a/src/file.c
+++ b/src/file.c
@@ -74,7 +74,7 @@ int getopt_long(int argc, char * const *argv, const char *optstring, const struc
 #include "patchlevel.h"
 
 #ifndef	lint
-FILE_RCSID("@(#)$File: file.c,v 1.125 2008/10/12 15:38:51 christos Exp $")
+FILE_RCSID("@(#)$File: file.c,v 1.126 2008/10/18 20:47:48 christos Exp $")
 #endif	/* lint */
 
 
@@ -193,9 +193,12 @@ main(int argc, char *argv[])
 				help();
 				break;
 			case 10:
-				flags |= MAGIC_MIME_TYPE;
+				flags |= MAGIC_APPLE;
 				break;
 			case 11:
+				flags |= MAGIC_MIME_TYPE;
+				break;
+			case 12:
 				flags |= MAGIC_MIME_ENCODING;
 				break;
 			}
diff --git a/src/file.h b/src/file.h
index 668bfb25..23088821 100644
--- a/src/file.h
+++ b/src/file.h
@@ -27,7 +27,7 @@
  */
 /*
  * file.h - definitions for file(1) program
- * @(#)$File: file.h,v 1.110 2008/10/12 15:38:52 christos Exp $
+ * @(#)$File: file.h,v 1.111 2008/10/18 20:47:48 christos Exp $
  */
 
 #ifndef __file_h__
@@ -103,8 +103,8 @@
 #define MAXstring 32		/* max leng of "string" types */
 
 #define MAGICNO		0xF11E041C
-#define VERSIONNO	6
-#define FILE_MAGICSIZE	(32 * 6)
+#define VERSIONNO	7
+#define FILE_MAGICSIZE	200
 
 #define	FILE_LOAD	0
 #define FILE_CHECK	1
@@ -258,10 +258,12 @@ struct magic {
 #define str_flags _u._s._flags
 	/* Words 9-16 */
 	union VALUETYPE value;		/* either number or string */
-	/* Words 17..31 */
+	/* Words 17-25 */
 	char desc[MAXDESC];	/* description */
-	/* Words 32..47 */
+	/* Words 26-34 */
 	char mimetype[MAXDESC]; /* MIME type */
+	/* Words 35-43 */
+	char apple[8];
 };
 
 #define BIT(A)   (1 << (A))
diff --git a/src/file_opts.h b/src/file_opts.h
index 46bc08ae..f38354ad 100644
--- a/src/file_opts.h
+++ b/src/file_opts.h
@@ -28,6 +28,7 @@ OPT('f', "files-from", 1, " FILE      read the filenames to be examined from FIL
 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("apple", 0, "            output the APPLE CREATOR/TYPE\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")
diff --git a/src/fsmagic.c b/src/fsmagic.c
index 0d055493..d643f1f1 100644
--- a/src/fsmagic.c
+++ b/src/fsmagic.c
@@ -57,7 +57,7 @@
 #undef HAVE_MAJOR
 
 #ifndef	lint
-FILE_RCSID("@(#)$File: fsmagic.c,v 1.53 2008/10/16 16:31:16 christos Exp $")
+FILE_RCSID("@(#)$File: fsmagic.c,v 1.54 2008/10/18 20:47:48 christos Exp $")
 #endif	/* lint */
 
 private int
@@ -106,6 +106,8 @@ file_fsmagic(struct magic_set *ms, const char *fn, struct stat *sb)
 	struct stat tstatbuf;
 #endif
 
+	if (ms->flags & MAGIC_APPLE)
+		return 0;
 	if (fn == NULL)
 		return 0;
 
diff --git a/src/is_tar.c b/src/is_tar.c
index a9311117..98ff7371 100644
--- a/src/is_tar.c
+++ b/src/is_tar.c
@@ -45,7 +45,7 @@
 #include "tar.h"
 
 #ifndef lint
-FILE_RCSID("@(#)$File: is_tar.c,v 1.31 2008/02/04 20:51:17 christos Exp $")
+FILE_RCSID("@(#)$File: is_tar.c,v 1.32 2008/10/18 20:47:48 christos Exp $")
 #endif
 
 #define	isodigit(c)	( ((c) >= '0') && ((c) <= '7') )
@@ -66,13 +66,14 @@ 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 tar;
 	int mime = ms->flags & MAGIC_MIME;
 
-	if (tar < 1 || tar > 3)
+	if ((ms->flags & MAGIC_APPLE) != 0 || mime == MAGIC_MIME_ENCODING)
 		return 0;
 
