11 months agopo: Auto update
Werner Koch [Mon, 11 Jun 2018 08:15:44 +0000 (10:15 +0200)]
po: Auto update


11 months agogpg: Sanitize diagnostic with the original file name.
Werner Koch [Fri, 8 Jun 2018 08:45:21 +0000 (10:45 +0200)]
gpg: Sanitize diagnostic with the original file name.

* g10/mainproc.c (proc_plaintext): Sanitize verbose output.

This fixes a forgotten sanitation of user supplied data in a verbose
mode diagnostic.  The mention CVE is about using this to inject
status-fd lines into the stderr output.  Other harm good as well be
done.  Note that GPGME based applications are not affected because
GPGME does not fold status output into stderr.

CVE-id: CVE-2018-12020
GnuPG-bug-id: 4012
(cherry picked from commit 13f135c7a252cc46cff96e75968d92b6dc8dce1b)

13 months agog10: Push compress filter only if compressed.
NIIBE Yutaka [Fri, 13 Apr 2018 01:09:02 +0000 (10:09 +0900)]
g10: Push compress filter only if compressed.

* g10/compress.c (handle_compressed): Fix memory leak.


(backport from STABLE-BRANCH-2-2 commit:

All other calls of push_compress_filter checks ALGO,
so, do it here, too.

GnuPG-bug-id: 3898
Signed-off-by: NIIBE Yutaka <>
17 months agopo: Fix a fr string. Mark a string fuzzy in ro and sk.
Werner Koch [Tue, 19 Dec 2017 11:47:23 +0000 (12:47 +0100)]
po: Fix a fr string.  Mark a string fuzzy in ro and sk.


The French string has an extra %s which would result in garbage output
or segv.

I am not sure about the sk andro and thus better mark them as fuzzy.

GnuPG-bug-id: 3619
Signed-off-by: Werner Koch <>
17 months agopo: Update Japanese translation.
NIIBE Yutaka [Mon, 18 Dec 2017 05:09:53 +0000 (14:09 +0900)]
po: Update Japanese translation.

* po/ja.po: Fix message with no "%s".


Backport of master commit from: 77e2fcb4ffbad8577a2cf41f17bf92dec6a93ad8

The wrong message caused segmentation fault for key generation when
no expiration is specified.

GnuPG-bug-id: 3619
Signed-off-by: NIIBE Yutaka <>
17 months agog10: Fix regexp sanitization.
NIIBE Yutaka [Thu, 9 Nov 2017 05:03:22 +0000 (14:03 +0900)]
g10: Fix regexp sanitization.

* g10/trustdb.c (sanitize_regexp): Only escape operators.


Backport from master commit:

To sanitize a regular expression, quoting by backslash should be only
done for defined characters.  POSIX defines 12 characters including
dot and backslash.

Quoting other characters is wrong, in two ways; It may build an
operator like: \b, \s, \w when using GNU library.  Case ignored match
doesn't work, because quoting lower letter means literally and no
much to upper letter.

GnuPG-bug-id: 2923
Co-authored-by: Damien Goutte-Gattat <>
Signed-off-by: NIIBE Yutaka <>
18 months agopo/da: Fix Danish confusion between "compressed" and "compromised"
Daniel Kahn Gillmor [Sat, 11 Nov 2017 10:53:29 +0000 (18:53 +0800)]
po/da: Fix Danish confusion between "compressed" and "compromised"

In , Jonas Smedegaard reports:

> In option number 1, the word "komprimeret" means "compressed".
> I am pretty sure it should say "kompromitteret" instead, which means
> "compromised".

Debian-Bug-Id: 881393
Signed-off-by: Daniel Kahn Gillmor <>
18 months agoDo not use C99 feature.
Dario Niedermann [Fri, 10 Nov 2017 14:45:59 +0000 (15:45 +0100)]
Do not use C99 feature.

* cipher/rsa.c (secret): Move var decl to the beginning.

Trivial patch; ChangeLog written by wk.

Signed-off-by: Werner Koch <>
18 months agobuild: Don't use /dev/srandom on OpenBSD
Jeremie Courreges-Anglas [Wed, 1 Nov 2017 16:13:53 +0000 (17:13 +0100)]
build: Don't use /dev/srandom on OpenBSD

All /dev/*random devices have been equivalent since OpenBSD 4.9, on
purpose (/dev/random doesn't block).  /dev/srandom has been removed in
the OpenBSD 6.3 development cycle, /dev/arandom will likely follow.

Signed-off-by: Jeremie Courreges-Anglas <>
20 months agopo: update Dutch translation.
Frans Spiesschaert [Wed, 6 Sep 2017 20:28:29 +0000 (16:28 -0400)]
po: update Dutch translation.

Signed-off-by: Daniel Kahn Gillmor <>
21 months agodoc: Remove documentation for future option --faked-system-time.
Marcus Brinkmann [Fri, 4 Aug 2017 18:24:27 +0000 (20:24 +0200)]
doc: Remove documentation for future option --faked-system-time.

doc/gpg.texi: Remove documentation for --faked-system-time.

Signed-off-by: Marcus Brinkmann <>
GnuPG-bug-id: 3329

21 months agodebian: Remove packaging from upstream repository.
Daniel Kahn Gillmor [Wed, 2 Aug 2017 12:37:09 +0000 (08:37 -0400)]
debian: Remove packaging from upstream repository.

