header_verbose_list: show off_t size, not cast to (unsigned)

get_header_tar: support GNU tar's "base256" encoding function old new delta get_header_tar 1536 1600 +64 header_verbose_list 242 257 +15
author: Denis Vlasenko <vda.linux@googlemail.com> 2008-12-12 02:14:35 +0000
committer: Denis Vlasenko <vda.linux@googlemail.com> 2008-12-12 02:14:35 +0000
commit: ab6b446e9567db741b463f74afb7d98cc583c4b9 (patch)
tree: 5a58620c525bf524451adec2ab2121fea04f04f8 /archival/libunarchive
parent: 12abcb36c871f4c97d7aed3d9bd55e5477888bb7 (diff)
download: busybox-w32-ab6b446e9567db741b463f74afb7d98cc583c4b9.tar.gz
busybox-w32-ab6b446e9567db741b463f74afb7d98cc583c4b9.tar.bz2
busybox-w32-ab6b446e9567db741b463f74afb7d98cc583c4b9.zip
2 files changed, 74 insertions, 3 deletions
diff --git a/archival/libunarchive/get_header_tar.c b/archival/libunarchive/get_header_tar.c
index bf0f92b25..48fc23d8b 100644
--- a/archival/libunarchive/get_header_tar.c
+++ b/archival/libunarchive/get_header_tar.c
@@ -14,6 +14,74 @@
 #include "libbb.h"
 #include "unarchive.h"
+/*
+ * GNU tar uses "base-256 encoding" for very large numbers (>8 billion).
+ * Encoding is binary, with highest bit always set as a marker
+ * and sign in next-highest bit:
+ * 80 00 .. 00 - zero
+ * bf ff .. ff - largest positive number
+ * ff ff .. ff - minus 1
+ * c0 00 .. 00 - smallest negative number
+ *
+ * We expect it only in size field, where negative numbers don't make sense.
+ */
+static off_t getBase256_len12(const char *str)
+{
+        off_t value;
+        int len;
+        /* if (*str & 0x40) error; - caller prevents this */
+        if (sizeof(off_t) >= 12) {
+                /* Probably 128-bit (16 byte) off_t. Can be optimized. */
+                len = 12;
+                value = *str++ & 0x3f;
+                while (--len)
+                        value = (value << 8) + (unsigned char) *str++;
+                return value;
+        }
+#ifdef CHECK_FOR_OVERFLOW
+        /* Can be optimized to eat 32-bit chunks */
+        char c = *str++ & 0x3f;
+        len = 12;
+        while (1) {
+                if (c)
+                        bb_error_msg_and_die("overflow in base-256 encoded file size");
+                if (--len == sizeof(off_t)) 
+                        break;
+                c = *str++;
+        }
+#else
+        str += (12 - sizeof(off_t));
+#endif
+/* Now str points to sizeof(off_t) least significant bytes.
+ *
+ * Example of tar file with 8914993153 (0x213600001) byte file.
+ * Field starts at offset 7c:
+ * 00070  30 30 30 00 30 30 30 30  30 30 30 00 80 00 00 00  |000.0000000.....|
+ * 00080  00 00 00 02 13 60 00 01  31 31 31 32 30 33 33 36  |.....`..11120336|
+ *
+ * str is at offset 80 or 84 now (64-bit or 32-bit off_t).
