Integrated file-3.17

usr.bin/file/apprentice.c

@@ -33,20 +33,24 @@
 
 #ifndef	lint
 static char *moduleid = 
-	"@(#)$Id: apprentice.c,v 1.6 1995/03/25 22:35:35 christos Exp $";
+	"@(#)$Id: apprentice.c,v 1.7 1995/04/28 19:23:38 christos Exp $";
 #endif	/* lint */
 
 #define	EATAB {while (isascii((unsigned char) *l) && \
 		      isspace((unsigned char) *l))  ++l;}
+#define LOWCASE(l) (isupper((unsigned char) (l)) ? \
+			tolower((unsigned char) (l)) : (l))
 
 
 static int getvalue	__P((struct magic *, char **));
 static int hextoint	__P((int));
 static char *getstr	__P((char *, char *, int, int *));
 static int parse	__P((char *, int *, int));
+static void eatsize	__P((char **));
 
 static int maxmagic = 0;
 
+
 int
 apprentice(fn, check)
 char *fn;			/* name of magic file */
@@ -186,13 +190,16 @@ int *ndx, check;
 		 * read [.lbs][+-]nnnnn)
 		 */
 		if (*l == '.') {
-			switch (*++l) {
+			l++;
+			switch (LOWCASE(*l)) {
 			case 'l':
 				m->in.type = LONG;
 				break;
+			case 'h':
 			case 's':
 				m->in.type = SHORT;
 				break;
+			case 'c':
 			case 'b':
 				m->in.type = BYTE;
 				break;
@@ -279,6 +286,7 @@ int *ndx, check;
 	if (*l == '&') {
 		++l;
 		m->mask = signextend(m, strtoul(l, &l, 0));
+		eatsize(&l);
 	} else
 		m->mask = ~0L;
 	EATAB;
@@ -360,8 +368,10 @@ char **p;
 		*p = getstr(*p, m->value.s, sizeof(m->value.s), &slen);
 		m->vallen = slen;
 	} else
-		if (m->reln != 'x')
+		if (m->reln != 'x') {
 			m->value.l = signextend(m, strtoul(*p, p, 0));
+			eatsize(&p);
+		}
 	return 0;
 }
 
@@ -549,3 +559,30 @@ int len;
 		}
 	}
 }
+
+/*
+ * eatsize(): Eat the size spec from a number [eg. 10UL]
+ */
+static void
+eatsize(p)
+char **p;
+{
+	char *l = *p;
+
+	if (LOWCASE(*l) == 'u') 
+		l++;
+
+	switch (LOWCASE(*l)) {
+	case 'l':    /* long */
+	case 's':    /* short */
+	case 'h':    /* short */
+	case 'b':    /* char/byte */
+	case 'c':    /* char/byte */
+		l++;
+		/*FALLTHROUGH*/
+	default:
+		break;
+	}
+
+	*p = l;
+}

usr.bin/file/ascmagic.c

@@ -36,7 +36,7 @@
 
 #ifndef	lint
 static char *moduleid = 
-	"@(#)$Id: ascmagic.c,v 1.5 1995/03/25 22:35:37 christos Exp $";
+	"@(#)$Id: ascmagic.c,v 1.6 1995/04/28 19:23:43 christos Exp $";
 #endif	/* lint */
 
 			/* an optimisation over plain strcmp() */
@@ -77,13 +77,13 @@ int nbytes;	/* size actually read */
 		while (isascii(*tp) && isspace(*tp))
 			++tp;	/* skip leading whitespace */
 		if ((isascii(*tp) && (isalnum(*tp) || *tp=='\\') &&
-		    isascii(*(tp+1)) && (isalnum(*(tp+1)) || *tp=='"'))) {
+		    isascii(tp[1]) && (isalnum(tp[1]) || tp[1] == '"'))) {
 			ckfputs("troff or preprocessor input text", stdout);
 			return 1;
 		}
 	}
 	if ((*buf == 'c' || *buf == 'C') && 
-	    isascii(*(buf + 1)) && isspace(*(buf + 1))) {
+	    isascii(buf[1]) && isspace(buf[1])) {
 		ckfputs("fortran program text", stdout);
 		return 1;
 	}
@@ -93,7 +93,7 @@ int nbytes;	/* size actually read */
 	s = (unsigned char*) memcpy(nbuf, buf, nbytes);
 	s[nbytes] = '\0';
 	has_escapes = (memchr(s, '\033', nbytes) != NULL);
-	while ((token = strtok((char*)s, " \t\n\r\f")) != NULL) {
+	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)) {
@@ -108,7 +108,7 @@ int nbytes;	/* size actually read */
 
 
 	for (i = 0; i < nbytes; i++) {
-		if (!isascii(*(buf+i)))
+		if (!isascii(buf[i]))
 			return 0;	/* not all ascii */
 	}
 

usr.bin/file/file.c

@@ -26,7 +26,7 @@
  */
 #ifndef	lint
 static char *moduleid = 
-	"@(#)$Id: file.c,v 1.4 1995/03/25 22:35:46 christos Exp $";
+	"@(#)$Id: file.c,v 1.5 1995/04/28 19:23:46 christos Exp $";
 #endif	/* lint */
 
