* src/gpgme.h.in: Clarify that the trustlist function should not be
used.
* src/engine.c (_gpgme_engine_op_trustlist): Always return an error.
* src/engine-backend.h (struct engine_ops): Remove trustlist member.
* src/engine-gpg.c (gpg_trustlist): Remove.
(struct engine_ops): Remove that member. Also in all other engines.
* tests/gpg/t-trustlist.c: Remove.
* lang/python/tests/t-trustlist.py: Remove.
--
This never worked in reality because the required feature has been
removed from GnuPG version 1.3.2 soon after introduction of this
feature in gpgme - 17 years ago. It was anyway marked as
experimental. We keep the API and ABI, though.
GnuPG-bug-id: 4834
Signed-off-by: Werner Koch <wk@gnupg.org>
* tests/json/t-keylist-secret.out.json: Do not check for keygrip
of bravo key.
--
Since GnuPG 2.2.19 the keygrip is not emitted for public keys
in a keylisting --with-secret.
As the GPGME test suite should pass with several versions
the json test cannot test this as the suite does not allow
version checks for different output.
What the intended behavior is might still be up for discussion,
always adding --with-keygip is not a good solution because
it slows down keylistings for large keyrings.
GnuPG-Bug-Id: T4820
* tests/json/t-keylist-secret.out.json
--
Since at least 2.2.19 GnuPG no longer emits secret = true
for keys where it does not have a secret. To avoid
failures with older versions we remove the check altogether.
GnuPG-Bug-Id: T4820
* src/gpgme.h.in (struct _gpgme_user_id): Add field 'uidhash'.
* src/key.c (gpgme_key_unref): Free it.
* src/keylist.c (keylist_colon_handler): Set it.
* tests/run-keylist.c (main): Print it.
--
The uidhash value is part of gpg's output since the year 2005. This
now adds support to gpgme. The application for uidhash is to select
a user id in an edit interactor: Instead of giving the number of the
user id, the uidhash value can be be used to avoid tracking the user
id numbers.
Signed-off-by: Werner Koch <wk@gnupg.org>
* tests/json/t-decrypt-verify.out.json: Remove check with key of
04071FB807287134.
--
GnuPG 2.3 changes how public key decryption works; It collects list of
PUBKEY_ENC packets and then processes an ENCRYPTED* packet. At the
stage of processing an ENCRYPTED* packet, it checks if private keys
can decrypt any of PUBKEY_ENC packets. It does not necessarily
examine all of PUBKEY_ENC packets. Only when it finds a PUBKEY_ENC
packet cannot be decrypted, it reports STATUS_NO_SECKEY with keyid of
that packet.
GnuPG 2.2 or older tries PUBKEY_ENC packets in order, and the first
successful PUBKEY_ENC packet is used. Rest of PUBKEY_ENC packets are
not examined, but STATUS_NO_SECKEY with keyid is reported even when
there is a valid private key.
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* tests/run-threaded.c (random_data_close): Correct FD test.
--
An invalid DS is -1 and not 0; this fix is for correctness and does
not chnage anything in the real world.
Signed-off-by: Werner Koch <wk@gnupg.org>
* tests/run-threaded.c (main): Handle allow-del.
(allow-del): New. Variable to allow deletion of keys.
(import): Delete key after import if allow-del is set.
(delete_impres): Delete keys from an import result.
(delete_fpr): Delete a key by fingerprint.
--
This is intended to test write access and locking of the
keyring by repeatedly importing and deleting pubkeys.
It is an option because it might cause keyring corruption
etc. so it should be explicitly enabled.
* tests/gpg/cipher-3.asc: add an encrypted file containing signatures
(one from a known key, and one from an unknown key)
* tests/gpg/cipher-no-sig.asc: add an encrypted file containing no
signatures at all
--
We currently aren't using these files yet, but they should be useful
for exercising the library in unusual circumstances (where the bugs
are).