Debian packaging for GnuPG is handled in debian git repositories, and
doesn't belong here in the upstream repository.  The packaging was
significantly out of date anyway.

If you're looking for debian packaging for the 1.4 branch of GnuPG,
please use the following git remote:

Signed-off-by: Daniel Kahn Gillmor <>
21 months agopo: Update Danish translation
Joe Hansen [Thu, 28 May 2015 05:50:54 +0000 (01:50 -0400)]
po: Update Danish translation

Originally reported at:

Signed-off-by: Daniel Kahn Gillmor <>
21 months agopo: Update Dutch translation
Frans Spiesschaert [Wed, 2 Aug 2017 06:07:48 +0000 (02:07 -0400)]
po: Update Dutch translation

Debian-Bug-Id: 845695
Signed-off-by: Daniel Kahn Gillmor <>
21 months agopo: Update Spanish translation
Manuel Venturi Porras Peralta [Tue, 1 Aug 2017 16:45:41 +0000 (12:45 -0400)]
po: Update Spanish translation

Debian-Bug-Id: 814541
Signed-off-by: Daniel Kahn Gillmor <>
22 months agoPost release updates
Werner Koch [Wed, 19 Jul 2017 09:37:01 +0000 (11:37 +0200)]
Post release updates


22 months agoRelease 1.4.22 gnupg-1.4.22
Werner Koch [Wed, 19 Jul 2017 09:10:22 +0000 (11:10 +0200)]
Release 1.4.22

22 months agopo: Update Norwegian translation
Åka Sikrom [Wed, 19 Jul 2017 09:00:54 +0000 (11:00 +0200)]
po: Update Norwegian translation

Signed-off-by: Werner Koch <>
22 months agobuild: Avoid check gpg --version during make distcheck.
Werner Koch [Wed, 19 Jul 2017 08:55:44 +0000 (10:55 +0200)]
build: Avoid check gpg --version during make distcheck.


Signed-off-by: Werner Koch <>
22 months agoindent: Fix indentation of an if block.
Werner Koch [Wed, 19 Jul 2017 08:12:00 +0000 (10:12 +0200)]
indent: Fix indentation of an if block.


22 months agogpg: Fix memory leak.
NIIBE Yutaka [Fri, 7 Jul 2017 12:53:12 +0000 (21:53 +0900)]
gpg: Fix memory leak.

* g10/textfilter.c (copy_clearsig_text): Free the buffer.


Signed-off-by: NIIBE Yutaka <>
(backport from master commit:

22 months agorsa: Reduce secmem pressure.
NIIBE Yutaka [Fri, 7 Jul 2017 12:51:42 +0000 (21:51 +0900)]
rsa: Reduce secmem pressure.

* cipher/rsa.c (secret): Don't keep secmem.

Signed-off-by: NIIBE Yutaka <>
22 months agorsa: Allow different build directory.
NIIBE Yutaka [Fri, 7 Jul 2017 12:20:56 +0000 (21:20 +0900)]
rsa: Allow different build directory.

* cipher/ (AM_CPPFLAGS): Add mpi dirs.
* cipher/rsa.c: Change include file.

Signed-off-by: NIIBE Yutaka <>
22 months agorsa: Add exponent blinding.
Marcus Brinkmann [Fri, 7 Jul 2017 12:03:10 +0000 (21:03 +0900)]
rsa: Add exponent blinding.

* cipher/rsa.c (secret_core_crt): Blind secret D with randomized
nonce R for mpi_powm computation.


Backport of libgcrypt 8725c99ffa41778f382ca97233183bcd687bb0ce.

Signed-off-by: Marcus Brinkmann <>
22 months agompi: Minor fix for mpi_pow.
NIIBE Yutaka [Fri, 7 Jul 2017 06:12:00 +0000 (15:12 +0900)]
mpi: Minor fix for mpi_pow.

* mpi/mpi-pow.c (mpi_powm): Fix allocation size.

Signed-off-by: NIIBE Yutaka <>
22 months agompi: Same computation for square and multiply for mpi_pow.
NIIBE Yutaka [Fri, 7 Jul 2017 05:38:19 +0000 (14:38 +0900)]
mpi: Same computation for square and multiply for mpi_pow.

* mpi/mpi-pow.c (_gcry_mpi_powm): Compare msize for max_u_size.  Move
the assignment to base_u into the loop.  Copy content refered by RP to
BASE_U except the last of the loop.


Signed-off-by: NIIBE Yutaka <>
(backport commit of libgcrypt master:

22 months agompi: Simplify mpi_powm.
NIIBE Yutaka [Fri, 7 Jul 2017 05:26:39 +0000 (14:26 +0900)]
mpi: Simplify mpi_powm.

* mpi/mpi-pow.c (_gcry_mpi_powm): Simplify the loop.


(backport of libgcrypt master commit:

Signed-off-by: NIIBE Yutaka <>
22 months agompi: Fix ARM assembler in longlong.h.
Marcus Brinkmann [Tue, 4 Jul 2017 21:19:51 +0000 (23:19 +0200)]
mpi: Fix ARM assembler in longlong.h.

