From b9bb87b67e9197ead02eb4cb89793d3645497adf Mon Sep 17 00:00:00 2001
From: Christos Zoulas <christos@zoulas.com>
Date: Wed, 17 Oct 2007 19:33:31 +0000
Subject: [PATCH] - separate mime encoding from mime type - fix printing -\012
---
src/ascmagic.c | 27 ++++++++-----
src/compress.c | 20 ++++++----
src/file.c | 25 +++++++++---
src/fsmagic.c | 15 ++++---
src/funcs.c | 102 +++++++++++++++++++++++++++---------------------
src/is_tar.c | 38 +++++++++---------
src/magic.c | 20 +++-------
src/magic.h | 24 ++++++------
src/softmagic.c | 7 ++--
9 files changed, 158 insertions(+), 120 deletions(-)
diff --git a/src/ascmagic.c b/src/ascmagic.c
index 3d1ee2a6..b507f775 100644
--- a/src/ascmagic.c
+++ b/src/ascmagic.c
@@ -49,7 +49,7 @@
#include "names.h"
#ifndef lint
-FILE_RCSID("@(#)$File: ascmagic.c,v 1.50 2007/03/15 14:51:00 christos Exp $")
+FILE_RCSID("@(#)$File: ascmagic.c,v 1.51 2007/08/19 03:45:07 christos Exp $")
#endif /* lint */
typedef unsigned long unichar;
@@ -76,6 +76,7 @@ file_ascmagic(struct magic_set *ms, const unsigned char *buf, size_t nbytes)
size_t ulen;
struct names *p;
int rv = -1;
+ int mime = ms->flags & MAGIC_MIME;
const char *code = NULL;
const char *code_mime = NULL;
@@ -271,21 +272,27 @@ subtype_identified:
if (seen_cr && nbytes < HOWMANY)
n_cr++;
- if ((ms->flags & MAGIC_MIME)) {
- if (subtype_mime) {
- if (file_printf(ms, subtype_mime) == -1)
- goto done;
- } else {
- if (file_printf(ms, "text/plain") == -1)
- goto done;
+ if (mime) {
+ if (mime & MAGIC_MIME_TYPE) {
+ if (subtype_mime) {
+ if (file_printf(ms, subtype_mime) == -1)
+ goto done;
+ } else {
+ if (file_printf(ms, "text/plain") == -1)
+ goto done;
+ }
}
- if (code_mime) {
- if (file_printf(ms, " charset=") == -1)
+ if ((mime == 0 || mime == MAGIC_MIME) && code_mime) {
+ if ((mime & MAGIC_MIME_TYPE) &&
+ file_printf(ms, " charset=") == -1)
goto done;
if (file_printf(ms, code_mime) == -1)
goto done;
}
+
+ if (mime == MAGIC_MIME_ENCODING)
+ file_printf(ms, "binary");
} else {
if (file_printf(ms, code) == -1)
goto done;
diff --git a/src/compress.c b/src/compress.c
index 35451f98..5fc741f6 100644
--- a/src/compress.c
+++ b/src/compress.c
@@ -55,7 +55,7 @@
#ifndef lint
-FILE_RCSID("@(#)$File: compress.c,v 1.51 2007/03/05 02:41:29 christos Exp $")
+FILE_RCSID("@(#)$File: compress.c,v 1.52 2007/08/19 03:45:08 christos Exp $")
#endif
private struct {
@@ -98,6 +98,7 @@ file_zmagic(struct magic_set *ms, int fd, const char *name,
unsigned char *newbuf = NULL;
size_t i, nsz;
int rv = 0;
+ int mime = ms->flags & MAGIC_MIME;
if ((ms->flags & MAGIC_COMPRESS) == 0)
return 0;
@@ -112,13 +113,18 @@ file_zmagic(struct magic_set *ms, int fd, const char *name,
rv = -1;
if (file_buffer(ms, -1, name, newbuf, nsz) == -1)
goto error;
- if (file_printf(ms, (ms->flags & MAGIC_MIME) ?
- " compressed-encoding=" : " (") == -1)
- goto error;
- if (file_buffer(ms, -1, NULL, buf, nbytes) == -1)
+
+ if (mime == MAGIC_MIME || mime == 0) {
+ if (file_printf(ms, mime ?