 #include <stdio.h>
@@ -43,6 +43,10 @@ static char *moduleid =
 #endif
 #include <unistd.h>	/* for read() */
 
+#ifdef HAVE_ELF
+#include <elf.h>
+#endif
+
 #include "patchlevel.h"
 #include "file.h"
 
@@ -66,7 +70,7 @@ int			/* Misc globals				*/
 
 struct  magic *magic;	/* array of magic entries		*/
 
-char *magicfile = MAGIC;/* where magic be found 		*/
+char *magicfile;	/* where magic be found 		*/
 
 char *progname;		/* used throughout 			*/
 int lineno;		/* line number in the magic file	*/
@@ -90,6 +94,9 @@ char *argv[];
 	else
 		progname = argv[0];
 
+	if (!(magicfile = getenv("MAGIC")))
+		magicfile = MAGIC;
+
 	while ((c = getopt(argc, argv, "vcdf:Lm:z")) != EOF)
 		switch (c) {
 		case 'v':
@@ -254,15 +261,47 @@ int wid;
 	if (nbytes == 0) 
 		ckfputs("empty", stdout);
 	else {
-		buf[nbytes++] = '\0';	/* NULL terminate */
+		buf[nbytes++] = '\0';   /* NULL terminate */
-		if (nbytes < sizeof(union VALUETYPE)) {
-		    /* The following is to handle *very* short files */
-		    memset(buf + nbytes, 0, sizeof(union VALUETYPE) - nbytes);
-		    nbytes = sizeof(union VALUETYPE);
-		}
 		tryit(buf, nbytes, zflag);
 	}
 
+#ifdef HAVE_ELF
+	/*
+	 * ELF executables have multiple section headers in arbitrary
+	 * file locations and thus file(1) cannot determine it from easily.
+	 * Instead we traverse thru all section headers until a symbol table
+	 * one is found or else the binary is stripped.
+	 * XXX: This will not work for binaries of a different byteorder.
+	 *	Should come up with a better fix.
+	 */
+
+	if (nbytes > sizeof (Elf32_Ehdr) &&
+	    buf[EI_MAG0] == ELFMAG0 &&
+	    buf[EI_MAG1] == ELFMAG1 &&
+	    buf[EI_MAG2] == ELFMAG2 &&
+	    buf[EI_MAG3] == ELFMAG3 ) {
+
+		Elf32_Ehdr elfhdr;
+		int stripped = 1;
+ 
+		(void) memcpy(&elfhdr, buf, sizeof elfhdr);
+ 
+		if (lseek(fd, elfhdr.e_shoff, SEEK_SET)<0)
+		  error("lseek failed (%s).\n", strerror(errno));
+ 
+		for ( ; elfhdr.e_shnum ; elfhdr.e_shnum--) {
+			if (read(fd, buf, elfhdr.e_shentsize)<0)
+				error("read failed (%s).\n", strerror(errno));
+			if (((Elf32_Shdr *)&buf)->sh_type == SHT_SYMTAB) {
+				stripped = 0;
+				break;
+			}
+		}
+		if (stripped)
+		  (void) printf (" - stripped");
+	}
+#endif
+
 	if (inname != stdname) {
 		/*
 		 * Try to restore access, modification times if read it.

usr.bin/file/file.h

@@ -1,6 +1,6 @@
 /*
  * file.h - definitions for file(1) program
- * @(#)$Id: file.h,v 1.5 1995/03/25 22:35:51 christos Exp $
+ * @(#)$Id: file.h,v 1.6 1995/04/28 19:23:48 christos Exp $
  *
  * Copyright (c) Ian F. Darwin, 1987.
  * Written by Ian F. Darwin.
@@ -26,7 +26,9 @@
  * 4. This notice may not be removed or altered.
  */
 
-#define HOWMANY	8192		/* how much of the file to look at */
+#ifndef HOWMANY
+# define HOWMANY 8192		/* how much of the file to look at */
+#endif
 #define MAXMAGIS 1000		/* max entries in /etc/magic */
 #define MAXDESC	50		/* max leng of text description */
 #define MAXstring 32		/* max leng of "string" types */

usr.bin/file/fsmagic.c

@@ -46,7 +46,7 @@
 
 #ifndef	lint
 static char *moduleid = 
-	"@(#)$Id: fsmagic.c,v 1.6 1995/03/25 22:35:58 christos Exp $";
+	"@(#)$Id: fsmagic.c,v 1.7 1995/04/28 19:23:51 christos Exp $";
 #endif	/* lint */
 
 int

usr.bin/file/is_tar.c

@@ -5,7 +5,7 @@
  * Pubic Domain version written 26 Aug 1985 John Gilmore (ihnp4!hoptoad!gnu).
  *
  * @(#)list.c 1.18 9/23/86 Public Domain - gnu
- * $Id: is_tar.c,v 1.4 1995/03/25 22:36:07 christos Exp $
+ * $Id: is_tar.c,v 1.5 1995/04/28 19:23:54 christos Exp $
  *
  * Comments changed and some code/comments reformatted
  * for file command by Ian Darwin.

