Rodrigo Queiro wrote:
I discovered this because Python's tarfile module fails to open such files
with "invalid header", since it expects this field to contain an ASCII
number, as described in the docs:
That's a shortcoming in tar's documentation. Tar uses the GNU format by default,
which has a base-256 extension that supports negative timestamps. If you want
GNU tar to refuse to use this GNU extension, please use '-H ustar'.
I fixed the documentation bug by installing the attached patches. Most of them
are to bring Tar up-to-date with recently-released compilers and whatnot; patch
0003 fixes the documentation bug in question.
>From 1bdc8c22f59d44fcb7be2e735b4758e6d4b2dd8a Mon Sep 17 00:00:00 2001
From: Paul Eggert <egg...@cs.ucla.edu>
Date: Sat, 18 Nov 2017 08:39:33 -0800
Subject: [PATCH 1/5] Fix typo caught by GCC 7.2.1
* lib/wordsplit.c (wordsplit_perror): Add missing "break;".
---
lib/wordsplit.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/lib/wordsplit.c b/lib/wordsplit.c
index 07d0f8a..f2ecada 100644
--- a/lib/wordsplit.c
+++ b/lib/wordsplit.c
@@ -1584,6 +1584,7 @@ wordsplit_perror (struct wordsplit *wsp)
case WRDSE_NOSUPP:
wsp->ws_error (_("command substitution is not yet supported"));
+ break;
case WRDSE_USAGE:
wsp->ws_error (_("invalid wordsplit usage"));
--
2.14.3
>From 75de6ec6425ae6fd76f7a9fe92f8426f213d120b Mon Sep 17 00:00:00 2001
From: Paul Eggert <egg...@cs.ucla.edu>
Date: Sat, 18 Nov 2017 08:39:33 -0800
Subject: [PATCH 2/5] build: update gnulib submodule to latest
---
gnulib | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/gnulib b/gnulib
index e210a3c..91e8348 160000
--- a/gnulib
+++ b/gnulib
@@ -1 +1 @@
-Subproject commit e210a3cbaec0ee82a67ff8fc427e21bdd64dba1b
+Subproject commit 91e834891d14dd73788290293bde0e42415c9bdb
--
2.14.3
>From 063790ca3cca7824276c447f3d5914a9a761579f Mon Sep 17 00:00:00 2001
From: Paul Eggert <egg...@cs.ucla.edu>
Date: Sat, 18 Nov 2017 08:39:33 -0800
Subject: [PATCH 3/5] Document base-256 representation in GNU format
Problem reported by Rodrigo Queiro in:
https://lists.gnu.org/r/bug-tar/2017-11/msg00018.html
* doc/intern.texi (Standard, Extensions):
Document base-256 representations.
---
doc/intern.texi | 24 ++++++++++++++++++------
1 file changed, 18 insertions(+), 6 deletions(-)
diff --git a/doc/intern.texi b/doc/intern.texi
index 8d20567..5ef0ee8 100644
--- a/doc/intern.texi
+++ b/doc/intern.texi
@@ -87,6 +87,8 @@ The @code{name}, @code{linkname}, @code{magic}, @code{uname}, and
@code{gname} are null-terminated character strings. All other fields
are zero-filled octal numbers in ASCII. Each numeric field of width
@var{w} contains @var{w} minus 1 digits, and a null.
+(In the extended @acronym{GNU} format, the numeric fields can take
+other forms.)
The @code{name} field is the file name of the file, with directory names
(if any) preceding the file name, separated by slashes.
@@ -112,14 +114,12 @@ be ignored.
The @code{size} field is the size of the file in bytes; linked files
are archived with this field specified as zero.
-The @code{mtime} field is the data modification time of the file at
-the time it was archived. It is the ASCII representation of the octal
-value of the last time the file's contents were modified, represented
-as an integer number of
+The @code{mtime} field represents the data modification time of the file at
+the time it was archived. It represents the integer number of
seconds since January 1, 1970, 00:00 Coordinated Universal Time.