* mpi/longlong.h [__arm__] (add_ssaaaa, sub_ddmmss): Add __CLOBBER_CC.
[__arm__][__ARM_ARCH <= 3] (umul_ppmm): Add __AND_CLOBBER_CC.

This is a backport of libgcrypt 8aa4f2161 and 3b1cc9e6c.

Signed-off-by: Marcus Brinkmann <>
GnuPG-bug-id: 3182

22 months agodoc: Fix typo.
Marcus Brinkmann [Mon, 3 Jul 2017 09:02:42 +0000 (11:02 +0200)]
doc: Fix typo.

Signed-off-by: Marcus Brinkmann <>
GnuPG-bug-id: 3243

2 years agog10: Fix secmem leak.
Ineiev [Thu, 14 Apr 2016 15:42:25 +0000 (15:42 +0000)]
g10: Fix secmem leak.

* g10/keygen.c (proc_parameter_file): Fix secmem leak.


proc_parameter_file adds certain parameters to the list in the PARA
argument; however, these new entries are leaked because they
are added to head, while the PARA list is released by the caller
of proc_parameter_file.

GnuPG-bug-id: 1371
Signed-off-by: Ineiev <>
2 years agogpg: Fix exporting of zero length user ID packets.
Werner Koch [Thu, 30 Mar 2017 08:35:20 +0000 (10:35 +0200)]
gpg: Fix exporting of zero length user ID packets.

* g10/build-packet.c (do_user_id): Avoid indeterminate length header.

We are able to import such user ids but when exporting them the
exported data could not be imported again because the parser bails out
on invalid keyrings.  This is now fixed and should be backported.

Note that in 1.4 and 2.0 this is only an issue for attribute packets.
In 2.1 user IDs were also affected.a

Signed-off-by: Werner Koch <>
2 years agospelling: Correct achived to achieved.
Daniel Kahn Gillmor [Mon, 29 Aug 2016 14:36:33 +0000 (10:36 -0400)]
spelling: Correct achived to achieved.

Signed-off-by: Daniel Kahn Gillmor <>
2 years agotools: Fix option parsing for gpg-zip.
Neal H. Walfield [Mon, 29 Aug 2016 14:56:16 +0000 (10:56 -0400)]
tools: Fix option parsing for gpg-zip.

* tools/ Correctly set GPG when --gpg is specified.
Correctly set TAR when --tar is specified.  Pass TAR_ARGS to tar.

(cherry-picked by dkg from master branch's

Signed-off-by: Neal H. Walfield <>
Co-authored-by: Michael Mönch <>
GnuPG-bug-id 1351
GnuPG-bug-id 1442

2 years agopo: Update Norwegian translation
Åka Sikrom [Thu, 18 Aug 2016 14:46:37 +0000 (16:46 +0200)]
po: Update Norwegian translation


Minor change: Re-insert the old copyright year 2004. - wk

2 years agoPost release updates
Werner Koch [Wed, 17 Aug 2016 13:29:13 +0000 (15:29 +0200)]
Post release updates


2 years agoRelease 1.4.21 gnupg-1.4.21
Werner Koch [Wed, 17 Aug 2016 13:07:57 +0000 (15:07 +0200)]
Release 1.4.21

2 years agogpg: Add dummy option --with-subkey-fingerprint.
Werner Koch [Wed, 17 Aug 2016 12:50:35 +0000 (14:50 +0200)]
gpg: Add dummy option --with-subkey-fingerprint.

* g10/gpg.c (opts): Add dummy option.

Signed-off-by: Werner Koch <>
2 years agopo: Auto update
Werner Koch [Wed, 17 Aug 2016 12:41:16 +0000 (14:41 +0200)]
po: Auto update


2 years agobuild: Create a swdb file during "make distcheck".
Werner Koch [Wed, 17 Aug 2016 12:39:26 +0000 (14:39 +0200)]
build: Create a swdb file during "make distcheck".

* (distcheck-hook): New.

Signed-off-by: Werner Koch <>
2 years agobuild: Update config.{guess,sub} to {2016-05-15,2016-06-20}.
Werner Koch [Wed, 17 Aug 2016 12:36:25 +0000 (14:36 +0200)]
build: Update config.{guess,sub} to {2016-05-15,2016-06-20}.


2 years agopo: Update Russian translation
Ineiev [Wed, 17 Aug 2016 12:31:12 +0000 (14:31 +0200)]
po: Update Russian translation

2 years agorandom: Hash continuous areas in the csprng pool.
Werner Koch [Wed, 17 Aug 2016 09:15:50 +0000 (11:15 +0200)]
random: Hash continuous areas in the csprng pool.

* cipher/random.c (mix_pool): Store the first hash at the end of the

This fixes a long standing bug (since 1998) in Libgcrypt and GnuPG.
An attacker who obtains 580 bytes of the random number from the
standard RNG can trivially predict the next 20 bytes of output.

This bug does not affect the default generation of
keys because running gpg for key creation creates at most 2 keys from
the pool: For a single 4096 bit RSA key 512 byte of random are
required and thus for the second key (encryption subkey), 20 bytes
could be predicted from the the first key.  However, the security of
an OpenPGP key depends on the primary key (which was generated first)
and thus the 20 predictable bytes should not be a problem.  For the
default key length of 2048 bit nothing will be predictable.