usr.bin/file/magdir/audio

@@ -2,7 +2,8 @@
 #------------------------------------------------------------------------------
 # audio:  file(1) magic for sound formats
 #
-# from Jan Nicolai Langfeldt <janl@ifi.uio.no>,
+# Jan Nicolai Langfeldt (janl@ifi.uio.no), Dan Quinlan (quinlan@yggdrasil.com),
+# and others
 #
 
 # Sun/NeXT audio data
@@ -66,3 +67,12 @@
 >22	short		=2		Stereo
 >22	short		>2		%d Channels
 >24	long		>0		%d Hz
+
+# Extended MOD format (*.emd) (Greg Roelofs, newt@uchicago.edu); NOT TESTED
+# [based on posting 940824 by "Dirk/Elastik", husberg@lehtori.cc.tut.fi]
+0	string		EMOD		Extended MOD sound data,
+>4	byte&0xf0	x		version %d
+>4	byte&0x0f	x		\b.%d,
+>45	byte		x		%d instruments
+>83	byte		0		(module)
+>83	byte		1		(song)

usr.bin/file/magdir/compress

@@ -43,7 +43,10 @@
 >2	belong		=1		%d character originally
 #
 # This magic number is byte-order-independent.
+# Are there two types of old packed data?  One big endian, one little endian?
 #
+#0	beshort		017437		old packed data, big endian
+#0	leshort		017437		old packed data, little endian
 0	short		017437		old packed data
 
 # XXX - why *two* entries for "compacted data", one of which is

usr.bin/file/magdir/database

@@ -4,12 +4,27 @@
 #
 # extracted from header/code files by Graeme Wilford (eep2gw@ee.surrey.ac.uk)
 #
-0	long		0x053162	Berkeley db BTREE database
->4	byte		1		(version 1)
->4	byte		2		(version 2)
->4	byte		3		(version 3)
-0	long		0x061561	Berkeley db HASH database
->4	byte		1		(version 1)
->4	byte		2		(version 2)
->4	byte		3		(version 3)
 0	long		0x13579ace	GNU gdbm or ndbm database
+#
+0	belong	0x061561	Berkeley DB Hash file
+>4	belong	>0		(Version %d,
+>8	belong	1234		Little Endian,
+>8	belong	4321		Big Endian,
+>12	belong	x		Bucket Size %d,
+>16	belong	x		Directory Size %d,
+>20	belong	x		Segment Size %d,
+>24	belong	x		Segment Shift %d,
+>28	belong	x		Overflow Point %d,
+>32	belong	x		Last Freed %d,
+>36	belong	x		Max Bucket %d,
+>40	belong	x		High Mask 0x%x,
+>44	belong	x		Low Mask 0x%x,
+>48	belong	x		Fill Factor %d,
+>52	belong	x		Number of Keys %d)
+#
+0	belong	0x053162	Berkeley DB Btree file
+>4	belong	>0		(Version %d,
+>8	belong	x		Page Size %d,
+>12	belong	x		Free Page %d,
+>16	belong	x		Number of Records %d,
+>20	belong	x		Flags 0x%x)

usr.bin/file/magdir/elf

@@ -5,43 +5,47 @@
 # We have to check the byte order flag to see what byte order all the
 # other stuff in the header is in.
 #
-# MIPS, i486 added by Daniel Quinlan (quinlan@yggdrasil.com)
+# updated by Daniel Quinlan (quinlan@yggdrasil.com)
 0	string		\177ELF		ELF
 >4	byte		0		invalid class
 >4	byte		1		32-bit
 >4	byte		2		64-bit
 >5	byte		0		invalid byte order
 >5	byte		1		LSB
->>16	leshort		0		unknown type
+>>16	leshort		0		no file type,
->>16	leshort		1		relocatable
+>>16	leshort		1		relocatable,
->>16	leshort		2		executable
+>>16	leshort		2		executable,
->>16	leshort		3		dynamic lib
+>>16	leshort		3		shared object,
->>16	leshort		4		core file
+>>16	leshort		4		core file,
->>18	leshort		0		unknown machine
+>>16	leshort		&0xff00		processor-specific,
->>18	leshort		1		WE32100 and up
+>>18	leshort		0		no machine,
->>18	leshort		2		SPARC - invalid byte order
+>>18	leshort		1		AT&T WE32100,
->>18	leshort		3		i386 (386 and up)
+>>18	leshort		2		SPARC - invalid byte order,
->>18	leshort		4		M68000 - invalid byte order
+>>18	leshort		3		Intel 80386,
->>18	leshort		5		M88000 - invalud byte order
+>>18	leshort		4		Motorola 68000 - invalid byte order,
->>18	leshort		6		i486
+>>18	leshort		5		Motorola 88000 - invalid byte order,
->>18	leshort		7		i860
+>>18	leshort		6		Intel 80486,
->>18	leshort		8		MIPS
+>>18	leshort		7		Intel 80860,
->>20	lelong		1		Version 1
+>>18	leshort		8		MIPS RS3000,
+>>20	lelong		0		invalid version
+>>20	lelong		1		version 1
 >>36	lelong		1		MathCoPro/FPU/MAU Required
 >5	byte		2		MSB
->>16	beshort		0		unknown type
+>>16	beshort		0		no file type,
->>16	beshort		1		relocatable
+>>16	beshort		1		relocatable,
->>16	beshort		2		executable
+>>16	beshort		2		executable,
->>16	beshort		3		dynamic lib
+>>16	beshort		3		shared object,
->>16	beshort		4		core file
+>>16	beshort		4		core file,
->>18	beshort		0		unknown machine
+>>16	beshort		&0xff00		processor-specific,
->>18	beshort		1		WE32100 and up
+>>18	beshort		0		no machine,
->>18	beshort		2		SPARC
+>>18	beshort		1		AT&T WE32100,
->>18	beshort		3		i386 (386 and up) - invalid byte order
+>>18	beshort		2		SPARC,
->>18	beshort		4		M68000
+>>18	beshort		3		Intel 80386 - invalid byte order,
->>18	beshort		5		M88000
+>>18	beshort		4		Motorola 68000,
->>18	beshort		6		i486 - invalid byte order
+>>18	beshort		5		Motorola 88000,
->>18	beshort		7		i860
+>>18	beshort		6		Intel 80486 - invalid byte order,
->>18	beshort		8		MIPS
+>>18	beshort		7		Intel 80860,
->>20	belong		1		Version 1
+>>18	beshort		8		MIPS RS3000,
+>>20	belong		0		invalid version
+>>20	belong		1		version 1
 >>36	belong		1		MathCoPro/FPU/MAU Required