-The @code{chksum} field is the ASCII representation of the octal value
-of the simple sum of all bytes in the header block. Each 8-bit
+The @code{chksum} field represents
+the simple sum of all bytes in the header block. Each 8-bit
byte in the header is added to an unsigned integer, initialized to
zero, the precision of which shall be no less than seventeen bits.
When calculating the checksum, the @code{chksum} field is treated as
@@ -310,6 +310,18 @@ of an archive should have this type.
@end table
+For fields containing numbers or timestamps that are out of range for
+the basic format, the @acronym{GNU} format uses a base-256
+representation instead of an ASCII octal number. If the leading byte
+is 0xff (255), all the bytes of the field (including the leading byte)
+are concatenated in big-endian order, with the result being a negative
+number expressed in two's complement form. If the leading byte is
+0x80 (128), the non-leading bytes of the field are concatenating in
+big-endian order, with the result being a positive number expressed in
+binary form. Leading bytes other than 0xff, 0x80 and ASCII octal
+digits are reserved for future use, as are base-256 representations of
+values that would be in range for the basic format.
+
You may have trouble reading a @acronym{GNU} format archive on a
non-@acronym{GNU} system if the options @option{--incremental} (@option{-G}),
@option{--multi-volume} (@option{-M}), @option{--sparse} (@option{-S}), or @option{--label=@var{archive-label}} (@option{-V @var{archive-label}}) were
--
2.14.3
>From 07d1f8185d4056d2db4ca125fc2808f3cffb63d8 Mon Sep 17 00:00:00 2001
From: Paul Eggert <egg...@cs.ucla.edu>
Date: Sat, 18 Nov 2017 08:39:33 -0800
Subject: [PATCH 4/5] Port to Texinfo 6.4
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
* doc/tar.texi (Sparse Recovery): Omit â.â from anchor name,
as âmakeinfoâ now complains about it. All uses changed.
---
doc/sparse.texi | 4 ++--
doc/tar.texi | 26 +++++++++++++-------------
2 files changed, 15 insertions(+), 15 deletions(-)
diff --git a/doc/sparse.texi b/doc/sparse.texi
index d607fb7..84fe2b0 100644
--- a/doc/sparse.texi
+++ b/doc/sparse.texi
@@ -143,7 +143,7 @@ format, it will also extract a file containing extension header
attributes. This file can be used to expand the file to its original
state. However, posix-aware @command{tar}s will usually ignore the
unknown variables, which makes restoring the file more
-difficult. @xref{extracting sparse v.0.x, Extraction of sparse
+difficult. @xref{extracting sparse v0x, Extraction of sparse
members in v.0.0 format}, for the detailed description of how to
restore such members using non-GNU @command{tar}s.
@end enumerate
@@ -175,7 +175,7 @@ The real name of the sparse file is stored in the variable
@code{GNU.sparse.name}. Thus, those @command{tar} implementations
that are not aware of GNU extensions will at least extract the files
into separate directories, giving the user a possibility to expand it
-afterwards. @xref{extracting sparse v.0.x, Extraction of sparse
+afterwards. @xref{extracting sparse v0x, Extraction of sparse
members in v.0.1 format}, for the detailed description of how to
restore such members using non-GNU @command{tar}s.
diff --git a/doc/tar.texi b/doc/tar.texi
index ec657f5..ce8508f 100644
--- a/doc/tar.texi
+++ b/doc/tar.texi
@@ -1521,7 +1521,7 @@ the error message is produced because there is no member named
@file{folk}, only one named @file{home/myself/folk}.
If you are not sure of the exact file name, use @dfn{globbing
-patterns}, for example:
+patterns}, for example:
@smallexample
$ @kbd{tar --list --file=practice.tar --wildcards '*/folk'}
@@ -3155,7 +3155,7 @@ Disable extended attributes support. @xref{Extended File Attributes, xattrs}.
When @command{tar} is using the @option{--files-from} option, this option
instructs @command{tar} to expect file names terminated with
-@acronym{NUL}, and to process file names verbatim.
+@acronym{NUL}, and to process file names verbatim.
This means that @command{tar} correctly works with file names that
contain newlines or begin with a dash.