For the former default of DSA+Elgamal key it is complicate to give an
answer: For 2048 bit keys a pool of 30 non-secret candidate primes of
about 300 bits each are first created.  This reads at least 1140 bytes
from the pool and thus parts could be predicted.  At some point a 256
bit secret is read from the pool; which in the worst case might be
partly predictable.

The bug was found and reported by Felix Dörre and Vladimir Klebanov,
Karlsruhe Institute of Technology.  A paper describing the problem in
detail will shortly be published.

CVE-id: CVE-2016-6313
Signed-off-by: Werner Koch <>
2 years agocipher: Improve readability by using a macro.
Werner Koch [Wed, 17 Aug 2016 09:03:07 +0000 (11:03 +0200)]
cipher: Improve readability by using a macro.

* cipher/random.c (mix_pool): Use DIGESTLEN instead of 20.

Signed-off-by: Werner Koch <>
2 years agogpg: Avoid publishing the GnuPG version by default
Daniel Kahn Gillmor [Fri, 5 Aug 2016 14:46:43 +0000 (10:46 -0400)]
gpg: Avoid publishing the GnuPG version by default

* g10/gpg.c (main): initialize opt.emit_version to 0
* doc/gpg.texi: document different default for --emit-version


The version of GnuPG in use is not particularly helpful.  It is not
cryptographically verifiable, and it doesn't distinguish between
significant version differences like 2.0.x and 2.1.x.

Additionally, it leaks metadata that can be used to distinguish users
from one another, and can potentially be used to target specific
attacks if there are known behaviors that differ between major

It's probably better to take the more parsimonious approach to
metadata production by default.

(backport of master commit c9387e41db7520d176edd3d6613b85875bdeb32c)

Signed-off-by: Daniel Kahn Gillmor <>
2 years agoClean up "allow to"
Daniel Kahn Gillmor [Wed, 3 Aug 2016 15:56:56 +0000 (11:56 -0400)]
Clean up "allow to"

* README, cipher/cipher.c, cipher/pubkey.c, doc/gpg.texi: replace
  "allow to" with clearer text

In standard English, the normal construction is "${XXX} allows ${YYY}
to" -- that is, the subject (${XXX}) of the sentence is allowing the
object (${YYY}) to do something.  When the object is missing, the
phrasing sounds awkward, even if the object is implied by context.
There's almost always a better construction that isn't as awkward.

These changes should make the language a bit clearer.

Signed-off-by: Daniel Kahn Gillmor <>
2 years agoFix spelling: "occured" should be "occurred"
Daniel Kahn Gillmor [Wed, 3 Aug 2016 15:56:55 +0000 (11:56 -0400)]
Fix spelling: "occured" should be "occurred"

* checks/armor.test, cipher/des.c, g10/ccid-driver.c, g10/pkclist.c,
  util/regcomp.c, util/regex_internal.c: correct the spelling of
  "occured" to "occurred"

Signed-off-by: Daniel Kahn Gillmor <>
2 years agog10: Fix checking key for signature validation.
NIIBE Yutaka [Thu, 4 Aug 2016 07:21:39 +0000 (16:21 +0900)]
g10: Fix checking key for signature validation.

* g10/sig-check.c (signature_check2): Not only subkey, but also primary
key should have flags.valid=1.


(backport of master
commit 6f284e6ed63f514b15fe610f490ffcefc87a2164)

Signed-off-by: NIIBE Yutaka <>
2 years agoPartially revert "g10: Fix another race condition for trustdb access."
Justus Winter [Wed, 3 Aug 2016 15:17:50 +0000 (17:17 +0200)]
Partially revert "g10: Fix another race condition for trustdb access."

This amends db246f8b which accidentally included the compiled
translation files.

Signed-off-by: Justus Winter <>
2 years agogpgv: Tweak default options for extra security.
NIIBE Yutaka [Sat, 9 Jul 2016 01:20:02 +0000 (10:20 +0900)]
gpgv: Tweak default options for extra security.

* g10/gpgv.c (main): Set opt.no_sig _cache, so that it doesn't depend on
cached status.  Similarly, set opt.flags.require_cross_cert for backsig
validation for subkey signature.


(backport of master
commit e32c575e0f3704e7563048eea6d26844bdfc494b)

It is common that an organization distributes binary keyrings with
signature cache (Tag 12, Trust Packet) and people use gpgv to validate
signature with such keyrings.  In such a use case, it is possible that
the key validation itself is skipped.

For the purpose of gpgv validation of signatures, we should not depend
on signature cache in keyrings (if any), but we should validate the key
by its self signature for primary key, and back signature for subkey.

Signed-off-by: NIIBE Yutaka <>
2 years agog10: Fix keysize with --expert.
NIIBE Yutaka [Wed, 6 Jul 2016 02:45:05 +0000 (11:45 +0900)]
g10: Fix keysize with --expert.

* g10/keygen.c (ask_keysize): It's 768 only for DSA.


GnuPG-bug-id: 2238
Signed-off-by: NIIBE Yutaka <>
2 years agog10: Fix --list-packets.
NIIBE Yutaka [Tue, 28 Jun 2016 06:56:48 +0000 (15:56 +0900)]
g10: Fix --list-packets.