usr.bin/file/magdir/hp

@@ -16,27 +16,25 @@
 # UNIX environment atop the "SUN kernel"; dunno whether it was
 # big-endian or little-endian.
 #
-# I'm guessing that the 200 series was 68K-based; the 300 and 400 series
+# Daniel Quinlan (quinlan@yggdrasil.com): hp200 machines are 68010 based;
-# are.
+# hp300 are 68020+68881 based; hp400 are also 68k.  The following basic
+# HP magic is useful for reference, but using "long" magic is a better
+# practice in order to avoid collisions.
 #
-# Daniel Quinlan (quinlan@yggdrasil.com): hp200 machines are 68010
+# 0	beshort		200		hp200 (68010) BSD binary
-# based; hp300 are 68020+68881 based.  I think the following "basic"
+# 0	beshort		300		hp300 (68020+68881) BSD binary
-# magic should probably be integrated and the various flavors of
+# 0	beshort		0x20c		hp200/300 HP-UX binary
-# binaries be implemented with ">2" in case some flavors have been missed.
+# 0	beshort		0x20b		hp800 HP-UX binary
-# 0	beshort		200	hp200 (68010) BSD binary
-# 0	beshort		300	hp300 (68020+68881) BSD binary
-# 0	beshort		0x20c	hp200/300 HP-UX binary
-# 0	beshort		0x20b	hp800 HP-UX binary
 #
 # The "misc" stuff needs a byte order; the archives look suspiciously
 # like the old 177545 archives (0xff65 = 0177545).
 #
 #### Old Apollo stuff
 0	beshort		0627		Apollo m68k COFF executable
->18	beshort		^040000		not stripped 
+>18	beshort		^040000		not stripped
 >22	beshort		>0		- version %ld
 0	beshort		0624		apollo a88k COFF executable
->18	beshort		^040000		not stripped 
+>18	beshort		^040000		not stripped
 >22	beshort		>0		- version %ld
 0       long            01203604016     TML 0123 byte-order format
 0       long            01702407010     TML 1032 byte-order format

usr.bin/file/magdir/images

@@ -7,7 +7,7 @@
 # merging several one- and two-line files into here.
 #
 # XXX - byte order for GIF and TIFF fields?
-# [GRR:  TIFF allows both byte orders; GIF is probably little-endian]
+# [GRR:  TIFF allows both byte orders; GIF is little-endian]
 #
 
 # [GRR:  what the hell is this doing in here?]
@@ -21,28 +21,34 @@
 0	string		P5		PGM "rawbits" file
 0	string		P6		PPM "rawbits" file
 
-# NIFF (Navy Interchange File Format, a modification of TIFF)
+# NIFF (Navy Interchange File Format, a modification of TIFF) images
-# [GRR:  this *must* go before TIFF]
+# this *must* go before TIFF
 0	string		IIN1		NIFF raster data
 
-# TIFF and friends
+# TIFF images
 0	string		MM		TIFF file, big-endian
->2	short		>0		version %d
+>2	beshort		>0		- version %d
 0	string		II		TIFF file, little-endian
->2	short		>0		version %d
+>2	leshort		>0		- version %d
 