+ " compressed-encoding=" : " (") == -1)
+ goto error;
+ }
+
+ if ((mime == 0 || mime & MAGIC_MIME_ENCODING) &&
+ file_buffer(ms, -1, NULL, buf, nbytes) == -1)
goto error;
- if (!(ms->flags & MAGIC_MIME) &&
- file_printf(ms, ")") == -1)
+
+ if (!mime && file_printf(ms, ")") == -1)
goto error;
rv = 1;
break;
diff --git a/src/file.c b/src/file.c
index df1d9cb5..e6dc6eb4 100644
--- a/src/file.c
+++ b/src/file.c
@@ -71,7 +71,7 @@
#include "patchlevel.h"
#ifndef lint
-FILE_RCSID("@(#)$File: file.c,v 1.109 2007/03/15 14:50:34 christos Exp $")
+FILE_RCSID("@(#)$File: file.c,v 1.111 2007/05/08 14:44:18 christos Exp $")
#endif /* lint */
@@ -133,8 +133,12 @@ main(int argc, char *argv[])
int longindex;
static const struct option long_options[] =
{
- {"version", 0, 0, 'v'},
+ /* Put long-only options first */
{"help", 0, 0, 0},
+ {"mime-type", 0, 0, 0},
+ {"mime-encoding", 0, 0, 0},
+
+ {"version", 0, 0, 'v'},
{"brief", 0, 0, 'b'},
{"checking-printout", 0, 0, 'c'},
{"debug", 0, 0, 'd'},
@@ -220,9 +224,17 @@ main(int argc, char *argv[])
switch (c) {
#ifdef HAVE_GETOPT_LONG
case 0 :
- if (longindex == 1)
+ switch (longindex) {
+ case 0:
help();
- break;
+ break;
+ case 1:
+ flags |= MAGIC_MIME_TYPE;
+ break;
+ case 2:
+ flags |= MAGIC_MIME_ENCODING;
+ break;
+ }
#endif
case '0':
nulsep = 1;
@@ -567,7 +579,10 @@ help(void)
" ascii, apptype, elf, compress, soft, tar\n"
" -f, --files-from FILE read the filenames to be examined from FILE\n"
" -F, --separator string use string as separator instead of `:'\n"
-" -i, --mime output mime type strings\n"
+" -i, --mime output MIME type strings (--mime-type and\n"
+" --mime-encoding)\n"
+" --mime-type output the MIME type\n"
+" --mime-encoding output the MIME encoding\n"
" -k, --keep-going don't stop at the first match\n"
" -L, --dereference causes symlinks to be followed\n"
" -n, --no-buffer do not buffer output\n"
diff --git a/src/fsmagic.c b/src/fsmagic.c
index cd8f13e2..ec7d0c9c 100644
--- a/src/fsmagic.c
+++ b/src/fsmagic.c
@@ -57,13 +57,14 @@
#undef HAVE_MAJOR
#ifndef lint
-FILE_RCSID("@(#)$File: fsmagic.c,v 1.46 2005/06/25 15:52:14 christos Exp $")
+FILE_RCSID("@(#)$File: fsmagic.c,v 1.47 2007/01/12 17:38:28 christos Exp $")
#endif /* lint */
protected int
file_fsmagic(struct magic_set *ms, const char *fn, struct stat *sb)
{
int ret = 0;
+ int mime = ms->flags & MAGIC_MIME;
#ifdef S_IFLNK
char buf[BUFSIZ+4];
int nch;
@@ -95,11 +96,12 @@ file_fsmagic(struct magic_set *ms, const char *fn, struct stat *sb)
return 1;
}
- if ((ms->flags & MAGIC_MIME) != 0) {
+ if (mime) {
if ((sb->st_mode & S_IFMT) != S_IFREG) {
- if (file_printf(ms, "application/x-not-regular-file")
+ if ((mime & MAGIC_MIME_TYPE) &&
+ file_printf(ms, "application/x-not-regular-file")
== -1)
- return -1;
+ return -1;
return 1;
}
}
@@ -303,8 +305,9 @@ file_fsmagic(struct magic_set *ms, const char *fn, struct stat *sb)
* when we read the file.)
*/
if ((ms->flags & MAGIC_DEVICES) == 0 && sb->st_size == 0) {
- if (file_printf(ms, (ms->flags & MAGIC_MIME) ?