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
* tests/run-verify.c (print_result): Chack is_de_vs.
* tests/run-import.c (main): Fix bad parameter for format string.
Signed-off-by: Werner Koch <wk@gnupg.org>
* tests/Makefile.am (run-threaded): Add.
* tests/run-threaded.c: New.
--
This test is intended to help detect race conditions
or other multithread problems. It can also be used
to put the whole GnuPG system under extreme load.
* tests/gpg/t-support.h (check_gpg_version, compare_versions),
(parse_version_string, parse_version_number): Remove version
check code.
* tests/json/t-json.c (check_gpg_version, compare_versions),
(parse_version_string, parse_version_number): Add.
--
t-support header was not a good place for this. It should
go into a library e.g. gpgrt in the future. For now we
can keep it close to where it is needed.
* t-config-opt.in.json,
t-config-opt.out.json,
t-import.in.json,
t-import.out.json,
t-sign.in.json,
t-sign.out.json: New.
--
These should actually have been in a previous commit as
they are already registered.
* tests/json/t-json.c (test_contains): Try all siblings of
the haystack child.
--
This fixes the case where we don't match against the
first child because the order might have changed or
a new element was inserted at the top. Then we
have to try out also all the siblings.
* t-json.c (tests): Register new tests.
* t-createkey.in.json,
t-createkey.out.json,
t-decrypt-verify.in.json,
t-decrypt-verify.out.json,
t-decrypt.in.json,
t-decrypt.out.json,
t-delete.in.json,
t-delete.out.json,
t-encrypt-sign.in.json,
t-encrypt-sign.out.json,
t-encrypt.in.json,
t-encrypt.out.json,
t-export-secret-info.in.json,
t-export-secret-info.out.json,
t-export.in.json,
t-export.out.json,
t-json.c,
t-keylist-secret.in.json,
t-keylist-secret.out.json,
t-keylist.in.json,
t-keylist.out.json,
t-verify.in.json,
t-verify.out.json,
t-version.in.json,
t-version.out.json: New.
--
Except for help and getmore there is now a test for each command.
* tests/json/t-json.c (test_contains): When searching for "*" accept
every string.
--
This allows it to write tests that check for the existence of
a string value but don't care about the contents. E.g. for Data.
* tests/json/t-json.c (main): Skip if version is not at least
2.2.0
--
In the past new tests often failed when a modern GPGME was
used with old GnuPG's.
This should avoid the hassle for packagers without the
requirement that we test against old versions.
* t-support.h (parse_version_number, parse_version_string)
(compare_versions): New. Copy&Paste from src/version.c
(check_gpg_version): New helper to check for a gpg version.
--
This should make it easier to write tests that e.g.
rely on modern gnupg features.
* configure.ac: Configure makefile.
* tests/Makefile.am: Run json tests if gpg tests are run.
* tests/json/t-json.c: New testrunner for json tests.
* tests/json/t-config.in, tests/json/t-config.out: First test.
--
The idea of this test runner is that it only looks for parts
in the output. This should allow it to write robust tests
that check for the basics in the output but don't fail when
the output is extended or slightly changed.
* src/context.h (gpgme_context): Extend with trust_model.
* src/engine-gpg.c (engine_gpg): Extend with trust_model.
(gpg_set_engine_flags): Take trust_model from context.
(build_argv): Handle trust_model.
(gpg_release): Free trust_model.
* src/gpgme.c (gpgme_set_ctx_flag): Handle trust-model flag.
(gpgme_release): Release trust-model.
* doc/gpgme.texi: Document new flag for gpgme_set_ctx_flag.
(Context Flags): New subsection for the context flags.
* tests/run-keylist.c (show_usage, main): Add new --trust-model
parameter.
--
This gives a GPGME user fine grained control over the
trust-model.
