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

From: Rafał Mużyło <galtgendo@gmail.com>
Date: Jun 13, 2012, 14:47 UTC
Message-ID: <20120613144758.GA2438@blackspire>

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

Basically, while this is not a problem for GNU tar, the correct checksum should be computed using unsigned values.

Attached trivial testcase shows the difference.

Patch making the change shown in the testcase also attached.

#include <stdio.h> #include <string.h>

static unsigned int ustar_header_chksum(const void *buffer, int sign)
{
  const char *p = (const char *)buffer;
  unsigned int chksum = 0;
  while (p < (const char *)buffer + strlen(buffer))
  {
    if (sign) chksum += *p++; else chksum += (unsigned char)*p++;
  }
  return chksum;
}

int main(int argc, char** argv) { const char* teststring = "żółte źrebię"; printf("%u\n", ustar_header_chksum(teststring, 0)); printf("%u\n", ustar_header_chksum(teststring, 1)); return 0; }

--- archive-tar.c 2012-04-26 21:25:49.000000000 +0200 +++ archive-tar.c 2012-06-13 16:43:59.220945967 +0200

@@ -104,11 +104,11 @@ static unsigned int ustar_header_chksum(
 	char *p = (char *)header;
 	unsigned int chksum = 0;
 	while (p < header->chksum)
-		chksum += *p++;
+		chksum += (unsigned char)*p++;
 	chksum += sizeof(header->chksum) * ' ';
 	p += sizeof(header->chksum);
 	while (p < (char *)header + sizeof(struct ustar_header))
-		chksum += *p++;
+		chksum += (unsigned char)*p++;
 	return chksum;
 }