* g10/gpg.c (main): Call set_packet_list_mode after assignment of
* g10/mainproc.c (do_proc_packets): Don't stop processing with
--list-packets as the comment says.
* g10/options.h (list_packets): Fix the comment.
* g10/parse-packet.c: Fix the condition for opt.list_packets.


(backport from 2.0 commit 4f336ed780cc2783395f3ff2b12b3ebb8e097f7b
which is backport of master
commit 52f65281f9743c42a48bf5a3354c9ab0ecdb681a)

Debian-bug-id: 828109
Signed-off-by: NIIBE Yutaka <>
2 years agog10: Fix another race condition for trustdb access.
Niibe Yutaka [Tue, 14 Jun 2016 23:41:56 +0000 (08:41 +0900)]
g10: Fix another race condition for trustdb access.

* g10/tdbio.c (create_version_record): Call create_hashtable to always
make hashtable, together with the version record.
(get_trusthashrec): Remove call to create_hashtable.


GnuPG-bug-id: 1675
Thanks to Scott Moser to reproducible script and patience.

Signed-off-by: NIIBE Yutaka <>
(backport from master
 commit 35a3ce2acf78a95fecbccfd8db0560cca24232df)

2 years agodoc: Remove non-implemented option --skip-hidden-recipients.
Werner Koch [Wed, 8 Jun 2016 16:03:43 +0000 (18:03 +0200)]
doc: Remove non-implemented option --skip-hidden-recipients.


GnuPG-bug-id: 1394

Note that --try-secret-key was already removed with commit

Signed-off-by: Werner Koch <>
3 years agodoc: Explain that gpg-preset-passphrase can't be used.
Werner Koch [Fri, 22 Apr 2016 07:57:38 +0000 (09:57 +0200)]
doc: Explain that gpg-preset-passphrase can't be used.


3 years agopo: Fix misleading german translation.
Justus Winter [Fri, 1 Apr 2016 14:38:24 +0000 (16:38 +0200)]
po: Fix misleading german translation.

GnuPG-bug-id: 2239
Signed-off-by: Justus Winter <>
3 years agog10: Make sure to have the directory for trustdb.
NIIBE Yutaka [Fri, 12 Feb 2016 01:00:31 +0000 (10:00 +0900)]
g10: Make sure to have the directory for trustdb.

* g10/tdbio.c (tdbio_set_dbname): Return earlier if !CREATE.  Check
the directory and create it if none before calling take_write_lock.


Thanks to Marc Deslauriers for the bug report and his patch.

GnuPG-bug-id: 2246

Signed-off-by: NIIBE Yutaka <>
(backport from master
 commit 2f3e42047d17313eeb38d354048f343158402a8d)

3 years agoFix possible sign extension problem with newer compilers.
Werner Koch [Mon, 1 Feb 2016 17:06:14 +0000 (18:06 +0100)]
Fix possible sign extension problem with newer compilers.

* cipher/des.c (READ_64BIT_DATA): Cast to u32 before shifting by 24.
* cipher/blowfish.c (do_encrypt_block): Ditto.
(do_decrypt_block): Ditto.
* cipher/camellia.c (CAMELLIA_RR8): Ditto.
* cipher/cast5.c (do_encrypt_block): Ditto.
(do_decrypt_block): Ditto.
(do_cast_setkey): Ditto.
* cipher/twofish.c (INPACK): Ditto.
* util/iobuf.c (block_filter): Ditto.

For cipher/des.c
Reported-by: Balint Reczey <>
See commit 57af33d9e7c9b20b413b96882e670e75a67a5e65 for details.

Signed-off-by: Werner Koch <>
3 years agog10: Fix iobuf API of filter function for alignment.
NIIBE Yutaka [Tue, 26 Jan 2016 06:38:27 +0000 (15:38 +0900)]
g10: Fix iobuf API of filter function for alignment.

* include/iobuf.h (struct iobuf_struct): Remove DESC.
* util/iobuf.c (iobuf_desc): New.
(print_chain, iobuf_close, iobuf_open, iobuf_fdopen, iobuf_sockopen)
(iobuf_create, iobuf_append, iobuf_openrw, iobuf_ioctl)
(iobuf_push_filter2, pop_filter, underflow): Use iobuf_desc.
(file_filter, sock_filter, block_filter): Fill the description.
* g10/armor.c, g10/cipher.c, g10/compress-bz2.c, g10/compress.c,
g10/encode.c, g10/encr-data.c, g10/mdfilter.c, g10/pipemode.c,
g10/progress.c, g10/textfilter.c: Likewise.


Newer GCC warns against possible alignment difference of pointers.
This change can silence those warnings.

Signed-off-by: NIIBE Yutaka <>
(backported from 2.1 commit 3f52c7da3940ec06572270d511000dc7fe9c27d2)

3 years agoFix possible AIX problem with sysconf in rndunix.
Werner Koch [Fri, 15 Jan 2016 14:51:37 +0000 (15:51 +0100)]
Fix possible AIX problem with sysconf in rndunix.

* cipher/rndunix.c [HAVE_STDINT_H]: Include stdint.h.
(start_gatherer): Detect misbehaving sysconf.