Changing the trust model for only a single application depends
on:
GnuPG-Bug-Id: T4134
Maniphest Tasks: T4134
Differential Revision: https://dev.gnupg.org/D466
* lang/python/tests/Makefile.am,
lang/qt/tests/Makefile.am,
tests/Makefile.am,
tests/gpg/Makefile.am,
tests/gpgsm/Makefile.am,
tests/opassuan/Makefile.am (GNUPGHOME): Make variable explict.
--
If the build directory has too long path, gpgme could fail.
This is similar to
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=847206
In order to fix that, this patch extracts the GNUPGHOME variable
to be presented directly in the Makefile and thus overridable by
command line option.
A build system can then create a symlink to the GNUPGHOME directory
in /tmp and use that symlink as the GNUPGHOME directory
thus making the path very short.
GnuPG-Bug-Id: T4091
Patch provided by vlmarek
* src/engine-gpg.c (engine_gpg): Add diagnostics member.
(gpg_release): Release diagnostics data.
(gpg_new): Set up logger-fd and diagnostics.
(gpg_getauditlog): New. Copy diagnostics to a user data.
(engine_ops): Add getauditlog.
* src/engine-gpgsm.c (gpgsm_getauditlog): Return not implemented
for GPGME_AUDITLOG_DIAG.
* src/getauditlog.c (getauditlog_start): Don't reset engine
for diagnostics.
* src/gpgme.h.in (GPGME_AUDITLOG_DIAG): New.
(GPGME_AUDITLOG_DEFAULT): New alias to 0.
* tests/run-decrypt.c (show_usage, main): Add --diagnostics.
* doc/gpgme.texi(Additional Logs): Document getauditlog.
--
This enables users of GPGME to get more verbose information
from gpg which can assist users in figuring out a problem
that was before hidden behind a generalized error like
"Decryption Failed".
For GPGSM it is not yet available as it is problematic to
get it properly in server mode and GPGSM already had the
original audit log mechanism in place.
GPGME_AUDITLOG_DEFAULT was added for a more explicit
documentation.
* src/gpgme.h.in (_gpgme_op_decrypt_result): Add flag
legacy_cipher_nomdc.
* src/decrypt.c (parse_status_error): Set this flag.
* tests/run-decrypt.c (print_result): print it.
(main): Print the result even on error.
Signed-off-by: Werner Koch <wk@gnupg.org>
* tests/gpg/t-verify.c (check_result): Tweak for gnupg < 2.2.7.
--
The not yet releases 2.2.7-beta may print a full fingerprint in the
ERRSIG status. This is compliant with the dscription but the new
t-verify test case did not took in account that older GnuPG versions
print only a keyid.
Fixes-commit: b99502274a
GnUPG-bug-id: 3920
Signed-off-by: Werner Koch <wk@gnupg.org>
* src/op-support.c (_gpgme_parse_plaintext): Add arg r_mime.
* src/decrypt.c (_gpgme_decrypt_status_handler): Ser mime flag.
* src/verify.c (_gpgme_verify_status_handler): Ditto.
* src/gpgme.h.in (gpgme_op_verify_result_t): Append fields 'is_mime'
and '_unused'.
(gpgme_op_decrypt_result_t): New field 'is_mime'. Shrink '_unused'.
* tests/run-decrypt.c (print_result): Print MIME flag.
* tests/run-verify.c (print_result): Ditto.
--
Note that this flag (Liternal Data packet's 'm' mode) is only
specified in RFC-4880bis. To use it you currently need to add
"rfc4880bis" to the the gpg.conf.
Signed-off-by: Werner Koch <wk@gnupg.org>
* src/gpgme.h.in (gpgme_op_encrypt_ext_start) New.
(gpgme_op_encrypt_ext): New.
(gpgme_op_encrypt_sign_ext_start): New.
(gpgme_op_encrypt_sign_ext): New.
* src/libgpgme.vers, tests/run-encrypt.c: Add them.
* src/encrypt.c (encrypt_start): Add arg recpstring.
