]> git.ipfire.org Git - thirdparty/libarchive.git/log
thirdparty/libarchive.git
9 years agoadd missing _PC_NAME_MAX checks 663/head
Mike Frysinger [Sun, 28 Feb 2016 18:10:16 +0000 (13:10 -0500)] 
add missing _PC_NAME_MAX checks

This file looks for _PC_NAME_MAX in a few places, but not all.
Add some ifdefs in setup_current_filesystem too.

9 years agoMerge pull request #645 from andrewgregory/check_symlinks
Tim Kientzle [Fri, 26 Feb 2016 03:01:33 +0000 (19:01 -0800)] 
Merge pull request #645 from andrewgregory/check_symlinks

fix check_symlinks

9 years agoadd test for SECURE_SYMLINKS with absolute path 645/head
Andrew Gregory [Tue, 23 Feb 2016 02:17:20 +0000 (21:17 -0500)] 
add test for SECURE_SYMLINKS with absolute path

9 years agoSkip root directory symlink check
Andrew Gregory [Fri, 15 Jan 2016 07:39:41 +0000 (02:39 -0500)] 
Skip root directory symlink check

The first time check_symlinks is called on an absolute path it will use
the entry pathname directly, blanking the leading slash.  This leads to
calling lstat on an empty string, which returns ENOENT, terminating the
loop early and falsely marking the path as safe.

9 years agoadd test for SECURE_SYMLINKS for first child
Andrew Gregory [Tue, 23 Feb 2016 02:18:02 +0000 (21:18 -0500)] 
add test for SECURE_SYMLINKS for first child

9 years agoRestore modified path after lstat
Andrew Gregory [Fri, 15 Jan 2016 07:40:00 +0000 (02:40 -0500)] 
Restore modified path after lstat

check_symlinks is intended to check each component of a path, but failed
to restore the stripped trailing components after each iteration,
leaving a NUL byte in the middle of the path.

9 years agoFix for issue #623.
Tim Kientzle [Sun, 21 Feb 2016 20:34:57 +0000 (12:34 -0800)] 
Fix for issue #623.

Apparently, people have come to expect that the following
is sufficient to get bit-for-bit identical output from tar:
   * Same filenames
   * Same contents
   * Same uid, gid
   * Same mtime (forced via "touch -t <timestamp>")
   * Sorting entries

Bsdtar's "restricted pax" format violated this by including
ctime, atime, and birthtime (which are not updated by
'touch -t).  So we should only emit those additional time
values in the full pax format.

People who are really serious about generating bit-for-bit
identical archives should really build their own command-line
interface: You can still use libarchive to build the output,
but your custom CLI could sort the entries and strip everything
except a bare minimum of basic metadata.

9 years agoClarify the limit on number of filters
Tim Kientzle [Sun, 21 Feb 2016 19:43:03 +0000 (11:43 -0800)] 
Clarify the limit on number of filters

9 years agoFix globbing error v3.1.900a
Tim Kientzle [Sun, 21 Feb 2016 16:51:18 +0000 (08:51 -0800)] 
Fix globbing error

9 years agoUpdate NEWS with highlights from last two years of work
Tim Kientzle [Sun, 14 Feb 2016 06:03:35 +0000 (22:03 -0800)] 
Update NEWS with highlights from last two years of work

9 years agoMinor adjustments to DEBUG messages
Tim Kientzle [Sun, 14 Feb 2016 06:02:54 +0000 (22:02 -0800)] 
Minor adjustments to DEBUG messages

9 years agoIssue #659: Fix xar parsing of archives with appleextended ACLs
Tim Kientzle [Sun, 14 Feb 2016 05:52:24 +0000 (21:52 -0800)] 
Issue #659:  Fix xar parsing of archives with appleextended ACLs

The parser treated the <appleextended> tag as both known and
unknown, leaving the parser in a state that prevented it
from reading any further attributes.  Depending on the
order of attributes, this could result in a file with no
name, for instance.

9 years agoMerge pull request #650 from cinience/master
Tim Kientzle [Fri, 5 Feb 2016 06:07:37 +0000 (22:07 -0800)] 
Merge pull request #650 from cinience/master

Update the remaining calls to __archive_write_program_allocate to include the new parameter.