@@ -3896,7 +3896,7 @@ tar: Exiting with failure status due to previous errors
@end group
@end example
-The following table summarizes all position-sensitive options.
+The following table summarizes all position-sensitive options.
@table @option
@item --directory=@var{dir}
@@ -3915,24 +3915,24 @@ The following table summarizes all position-sensitive options.
@itemx --no-verbatim-files-from
@xref{verbatim-files-from}.
-@item --recursion
+@item --recursion
@itemx --no-recursion
@xref{recurse}.
-@item --anchored
-@itemx --no-anchored
+@item --anchored
+@itemx --no-anchored
@xref{anchored patterns}.
-@item --ignore-case
-@itemx --no-ignore-case
+@item --ignore-case
+@itemx --no-ignore-case
@xref{case-insensitive matches}.
@item --wildcards
@itemx --no-wildcards
@xref{controlling pattern-matching}.
-@item --wildcards-match-slash
-@itemx --no-wildcards-match-slash
+@item --wildcards-match-slash
+@itemx --no-wildcards-match-slash
@xref{controlling pattern-matching}.
@item --exclude
@@ -5596,7 +5596,7 @@ map file.
@subsection Extended File Attributes
Extended file attributes are name-value pairs that can be
-associated with each node in a file system. Despite the fact that
+associated with each node in a file system. Despite the fact that
POSIX.1e draft which proposed them has been withdrawn, the extended
file attributes are supported by many file systems. @GNUTAR{} can
store extended file attributes along with the files. This feature is
@@ -5711,7 +5711,7 @@ listing of ACL is printed after each file entry:
@group
-rw-r--r--+ smith/users 110 2016-03-16 16:07 file
a: user::rw-,user:gray:-w-,group::r--,mask::rw-,other::r--
-@end group
+@end group
@end example
@dfn{Security-Enhanced Linux} (@dfn{SELinux} for short) is a Linux
@@ -10856,7 +10856,7 @@ Done
@end group
@end smallexample
-@anchor{extracting sparse v.0.x}
+@anchor{extracting sparse v0x}
@cindex sparse files v.0.1, extracting with non-GNU tars
@cindex sparse files v.0.0, extracting with non-GNU tars
An @dfn{extended header} is a special @command{tar} archive header
--
2.14.3
>From 6b493e913328b1bc8f7cc5a913a9d2f4d3af59eb Mon Sep 17 00:00:00 2001
From: Paul Eggert <egg...@cs.ucla.edu>
Date: Sat, 18 Nov 2017 08:39:33 -0800
Subject: [PATCH 5/5] Port to gcc -Wimplicit-fallthrough=5
* src/common.h (FALLTHROUGH): New macro, for use with gcc
-Wimplicit-fallthrough=5, which is now the default when used with
Gnulib after commit 2017-05-16T16:23:52!egg...@cs.ucla.edu
and with --enable-gcc-warnings
---
src/buffer.c | 12 +++++-------
src/common.h | 6 ++++++
src/compare.c | 3 +--
src/delete.c | 11 ++++-------
src/extract.c | 12 +++++-------
src/list.c | 9 ++++-----
src/names.c | 24 ++++++++++++------------
src/sparse.c | 3 ++-
src/tar.c | 9 ++++-----
src/update.c | 6 ++----
10 files changed, 45 insertions(+), 50 deletions(-)
diff --git a/src/buffer.c b/src/buffer.c
index 6f96c2f..51f299f 100644
--- a/src/buffer.c
+++ b/src/buffer.c
@@ -106,7 +106,7 @@ bool write_archive_to_stdout;
/* When creating a multi-volume archive, each 'bufmap' represents
a member stored (perhaps partly) in the current record buffer.
Bufmaps are form a single-linked list in chronological order.