(gpgme_op_encrypt): Factor code out to ...
(gpgme_op_encrypt_ext): new function with new arg recpstring.
(gpgme_op_encrypt_start): Factor code out to ...
(gpgme_op_encrypt_ext_start): new function with new arg recpstring.
* src/encrypt-sign.c (encrypt_sign_start): Add arg recpstring.
(gpgme_op_encrypt_sign): Factor code out to ...
(gpgme_op_encrypt_sign_ext): new function with new arg recpstring.
(gpgme_op_encrypt_sign_start): Factor code out to ...
(gpgme_op_encrypt_sign_ext_start): new function with new arg
recpstring.
* src/engine-backend.h (struct engine_ops): Change fields encrypt and
encrypt_sign.
* src/engine.c (_gpgme_engine_op_encrypt): Add arg recpstring and pass
to engine.
(_gpgme_engine_op_encrypt_sign): Ditto.
* src/engine-gpg.c (append_args_from_recipients_string): New.
(gpg_encrypt): Add arg recpstring and call new function as needed.
(gpg_encrypt_sign): Ditto.
* src/engine-gpgsm.c (set_recipients_from_string): New.
(gpgsm_encrypt): Add arg recpstring and call new function as needed.
* src/engine-uiserver.c (set_recipients_from_string): New.
(uiserver_encrypt): Add arg recpstring and call new function as
needed.
* tests/run-encrypt.c (xstrdup): New.
(main): Add option --keystring.
* src/gpgme-json.c (get_keys): Simplify.
(op_encrypt): Modify to make use of the extended encrypt function.
--
This new feature can be used to avoid the need for a key lookup and
thus several extra calls to the backend.
Note that run-test uses a semicolon as delimiter because that make
testing the feature on the command line much easier.
Signed-off-by: Werner Koch <wk@gnupg.org>
* src/op-support.c (_gpgme_parse_failure): Ignore failures with
location "gpg-exit".
* tests/gpg/t-verify.c (main): Adjust for the now working checking of
the second key.
Signed-off-by: Werner Koch <wk@gnupg.org>
* tests/gpg/t-verify.c (PGM): New. Use it instead of __FILE__.
(test_sig1_plus_unknown_key): New test signature.
(check_result): Allow checking of several signatures.
(main): Check a signature with a know and an unknown key.
Signed-off-by: Werner Koch <wk@gnupg.org>
* src/gpgme.h.in (gpgme_import_result_t): Extend with skipped_v3_keys.
* src/import.c (gpgme_op_import_result): Extend debug with new field.
(parse_import_res): Parse skipped_v3_keys.
* tests/gpg/t-support.h, tests/run-support.h (print_import_result):
Print skipped_v3_keys.
--
This makes it possible to handle this in a GUI in a future
version.
GnuPG-Bug-Id: T3776
* tests/gpg/Makefile.am: Don't allow target with '/'.
* tests/gpgsm/Makefile.am: Ditto.
--
BSD Make doesn't allow a target with '/'. We still have such a target
in lang/python/Makefile.am, but it's for maintainer only, so,
assumption to GNU Make is OK there.
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* tests/gpg/Makefile.am, tests/gpgsm/Makefile.am: Remove ./.
--
GNU Make is powerful enough for handling and interpreting of
pathname as target, but BSD Make is not.
GnuPG-bug-id: 3056
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* src/data.c (_gpgme_data_new): Check for failed selftest.
* tests/run-support.h (make_filename): Print a message on mallooc
failure.
(init_gpgme_basic): New.
* tests/run-identify.c (main): Call init_gpgme_basic becuase we do not
need to setup a protocol.
* tests/t-data.c: Define PGM and include run-support.h.
(make_filename): Remove.
(main): Call init_gpgme_basic.
--
Note: This patch may break applications which used gpgme_data but no
gpgme_new without having called the required gpgme_check_version.