9 years agoFix bug for __archive_write_program_allocate 650/head
cinience [Mon, 1 Feb 2016 08:54:29 +0000 (16:54 +0800)] 
Fix bug for __archive_write_program_allocate

9 years agoFix bug for __archive_write_program_allocate
cinience [Mon, 1 Feb 2016 08:52:48 +0000 (16:52 +0800)] 
Fix bug for __archive_write_program_allocate

9 years agoSet version number to 3.1.900a
Tim Kientzle [Mon, 1 Feb 2016 05:47:21 +0000 (21:47 -0800)] 
Set version number to 3.1.900a

9 years agoRead very large RAR headers incrementally
Tim Kientzle [Mon, 1 Feb 2016 05:43:23 +0000 (21:43 -0800)] 
Read very large RAR headers incrementally

Formerly, libarchive tried to read the entire header into
memory at once, which created problems for malformed RAR
files with 4GB header sizes.  This was causing occasional
crashes of the test suite on memory-limited systems.

9 years agoRework LZ4 option tests to work with various implementations of posix_spawnp
Tim Kientzle [Sun, 31 Jan 2016 21:22:47 +0000 (13:22 -0800)] 
Rework LZ4 option tests to work with various implementations of posix_spawnp

9 years agoUpdate lz4 test to watch for the reworded error messages
Tim Kientzle [Sun, 31 Jan 2016 20:33:36 +0000 (12:33 -0800)] 
Update lz4 test to watch for the reworded error messages

9 years agoImprove error messages by including the name of the external program being used
Tim Kientzle [Sun, 31 Jan 2016 20:33:03 +0000 (12:33 -0800)] 
Improve error messages by including the name of the external program being used

9 years agoBring along a trailing null when guessing the path to the program under test
Tim Kientzle [Sun, 31 Jan 2016 20:24:45 +0000 (12:24 -0800)] 
Bring along a trailing null when guessing the path to the program under test

9 years agoClarify error messages when using external compression programs
Tim Kientzle [Sun, 31 Jan 2016 19:08:59 +0000 (11:08 -0800)] 
Clarify error messages when using external compression programs

9 years agoFix LZ4 tests when liblz4 is not present
Tim Kientzle [Sat, 30 Jan 2016 19:26:02 +0000 (11:26 -0800)] 
Fix LZ4 tests when liblz4 is not present

9 years agoTrack changes in the lz4 command-line tool
Tim Kientzle [Sat, 30 Jan 2016 19:14:19 +0000 (11:14 -0800)] 
Track changes in the lz4 command-line tool

9 years agoReally sync up the test harnesses
Tim Kientzle [Fri, 29 Jan 2016 05:33:53 +0000 (21:33 -0800)] 
Really sync up the test harnesses

9 years agoFix two broken UU files
Tim Kientzle [Fri, 29 Jan 2016 05:25:42 +0000 (21:25 -0800)] 
Fix two broken UU files

9 years agoSync up the test harnesses; add error checks to the UUdecode utility
Tim Kientzle [Fri, 29 Jan 2016 05:25:25 +0000 (21:25 -0800)] 
Sync up the test harnesses; add error checks to the UUdecode utility

9 years agoFix test on systems without zlib
Tim Kientzle [Thu, 28 Jan 2016 05:16:01 +0000 (21:16 -0800)] 
Fix test on systems without zlib

9 years agoUpdate the version string correctly
Tim Kientzle [Sun, 24 Jan 2016 02:26:46 +0000 (18:26 -0800)] 
Update the version string correctly

9 years agoFix cmake tests for nettle and OpenSSL support
Tim Kientzle [Mon, 11 Jan 2016 13:51:28 +0000 (05:51 -0800)] 
Fix cmake tests for nettle and OpenSSL support

9 years agoMerge branch 'master' of github.com:libarchive/libarchive
Tim Kientzle [Tue, 19 Jan 2016 13:35:11 +0000 (05:35 -0800)] 
Merge branch 'master' of github.com:libarchive/libarchive

9 years agoFix nettle library probe
Tim Kientzle [Tue, 19 Jan 2016 13:32:56 +0000 (05:32 -0800)] 
Fix nettle library probe