-# possible GIF replacements; none yet released!
+# PNG [Portable Network Graphics, or "PNG's Not GIF"] images
 # (Greg Roelofs, newt@uchicago.edu)
 #
-# GRR 950115:  this was mine ("Zip GIF"):
+# 137 P N G \r \n ^Z \n [4-byte length] H E A D [HEAD data] [HEAD crc] ...
-0	string		GIF94z		ZIF image (GIF+deflate alpha)
 #
-# GRR 950115:  this is Jeremy Wohl's Free Graphics Format (better):
+0	string		\x89PNG		PNG image,
-0	string		FGF95a		FGF image (GIF+deflate beta)
+>4	belong		!0x0d0a1a0a	CORRUPTED,
-#
+>16	belong		x		%ld x
-# GRR 950115:  this is Thomas Boutell's Portable Bitmap Format proposal
+>20	belong		x		%ld,
-# (best; not yet implemented):
+>24	byte		x		%d-bit
-0	string		.PBF		PBF image (deflate compression)
+>25	byte		0		grayscale,
+>25	byte		2		\b/color RGB,
+>25	byte		3		colormap,
+>25	byte		4		gray+alpha,
+>25	byte		6		\b/color RGBA,
+#>26	byte		0		deflate/32K,
+>28	byte		0		non-interlaced
+>28	byte		1		interlaced
 
 # GIF
 0	string		GIF		GIF image
@@ -51,7 +57,7 @@
 >6	leshort		>0		%hd x
 >8	leshort		>0		%hd,
 #>10	byte		&0x80		color mapped,
-# GRR 950118:  the following is not accurate for xv-created GIFs:
+# GRR 950330:  the following is not accurate for most GIFs:
 #>10	byte		&0x40		interlaced,
 >10	byte&0x07	=0x00		2 colors
 >10	byte&0x07	=0x01		4 colors
@@ -186,18 +192,24 @@
 #>28	belong		>0		colormap is %d bytes long
 
 # Daniel Quinlan (quinlan@yggdrasil.com) -- from an SGI machine
-0	string	IT01			FIT image file
+# There may be a byte swapped version of SGI imagelib images, but
->4	belong	x			(%d x
+# I haven't seen any evidence of programs that support it.
->8	belong	x			%d x
+0	beshort		000732		SGI imagelib image
->12	belong	x			%d)
+>6	beshort		x		\b, %d x
+>8	beshort		x		%d
 #
-0	string	IT02			FIT image file
+0	string		IT01		FIT image file
->4	belong	x			(%d x
+>4	belong		x		\b, %d x
->8	belong	x			%d x
+>8	belong		x		%d x
->12	belong	x			%d)
+>12	belong		x		%d
 #
-2048	string	PCD_IPI			Kodak Photo CD image pack file
+0	string		IT02		FIT image file
-0	string	PCD_OPA			Kodak Photo CD overview pack file
+>4	belong		x		\b, %d x
+>8	belong		x		%d x
+>12	belong		x		%d
+#
+2048	string		PCD_IPI		Kodak Photo CD image pack file
+0	string		PCD_OPA		Kodak Photo CD overview pack file
 
 # Jeff Uphoff <juphoff@tarsier.cv.nrao.edu>
 # FITS is the Flexible Image Transport System, the de facto standard for

usr.bin/file/magdir/interleaf

@@ -6,4 +6,3 @@
 0	string		=<!OPS		Interleaf document text
 >5	string		,\ Version\ 	(version
 >>14	string		>\0		%s)
-