GnuPG-bug-id: 1778
for the reason of this patch. There is no concrete bug report but this
chnage should not harm.

Signed-off-by: Werner Koch <>
3 years agoFix to support git worktree.
NIIBE Yutaka [Wed, 13 Jan 2016 08:22:37 +0000 (17:22 +0900)]
Fix to support git worktree.

* Use -e for testing .git.

Signed-off-by: NIIBE Yutaka <>
(backport commit of 96237b9a63a50aed1884cb06f84279b977d6a8fa)

3 years agopo: Update Japanese translation.
NIIBE Yutaka [Mon, 21 Dec 2015 12:27:53 +0000 (21:27 +0900)]
po: Update Japanese translation.

3 years agoPost release updates.
Werner Koch [Sun, 20 Dec 2015 08:01:18 +0000 (09:01 +0100)]
Post release updates.


3 years agoRelease 1.4.20 gnupg-1.4.20
Werner Koch [Sat, 19 Dec 2015 14:42:22 +0000 (15:42 +0100)]
Release 1.4.20

3 years agow32: Avoid warning when using newer mingw versions.
Werner Koch [Sat, 19 Dec 2015 15:54:57 +0000 (16:54 +0100)]
w32: Avoid warning when using newer mingw versions.

* g10/tdbio.c (ftruncate): Do not define if already defined.

Signed-off-by: Werner Koch <>
3 years agopo: Auto-update
Werner Koch [Sat, 19 Dec 2015 14:39:45 +0000 (15:39 +0100)]
po: Auto-update


3 years agopo: Update German translation
Werner Koch [Sat, 19 Dec 2015 15:53:45 +0000 (16:53 +0100)]
po: Update German translation


Signed-off-by: Werner Koch <>
3 years agopo: Update Russian translation
Ineiev [Sat, 19 Dec 2015 14:32:58 +0000 (15:32 +0100)]
po: Update Russian translation


Signed-off-by: Werner Koch <>
3 years agogpg: Add option --weak-digest to gpg and gpgv.
Daniel Kahn Gillmor [Tue, 27 Oct 2015 04:01:32 +0000 (00:01 -0400)]
gpg: Add option --weak-digest to gpg and gpgv.

* g10/options.h: Add weak_digests linked list to opts.
* g10/main.h: Declare weakhash linked list struct and
additional_weak_digest() function to insert newly-declared weak
digests into opts.
* g10/misc.c: (additional_weak_digest): New function.
(print_digest_algo_note): Check for deprecated digests.
* g10/sig-check.c: (do_check): Reject all weak digests.
* g10/gpg.c: Add --weak-digest option to gpg.
* doc/gpg.texi: Document gpg --weak-digest option.
* g10/gpgv.c: Add --weak-digest option to gpgv.
* doc/gpgv.texi: Document gpgv --weak-digest option.

gpg and gpgv treat signatures made over MD5 as unreliable, unless the
user supplies --allow-weak-digests to gpg.  Signatures over any other
digest are considered acceptable.

Despite SHA-1 being a mandatory-to-implement digest algorithm in RFC
4880, the collision-resistance of SHA-1 is weaker than anyone would
like it to be.

Some operators of high-value targets that depend on OpenPGP signatures
may wish to require their signers to use a stronger digest algorithm
than SHA1, even if the OpenPGP ecosystem at large cannot deprecate
SHA1 entirely today.

This changeset adds a new "--weak-digest DIGEST" option for both gpg
and gpgv, which makes it straightforward for anyone to treat any
signature or certification made over the specified digest as

This option can be supplied multiple times if the operator wishes to
deprecate multiple digest algorithms, and will be ignored completely
if the operator supplies --allow-weak-digests (as before).

MD5 is always considered weak, regardless of any further
--weak-digest options supplied.

Signed-off-by: Daniel Kahn Gillmor <>
(this is a rough cherry-pick of applying the following commits to

3 years agogpg: Reject signatures made with MD5.
Werner Koch [Tue, 27 Oct 2015 04:01:31 +0000 (00:01 -0400)]
gpg: Reject signatures made with MD5.

* g10/gpg.c: Add option --allow-weak-digest-algos.
(main): Set option also in PGP2 mode.
* g10/options.h (struct opt): Add flags.allow_weak_digest_algos.
* g10/sig-check.c (do_check): Reject MD5 signatures.
* tests/openpgp/gpg.conf.tmpl: Add allow_weak_digest_algos.
(cherry picked from commit f90cfe6b66269de0154d810c5cee1fe9a5af475c)

Resolved conflicts:
g10/gpg.c - adjust.
tests/openpgp/ - no changes

3 years agogpg: Change default cipher for --symmetric from CAST5 to AES-128.
Werner Koch [Mon, 18 Aug 2014 09:45:00 +0000 (11:45 +0200)]
gpg: Change default cipher for --symmetric from CAST5 to AES-128.

* g10/main.h (DEFAULT_CIPHER_ALGO): Change to AES or CAST5 or 3DES
depending on configure options.
* g10/gpg.c (main): Set opt.s2k_cipher_algo to DEFAULT_CIPHER_ALGO.