+ * We (ab)use the fact that value happens to be aligned,
+ * and fetch it in one go:
+ */
+        if (sizeof(off_t) == 8) {
+                value = *(off_t*)str;
+                value = SWAP_BE64(value);
+        } else if (sizeof(off_t) == 4) {
+                value = *(off_t*)str;
+                value = SWAP_BE32(value);
+        } else {
+                value = 0;
+                len = sizeof(off_t);
+                while (--len)
+                        value = (value << 8) + (unsigned char) *str++;
+        }
+        return value;
+}
 /* NB: _DESTROYS_ str[len] character! */
 static unsigned long long getOctal(char *str, int len)
 {
@@ -234,7 +302,10 @@ char FAST_FUNC get_header_tar(archive_handle_t *archive_handle)
        file_header->gname = tar.gname[0] ? xstrndup(tar.gname, sizeof(tar.gname)) : NULL;
 #endif
        file_header->mtime = GET_OCTAL(tar.mtime);
-        file_header->size = GET_OCTAL(tar.size);
+        /* Size field: handle GNU tar's "base256 encoding" */
+        file_header->size = (*tar.size & 0xc0) == 0x80 /* positive base256? */
+                        ? getBase256_len12(tar.size)
+                        : GET_OCTAL(tar.size);
        file_header->gid = GET_OCTAL(tar.gid);
        file_header->uid = GET_OCTAL(tar.uid);
        /* Set bits 0-11 of the files mode */
diff --git a/archival/libunarchive/header_verbose_list.c b/archival/libunarchive/header_verbose_list.c
index f059dd981..dc3100361 100644
--- a/archival/libunarchive/header_verbose_list.c
+++ b/archival/libunarchive/header_verbose_list.c
@@ -24,11 +24,11 @@ void FAST_FUNC header_verbose_list(const file_header_t *file_header)
                snprintf(gid, sizeof(gid), "%u", (unsigned)file_header->gid);
                group = gid;
        }
-        printf("%s %s/%s %9u %4u-%02u-%02u %02u:%02u:%02u %s",
+        printf("%s %s/%s %9"OFF_FMT"u %4u-%02u-%02u %02u:%02u:%02u %s",
                bb_mode_string(file_header->mode),
                user,
                group,
-                (unsigned int) file_header->size,
+                file_header->size,
                1900 + mtime->tm_year,
                1 + mtime->tm_mon,
                mtime->tm_mday,
author	Denis Vlasenko <vda.linux@googlemail.com>	2008-12-12 02:14:35 +0000
committer	Denis Vlasenko <vda.linux@googlemail.com>	2008-12-12 02:14:35 +0000
commit	ab6b446e9567db741b463f74afb7d98cc583c4b9 (patch)
tree	5a58620c525bf524451adec2ab2121fea04f04f8 /archival/libunarchive
parent	12abcb36c871f4c97d7aed3d9bd55e5477888bb7 (diff)
download	busybox-w32-ab6b446e9567db741b463f74afb7d98cc583c4b9.tar.gz busybox-w32-ab6b446e9567db741b463f74afb7d98cc583c4b9.tar.bz2 busybox-w32-ab6b446e9567db741b463f74afb7d98cc583c4b9.zip

diff --git a/archival/libunarchive/get_header_tar.c b/archival/libunarchive/get_header_tar.c index bf0f92b25..48fc23d8b 100644 --- a/archival/libunarchive/get_header_tar.c +++ b/archival/libunarchive/get_header_tar.c
@@ -14,6 +14,74 @@
14	#include "libbb.h"	14	#include "libbb.h"
15	#include "unarchive.h"	15	#include "unarchive.h"
16		16
		17	/*
		18	* GNU tar uses "base-256 encoding" for very large numbers (>8 billion).
		19	* Encoding is binary, with highest bit always set as a marker
		20	* and sign in next-highest bit:
		21	* 80 00 .. 00 - zero
		22	* bf ff .. ff - largest positive number
		23	* ff ff .. ff - minus 1
		24	* c0 00 .. 00 - smallest negative number
		25	*
		26	* We expect it only in size field, where negative numbers don't make sense.
		27	*/
		28	static off_t getBase256_len12(const char *str)
		29	{
		30	off_t value;
		31	int len;
		32
		33	/* if (str & 0x40) error; - caller prevents this /
		34
		35	if (sizeof(off_t) >= 12) {
		36	/* Probably 128-bit (16 byte) off_t. Can be optimized. */
		37	len = 12;
		38	value = *str++ & 0x3f;
		39	while (--len)
		40	value = (value << 8) + (unsigned char) *str++;
		41	return value;
		42	}
		43
		44	#ifdef CHECK_FOR_OVERFLOW
		45	/* Can be optimized to eat 32-bit chunks */
		46	char c = *str++ & 0x3f;
		47	len = 12;
		48	while (1) {
		49	if (c)
		50	bb_error_msg_and_die("overflow in base-256 encoded file size");
		51	if (--len == sizeof(off_t))
		52	break;
		53	c = *str++;
		54	}
		55	#else
		56	str += (12 - sizeof(off_t));
		57	#endif
		58
		59	/* Now str points to sizeof(off_t) least significant bytes.
