tar: optional autodetection of gz/bz2 compressed tarballs.

+130 bytes. Closes bug 992.

3 files changed, 92 insertions, 31 deletions

diff --git a/archival/Config.in b/archival/Config.in
index 160c54d64..6f803f4fe 100644
--- a/archival/Config.in
+++ b/archival/Config.in
@@ -166,6 +166,14 @@ config FEATURE_TAR_CREATE
 	  If you enable this option you'll be able to create
 	  tar archives using the `-c' option.
 
+config FEATURE_TAR_GZIP
+	bool "Enable -z option"
+	default y
+	depends on TAR
+	help
+	  If you enable this option tar will be able to call gzip,
+	  when creating or extracting tar gziped archives.
+
 config FEATURE_TAR_BZIP2
 	bool "Enable -j option to handle .tar.bz2 files"
 	default n
@@ -182,29 +190,29 @@ config FEATURE_TAR_LZMA
 	  If you enable this option you'll be able to extract
 	  archives compressed with lzma.
 
-config FEATURE_TAR_FROM
-	bool "Enable -X (exclude from) and -T (include from) options)"
+config FEATURE_TAR_COMPRESS
+	bool "Enable -Z option"
 	default n
 	depends on TAR
 	help
-	  If you enable this option you'll be able to specify
-	  a list of files to include or exclude from an archive.
+	  If you enable this option tar will be able to call uncompress,
+	  when extracting .tar.Z archives.
 
-config FEATURE_TAR_GZIP
-	bool "Enable -z option"
-	default y
-	depends on TAR
+config FEATURE_TAR_AUTODETECT
+	bool "Let tar autodetect gz/bz2 compresses tarballs"
+	default n
+	depends on FEATURE_TAR_GZIP || FEATURE_TAR_BZIP2
 	help
-	  If you enable this option tar will be able to call gzip,
-	  when creating or extracting tar gziped archives.
+	  With this option tar can automatically detect gzip/bzip2 compressed
+	  tarballs. Currently it works only on seekable streams.
 
-config FEATURE_TAR_COMPRESS
-	bool "Enable -Z option"
+config FEATURE_TAR_FROM
+	bool "Enable -X (exclude from) and -T (include from) options)"
 	default n
 	depends on TAR
 	help
-	  If you enable this option tar will be able to call uncompress,
-	  when extracting .tar.Z archives.
+	  If you enable this option you'll be able to specify
+	  a list of files to include or exclude from an archive.
 
 config FEATURE_TAR_OLDGNU_COMPATIBILITY
 	bool "Enable support for old tar header format"
diff --git a/archival/libunarchive/get_header_tar.c b/archival/libunarchive/get_header_tar.c
index 893cd5b79..54c8f7665 100644
--- a/archival/libunarchive/get_header_tar.c
+++ b/archival/libunarchive/get_header_tar.c
@@ -46,6 +46,9 @@ void BUG_tar_header_size(void);
 char get_header_tar(archive_handle_t *archive_handle)
 {
 	static smallint end;
+#if ENABLE_FEATURE_TAR_AUTODETECT
+	static smallint not_first;
+#endif
 
 	file_header_t *file_header = archive_handle->file_header;
 	struct {
@@ -115,7 +118,7 @@ char get_header_tar(archive_handle_t *archive_handle)
 			 * Read until the end to empty the pipe from gz or bz2
 			 */
 			while (full_read(archive_handle->src_fd, &tar, 512) == 512)
-				/* repeat */;
+				continue;
 			return EXIT_FAILURE;
 		}
 		end = 1;
@@ -123,16 +126,49 @@ char get_header_tar(archive_handle_t *archive_handle)
 	}
 	end = 0;
 
-	/* Check header has valid magic, "ustar" is for the proper tar
-	 * 0's are for the old tar format
-	 */
-	if (strncmp(tar.magic, "ustar", 5) != 0) {
-#if ENABLE_FEATURE_TAR_OLDGNU_COMPATIBILITY
-		if (memcmp(tar.magic, "\0\0\0\0", 5) != 0)
+	/* Check header has valid magic, "ustar" is for the proper tar,
+	 * five NULs are for the old tar format  */
+	if (strncmp(tar.magic, "ustar", 5) != 0
+	 && (!ENABLE_FEATURE_TAR_OLDGNU_COMPATIBILITY
+	     || memcmp(tar.magic, "\0\0\0\0", 5) != 0)
+	) {
+#if ENABLE_FEATURE_TAR_AUTODETECT
+		char (*get_header_ptr)(archive_handle_t *);
+
+		/* tar gz/bz autodetect: check for gz/bz2 magic.
+		 * If it is the very first block, and we see the magic,
+		 * we can switch to get_header_tar_gz/bz2/lzma().
+		 * Needs seekable fd. I wish recv(MSG_PEEK) would work
+		 * on any fd... */
+		if (not_first)
+			goto err;
+#if ENABLE_FEATURE_TAR_GZIP
+		if (tar.name[0] == 0x1f && tar.name[1] == 0x8b) { /* gzip */
+			get_header_ptr = get_header_tar_gz;
+		} else
 #endif
-			bb_error_msg_and_die("invalid tar magic");
+#if ENABLE_FEATURE_TAR_BZIP2
+		if (tar.name[0] == 'B' && tar.name[1] == 'Z'
+		 && tar.name[2] == 'h' && isdigit(tar.name[3])
+		) { /* bzip2 */
+			get_header_ptr = get_header_tar_bz2;
+		} else
+#endif
+			goto err;
+		if (lseek(archive_handle->src_fd, -512, SEEK_CUR) != 0)
+			goto err;
+		while (get_header_ptr(archive_handle) == EXIT_SUCCESS)
+			continue;
+		return EXIT_FAILURE;
+ err:
+#endif /* FEATURE_TAR_AUTODETECT */
+		bb_error_msg_and_die("invalid tar magic");
 	}
 