GPGME can be used without a protocol for example to work with the data
object abstraction. Thus a call to gpgme_data_new also needs to check
the result of the core selftests - including on whether
gpgme_check_version has been called.
Signed-off-by: Werner Koch <wk@gnupg.org>
* tests/gpgsm/Makefile.am (GPG_AGENT): Set var. It is used later.
--
Note that the var is used by the gpgsm.conf target.
Reported-by: Alon Bar-Lev <alon.barlev@gmail.com>
Signed-off-by: Werner Koch <wk@gnupg.org>
* lang/python/tests/Makefile.am: Distinguish target and path.
* tests/gpg/Makefile.am: Ditto.
* tests/gpgsm/Makefile.am: Ditto.
--
GNU Make is powerful enough to match path to target (and vice versa),
but BSD make is not.
GnuPG-bug-id: 3056
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* src/gpgme.c (gpgme_set_ctx_flag, gpgme_get_ctx_flag): New flag
"auto-key-retrieve".
* src/context.h (gpgme_context): New field auto_key_retrieve.
* src/engine-backend.h (struct engine_ops): Add arg auto_key_retrieve
to field 'decrypt'.
* src/engine-gpg.c (gpg_decrypt): Add arg auto_key_retrieve and pass
option --auto-key-retrieve to gpg. Adjust all callers.
(gpg_verify): Ditto.
* src/engine-gpgsm.c (gpgsm_decrypt): Add dummy arg auto_key_retrieve.
* src/engine-uiserver.c (uiserver_decrypt): Ditto.
* tests/run-verify.c (main): Add option --auto-key-retrieve.
--
This makes the --auto-key-retrieve option available in the GPGME API.
Test plan:
Run
GPGME_DEBUG=9:out tests/run-verify SIGNEDFILE
with and without its new option --auto-key-retrieve and check in the
trace stored in "out" whether --auto-key-retrieve was passed to gpg.
Signed-off-by: Werner Koch <wk@gnupg.org>
* tests/gpg/t-gpgconf.c (main): Fix warnings.
* src/engine-gpgconf.c (struct gpgconf_config_dir_s): New struct.
(gpgconf_config_dir_cb, gpgconf_conf_dir) Use it to fix warning.
Signed-off-by: Marcus Brinkmann <mb@g10code.com>
* tests/gpgsm/Makefile.am (gpgsm.conf): Add agent-program directive.
* tests/start-stop-agent: Update agent-program directive.
--
Update the 'agent-program' configuration directive to point to the
same agent that we are starting. Previously, it was possible that a
different agent was started if 'make check' was run with a different
PATH.
Signed-off-by: Justus Winter <justus@g10code.com>
* tests/gpgsm/Makefile.am (check-local): Drop rule and the dependency
on it.
--
Previously, the test environment was created during 'make check'.
Nowadays we create it using 'BUILT_SOURCES' during 'make all'. Drop
remnant of the previous method.
Signed-off-by: Justus Winter <justus@g10code.com>
* lang/python/tests/Makefile.am: Kill all previously running daemons
before creating the private key store.
* lang/qt/tests/Makefile.am: Likewise.
* tests/gpg/Makefile.am: Likewise.
* tests/gpgsm/Makefile.am: Likewise.
--
Now that the daemons sockets are no longer created in the GNUPGHOME,
we cannot rely on cleaning the build directory to make sure they are
shut down. Therefore, we explicitly kill any running daemons when
creating the test environment.
Signed-off-by: Justus Winter <justus@g10code.com>
* lang/python/tests/Makefile.am: Create test environment as part of
'make all'.
* tests/gpg/Makefile.am: Make sure the private keystore is created
first.
* tests/gpgsm/Makefile.am: Create test environment as part of
'make all'. Make sure the private keystore is created
first.
Signed-off-by: Justus Winter <justus@g10code.com>
* tests/gpg/Makefile.am: Use BUILT_SOURCES instead of check-local
and initial.test.
* lang/qt/tests/Makefile.am: Ditto.