		60	*
		61	* Example of tar file with 8914993153 (0x213600001) byte file.
		62	* Field starts at offset 7c:
		63	* 00070 30 30 30 00 30 30 30 30 30 30 30 00 80 00 00 00 \|000.0000000.....\|
		64	* 00080 00 00 00 02 13 60 00 01 31 31 31 32 30 33 33 36 \|.....`..11120336\|
		65	*
		66	* str is at offset 80 or 84 now (64-bit or 32-bit off_t).
		67	* We (ab)use the fact that value happens to be aligned,
		68	* and fetch it in one go:
		69	*/
		70	if (sizeof(off_t) == 8) {
		71	value = (off_t)str;
		72	value = SWAP_BE64(value);
		73	} else if (sizeof(off_t) == 4) {
		74	value = (off_t)str;
		75	value = SWAP_BE32(value);
		76	} else {
		77	value = 0;
		78	len = sizeof(off_t);
		79	while (--len)
		80	value = (value << 8) + (unsigned char) *str++;
		81	}
		82	return value;
		83	}
		84
17	/* NB: _DESTROYS_ str[len] character! */	85	/* NB: _DESTROYS_ str[len] character! */
18	static unsigned long long getOctal(char *str, int len)	86	static unsigned long long getOctal(char *str, int len)
19	{	87	{
@@ -234,7 +302,10 @@ char FAST_FUNC get_header_tar(archive_handle_t *archive_handle)
234	file_header->gname = tar.gname[0] ? xstrndup(tar.gname, sizeof(tar.gname)) : NULL;	302	file_header->gname = tar.gname[0] ? xstrndup(tar.gname, sizeof(tar.gname)) : NULL;
235	#endif	303	#endif
236	file_header->mtime = GET_OCTAL(tar.mtime);	304	file_header->mtime = GET_OCTAL(tar.mtime);
237	file_header->size = GET_OCTAL(tar.size);	305	/* Size field: handle GNU tar's "base256 encoding" */
		306	file_header->size = (tar.size & 0xc0) == 0x80 / positive base256? */
		307	? getBase256_len12(tar.size)
		308	: GET_OCTAL(tar.size);
238	file_header->gid = GET_OCTAL(tar.gid);	309	file_header->gid = GET_OCTAL(tar.gid);
239	file_header->uid = GET_OCTAL(tar.uid);	310	file_header->uid = GET_OCTAL(tar.uid);
240	/* Set bits 0-11 of the files mode */	311	/* Set bits 0-11 of the files mode */


diff --git a/archival/libunarchive/header_verbose_list.c b/archival/libunarchive/header_verbose_list.c index f059dd981..dc3100361 100644 --- a/archival/libunarchive/header_verbose_list.c +++ b/archival/libunarchive/header_verbose_list.c
@@ -24,11 +24,11 @@ void FAST_FUNC header_verbose_list(const file_header_t *file_header)
24	snprintf(gid, sizeof(gid), "%u", (unsigned)file_header->gid);	24	snprintf(gid, sizeof(gid), "%u", (unsigned)file_header->gid);
25	group = gid;	25	group = gid;
26	}	26	}
27	printf("%s %s/%s %9u %4u-%02u-%02u %02u:%02u:%02u %s",	27	printf("%s %s/%s %9"OFF_FMT"u %4u-%02u-%02u %02u:%02u:%02u %s",
28	bb_mode_string(file_header->mode),	28	bb_mode_string(file_header->mode),
29	user,	29	user,
30	group,	30	group,
31	(unsigned int) file_header->size,	31	file_header->size,
32	1900 + mtime->tm_year,	32	1900 + mtime->tm_year,
33	1 + mtime->tm_mon,	33	1 + mtime->tm_mon,
34	mtime->tm_mday,	34	mtime->tm_mday,