usr.bin/file/magdir/ispell

@@ -2,23 +2,53 @@
 #------------------------------------------------------------------------------
 # ispell:  file(1) magic for ispell
 #
-# XXX - byte order?
+# Ispell 3.0 has a magic of 0x9601 and ispell 3.1 has 0x9602.  This magic
+# will match 0x9600 through 0x9603 in *both* little endian and big endian.
+# (No other current magic entries collide.)
 #
-0	short	0xffff9601	ispell hash file
+# Updated by Daniel Quinlan (quinlan@yggdrasil.com)
->2	short	0x00	- 8-bit, no capitalization, 26 flags
+#
->2	short	0x01	- 7-bit, no capitalization, 26 flags 
+0	leshort&0xFFFC	0x9600		little endian ispell
->2	short	0x02	- 8-bit, capitalization, 26 flags
+>0	byte		0		hash file (?),
->2	short	0x03	- 7-bit, capitalization, 26 flags 
+>0	byte		1		3.0 hash file,
->2	short	0x04	- 8-bit, no capitalization, 52 flags
+>0	byte		2		3.1 hash file,
->2	short	0x05	- 7-bit, no capitalization, 52 flags 
+>0	byte		3		hash file (?),
->2	short	0x06	- 8-bit, capitalization, 52 flags
+>2	leshort		0x00		8-bit, no capitalization, 26 flags
->2	short	0x07	- 7-bit, capitalization, 52 flags 
+>2	leshort		0x01		7-bit, no capitalization, 26 flags
->2	short	0x08	- 8-bit, no capitalization, 128 flags
+>2	leshort		0x02		8-bit, capitalization, 26 flags
->2	short	0x09	- 7-bit, no capitalization, 128 flags 
+>2	leshort		0x03		7-bit, capitalization, 26 flags
->2	short	0x0A	- 8-bit, capitalization, 128 flags
+>2	leshort		0x04		8-bit, no capitalization, 52 flags
->2	short	0x0B	- 7-bit, capitalization, 128 flags 
+>2	leshort		0x05		7-bit, no capitalization, 52 flags
->2	short	0x0C	- 8-bit, no capitalization, 256 flags
+>2	leshort		0x06		8-bit, capitalization, 52 flags
->2	short	0x0D	- 7-bit, no capitalization, 256 flags 
+>2	leshort		0x07		7-bit, capitalization, 52 flags
->2	short	0x0E	- 8-bit, capitalization, 256 flags
+>2	leshort		0x08		8-bit, no capitalization, 128 flags
->2	short	0x0F	- 7-bit, capitalization, 256 flags 
+>2	leshort		0x09		7-bit, no capitalization, 128 flags
->4	short	>0	and %d string characters
+>2	leshort		0x0A		8-bit, capitalization, 128 flags
+>2	leshort		0x0B		7-bit, capitalization, 128 flags
+>2	leshort		0x0C		8-bit, no capitalization, 256 flags
+>2	leshort		0x0D		7-bit, no capitalization, 256 flags
+>2	leshort		0x0E		8-bit, capitalization, 256 flags
+>2	leshort		0x0F		7-bit, capitalization, 256 flags
+>4	leshort		>0		and %d string characters
+0	beshort&0xFFFC	0x9600		big endian ispell
+>1	byte		0		hash file (?),
+>1	byte		1		3.0 hash file,
+>1	byte		2		3.1 hash file,
+>1	byte		3		hash file (?),
+>2	beshort		0x00		8-bit, no capitalization, 26 flags
+>2	beshort		0x01		7-bit, no capitalization, 26 flags
+>2	beshort		0x02		8-bit, capitalization, 26 flags
+>2	beshort		0x03		7-bit, capitalization, 26 flags
+>2	beshort		0x04		8-bit, no capitalization, 52 flags
+>2	beshort		0x05		7-bit, no capitalization, 52 flags
+>2	beshort		0x06		8-bit, capitalization, 52 flags
+>2	beshort		0x07		7-bit, capitalization, 52 flags
+>2	beshort		0x08		8-bit, no capitalization, 128 flags
+>2	beshort		0x09		7-bit, no capitalization, 128 flags
+>2	beshort		0x0A		8-bit, capitalization, 128 flags
+>2	beshort		0x0B		7-bit, capitalization, 128 flags
+>2	beshort		0x0C		8-bit, no capitalization, 256 flags
+>2	beshort		0x0D		7-bit, no capitalization, 256 flags
+>2	beshort		0x0E		8-bit, capitalization, 256 flags
+>2	beshort		0x0F		7-bit, capitalization, 256 flags
+>4	beshort		>0		and %d string characters

usr.bin/file/magdir/linux

@@ -2,44 +2,50 @@
 #------------------------------------------------------------------------------
 # linux:  file(1) magic for Linux files
 #
-# Values for Linux/i386 binaries, from Rik Faith <faith@cs.unc.edu>,
+# Values for Linux/i386 binaries, from Daniel Quinlan <quinlan@yggdrasil.com>
-# Peter Tobias <tobias@server.et-inf.fho-emden.de>, and Daniel Quinlan
+# The following basic Linux magic is useful for reference, but using
-# <quinlan@yggdrasil.com>
+# "long" magic is a better practice in order to avoid collisions.
+#
+# 2	leshort		100		Linux/i386
+# >0	leshort		0407		impure executable (OMAGIC)
+# >0	leshort		0410		pure executable (NMAGIC)
+# >0	leshort		0413		demand-paged executable (ZMAGIC)
+# >0	leshort		0314		demand-paged executable (QMAGIC)
+#
+0	lelong		0x00640107	Linux/i386 impure executable (OMAGIC)
+>16	lelong		0		- stripped
+0	lelong		0x00640108	Linux/i386 pure executable (NMAGIC)
+>16	lelong		0		- stripped
+0	lelong		0x0064010b	Linux/i386 demand-paged executable (ZMAGIC)
+>16	lelong		0		- stripped
+0	lelong		0x006400cc	Linux/i386 demand-paged executable (QMAGIC)
+>16	lelong		0		- stripped
 #
-2	leshort		100		Linux/i386
->0	leshort		0407		impure executable (OMAGIC)
->0	leshort		0410		pure executable (NMAGIC)
->0	leshort		0413		demand-paged executable (ZMAGIC)
->0	leshort		0314		demand-paged executable (QMAGIC)
->16	lelong		>0		not stripped
->16	lelong		0		stripped
->0	string		Jump		jump
-# object files
-# first entry is absolutely correct, but may conflict with PDP-11 executable
-#0	leshort		0407		Linux/i386 object file
 0	string		\007\001\000	Linux/i386 object file
->20	long		>0x1020		- DLL library
+>20	lelong		>0x1020		- DLL library
 # message catalogs, from Mitchum DSouza <m.dsouza@mrc-apu.cam.ac.uk>
 0	string		*nazgul*	compiled message catalog
->8	long		>0		- version %ld
+>8	lelong		>0		- version %ld
 # core dump file, from Bill Reynolds <bill@goshawk.lanl.gov>
 216	lelong		0421		Linux/i386 core file
 >220	string		>\0		of '%s'
+>200	lelong		>0		(signal %d)
 #
 # LILO boot/chain loaders, from Daniel Quinlan <quinlan@yggdrasil.com>
 # this can be overridden by the DOS executable (COM) entry
-# XXX - moved to "dos"
+2	string		LILO		Linux/i386 LILO boot/chain loader
+#
 # Debian Packages, from Peter Tobias <tobias@server.et-inf.fho-emden.de>
 0	string		0.9
 >8	byte		0x0a		Debian Binary Package -
 >>3	byte		>0		created by dpkg 0.9%c
 >>4	byte		>0		pl%c
 # PSF fonts, from H. Peter Anvin <hpa@yggdrasil.com>
-0	leshort		0x0436		Pc Screen Font data
+0	leshort		0x0436		Linux/i386 PC Screen Font data
->2	byte		0		- 256 characters, no directory
+>2	byte		0		- 256 characters, no directory,
->2	byte		1		- 512 characters, no directory
+>2	byte		1		- 512 characters, no directory,
->2	byte		2		- 256 characters, Unicode directory
+>2	byte		2		- 256 characters, Unicode directory,
->2	byte		3		- 512 characters, Unicode directory
+>2	byte		3		- 512 characters, Unicode directory,
->3	byte		>0		- 8x%d
+>3	byte		>0		8x%d
 # Linux swap file, from Daniel Quinlan <quinlan@yggdrasil.com>
 4086	string		SWAP-SPACE	Linux/i386 swap file