 = Only use it if we can find the library *and* the headers
 = When probing the library, try to link a function that's
   actually in the library (not 'main')

9 years agoMerge pull request #644 from hinrichsenhans/vs2015
Tim Kientzle [Thu, 14 Jan 2016 05:34:31 +0000 (21:34 -0800)] 
Merge pull request #644 from hinrichsenhans/vs2015

Visual Studio 2015 Changes

9 years agoVisual Studio 2015 Changes 644/head
Hans Hinrichsen [Wed, 6 Jan 2016 14:58:53 +0000 (08:58 -0600)] 
Visual Studio 2015 Changes

snprintf shouldn't be defined in a macro in VS2015 and above due to its C99 changes

9 years agoInclude missing reference files for tests
Tim Kientzle [Mon, 11 Jan 2016 01:57:45 +0000 (17:57 -0800)] 
Include missing reference files for tests

9 years agoInclude overlooked header file in distro
Tim Kientzle [Mon, 11 Jan 2016 01:48:34 +0000 (17:48 -0800)] 
Include overlooked header file in distro

9 years agoRemove Zip UTF8 tests from build
Tim Kientzle [Mon, 11 Jan 2016 00:21:54 +0000 (16:21 -0800)] 
Remove Zip UTF8 tests from build

9 years agoRemove Zip UTF8 tests from master.
Tim Kientzle [Sun, 10 Jan 2016 19:29:44 +0000 (11:29 -0800)] 
Remove Zip UTF8 tests from master.

Ongoing work on Zip UTF8 support is occurring in the 'utf8' branch.

9 years agoMention V7, a few wording tweaks
Tim Kientzle [Sun, 10 Jan 2016 19:21:07 +0000 (11:21 -0800)] 
Mention V7, a few wording tweaks

9 years agoRemove an extraneous .El
Tim Kientzle [Sun, 10 Jan 2016 19:10:28 +0000 (11:10 -0800)] 
Remove an extraneous .El

9 years agoDocument the 'v7' tar format support.
Tim Kientzle [Sun, 10 Jan 2016 18:59:29 +0000 (10:59 -0800)] 
Document the 'v7' tar format support.

9 years agoMerge pull request #638 from hinrichsenhans/test_leading_slash_fix
Tim Kientzle [Fri, 8 Jan 2016 06:25:19 +0000 (22:25 -0800)] 
Merge pull request #638 from hinrichsenhans/test_leading_slash_fix

Test leading slash fix

9 years agoMerge pull request #641 from watercrossing/fixShadowTime
Tim Kientzle [Fri, 8 Jan 2016 06:22:54 +0000 (22:22 -0800)] 
Merge pull request #641 from watercrossing/fixShadowTime

Fix `time` shadowing in archive_write_set_format_warc.c

9 years agoFix `time` shadowing in archive_write_set_format_warc.c 641/head
watercrossing [Thu, 7 Jan 2016 15:32:44 +0000 (15:32 +0000)] 
Fix `time` shadowing in archive_write_set_format_warc.c

9 years agotest_leading_slash compatibility with Windows 638/head
Hans Hinrichsen [Thu, 7 Jan 2016 02:52:36 +0000 (20:52 -0600)] 
test_leading_slash compatibility with Windows

Add '.exe' to the binary name to ensure the text matching works

9 years agoMerge pull request #634 from scfarley/master
Tim Kientzle [Sun, 3 Jan 2016 23:05:43 +0000 (15:05 -0800)] 
Merge pull request #634 from scfarley/master

Apply global flag in substitution

9 years agoApply global flag in substitution 634/head
Sean Farley [Sun, 3 Jan 2016 21:42:00 +0000 (16:42 -0500)] 
Apply global flag in substitution

Add loop to move through a path name looking for additional matches if
the global flag is provided for a substitution.  The loop breaks when no
more matches are found for the rule.  Previously, the global flag would
end up advancing to the next rule instead of searching farther into the
path for another match.

