* g10/mainproc.c (proc_encrypted): Force a decryption failure if any
error has been seen.
* g10/decrypt-data.c (aead_checktag): Issue an ERROR line.
--
GnuPG-bug-id: 7042
Note that gpg in any case returns a failure exit code but due to
double forking GPGME would not see it.
* tools/gpg-check-pattern.c (read_file): Check length before calling
fread.
--
The problem with an empty file is that es_fread is called to read one
element of length zero which seems to be undefined behaviour and
results in ENOENT on my test box.
* tools/gpg-wks-client.c (mirror_one_key): Test for no domain
specified.
--
The code did not really work if no domain was given. It worked but
filtered out all keys so that no key was actually exported.
* g10/trustdb.c: Move some definitions around.
(user_utk_list): Rename to trusted_key_list. Change all users.
(any_trusted_key_seen): New.
(tdb_register_trusted_key): Set it here. Handle the new value "none".
(verify_own_keys): Do not delete a trusted key from the trustdb if a
trusted-key option was not used.
--
GnuPG-bug-id: 7025
* dirmngr/http.c (run_proxy_connect): Don't set keep_alive, since it
causes resource leak of FP_WRITE.
Don't try to read response body to fix the hang.
--
GnuPG-bug-id: 6997
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* dirmngr/http.c (proxy_get_token, run_proxy_connect): Always
available regardless of USE_TLS.
(send_request): Remove USE_TLS.
--
Since quite some time building w/o TLS won't work.
GnuPG-bug-id: 6997
* configure.ac: Require libgpg-error 1.38 and libksba 1.4.0
* common/util.h: Remove error number substitutes.
--
These versions were already needed for a successful compilation,
but it was not reflected by configure.
GnuPG-bug-id: 6974
Signed-off-by: Ángel González <angel@pgp.16bits.net>
util.h patch by wk.
* scd/app-openpgp.c (do_change_pin): Fix PIN length check. Add "R"
flag to the reset code prompt.
--
When using the reset code it was not possible to set a PIN of length
6. The "R" flags fixes a funny prompt.
Fixes-commit: efe325ffdf
* scd/app-openpgp.c (do_change_pin): Make sure new PIN length
is longer than MINLEN.
--
GnuPG-bug-id: 6843
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
(cherry picked from commit 2376cdff13)
* scd/app-openpgp.c (wipe_and_free_string, wipe_and_free): Enable
functions.
(pin2hash_if_kdf): Change interface. The input PIN is not anymore
changed. Further there are no more assumptions about the length of
the provided buffer.
(verify_a_chv): Adjust for changed pin2hash_if_kdf.
(verify_chv2): Ditto
(verify_chv3): Ditto.
(do_change_pin): Ditto.
(do_sign): Ditto.
--
Note that this a part of the patch
63bda3aad8 which we used in 2.4 to
implement a PIN cache. For easier backporting we need to add this
here.
* agent/command.c (cmd_learn): Add option --reallyforce.
* agent/findkey.c (agent_write_private_key): Implement reallyforce.
Also add arg reallyforce and pass it along the call chain.
* g10/call-agent.c (agent_scd_learn): Pass --reallyforce with a
special force value.
* g10/keygen.c (card_store_key_with_backup): Use that force value.
--
This was a regression in 2.2.42. We took the easy path to fix it by
getting the behaviour back to what we did prior to 2.2.42. With GnuPG
2.4.4 we use an entire different and safer approach by introducing an
ephemeral private key store.
GnuPG-bug-id: 6944
* g10/sign.c (do_sign): Skip compliance check for revocation certs.
--
It just does not make sense to inhibit the creation of revocations
depending on the compliance mode. We do this only for key revocation
but not for another kind of revocation because the rationale for uid
or subkey revocation is more complicated to explain.
* common/homedir.c (copy_dir_with_fixup) [W32]: Fold double
backslashes.
--
This is in general no problem but when we hash or compare the directory
to test whether tit is the standard home directory, we may use a
different socket file and thus a second instance of a daemon.
GnuPG-bug-id: 6833
* scd/app-openpgp.c (ecc_writekey): Remove the useless check and print
a diagnostic if the default params are used.
--
Note that here in 2.2 we use different default ECDH parameters than in
2.4 (AES192 instead of AES256 for 384 bit curves).
GnuPG-bug-id: 6378
* agent/findkey.c (read_key_file): Add optional arg r_orig_key_value
to return the old Key value. Change all callers.
(agent_write_private_key): Detect whether the Key entry was really
changed.
--
GnuPG-bug-id: 6829
* common/name-value.c (struct name_value_container): Add flag
"modified".
(nvc_modified): New.
(nvc_new): Set flag.
(_nvc_add): Set flag.
(nvc_delete): Set flag.
(nve_set): Add arg PK. Change the caller. Check whether to change at
all.
* agent/findkey.c (agent_write_private_key): Update only if modified.
--
This helps software which uses a file system watcher to track changes
to private keys. In particular smartcard triggered changes are a
problem for such software because this may at worst trigger another
smartcard read.
GnuPG-bug-id: 6829
* g10/call-agent.c (agent_get_passphrase): Call
assuan_begin_confidential and assuan_end_confidential.
* sm/call-agent.c (gpgsm_agent_ask_passphrase): Ditto.
--
GnuPG-bug-id: 6654
The drawback of this solution is that we don't see any IPC lines from
the assuan_transact. Everything else would require larger changes to
libassuan.
* sm/encrypt.c (gpgsm_encrypt): Unify use of RC and ERR.
* sm/sign.c (gpgsm_sign): ditto.
--
Initially we didn't used the gpg_error_t thingy and while migrating
we sometimes used RC and ERR for tracking the error. This is pretty
error prone and thus we better remove it (after 20 years).
* g10/call-agent.c (agent_keytocard): Add arg ecdh_param_str.
* g10/keyid.c (ecdh_param_str_from_pk): New.
* g10/card-util.c (card_store_subkey): Pass ECDH params to writekey.
* g10/keygen.c (card_store_key_with_backup): Ditto.
--
Backported from 2.4 - here the gpg part.
See-commit: c03ba92576
This is related to
GnuPG-bug-id: 6378
* agent/command.c (KEYTOCARD_TIMESTAMP_FORMAT): Remove and use format
string direct.
(cmd_keytocard): Change timestamp to an u64 and use the new u64 parser
functions. Use split_fields. Add ecdh parameter stuff. Take the
default timestamp from the keyfile.
* agent/findkey.c (agent_key_from_file): Add arg timestamp and set it.
Adjust all callers.
--
This is backport from 2.4 but only the agent part. We consider it
more relibale if we also pass the ECDH parameters along in 2.2.
Adding the timestamp stuff should not harm either. Because we now
have the u64 time string parser, we use them here.
See-commit: c03ba92576
See-commit: c795be79c1
This patch is somewhat related to:
GnuPG-bug-id: 6378, 5538
* common/mischelp.c (timegm): Move to ...
* common/gettime.c (timegm): here. On Windows use timegm_u32.
(timegm_u32): New.
(isotime2epoch): Factor code out to ...
(isotime_make_tm): new helper.
(isotime2epoch_u64): New.
(_win32_timegm): Remove duplicated code.
(parse_timestamp): Use of timegm.
(scan_isodatestr): Fallback to isotime2epoch_u64.
--
This mainly helps on 32 bit Windows. For Unix we assume everyone is
using 64 bit or shall wait until the libc hackers finally provide a
time64_t.
GnuPG-bug-id: 6736