-
+
After flushing the record to the output media, all bufmaps that
represent fully written members are removed from the list, the
nblocks and sizeleft values in the bufmap_head and start values
@@ -1004,7 +1004,7 @@ void
flush_archive (void)
{
size_t buffer_level;
-
+
if (access_mode == ACCESS_READ && time_to_start_writing)
{
access_mode = ACCESS_WRITE;
@@ -1296,8 +1296,7 @@ change_tape_menu (FILE *read_file)
sys_spawn_shell ();
break;
}
- /* FALL THROUGH */
-
+ FALLTHROUGH;
default:
fprintf (stderr, _("Invalid input. Type ? for help.\n"));
}
@@ -1506,8 +1505,7 @@ try_new_volume (void)
header = find_next_block ();
if (header->header.typeflag != GNUTYPE_MULTIVOL)
break;
- /* FALL THROUGH */
-
+ FALLTHROUGH;
case GNUTYPE_MULTIVOL:
if (!read_header0 (&dummy))
return false;
@@ -1532,7 +1530,7 @@ try_new_volume (void)
quote (bufmap_head->file_name)));
return false;
}
-
+
if (strcmp (continued_file_name, bufmap_head->file_name))
{
if ((archive_format == GNU_FORMAT || archive_format == OLDGNU_FORMAT)
diff --git a/src/common.h b/src/common.h
index 964f0b6..bbe167e 100644
--- a/src/common.h
+++ b/src/common.h
@@ -44,6 +44,12 @@
# define GLOBAL extern
#endif
+#if 7 <= __GNUC__
+# define FALLTHROUGH __attribute__ ((__fallthrough__))
+#else
+# define FALLTHROUGH ((void) 0)
+#endif
+
#define TAREXIT_SUCCESS PAXEXIT_SUCCESS
#define TAREXIT_DIFFERS PAXEXIT_DIFFERS
#define TAREXIT_FAILURE PAXEXIT_FAILURE
diff --git a/src/compare.c b/src/compare.c
index 8d609e8..fee1372 100644
--- a/src/compare.c
+++ b/src/compare.c
@@ -480,8 +480,7 @@ diff_archive (void)
ERROR ((0, 0, _("%s: Unknown file type '%c', diffed as normal file"),
quotearg_colon (current_stat_info.file_name),
current_header->header.typeflag));
- /* Fall through. */
-
+ FALLTHROUGH;
case AREGTYPE:
case REGTYPE:
case GNUTYPE_SPARSE:
diff --git a/src/delete.c b/src/delete.c
index 0b3c27f..9044ba2 100644
--- a/src/delete.c
+++ b/src/delete.c
@@ -187,8 +187,7 @@ delete_archive_members (void)
skip_member ();
break;
}
-
- /* Fall through. */
+ FALLTHROUGH;
case HEADER_SUCCESS_EXTENDED:
logical_status = status;
break;
@@ -199,7 +198,7 @@ delete_archive_members (void)
set_next_block_after (current_header);
break;
}
- /* Fall through. */
+ FALLTHROUGH;
case HEADER_END_OF_FILE:
logical_status = HEADER_END_OF_FILE;
break;
@@ -210,14 +209,12 @@ delete_archive_members (void)
{
case HEADER_STILL_UNREAD:
WARN ((0, 0, _("This does not look like a tar archive")));
- /* Fall through. */
-
+ FALLTHROUGH;
case HEADER_SUCCESS:
case HEADER_SUCCESS_EXTENDED:
case HEADER_ZERO_BLOCK:
ERROR ((0, 0, _("Skipping to next header")));
- /* Fall through. */
-
+ FALLTHROUGH;
case HEADER_FAILURE:
break;
diff --git a/src/extract.c b/src/extract.c
index 36919da..395db55 100644
--- a/src/extract.c
+++ b/src/extract.c
@@ -394,7 +394,7 @@ set_stat (char const *file_name,
}
/* Find the direct ancestor of FILE_NAME in the delayed_set_stat list.
- */
+ */
static struct delayed_set_stat *
find_direct_ancestor (char const *file_name)
{
@@ -758,10 +758,9 @@ maybe_recoverable (char *file_name, bool regular, bool *interdir_made)
break;
stp = &st;
}
-
/* The caller tried to open a symbolic link with O_NOFOLLOW.