usr.bin/file/magdir/rtf

@@ -1,13 +1,12 @@
 
 #------------------------------------------------------------------------------
-# rtf:	file(1) magic for RTF (Rich text format)
+# rtf:	file(1) magic for Rich Text Format (RTF)
 #
-# This information was gleaned from the version 1 documentation by
+# Duncan P. Simpson, D.P.Simpson@dcs.warwick.ac.uk
-# D.P.Simpson@dcs.warwick.ac.uk (Duncan P Simpson)
 #
-0	string		{\\rtf		Rich Text Format data (version
+0	string		{\\rtf		Rich Text Format data,
->5	byte		x		%c,
+>5	byte		x		version %c,
->6	string		\\mac		Apple Macintosh characters)
+>6	string		\\ansi		ANSI
->6	string		\\ansi		ANSI characters)
+>6	string		\\mac		Apple Macintosh
->6	string		\\pc		IBM PC characters)
+>6	string		\\pc		IBM PC code page 437
->6	string		\\pca		IBM PS/2 characters)
+>6	string		\\pca		IBM PC code page 850

usr.bin/file/magdir/sgi

@@ -152,13 +152,6 @@
 0	string	SGIAUDIT	SGI Audit file
 >8	byte	x		- version %d
 >9	byte	x		.%ld
-#
-0	beshort	000732		SGI imagelib image
->6	beshort	x		(%d x
->8	beshort	x		%d)
-0	beshort	0155001		SGI imagelib image byte-swapped
-0	beshort	017436		packed data
-0	beshort	017037		packed data (byte swapped)
 # Are these three SGI-based file types or general ones?
 0	string	WNGZWZSC	Wingz compiled script
 0	string	WNGZWZSS	Wingz spreadsheet

usr.bin/file/magdir/sgml

@@ -2,7 +2,7 @@
 #------------------------------------------------------------------------------
 # sgml:  file(1) magic for Standard(?) Generalized Mark-up Language
 #
-# $Id: sgml,v 1.2 1995/03/25 22:39:42 christos Exp $
+# $Id: sgml,v 1.3 1995/04/28 19:24:37 christos Exp $
 # SGML goop, mostly from rph@sq.
 0	string		\<!DOCTYPE	Exported SGML document
 0	string		\<!doctype	Exported SGML document

usr.bin/file/magdir/softquad

@@ -2,7 +2,7 @@
 #------------------------------------------------------------------------------
 # softquad:  file(1) magic for SoftQuad Publishing Software
 #
-# $Id: softquad,v 1.3 1995/03/25 22:39:43 christos Exp $
+# $Id: softquad,v 1.4 1995/04/28 19:24:38 christos Exp $
 # Author/Editor and RulesBuilder
 #
 # XXX - byte order?

usr.bin/file/magdir/varied.out

@@ -7,3 +7,10 @@
 #	and deleted if they duplicate other entries.
 #
 0	short		0610		Perkin-Elmer executable
+# AMD 29K
+0	beshort		0572		amd 29k coff noprebar executable
+0	beshort		01572		amd 29k coff prebar executable
+0	beshort		0160007		amd 29k coff archive
+# Cray
+6	beshort		0407		unicos (cray) executable
+

usr.bin/file/magdir/vms

@@ -3,7 +3,6 @@
 # vms:  file(1) magic for VMS executables (experimental)
 #
 # VMS .exe formats, both VAX and AXP (Greg Roelofs, newt@uchicago.edu)
-# This file should be renamed to "vms" eventually...
 