- "application/x-empty" : "empty") == -1)
+ if ((!mime || (mime & MAGIC_MIME_TYPE)) &&
+ file_printf(ms, mime ? "application/x-empty" :
+ "empty") == -1)
return -1;
return 1;
}
diff --git a/src/funcs.c b/src/funcs.c
index 3ffcaa8a..5885ff39 100644
--- a/src/funcs.c
+++ b/src/funcs.c
@@ -48,7 +48,7 @@
#endif
#ifndef lint
-FILE_RCSID("@(#)$File: funcs.c,v 1.32 2007/05/24 17:22:27 christos Exp $")
+FILE_RCSID("@(#)$File: funcs.c,v 1.33 2007/06/15 00:01:15 christos Exp $")
#endif /* lint */
#ifndef HAVE_VSNPRINTF
@@ -164,59 +164,73 @@ protected int
file_buffer(struct magic_set *ms, int fd, const char *inname, const void *buf,
size_t nb)
{
- int m;
+ int m;
+ int mime = ms->flags & MAGIC_MIME;
+
+ if (nb == 0) {
+ if ((!mime || (mime & MAGIC_MIME_TYPE)) &&
+ file_printf(ms, mime ? "application/x-empty" :
+ "empty") == -1)
+ return -1;
+ return 1;
+ } else if (nb == 1) {
+ if ((!mime || (mime & MAGIC_MIME_TYPE)) &&
+ file_printf(ms, mime ? "application/octet-stream" :
+ "very short file (no magic)") == -1)
+ return -1;
+ return 1;
+ }
#ifdef __EMX__
- if ((ms->flags & MAGIC_NO_CHECK_APPTYPE) == 0 && inname) {
- switch (file_os2_apptype(ms, inname, buf, nb)) {
- case -1:
- return -1;
- case 0:
- break;
- default:
- return 1;
+ if ((ms->flags & MAGIC_NO_CHECK_APPTYPE) == 0 && inname) {
+ switch (file_os2_apptype(ms, inname, buf, nb)) {
+ case -1:
+ return -1;
+ case 0:
+ break;
+ default:
+ return 1;
+ }
}
- }
#endif
- /* try compression stuff */
- if ((ms->flags & MAGIC_NO_CHECK_COMPRESS) != 0 ||
- (m = file_zmagic(ms, fd, inname, buf, nb)) == 0) {
- /* Check if we have a tar file */
- if ((ms->flags & MAGIC_NO_CHECK_TAR) != 0 ||
- (m = file_is_tar(ms, buf, nb)) == 0) {
- /* try tests in /etc/magic (or surrogate magic file) */
- if ((ms->flags & MAGIC_NO_CHECK_SOFT) != 0 ||
- (m = file_softmagic(ms, buf, nb)) == 0) {
- /* try known keywords, check whether it is ASCII */
- if ((ms->flags & MAGIC_NO_CHECK_ASCII) != 0 ||
- (m = file_ascmagic(ms, buf, nb)) == 0) {
- /* abandon hope, all ye who remain here */
- if (file_printf(ms, ms->flags & MAGIC_MIME ?
- (nb ? "application/octet-stream" :
- "application/empty") :
- (nb ? "data" :
- "empty")) == -1)
- return -1;
- m = 1;
+ /* try compression stuff */
+ if ((ms->flags & MAGIC_NO_CHECK_COMPRESS) != 0 ||
+ (m = file_zmagic(ms, fd, inname, buf, nb)) == 0) {
+ /* Check if we have a tar file */
+ if ((ms->flags & MAGIC_NO_CHECK_TAR) != 0 ||
+ (m = file_is_tar(ms, buf, nb)) == 0) {
+ /* try tests in /etc/magic (or surrogate magic file) */
+ if ((ms->flags & MAGIC_NO_CHECK_SOFT) != 0 ||
+ (m = file_softmagic(ms, buf, nb)) == 0) {
+ /* try known keywords, check whether it is ASCII */
+ if ((ms->flags & MAGIC_NO_CHECK_ASCII) != 0 ||
+ (m = file_ascmagic(ms, buf, nb)) == 0) {
+ /* abandon hope, all ye who remain here */
+ if ((!mime || (mime & MAGIC_MIME_TYPE)) &&
+ file_printf(ms, mime ? "application/octet-stream" :
+ "data") == -1)
+ return -1;
+ m = 1;
+ }
}
}
}
- }
#ifdef BUILTIN_ELF
- if ((ms->flags & MAGIC_NO_CHECK_ELF) == 0 && m == 1 && nb > 5 && fd != -1) {
- /*
- * We matched something in the file, so this *might*
- * be an ELF file, and the file is at least 5 bytes
- * long, so if it's an ELF file it has at least one
- * byte past the ELF magic number - try extracting
- * information from the ELF headers that cannot easily
- * be extracted with rules in the magic file.