Fall through, treating it as an already-existing file. */
-
+ FALLTHROUGH;
case EEXIST:
/* Remove an old file, if the options allow this. */
@@ -778,8 +777,7 @@ maybe_recoverable (char *file_name, bool regular, bool *interdir_made)
case KEEP_NEWER_FILES:
if (file_newer_p (file_name, stp, ¤t_stat_info))
break;
- /* FALL THROUGH */
-
+ FALLTHROUGH;
case DEFAULT_OLD_FILES:
case NO_OVERWRITE_DIR_OLD_FILES:
case OVERWRITE_OLD_FILES:
@@ -939,7 +937,7 @@ is_directory_link (const char *file_name)
struct stat st;
int e = errno;
int res;
-
+
res = (fstatat (chdir_fd, file_name, &st, AT_SYMLINK_NOFOLLOW) == 0 &&
S_ISLNK (st.st_mode) &&
fstatat (chdir_fd, file_name, &st, 0) == 0 &&
@@ -1011,7 +1009,7 @@ extract_dir (char *file_name, int typeflag)
if (keep_directory_symlink_option && is_directory_link (file_name))
return 0;
-
+
if (deref_stat (file_name, &st) == 0)
{
current_mode = st.st_mode;
diff --git a/src/list.c b/src/list.c
index 84e73ac..14388a5 100644
--- a/src/list.c
+++ b/src/list.c
@@ -120,7 +120,7 @@ enforce_one_top_level (char **pfile_name)
{
char *file_name = *pfile_name;
char *p;
-
+
for (p = file_name; *p && (ISSLASH (*p) || *p == '.'); p++)
;
@@ -132,7 +132,7 @@ enforce_one_top_level (char **pfile_name)
if (ISSLASH (p[pos]) || p[pos] == 0)
return;
}
-
+
*pfile_name = make_file_name (one_top_level_dir, file_name);
normalize_filename_x (*pfile_name);
}
@@ -218,7 +218,7 @@ read_and (void (*do_something) (void))
if (show_omitted_dirs_option)
WARN ((0, 0, _("%s: Omitting"),
quotearg_colon (current_stat_info.file_name)));
- /* Fall through. */
+ FALLTHROUGH;
default:
skip_member ();
continue;
@@ -273,8 +273,7 @@ read_and (void (*do_something) (void))
{
case HEADER_STILL_UNREAD:
ERROR ((0, 0, _("This does not look like a tar archive")));
- /* Fall through. */
-
+ FALLTHROUGH;
case HEADER_ZERO_BLOCK:
case HEADER_SUCCESS:
if (block_number_option)
diff --git a/src/names.c b/src/names.c
index bd1e44e..f6ad9fe 100644
--- a/src/names.c
+++ b/src/names.c
@@ -146,7 +146,7 @@ static struct argp_option names_options[] = {
{"no-wildcards-match-slash", NO_WILDCARDS_MATCH_SLASH_OPTION, 0, 0,
N_("wildcards do not match '/'"), GRID+1 },
#undef GRID
-
+
{NULL}
};
@@ -160,7 +160,7 @@ file_selection_option (int key)
if (p->key == key)
return p;
return NULL;
-}
+}
static char const *
file_selection_option_name (int key)
@@ -173,7 +173,7 @@ static bool
is_file_selection_option (int key)
{
return file_selection_option (key) != NULL;
-}
+}
/* Either NL or NUL, as decided by the --null option. */
static char filename_terminator = '\n';
@@ -221,7 +221,7 @@ static int matching_flags = 0;
/* exclude_fnmatch options */
static int include_anchored = EXCLUDE_ANCHORED;
/* Pattern anchoring options used for file inclusion */
-
+
#define EXCLUDE_OPTIONS \
(((wildcards != disable_wildcards) ? EXCLUDE_WILDCARDS : 0) \
| matching_flags \
@@ -696,7 +696,7 @@ name_list_adjust (void)
For simplicity, only a tail pointer of the list is maintained.