 # GRR 950122:  I'm just guessing on these, based on inspection of the headers
 # of three executables each for Alpha and VAX architectures.  The VAX files

usr.bin/file/names.h

@@ -10,7 +10,7 @@
  *
  * See LEGAL.NOTICE
  *
- * $Id: names.h,v 1.3 1995/03/25 22:36:22 christos Exp $
+ * $Id: names.h,v 1.4 1995/04/28 19:23:56 christos Exp $
  */
 
 /* these types are used to index the table 'types': keep em in sync! */
@@ -25,13 +25,13 @@
 #define	L_NEWS	8		/* Usenet Netnews */
 
 static char *types[] = {
-	"c program text",
+	"C program text",
-	"fortran program text",
+	"FORTRAN program text",
 	"make commands text" ,
-	"pl/1 program text",
+	"PL/1 program text",
 	"assembler program text",
 	"English text",
-	"pascal program text",
+	"Pascal program text",
 	"mail text",
 	"news text",
 	"can't happen error on names.h/types",
@@ -43,7 +43,7 @@ static struct names {
 } names[] = {
 	/* These must be sorted by eye for optimal hit rate */
 	/* Add to this list only after substantial meditation */
-	{"/*",		L_C},	/* must preced "The", "the", etc. */
+	{"/*",		L_C},	/* must precede "The", "the", etc. */
 	{"#include",	L_C},
 	{"char",	L_C},
 	{"The",		L_ENG},
@@ -75,6 +75,7 @@ static struct names {
 	{".byte",	L_MACH},
 	{".even",	L_MACH},
 	{".globl",	L_MACH},
+	{".text",	L_MACH},
 	{"clr",		L_MACH},
 	{"(input,",	L_PAS},
 	{"dcl",		L_PLI},

usr.bin/file/patchlevel.h

@@ -1,13 +1,21 @@
 #define	FILE_VERSION_MAJOR	3
-#define	patchlevel		16
+#define	patchlevel		17
 
 /*
  * Patchlevel file for Ian Darwin's MAGIC command.
- * $Id: patchlevel.h,v 1.2 1995/03/25 22:36:27 christos Exp $
+ * $Id: patchlevel.h,v 1.3 1995/04/28 19:23:58 christos Exp $
  *
  * $Log: patchlevel.h,v $
- * Revision 1.2  1995/03/25 22:36:27  christos
+ * Revision 1.3  1995/04/28 19:23:58  christos
- * Updated to file-3.16; fixes PR867
+ * Integrated file-3.17
+ *
+ * Revision 1.17  1995/04/28  17:29:13  christos
+ * - Incorrect nroff detection fix from der Mouse
+ * - Lost and incorrect magic entries.
+ * - Added ELF stripped binary detection [in C; ugh]
+ * - Look for $MAGIC to find the magic file.
+ * - Eat trailing size specifications from numbers i.e. ignore 10L
+ * - More fixes for very short files
  *
  * Revision 1.16  1995/03/25  22:06:45  christos
  * - use strtoul() where it exists.

usr.bin/file/print.c

@@ -40,7 +40,7 @@
 
 #ifndef lint
 static char *moduleid =
-	"@(#)$Id: print.c,v 1.6 1995/03/25 22:36:34 christos Exp $";
+	"@(#)$Id: print.c,v 1.7 1995/04/28 19:24:01 christos Exp $";
 #endif  /* lint */
 
 #define SZOF(a)	(sizeof(a) / sizeof(a[0]))

usr.bin/file/softmagic.c

@@ -34,7 +34,7 @@
 
 #ifndef	lint
 static char *moduleid = 
-	"@(#)$Id: softmagic.c,v 1.6 1995/03/25 22:36:39 christos Exp $";
+	"@(#)$Id: softmagic.c,v 1.7 1995/04/28 19:24:04 christos Exp $";
 #endif	/* lint */
 
 static int match	__P((unsigned char *, int));
@@ -287,11 +287,20 @@ struct magic *m;
 int nbytes;
 {
 	long offset = m->offset;
-	if (offset + sizeof(union VALUETYPE) > nbytes)
+	long diff = nbytes - (offset + sizeof(union VALUETYPE));
-	    return 0;
+	if (diff >= 0)
+		memcpy(p, s + offset, sizeof(union VALUETYPE));
+	else {
+		/* Not enough space; zeropad */
+		long have = sizeof(union VALUETYPE) + diff;
+		if (have > 0)
+			memcpy(p, s + offset, have);
+		else
+			have = 0;
 
+		memset(p + have, 0, sizeof(union VALUETYPE) - have);
+	}
 
-	memcpy(p, s + offset, sizeof(union VALUETYPE));
 
 	if (debug) {
 		mdebug(offset, (char *) p, sizeof(union VALUETYPE));