(cherry picked from commit 57df1121c18b004dd763b35eabf7b51fc9e8ec38)
Signed-off-by: Werner Koch <>
3 years agoPass DBUS_SESSION_BUS_ADDRESS for gnome3
Daniel Kahn Gillmor [Tue, 18 Aug 2015 11:34:57 +0000 (13:34 +0200)]

* g10/passphrase.c (stdenvnames): Add DBUS_SESSION_BUS_ADDRESS.

pinentry-gnome3 talks to the gcr prompter via dbus.  Without this
environment variable, it can't find the correct session to talk to.

3 years agogpg: Avoid cluttering stdout with trustdb info in verbose mode.
Werner Koch [Fri, 8 May 2015 14:30:04 +0000 (16:30 +0200)]
gpg: Avoid cluttering stdout with trustdb info in verbose mode.

* g10/trustdb.c (validate_keys): Call dump_key_array only in debug

I guess that is a left-over from an early attempt to output
information on the trustdb for use by other tools.  Maybe related to
the former --list-trust-path command.  Sending it to stdout is
probably useful so we do this now only in debug mode.

Signed-off-by: Werner Koch <>
Backported to STABLE-BRANCH-1-4 from
b03a2647299a6c8764a2574590cbaccdff9e497d by dkg

3 years agogpg: Silence a compiler warning.
Werner Koch [Thu, 18 Sep 2014 13:08:51 +0000 (15:08 +0200)]
gpg: Silence a compiler warning.

* g10/parse-packet.c (enum_sig_subpkt): Replace hack.


GCC 5 failure reported by Kevin Locke <>

(backport from master commit 6a0c3fa19cfcdd590b96691e8a8ffb48fb5e0ec4)

3 years agopo: Update Japanese translation.
NIIBE Yutaka [Thu, 17 Sep 2015 08:05:37 +0000 (17:05 +0900)]
po: Update Japanese translation.

3 years agopo: Fix Spanish translation
NIIBE Yutaka [Tue, 8 Sep 2015 00:14:20 +0000 (09:14 +0900)]
po: Fix Spanish translation

3 years agoObsolete option --no-sig-create-check.
Werner Koch [Mon, 31 Aug 2015 21:55:16 +0000 (23:55 +0200)]
Obsolete option --no-sig-create-check.

* cipher/rsa.c (rsa_sign): Verify after sign.
* g10/gpg.c (opts): Make --no-sig-create-check a NOP.
* g10/options.h (opt): Remove field "no_sig_create_check".
* g10/sign.c (do_sign): Do check only for DSA.

Signed-off-by: Werner Koch <>
3 years agog10: Fix a race condition initially creating trustdb.
NIIBE Yutaka [Tue, 16 Jun 2015 03:01:28 +0000 (12:01 +0900)]
g10: Fix a race condition initially creating trustdb.

* g10/tdbio.c (take_write_lock, tdbio_set_dbname): Fix message.

3 years agog10: Fix a race condition initially creating trustdb.
NIIBE Yutaka [Mon, 15 Jun 2015 05:38:05 +0000 (14:38 +0900)]
g10: Fix a race condition initially creating trustdb.

* g10/tdbio.c (take_write_lock, release_write_lock): New.
(put_record_into_cache, tdbio_sync, tdbio_end_transaction): Use
new lock functions.
(tdbio_set_dbname): Fix the race.
(open_db): Don't call dotlock_create.


(backported from 2.1 commit fe5c6edaed78839303d67e01e141cfc6b5de9aec)
GnuPG-bug-id: 1675

4 years agog10: detects public key encryption packet error properly.
NIIBE Yutaka [Tue, 19 May 2015 01:47:42 +0000 (10:47 +0900)]
g10: detects public key encryption packet error properly.

g10/mainproc.c (proc_pubkey_enc): Only allow relevant algorithms for


(backported from 2.1 commit c771963140cad7c1c25349bcde27e427effc0058)

4 years agog10: Improve handling of no corresponding public key.
NIIBE Yutaka [Tue, 19 May 2015 01:14:09 +0000 (10:14 +0900)]
g10: Improve handling of no corresponding public key.

* g10/getkey.c (get_seckey): Return G10ERR_NO_PUBKEY when it's not
exact match.


In the situation of corrupted .gnupg/ where only private subkey is
available but no corresponding public key of the subkey, the code
returned public primary key which caused mysterious error (for a
user).  This fix detects an error earlier.

GnuPG-bug-id: 1422
Debian-Bug-Id: #638619

4 years agog10: fix cmp_public_key and cmp_secret_keys.
NIIBE Yutaka [Thu, 30 Apr 2015 08:20:08 +0000 (17:20 +0900)]
g10: fix cmp_public_key and cmp_secret_keys.

* g10/free-packet.c (cmp_public_keys, cmp_secret_keys): Compare opaque
data at the first entry of the array when it's unknown algo.
* mpi/mpi-cmp.c (mpi_cmp): Backport libgcrypt 1.5.0's semantics.


(backported from 2.0 commit 43429c7869152f301157e4b24790b3801dce0f0a)

GnuPG-bug-id: 1962

4 years agogpg: Fix DoS while parsing mangled secret key packets.
Werner Koch [Sun, 5 Apr 2015 17:42:59 +0000 (19:42 +0200)]
gpg: Fix DoS while parsing mangled secret key packets.