9 years agoMerge pull request #607 from betafive/pbarker/issue578
Tim Kientzle [Sun, 3 Jan 2016 19:01:19 +0000 (11:01 -0800)] 
Merge pull request #607 from betafive/pbarker/issue578

Issue 578: Close temp_fd in 7zip and xar writers

9 years agoIssue #631: Move 'r' into an inner scope to fix shadowing
Tim Kientzle [Fri, 1 Jan 2016 22:25:16 +0000 (14:25 -0800)] 
Issue #631:  Move 'r' into an inner scope to fix shadowing

9 years agoMerge pull request #1 from libarchive/master
Hans Hinrichsen [Thu, 31 Dec 2015 03:36:12 +0000 (21:36 -0600)] 
Merge pull request #1 from libarchive/master

Updating with current libarchive development

9 years agoMerge pull request #628 from obache/neatsrc/issue-627
Tim Kientzle [Sun, 27 Dec 2015 20:00:12 +0000 (12:00 -0800)] 
Merge pull request #628 from obache/neatsrc/issue-627

Ensure to declare UINTMAX_MAX, INTMAX_MAX and INTMAX_MIN

9 years agoEnsure to declare UINTMAX_MAX, INTMAX_MAX and INTMAX_MIN 628/head
OBATA Akio [Sun, 27 Dec 2015 06:54:20 +0000 (15:54 +0900)] 
Ensure to declare UINTMAX_MAX, INTMAX_MAX and INTMAX_MIN

9 years agoFix typo
Joerg Sonnenberger [Sat, 19 Dec 2015 15:20:02 +0000 (16:20 +0100)] 
Fix typo

9 years agoAdd in two missing test cases.
Tim Kientzle [Fri, 20 Nov 2015 05:09:08 +0000 (21:09 -0800)] 
Add in two missing test cases.

9 years agoRework to avoid some C99-isms. <sigh>
Tim Kientzle [Fri, 20 Nov 2015 05:07:56 +0000 (21:07 -0800)] 
Rework to avoid some C99-isms.  <sigh>

9 years agoUse CommonCrypto APIs on OS X only when available 612/head
Brad King [Wed, 21 Oct 2015 18:46:37 +0000 (14:46 -0400)] 
Use CommonCrypto APIs on OS X only when available

Use each CommonCrypto API only when using an OS X SDK version new enough
to provide it.

9 years agoDo not use pthread.h API without header
Brad King [Mon, 26 Oct 2015 14:13:24 +0000 (10:13 -0400)] 
Do not use pthread.h API without header

9 years agoDefine O_CLOEXEC when missing in archive_random.c
Brad King [Wed, 21 Oct 2015 18:16:21 +0000 (14:16 -0400)] 
Define O_CLOEXEC when missing in archive_random.c

We do this several other sources already.

9 years agoForward declare __archive_get_date in a dedicated internal header
Brad King [Wed, 21 Oct 2015 16:02:43 +0000 (12:02 -0400)] 
Forward declare __archive_get_date in a dedicated internal header

Avoid repeating the signature declaration in every source that uses it.
While at it, constify internal __archive_get_date implementation to
keep signatures consistent.

9 years agoUse macros for ZIP format integer constants
Brad King [Wed, 21 Oct 2015 15:56:28 +0000 (11:56 -0400)] 
Use macros for ZIP format integer constants

Define ZIP_4GB_MAX and ZIP_4GB_MAX_UNCOMPRESSED macros to hold the "LL"
integer literals.  This improves readability and improves portability of
the integer literals.

9 years agoInvert logic of ternary pointer checks
Brad King [Wed, 21 Oct 2015 15:59:45 +0000 (11:59 -0400)] 
Invert logic of ternary pointer checks

A PGI compiler warning is triggered by expressions like

  ptr == NULL ? NULL : ptr

that the PGI compiler handles incorrectly.  It chooses the pointer type
of the first option (void*) and warns about conversion of the second
without a cast.  Flip the expression logic to

  ptr != NULL ? ptr : NULL

to help the compiler choose the proper result type.