+#if ENABLE_FEATURE_TAR_AUTODETECT
+	not_first = 1;
+#endif
+
 	/* Do checksum on headers.
 	 * POSIX says that checksum is done on unsigned bytes, but
 	 * Sun and HP-UX gets it wrong... more details in
diff --git a/archival/tar.c b/archival/tar.c
index 4ec454b88..e790f28ea 100644
--- a/archival/tar.c
+++ b/archival/tar.c
@@ -33,8 +33,17 @@
 #define FNM_LEADING_DIR 0
 #endif
 
+
 #define block_buf bb_common_bufsiz1
 
+
+#if !ENABLE_FEATURE_TAR_GZIP && !ENABLE_FEATURE_TAR_BZIP2
+/* Do not pass gzip flag to writeTarFile() */
+#define writeTarFile(tar_fd, verboseFlag, dereferenceFlag, include, exclude, gzip) \
+	writeTarFile(tar_fd, verboseFlag, dereferenceFlag, include, exclude)
+#endif
+
+
 #if ENABLE_FEATURE_TAR_CREATE
 
 /* Tar file constants  */
@@ -514,18 +523,23 @@ static int writeTarFile(const int tar_fd, const int verboseFlag,
 	if (fstat(tbInfo.tarFd, &tbInfo.statBuf) < 0)
 		bb_perror_msg_and_die("cannot stat tar file");
 
-	if ((ENABLE_FEATURE_TAR_GZIP || ENABLE_FEATURE_TAR_BZIP2) && gzip) {
-// On Linux, vfork never unpauses parent early, although standard
-// allows for that. Do we want to waste bytes checking for it?
+#if ENABLE_FEATURE_TAR_GZIP || ENABLE_FEATURE_TAR_BZIP2
+	if (gzip) {
+#if ENABLE_FEATURE_TAR_GZIP && ENABLE_FEATURE_TAR_BZIP2
+		const char *zip_exec = (gzip == 1) ? "gzip" : "bzip2";
+#elif ENABLE_FEATURE_TAR_GZIP
+		const char *zip_exec = "gzip";
+#else /* only ENABLE_FEATURE_TAR_BZIP2 */
+		const char *zip_exec = "bzip2";
+#endif
+	// On Linux, vfork never unpauses parent early, although standard
+	// allows for that. Do we want to waste bytes checking for it?
 #define WAIT_FOR_CHILD 0
-
 		volatile int vfork_exec_errno = 0;
 #if WAIT_FOR_CHILD
 		struct fd_pair gzipStatusPipe;
 #endif
 		struct fd_pair gzipDataPipe;
-		const char *zip_exec = (gzip == 1) ? "gzip" : "bzip2";
-
 		xpiped_pair(gzipDataPipe);
 #if WAIT_FOR_CHILD
 		xpiped_pair(gzipStatusPipe);
@@ -584,6 +598,7 @@ static int writeTarFile(const int tar_fd, const int verboseFlag,
 			bb_perror_msg_and_die("cannot exec %s", zip_exec);
 		}
 	}
+#endif
 
 	tbInfo.excludeList = exclude;
 
@@ -934,11 +949,13 @@ int tar_main(int argc, char **argv)
 
 	/* create an archive */
 	if (opt & OPT_CREATE) {
+#if ENABLE_FEATURE_TAR_GZIP || ENABLE_FEATURE_TAR_BZIP2
 		int zipMode = 0;
-		if (ENABLE_FEATURE_TAR_GZIP && get_header_ptr == get_header_tar_gz)
+		if (ENABLE_FEATURE_TAR_GZIP && (opt & OPT_GZIP))
 			zipMode = 1;
-		if (ENABLE_FEATURE_TAR_BZIP2 && get_header_ptr == get_header_tar_bz2)
+		if (ENABLE_FEATURE_TAR_BZIP2 && (opt & OPT_BZIP2))
 			zipMode = 2;
+#endif
 		/* NB: writeTarFile() closes tar_handle->src_fd */
 		return writeTarFile(tar_handle->src_fd, verboseFlag, opt & OPT_DEREFERENCE,
 				tar_handle->accept,
@@ -946,7 +963,7 @@ int tar_main(int argc, char **argv)
 	}
 
 	while (get_header_ptr(tar_handle) == EXIT_SUCCESS)
-		/* nothing */;
+		continue;
 
 	/* Check that every file that should have been extracted was */
 	while (tar_handle->accept) {