- */
- (void)file_tryelf(ms, fd, buf, nb);
- }
+ if ((ms->flags & MAGIC_NO_CHECK_ELF) == 0 && m == 1 &&
+ nb > 5 && fd != -1) {
+ /*
+ * We matched something in the file, so this *might*
+ * be an ELF file, and the file is at least 5 bytes
+ * long, so if it's an ELF file it has at least one
+ * byte past the ELF magic number - try extracting
+ * information from the ELF headers that cannot easily
+ * be extracted with rules in the magic file.
+ */
+ (void)file_tryelf(ms, fd, buf, nb);
+ }
#endif
- return m;
+ return m;
}
#endif
diff --git a/src/is_tar.c b/src/is_tar.c
index 10c8b05a..cbaba63c 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 1.27 2007/01/12 17:38:28 christos Exp $")
+FILE_RCSID("@(#)$File: is_tar.c,v 1.28 2007/08/19 03:45:08 christos Exp $")
#endif
#define isodigit(c) ( ((c) >= '0') && ((c) <= '7') )
@@ -53,6 +53,12 @@ FILE_RCSID("@(#)$File: is_tar.c,v 1.27 2007/01/12 17:38:28 christos Exp $")
private int is_tar(const unsigned char *, size_t);
private int from_oct(int, const char *); /* Decode octal number */
+static const char *tartype[] = {
+ "tar archive",
+ "POSIX tar archive",
+ "POSIX tar archive (GNU)",
+};
+
protected int
file_is_tar(struct magic_set *ms, const unsigned char *buf, size_t nbytes)
{
@@ -60,25 +66,19 @@ file_is_tar(struct magic_set *ms, const unsigned char *buf, size_t nbytes)
* Do the tar test first, because if the first file in the tar
* archive starts with a dot, we can confuse it with an nroff file.
*/
- switch (is_tar(buf, nbytes)) {
- case 1:
- if (file_printf(ms, (ms->flags & MAGIC_MIME) ?
- "application/x-tar" : "tar archive") == -1)
- return -1;
- return 1;
- case 2:
- if (file_printf(ms, (ms->flags & MAGIC_MIME) ?
- "application/x-tar" : "POSIX tar archive") == -1)
- return -1;
- return 1;
- case 3:
- if (file_printf(ms, (ms->flags & MAGIC_MIME) ?
- "application/x-tar" : "POSIX tar archive (GNU)") == -1)
- return -1;
- return 1;
- default:
+ int tar = is_tar(buf, nbytes);
+ int mime = ms->flags & MAGIC_MIME;
+
+ if (tar < 1 || tar > 3)
return 0;
- }
+
+ if (mime == MAGIC_MIME_ENCODING)
+ return 0;
+
+ if (file_printf(ms, mime ? "application/x-tar" :
+ tartype[tar - 1]) == -1)
+ return -1;
+ return 1;
}
/*
diff --git a/src/magic.c b/src/magic.c
index c6fd9ad8..11cfb20e 100644
--- a/src/magic.c
+++ b/src/magic.c
@@ -63,7 +63,7 @@
#include "patchlevel.h"
#ifndef lint
-FILE_RCSID("@(#)$File: magic.c,v 1.42 2007/08/19 03:45:08 christos Exp $")
+FILE_RCSID("@(#)$File: magic.c,v 1.43 2007/09/26 20:45:26 christos Exp $")
#endif /* lint */
#ifdef __EMX__
@@ -260,6 +260,7 @@ file_or_fd(struct magic_set *ms, const char *inname, int fd)
struct stat sb;
ssize_t nbytes = 0; /* number of bytes read from a datafile */
int ispipe = 0;
+ int mime = ms->flags & MAGIC_MIME;
/*
* one extra for terminating '\0', and
@@ -343,20 +344,9 @@ file_or_fd(struct magic_set *ms, const char *inname, int fd)
}
}
- if (nbytes == 0) {
- if (file_printf(ms, (ms->flags & MAGIC_MIME) ?