9 years agoImprove readability of ternary expression checking for empty string
Brad King [Wed, 21 Oct 2015 15:59:45 +0000 (11:59 -0400)] 
Improve readability of ternary expression checking for empty string

In expressions of the form

    m != NULL && m[0] == '\0' ? NULL : m

the goal is to get a `NULL` pointer if `m` is either `NULL` or an empty
string.  This is not clear because in the `m == NULL` case we use the
`m` side to get `NULL` instead of an explicit `NULL`.  Clarify the
intent by using the form

    (m != NULL && m[0] != '\0') ? m : NULL;

instead.

9 years agoGuard inclusion of bcrypt.h with HAVE_BCRYPT_H
Brad King [Wed, 21 Oct 2015 15:53:50 +0000 (11:53 -0400)] 
Guard inclusion of bcrypt.h with HAVE_BCRYPT_H

We already use this guard everywhere except one inclusion.

9 years agoImprove argument names of xmemmem
Brad King [Thu, 22 Oct 2015 14:02:14 +0000 (10:02 -0400)] 
Improve argument names of xmemmem

AIX system headers define the symbol 'hz' as a preprocessor macro.
Rename the arguments to avoid this name and be more readable.

Suggested-by: Joerg Sonnenberger
9 years agoAvoid non-portable u_int8_t
Brad King [Thu, 22 Oct 2015 13:42:23 +0000 (09:42 -0400)] 
Avoid non-portable u_int8_t

Use uint8_t that we prepare in config.h instead.

9 years agoAvoid trailing `,` in enum for ISO C90 compatiblity
Brad King [Wed, 21 Oct 2015 16:00:23 +0000 (12:00 -0400)] 
Avoid trailing `,` in enum for ISO C90 compatiblity

9 years agoWorkaround missing byte swap symbol on VS 7.1 Debug builds
Brad King [Wed, 21 Oct 2015 16:00:41 +0000 (12:00 -0400)] 
Workaround missing byte swap symbol on VS 7.1 Debug builds

The VS 7.1 MS C Debug runtime library does not export _byteswap_ushort.

9 years agoPort main library code to OSF operating system
Brad King [Wed, 21 Oct 2015 15:54:24 +0000 (11:54 -0400)] 
Port main library code to OSF operating system

Add preprocessor conditionals using '__osf__' as needed to build the
main library.  Do not try to port the tests or executables yet.

9 years agoUpdate 'inline' portability for old IBM and Sun compilers
Brad King [Wed, 21 Oct 2015 15:51:56 +0000 (11:51 -0400)] 
Update 'inline' portability for old IBM and Sun compilers

9 years agoAvoid using 'uint8_t' as bitfield type
Brad King [Thu, 22 Oct 2015 14:12:21 +0000 (10:12 -0400)] 
Avoid using 'uint8_t' as bitfield type

The IBM XL compiler does not like it.  Use plain 'unsigned' instead.

9 years agoAvoid using name 'u_long'
Brad King [Wed, 21 Oct 2015 16:03:09 +0000 (12:03 -0400)] 
Avoid using name 'u_long'

The system headers on some platforms define a 'u_long' type so we cannot
use the name.  Spell out 'unsigned long' instead.

9 years agoAvoid left-shift overflow of signed integer
Brad King [Wed, 21 Oct 2015 16:02:21 +0000 (12:02 -0400)] 
Avoid left-shift overflow of signed integer

In libarchive/archive_write_set_format_zip.c there are two calls to
archive_le32enc whose second argument is of the form

 archive_entry_mode(zip->entry) << 16

However, the return type from archive_entry_mode may be a signed integer
so the shift may overflow.  Since the second argument of archive_le32enc
expects uint32_t anyway, simply cast to that prior to shifting.

9 years agoFix typo in archive_write_add_filter_lzop.c comment
Brad King [Wed, 21 Oct 2015 15:58:51 +0000 (11:58 -0400)] 
Fix typo in archive_write_add_filter_lzop.c comment

9 years agoFix typo in archive_write_set_format_mtree.c comment
Brad King [Wed, 21 Oct 2015 15:58:28 +0000 (11:58 -0400)] 
Fix typo in archive_write_set_format_mtree.c comment

9 years agoFix typo in archive_write_set_options.3
Brad King [Wed, 21 Oct 2015 15:55:07 +0000 (11:55 -0400)] 
Fix typo in archive_write_set_options.3

9 years agoDefine ARCHIVE_VERSION_NUMBER before first use
Brad King [Wed, 21 Oct 2015 15:47:41 +0000 (11:47 -0400)] 
Define ARCHIVE_VERSION_NUMBER before first use

9 years agoThese tests only run on Mac; put all the code behind conditionals.
Tim Kientzle [Sat, 24 Oct 2015 11:10:08 +0000 (04:10 -0700)] 
These tests only run on Mac; put all the code behind conditionals.

