[Bug-tar] Extracting archives with sparse files still broken in 1.27

[Mark]

Hi,

Just a heads-up... I just downloaded GNU tar 1.27 and noticed that the
extraction bug for archives containing sparse files is still present. :(

The bug shows up when attempting to extract archives which were created by
star with its "artype=xustar" option.

Sergey Poznyakoff originally posted a patch for the bug in February 2011, see
 http://lists.gnu.org/archive/html/bug-tar/2011-02/msg00011.html

I applied that patch to the tar 1.27 source and it does fix the bug.

Without the patch, tar gets very confused and can create zero-length files
with "garbage" names. Here's an example of the output of the unpatched tar
when trying to extract a particular archive:

$ tar -xf test_star.tar
tar: Skipping to next header
tar: Substituting `.' for empty member name
tar: .: Cannot open: Is a directory
tar: Skipping to next header
tar: Skipping to next header
tar: Archive contains '\0009\026B\002\002\001\001H\0`\001' where numeric
off_t value expected
tar: Substituting `.' for empty member name
tar: .: Cannot open: Is a directory
tar: Skipping to next header
tar: Skipping to next header
tar: Substituting `.' for empty member name
tar: .: Cannot open: Is a directory
tar: Skipping to next header
tar: Skipping to next header
tar: Skipping to next header
tar: Substituting `.' for empty member name
tar: .: Cannot open: Is a directory
tar: Skipping to next header
tar: Substituting `.' for empty member name
tar: .: Cannot open: Is a directory
tar: Skipping to next header
tar: Substituting `.' for empty member name
tar: .: Cannot open: Is a directory
tar: Skipping to next header
tar: Skipping to next header
tar: Skipping to next header
tar: Skipping to next header
tar: Exiting with failure status due to previous errors



-- Mark