* g10/parse-packet.c (parse_key): Check PKTLEN before calling mpi_read
et al.

Due to the missing length checks PKTLEN may turn negative.  Because
PKTLEN is an unsigned int the malloc in read_rest would try to malloc
a too large number and terminate the process with "error reading rest
of packet: Cannot allocate memory".

Reported-by: Hanno Böck.
Signed-off-by: Werner Koch <>
(backported from 2.0 commit 0aac920f23fd07e152fdb7385299c92bb9a4ade3)

4 years agogpg: Remove left-over debug message.
Werner Koch [Thu, 26 Feb 2015 08:38:58 +0000 (09:38 +0100)]
gpg: Remove left-over debug message.

* g10/armor.c (check_input): Remove log_debug.

4 years agoPost release updates
Werner Koch [Fri, 27 Feb 2015 09:01:08 +0000 (10:01 +0100)]
Post release updates


4 years agoRelease 1.4.19. gnupg-1.4.19
Werner Koch [Fri, 27 Feb 2015 08:33:58 +0000 (09:33 +0100)]
Release 1.4.19.

4 years agopo: Auto-update
Werner Koch [Fri, 27 Feb 2015 08:06:46 +0000 (09:06 +0100)]
po: Auto-update


4 years agopo: Update German translation
Werner Koch [Fri, 27 Feb 2015 08:05:21 +0000 (09:05 +0100)]
po: Update German translation

4 years agopo: Update French translation
David Prévot [Mon, 3 Nov 2014 03:24:29 +0000 (23:24 -0400)]
po: Update French translation

4 years agopo: Update Czech translation
Roman Pavlik [Tue, 25 Nov 2014 20:22:30 +0000 (16:22 -0400)]
po: Update Czech translation

4 years agopo: Update Dutch translation
Frans Spiesschaert [Mon, 24 Nov 2014 20:28:08 +0000 (16:28 -0400)]
po: Update Dutch translation

4 years agopo: Update Spanish translation
Manuel \"Venturi\" Porras Peralta [Sat, 22 Nov 2014 19:42:36 +0000 (15:42 -0400)]
po: Update Spanish translation

4 years agopo: Update Polish translation
Jakub Bogusz [Fri, 21 Nov 2014 21:06:59 +0000 (17:06 -0400)]
po: Update Polish translation

4 years agopo: Update Russian translation
Ineiev [Thu, 26 Feb 2015 20:06:38 +0000 (21:06 +0100)]
po: Update Russian translation

4 years agopo: Update Ukrainian translation
Yuri Chornoivan [Thu, 26 Feb 2015 20:04:57 +0000 (21:04 +0100)]
po: Update Ukrainian translation

4 years agopo: Update Italian translation.
Milo Casagrande [Thu, 26 Feb 2015 20:02:57 +0000 (21:02 +0100)]
po: Update Italian translation.

4 years agoUpdate Chinese (traditional) translation
Jedi Lin [Fri, 21 Nov 2014 21:42:44 +0000 (17:42 -0400)]
Update Chinese (traditional) translation

4 years agoFix for building without DNS support.
Werner Koch [Thu, 26 Feb 2015 19:00:17 +0000 (20:00 +0100)]
Fix for building without DNS support.

* util/cert.c (get_cert) [!USE_DNS_CERT]: Add want_ipgp.

4 years agopo,intl: Update to 0.19.3.
Werner Koch [Thu, 26 Feb 2015 18:01:30 +0000 (19:01 +0100)]
po,intl: Update to 0.19.3.

4 years agoSwitch to a hash and CERT record based PKA system.
Werner Koch [Thu, 26 Feb 2015 15:34:44 +0000 (16:34 +0100)]
Switch to a hash and CERT record based PKA system.

* util/pka.c: Rewrite.
(get_pka_info): Add arg fprbuflen.  Change callers to pass this.
* util/strgutil.c (ascii_strlwr): New.
* Remove option --disable-dns-pka.
(USE_DNS_PKA): Remove ac_define.
* g10/getkey.c (parse_auto_key_locate): Always include PKA.

Note that although PKA is now always build, it will only work if
support for looking up via DNS has not been disabled.

The new PKA only works with the IPGP DNS certtype and shall be used
only to retrieve the fingerprint and optional the key for the first
time.  Due to the security problems with DNSSEC the former assumption
to validate the key using DNSSEC is not anymore justified.  Instead an
additional layer (e.g. Trust-On-First-Use) needs to be implemented to
track change to the key.  Having a solid way of getting a key matching
a mail address is however a must have.

More work needs to go into a redefinition of the --verify-options
pka-lookups and pka-trust-increase.  The auto-key-locate mechanism
should also be able to continue key fetching with another method once
the fingerprint has been retrieved with PKA.

Signed-off-by: Werner Koch <>
This is a backport from master.
(backported from commit 2fc27c8696f5cf2ddf3212397ea49bff115d617b)

4 years agoMove two functions from g10/ to util/.
Werner Koch [Thu, 26 Feb 2015 15:08:02 +0000 (16:08 +0100)]
Move two functions from g10/ to util/.

* g10/misc.c (has_invalid_email_chars, is_valid_mailbox): Move to ...
* util/strgutil.c: here.

Signed-off-by: Werner Koch <>