9 years agoIt seems we can no longer rely on getting WARN back when
Tim Kientzle [Fri, 23 Oct 2015 15:11:44 +0000 (08:11 -0700)] 
It seems we can no longer rely on getting WARN back when
character-set conversion is inexact.  Change this test to
reflect the current realities.

9 years agoDon't try to be smart about probing the min/max tim_t values.
Tim Kientzle [Fri, 23 Oct 2015 04:43:07 +0000 (21:43 -0700)] 
Don't try to be smart about probing the min/max tim_t values.
Just assume that a signed time_t is really a 64-bit or 32-bit integer.

9 years agoIf 7zip is not supported, skip the encryption detection test.
Tim Kientzle [Wed, 21 Oct 2015 08:47:34 +0000 (01:47 -0700)] 
If 7zip is not supported, skip the encryption detection test.

9 years agoWhen comparing ACLs for these tests, ignore 'Guest' account
Tim Kientzle [Wed, 21 Oct 2015 08:08:20 +0000 (01:08 -0700)] 
When comparing ACLs for these tests, ignore 'Guest' account
information that may vary between machines.

9 years agoMerge pull request #608 from Tarsnap/fix-automake-whitespace
Tim Kientzle [Wed, 30 Sep 2015 03:44:33 +0000 (20:44 -0700)] 
Merge pull request #608 from Tarsnap/fix-automake-whitespace

Fix trailing whitespace in Makefile.am

9 years agoFix trailing whitespace in Makefile.am 608/head
Graham Percival [Tue, 29 Sep 2015 20:33:33 +0000 (13:33 -0700)] 
Fix trailing whitespace in Makefile.am

Previously, running the build process gave a warning:

$ sh build/autogen.sh
...
Makefile.am:356: warning: whitespace following trailing backslash
...

This removes that warning.

9 years agoIssue 578: Close temp_fd in 7zip and xar writers 607/head
Paul Barker [Sat, 26 Sep 2015 18:38:13 +0000 (19:38 +0100)] 
Issue 578: Close temp_fd in 7zip and xar writers

The ISO9660 writer uses a temp_fd and closes it in the archive_free handler. The
same logic is added to the archive_free handlers for the 7zip and xar writers.

Signed-off-by: Paul Barker <paul@paulbarker.me.uk>
9 years agoMerge pull request #606 from betafive/pbarker/minitar
Tim Kientzle [Sat, 26 Sep 2015 17:16:28 +0000 (10:16 -0700)] 
Merge pull request #606 from betafive/pbarker/minitar

Cleanup error message printing in minitar example

9 years agoCleanup error message printing in minitar example 606/head
Paul Barker [Wed, 2 Sep 2015 09:53:15 +0000 (11:53 +0200)] 
Cleanup error message printing in minitar example

* An error message should not be printed if EOF is reached in copy_data. However
  an error message should be printed if any other error is encountered.

* A newline should be printed in extract if an error message was printed.

* If a file name is printed in extract (verbose operation) it should be followed
  by a space incase there is an error message.

Signed-off-by: Paul Barker <paul@paulbarker.me.uk>
9 years agoMerge pull request #601 from Tarsnap/strdup-allocation-failure-check
Tim Kientzle [Thu, 10 Sep 2015 04:20:28 +0000 (21:20 -0700)] 
Merge pull request #601 from Tarsnap/strdup-allocation-failure-check

Check for strdup() allocation failure

9 years agoCheck for strdup() allocation failure 601/head
Graham Percival [Mon, 7 Sep 2015 21:31:49 +0000 (14:31 -0700)] 
Check for strdup() allocation failure

Sponsored by: Tarsnap Backup Inc.

