Re: [Bug-tar] POSIX_ARCHIVE & big sparse files (>= 2^33 B excluding holes)

[Paul Eggert]

Thanks for catching that bug.  We don't yet have a facility
for large test cases, so I guess we can put that on the todo list.
Your patch doesn't feel quite right, as there's a similar
issue in pax_dump_header_0, and also there's a problem if
the shrunken size is less than 8 GiB but the real size is not.
I pushed the following patch instead: does it fix things for you?

>From ba2d1f62c731a58a80ef4a2e33b0b1528a605cae Mon Sep 17 00:00:00 2001
From: Paul Eggert <address@hidden>
Date: Sat, 26 Jan 2013 09:52:55 -0800
Subject: [PATCH] tar: fix bug with sparse files with effective size of 8 GiB
 or more

Reported by Pavel Raiskup in
<http://lists.gnu.org/archive/html/bug-tar/2013-01/msg00001.html>.
* NEWS: Document the fix.
* src/sparse.c (pax_start_header): New function.
(pax_dump_header_0, pax_dump_header_1): Use it.
---
 NEWS         |  9 ++++++++-
 src/sparse.c | 20 ++++++++++++++------
 2 files changed, 22 insertions(+), 7 deletions(-)

diff --git a/NEWS b/NEWS
index e372663..29b4486 100644
--- a/NEWS
+++ b/NEWS
@@ -1,9 +1,16 @@
-GNU tar NEWS - User visible changes. 2013-01-01
+GNU tar NEWS - User visible changes. 2013-01-26
 Please send GNU tar bug reports to <address@hidden>
 
 
 version 1.26.90 (Git)
 
+* Bug fixes
+
+** Sparse files with large data
+
+When creating a PAX-format archive, tar no longer arbitrarily restricts
+the size of the representation of a sparse file to be less than 8 GiB.
+
 * Quoting
 
 In the default C locale, diagnostics and output of 'tar' have been
diff --git a/src/sparse.c b/src/sparse.c
index b93cdc2..417d4ff 100644
--- a/src/sparse.c
+++ b/src/sparse.c
@@ -917,6 +917,18 @@ pax_sparse_member_p (struct tar_sparse_file *file)
           || file->stat_info->sparse_major > 0;
 }
 
+/* Start a header that uses the effective (shrunken) file size.  */
+static union block *
+pax_start_header (struct tar_stat_info *st)
+{
+  off_t realsize = st->stat.st_size;
+  union block *blk;
+  st->stat.st_size = st->archive_file_size;
+  blk = start_header (st);
+  st->stat.st_size = realsize;
+  return blk;
+}
+
 static bool
 pax_dump_header_0 (struct tar_sparse_file *file)
 {
@@ -966,9 +978,7 @@ pax_dump_header_0 (struct tar_sparse_file *file)
          return false;
        }
     }
-  blk = start_header (file->stat_info);
-  /* Store the effective (shrunken) file size */
-  OFF_TO_CHARS (file->stat_info->archive_file_size, blk->header.size);
+  blk = pax_start_header (file->stat_info);
   finish_header (file->stat_info, blk, block_ordinal);
   if (save_file_name)
     {
@@ -1033,9 +1043,7 @@ pax_dump_header_1 (struct tar_sparse_file *file)
   if (strlen (file->stat_info->file_name) > NAME_FIELD_SIZE)
     file->stat_info->file_name[NAME_FIELD_SIZE] = 0;
 
-  blk = start_header (file->stat_info);
-  /* Store the effective (shrunken) file size */
-  OFF_TO_CHARS (file->stat_info->archive_file_size, blk->header.size);
+  blk = pax_start_header (file->stat_info);
   finish_header (file->stat_info, blk, block_ordinal);
   free (file->stat_info->file_name);
   file->stat_info->file_name = save_file_name;
-- 
1.7.11.7