-	if (mime == MAGIC_MIME_ENCODING)
+	tar = is_tar(buf, nbytes);
+	if (tar < 1 || tar > 3)
 		return 0;
 
 	if (file_printf(ms, mime ? "application/x-tar" :
diff --git a/src/magic.h b/src/magic.h
index 2c085701..5ff78a29 100644
--- a/src/magic.h
+++ b/src/magic.h
@@ -42,6 +42,7 @@
 #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_APPLE		0x000800 /* Return APPLE CREATOR/TYPE */
 #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 */
diff --git a/src/readcdf.c b/src/readcdf.c
index d4de5d2f..5908473f 100644
--- a/src/readcdf.c
+++ b/src/readcdf.c
@@ -26,7 +26,7 @@
 #include "file.h"
 
 #ifndef lint
-FILE_RCSID("@(#)$File: readcdf.c,v 1.6 2008/10/14 10:38:22 christos Exp $")
+FILE_RCSID("@(#)$File: readcdf.c,v 1.7 2008/10/18 20:47:48 christos Exp $")
 #endif
 
 #include <stdio.h>
@@ -191,6 +191,8 @@ file_trycdf(struct magic_set *ms, int fd, const unsigned char *buf,
 	(void)&nbytes;
 	(void)&buf;
 
+	if (ms->flags & MAGIC_APPLE)
+		return 0;
 	if (cdf_read_header(fd, &h) == -1)
 		return 0;
 #ifdef CDF_DEBUG
diff --git a/src/readelf.c b/src/readelf.c
index 57aa200d..31b36311 100644
--- a/src/readelf.c
+++ b/src/readelf.c
@@ -38,7 +38,7 @@
 #include "magic.h"
 
 #ifndef lint
-FILE_RCSID("@(#)$File: readelf.c,v 1.79 2008/10/12 19:06:36 christos Exp $")
+FILE_RCSID("@(#)$File: readelf.c,v 1.80 2008/10/18 20:47:48 christos Exp $")
 #endif
 
 #ifdef	ELFCORE
@@ -1141,7 +1141,7 @@ file_tryelf(struct magic_set *ms, int fd, const unsigned char *buf,
 	Elf64_Ehdr elf64hdr;
 	uint16_t type;
 
-	if (ms->flags & MAGIC_MIME)
+	if (ms->flags & (MAGIC_MIME|MAGIC_APPLE))
 		return 0;
 	/*
 	 * ELF executables have multiple section headers in arbitrary
diff --git a/src/softmagic.c b/src/softmagic.c
index ac2b8bef..5eab3f68 100644
--- a/src/softmagic.c
+++ b/src/softmagic.c
@@ -38,7 +38,7 @@
 
 
 #ifndef	lint
-FILE_RCSID("@(#)$File: softmagic.c,v 1.123 2008/10/16 16:31:16 christos Exp $")
+FILE_RCSID("@(#)$File: softmagic.c,v 1.124 2008/10/18 20:47:48 christos Exp $")
 #endif	/* lint */
 
 private int match(struct magic_set *, struct magic *, uint32_t,
@@ -52,7 +52,7 @@ 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 int handle_mime(struct magic_set *, struct magic *);
+private int handle_annotation(struct magic_set *, struct magic *);
 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 *);
@@ -168,7 +168,7 @@ match(struct magic_set *ms, struct magic *magic, uint32_t nmagic,
 		if (*m->desc) {
 			need_separator = 1;
 			printed_something = 1;
-			if ((e = handle_mime(ms, m)) != 0)
+			if ((e = handle_annotation(ms, m)) != 0)
 				return e;
 			if (print_sep(ms, firstline) == -1)
 				return -1;
@@ -236,7 +236,7 @@ match(struct magic_set *ms, struct magic *magic, uint32_t nmagic,
 				 */
 				if (*m->desc) {
 					printed_something = 1;
-					if ((e = handle_mime(ms, m)) != 0)
+					if ((e = handle_annotation(ms, m)) != 0)
 						return e;
 					if (print_sep(ms, firstline) == -1)
 						return -1;
@@ -1592,7 +1592,8 @@ magiccheck(struct magic_set *ms, struct magic *m)
 	
 		default:
 			matched = 0;
-			file_magerror(ms, "cannot happen with float: invalid relation `%c'", m->reln);
+			file_magerror(ms, "cannot happen with float: invalid relation `%c'",
+			    m->reln);
 			return -1;
 		}
 		return matched;
@@ -1820,10 +1821,17 @@ magiccheck(struct magic_set *ms, struct magic *m)
 }
 
 private int
-handle_mime(struct magic_set *ms, struct magic *m)
+handle_annotation(struct magic_set *ms, struct magic *m)
 {
-	if (ms->flags & MAGIC_MIME_TYPE)
-		return file_printf(ms, "%s", m->mimetype);
+	if (ms->flags & MAGIC_APPLE) {
+		if (file_printf(ms, "%.8s", m->apple) == -1)
+			return -1;
+		return 1;
+	}
+	if (ms->flags & MAGIC_MIME_TYPE) {
+		if (file_printf(ms, "%.8s", m->mimetype) == -1)
+		return 1;
+	}
 	return 0;
 }