*/
-
+
struct name_elt *unconsumed_option_tail;
/* Push an option to the list */
@@ -729,7 +729,7 @@ unconsumed_option_report (void)
if (unconsumed_option_tail)
{
struct name_elt *elt;
-
+
ERROR ((0, 0, _("The following options were used after any non-optional arguments in archive create or update mode. These options are positional and affect only arguments that follow them. Please, rearrange them properly.")));
elt = unconsumed_option_tail;
@@ -753,13 +753,13 @@ unconsumed_option_report (void)
ERROR ((0, 0, _("--%s has no effect"),
file_selection_option_name (elt->v.opt.option)));
break;
-
+
default:
break;
}
elt = elt->next;
}
-
+
unconsumed_option_free ();
}
}
@@ -967,7 +967,7 @@ handle_option (const char *str, struct name_elt const *ent)
struct wordsplit ws;
int i;
struct option_locus loc;
-
+
while (*str && isspace (*str))
++str;
if (*str != '-')
@@ -1025,7 +1025,7 @@ read_next_name (struct name_elt *ent, struct name_elt *ret)
(0, 0, N_("%s: file name read contains nul character"),
quotearg_colon (ent->v.file.name)));
ent->v.file.term = 0;
- /* fall through */
+ FALLTHROUGH;
case file_list_success:
if (!ent->v.file.verbatim)
{
@@ -1110,7 +1110,7 @@ name_next_elt (int change_dirs)
name_list_advance ();
break;
}
- /* fall through */
+ FALLTHROUGH;
case NELT_NAME:
copy_name (ep);
if (unquote_option)
@@ -1128,7 +1128,7 @@ name_next_elt (int change_dirs)
}
unconsumed_option_report ();
-
+
return NULL;
}
diff --git a/src/sparse.c b/src/sparse.c
index b3a3fd3..d41c0ea 100644
--- a/src/sparse.c
+++ b/src/sparse.c
@@ -361,11 +361,12 @@ sparse_scan_file (struct tar_sparse_file *file)
/* fall back to "raw" for this and all other files */
hole_detection = HOLE_DETECTION_RAW;
#endif
+ FALLTHROUGH;
case HOLE_DETECTION_RAW:
if (sparse_scan_file_raw (file))
return true;
}
-
+
return false;
}
diff --git a/src/tar.c b/src/tar.c
index 07a6995..3f844a8 100644
--- a/src/tar.c
+++ b/src/tar.c
@@ -1241,7 +1241,7 @@ parse_owner_group (char *arg, uintmax_t field_max, char const **name_option)
u = u1;
break;
}
- /* Fall through. */
+ FALLTHROUGH;
case LONGINT_OVERFLOW:
invalid_num = arg;
break;
@@ -1396,8 +1396,7 @@ parse_opt (int key, char *arg, struct argp_state *state)
optloc_save (OC_LISTED_INCREMENTAL, args->loc);
listed_incremental_option = arg;
after_date_option = true;
- /* Fall through. */
-
+ FALLTHROUGH;
case 'G':
/* We are making an incremental dump (FIXME: are we?); save
directories at the beginning of the archive, and include in each
@@ -1522,8 +1521,7 @@ parse_opt (int key, char *arg, struct argp_state *state)
case 'N':
after_date_option = true;
- /* Fall through. */
-
+ FALLTHROUGH;
case NEWER_MTIME_OPTION:
if (TIME_OPTION_INITIALIZED (newer_mtime_option))
USAGE_ERROR ((0, 0, _("More than one threshold date")));
@@ -2066,6 +2064,7 @@ parse_opt (int key, char *arg, struct argp_state *state)
argp_error (state,
_("Options '-[0-7][lmh]' not supported by *this* tar"));
+ exit (EX_USAGE);
#endif /* not DEVICE_PREFIX */
diff --git a/src/update.c b/src/update.c
index 73fbbe1..2f823e4 100644
--- a/src/update.c
+++ b/src/update.c
@@ -186,13 +186,11 @@ update_archive (void)
{
case HEADER_STILL_UNREAD:
WARN ((0, 0, _("This does not look like a tar archive")));
- /* Fall through. */
-
+ FALLTHROUGH;
case HEADER_SUCCESS:
case HEADER_ZERO_BLOCK:
ERROR ((0, 0, _("Skipping to next header")));
- /* Fall through. */
-
+ FALLTHROUGH;
case HEADER_FAILURE:
break;
--
2.14.3
