Re: A slight inconvenience with 'git archive --format=tar'

Rafał Mużyło [off-list ref] writes:

I just stumbled upon this while checking a few mailing lists.
I haven't found any mails about in in the archives yet, so I assume, that
no mail have been written yet.

The problem is described here:
http://sourceforge.net/projects/sevenzip/forums/forum/45798/topic/5322604

Thanks.  It sounds a bit more than "slight inconvenience" to me ;-)

-- >8 --
Date: Wed, 13 Jun 2012 10:42:25 -0700
Subject: [PATCH] archive: ustar header checksum is computed unsigned

POSIX.1 (pax) is pretty clear on this:

  The chksum field shall be the ISO/IEC 646:1991 standard IRV
  representation of the octal value of the simple sum of all octets
  in the header logical record. Each octet in the header shall be
  treated as an unsigned value. These values shall be added to an
  unsigned integer, initialized to zero, the precision of which is
  not less than 17 bits. When calculating the checksum, the chksum
  field is treated as if it were all <space> characters.

so is GNU:

  http://www.gnu.org/software/tar/manual/html_node/Checksumming.html

Found by 7zip folks and reported by Rafał Mużyło.

Signed-off-by: Junio C Hamano <redacted>
---
 archive-tar.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/archive-tar.c b/archive-tar.c
index dc91c6b..0ba3f25 100644
--- a/archive-tar.c
+++ b/archive-tar.c

@@ -139,13 +139,13 @@ static void strbuf_append_ext_header(struct strbuf *sb, const char *keyword,
 
 static unsigned int ustar_header_chksum(const struct ustar_header *header)
 {
-	const char *p = (const char *)header;
+	const unsigned char *p = (const unsigned char *)header;
 	unsigned int chksum = 0;
-	while (p < header->chksum)
+	while (p < (const unsigned char *)header->chksum)
 		chksum += *p++;
 	chksum += sizeof(header->chksum) * ' ';
 	p += sizeof(header->chksum);
-	while (p < (const char *)header + sizeof(struct ustar_header))
+	while (p < (const unsigned char *)header + sizeof(struct ustar_header))
 		chksum += *p++;
 	return chksum;
 }

-- 
1.7.11.rc3.25.g4c2075b

Am 13.06.2012 19:45, schrieb Junio C Hamano:

Rafał Mużyło [off-list ref] writes:

quoted

I just stumbled upon this while checking a few mailing lists.
I haven't found any mails about in in the archives yet, so I assume, that
no mail have been written yet.

The problem is described here:
http://sourceforge.net/projects/sevenzip/forums/forum/45798/topic/5322604

Thanks.  It sounds a bit more than "slight inconvenience" to me ;-)

Indeed, but two mitigating factors, if you will, are that this only 
affects files whose path or link target contain non-ASCII characters, 
and -- as Rafał wrote -- that GNU tar silently accepts signed checksums 
as well, which makes it a bit difficult to test.

René