9 years agoRework the la_ssize_t and la_int64_t definitions to avoid
Tim Kientzle [Sun, 6 Sep 2015 05:43:09 +0000 (22:43 -0700)] 
Rework the la_ssize_t and la_int64_t definitions to avoid
redefining la_int64_t if both archive.h and archive_entry.h are
included.  (This avoids a warning from clang that redefining
a typedef is a C11 feature.)

9 years agoMerge pull request #599 from Tarsnap/add-missing-tests-to-automake
Tim Kientzle [Fri, 4 Sep 2015 16:15:12 +0000 (09:15 -0700)] 
Merge pull request #599 from Tarsnap/add-missing-tests-to-automake

add missing tests to automake

9 years agoMerge pull request #596 from Tarsnap/warn-hardlink-target-failure
Tim Kientzle [Fri, 4 Sep 2015 04:49:40 +0000 (21:49 -0700)] 
Merge pull request #596 from Tarsnap/warn-hardlink-target-failure

Warn if target of hard-link is not present

9 years agoAdd test for #595 "warn missing hardlink target" 596/head
Graham Percival [Thu, 3 Sep 2015 03:17:06 +0000 (20:17 -0700)] 
Add test for #595 "warn missing hardlink target"

9 years agoWarn if target of hard-link is not present
Graham Percival [Sun, 30 Aug 2015 17:53:12 +0000 (10:53 -0700)] 
Warn if target of hard-link is not present

To reproduce,

        touch a
        ln a b
        ./bsdtar -c -f warn-hard-links.tar a b
        rm a b
        ./bsdtar -x -f warn-hard-links.tar b

should produce a warning message about failing to create 'b' because the
hard-link target 'a' does not exist.  Previously, it did not give any hints
about why 'b' could not be created.

9 years agoadd missing tests to automake 599/head
Graham Percival [Thu, 3 Sep 2015 02:09:06 +0000 (19:09 -0700)] 
add missing tests to automake

9 years agoIssue 593: Add .travis.yml for Travis-CI
Tim Kientzle [Thu, 27 Aug 2015 15:42:02 +0000 (08:42 -0700)] 
Issue 593:  Add .travis.yml for Travis-CI

This will help people who want to use Travis-CI to follow
libarchive development.  Including Travis-CI status in the
README.md is a separate issue that requires one of the
libarchive maintainers to commit to managing it.

9 years agoIssue 591
Tim Kientzle [Thu, 27 Aug 2015 15:35:59 +0000 (08:35 -0700)] 
Issue 591

9 years agoIssue 547: problems with compress bidder
Tim Kientzle [Sun, 9 Aug 2015 04:47:43 +0000 (21:47 -0700)] 
Issue 547:  problems with compress bidder

The code previously was not very careful about verifying the
compression parameters.  This led to cases where it failed to
reject invalid compressed data at the beginning.  The invalid
left shift was one symptom of this.

The code is now more careful:  It verifies that the compression
parameter byte exists and verifies that the maximum code size
is <= 16 bits.

This also includes some new tests to verify that truncated or
otherwise invalid compressed data is rejected.

9 years agoIssue 551: Fix the best-effort UTF8 conversion
Tim Kientzle [Sun, 9 Aug 2015 03:52:19 +0000 (20:52 -0700)] 
Issue 551:  Fix the best-effort UTF8 conversion

If a valid character set conversion is impossible, the code falls back
to a best-effort conversion that preserves ASCII bytes and converts
the rest to Unicode Replacement Characters (if the output is UTF8)
or '?' (otherwise).  This code did not correctly track the remaining
bytes in the output buffer; I've replaced this with slower and simpler
code that utilizes the safe string append functions.

10 years agoIssue #582: reject sparse blocks with negative size or offset, detect overflow when...
Tim Kientzle [Mon, 27 Jul 2015 00:09:22 +0000 (17:09 -0700)] 
Issue #582: reject sparse blocks with negative size or offset, detect overflow when tracking sparse blocks

10 years agoAdd archive_read_next_header support to archive_read_disk traversals, in addition...
Tim Kientzle [Sun, 26 Jul 2015 23:40:11 +0000 (16:40 -0700)] 
Add archive_read_next_header support to archive_read_disk traversals, in addition to archive_read_next_header2