- "application/x-empty" : "empty") == -1)
- goto done;
- } else if (nbytes == 1) {
- if (file_printf(ms, (ms->flags & MAGIC_MIME) ?
- "application/octet-stream" : "very short file (no magic)")
- == -1)
- goto done;
- } else {
- (void)memset(buf + nbytes, 0, SLOP); /* NUL terminate */
- if (file_buffer(ms, fd, inname, buf, (size_t)nbytes) == -1)
- goto done;
- }
+ (void)memset(buf + nbytes, 0, SLOP); /* NUL terminate */
+ if (file_buffer(ms, fd, inname, buf, (size_t)nbytes) == -1)
+ goto done;
rv = 0;
done:
free(buf);
diff --git a/src/magic.h b/src/magic.h
index 483cac3f..96aa2f68 100644
--- a/src/magic.h
+++ b/src/magic.h
@@ -34,21 +34,23 @@
#define MAGIC_SYMLINK 0x000002 /* Follow symlinks */
#define MAGIC_COMPRESS 0x000004 /* Check inside compressed files */
#define MAGIC_DEVICES 0x000008 /* Look at the contents of devices */
-#define MAGIC_MIME 0x000010 /* Return a mime string */
+#define MAGIC_MIME_TYPE 0x000010 /* Return only the MIME type */
#define MAGIC_CONTINUE 0x000020 /* Return all matches */
#define MAGIC_CHECK 0x000040 /* Print warnings to stderr */
#define MAGIC_PRESERVE_ATIME 0x000080 /* Restore access time on exit */
-#define MAGIC_RAW 0x000100 /* Don't translate unprintable chars */
+#define MAGIC_RAW 0x000100 /* Don't translate unprint chars */
#define MAGIC_ERROR 0x000200 /* Handle ENOENT etc as real errors */
-#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 */
-#define MAGIC_NO_CHECK_APPTYPE 0x008000 /* Don't check application type */
-#define MAGIC_NO_CHECK_ELF 0x010000 /* Don't check for elf details */
-#define MAGIC_NO_CHECK_ASCII 0x020000 /* Don't check for ascii files */
-#define MAGIC_NO_CHECK_TROFF 0x040000 /* Don't check ascii/troff */
-#define MAGIC_NO_CHECK_FORTRAN 0x080000 /* Don't check ascii/fortran */
-#define MAGIC_NO_CHECK_TOKENS 0x100000 /* Don't check ascii/tokens */
+#define MAGIC_MIME_ENCODING 0x000400 /* Return only the MIME encoding */
+#define MAGIC_MIME (MAGIC_MIME_TYPE|MAGIC_MIME_ENCODING)
+#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 */
+#define MAGIC_NO_CHECK_APPTYPE 0x008000 /* Don't check application type */
+#define MAGIC_NO_CHECK_ELF 0x010000 /* Don't check for elf details */
+#define MAGIC_NO_CHECK_ASCII 0x020000 /* Don't check for ascii files */
+#define MAGIC_NO_CHECK_TROFF 0x040000 /* Don't check ascii/troff */
+#define MAGIC_NO_CHECK_FORTRAN 0x080000 /* Don't check ascii/fortran */
+#define MAGIC_NO_CHECK_TOKENS 0x100000 /* Don't check ascii/tokens */
#ifdef __cplusplus
extern "C" {
diff --git a/src/softmagic.c b/src/softmagic.c
index a0e709f3..a0548209 100644
--- a/src/softmagic.c
+++ b/src/softmagic.c
@@ -38,7 +38,7 @@
#ifndef lint
-FILE_RCSID("@(#)$File: softmagic.c,v 1.99 2007/05/08 14:44:18 christos Exp $")
+FILE_RCSID("@(#)$File: softmagic.c,v 1.100 2007/09/26 20:19:05 christos Exp $")
#endif /* lint */
private int match(struct magic_set *, struct magic *, uint32_t,
@@ -254,9 +254,10 @@ match(struct magic_set *ms, struct magic *magic, uint32_t nmagic,
break;
}
}
- firstline = 0;
- if (printed_something)
+ if (printed_something) {
+ firstline = 0;
returnval = 1;
+ }
if ((ms->flags & MAGIC_CONTINUE) == 0 && printed_something) {
return 1; /* don't keep searching */
}
--
2.40.0