--
This fixes "make dist" failure when source tree is clean:
git clean -dxf
autoreconf -ivf
./configure
make dist
BUILT_SOURCES should be used when file as generated without explicit
dependency. The check-local is all-am dependency, this means that it
will be resolved also in "make dist".
Signed-off-by: Alon Bar-Lev <alon.barlev@gmail.com>
* tests/gpg/Makefile.am (tests_unix): Add 't-thread-keylist' and
't-thread-keylist-verify'.
(c_tests): Drop them here.
--
The tests use pthreads and we suspect problems with MinGW's wrappers.
Signed-off-by: Justus Winter <justus@g10code.com>
* src/gpgme.h.in (GPGME_ENCRYPT_WRAP): New const.
(gpgme_decrypt_flags_t): New enum.
(GPGME_DECRYPT_VERIFY): New const
(GPGME_DECRYPT_UNWRAP): New const
(gpgme_op_decrypt_ext_start): New func.
(gpgme_op_decrypt_ext): New func.
* src/decrypt-verify.c (gpgme_op_decrypt_ext_start): New.
(gpgme_op_decrypt_ext): New.
(decrypt_verify_start): Add arg FLAGS. Replace call to
engine_op_decrypt_verify by the plain decrypt with the flag set.
(gpgme_op_decrypt_verify_start): Pass the flag.
(gpgme_op_decrypt_verify): Pass the flag.
* src/decrypt.c (decrypt_start): Rename to ...
(_gpgme_decrypt_start): this. Add arg FLAGS. Pass FLAGS to
engine_op_decrypt.
(gpgme_op_decrypt_start): Adjust for chnage pass 0 for FLAG.
(gpgme_op_decrypt_start): Ditto.
* src/engine.c (_gpgme_engine_op_decrypt_verify): Remove.
(_gpgme_engine_op_decrypt): Add arg FLAGS.
* src/gpgme.def, src/libgpgme.vers: Add new functions.
* src/engine-backend.h (struct engine_ops): Remove member
'decrypt_verify'. Add FLAGS to 'decrypt'. Adjust all initialization.
* src/engine-uiserver.c (uiserver_decrypt): Remove.
(uiserver_decrypt_verify): Remove.
(_uiserver_decrypt): Rename to ...
(uiserver_decrypt): this. Replace arg VERIFY by new arg FLAGS.
* src/engine-gpg.c (gpg_decrypt): Support GPGME_DECRYPT_UNWRAP.
(gpg_encrypt): Support GPGME_ENCRYPT_WRAP.
* tests/run-decrypt.c (main): New option --unwrap.
* tests/run-encrypt.c (main): New option --wrap.
--
Manual testing of that wrap/unwrap feature can be done this way:
./run-encrypt --verbose --key Alice /etc/motd > x
./run-decrypt --verbose --unwrap x > y
./run-encrypt --verbose --key Bob --wrap y > z
1. The message was first encrypted to Alice.
2. Alice decrypts the message receiving a valid OpenPGP message.
3. Alice encrypt that message to Bob
This will also work with encrypted and signed messages; the signature
will be kept intact during re-encryption. Requires GnuPG 2.1.12.
Signed-off-by: Werner Koch <wk@gnupg.org>
* src/dirinfo.c (WANT_GPG_WKS_CLIENT_NAME): New const.
(struct dirinfo): New field 'gpg_wks_client_name'.
(get_gpgconf_item): Build it on demand.
(gpgme_get_dirinfo): New value "gpg-wks-client-name" for WHAT.
Signed-off-by: Werner Koch <wk@gnupg.org>
* src/gpgme.h.in (GPGME_ENCRYPT_THROW_KEYIDS): New flag.
* src/engine-gpg.c (gpg_encrypt): Implement flag
(gpg_encrypt_sign): Implement flag.
* tests/run-encrypt.c (main): New option --throw-keyids.
--
It would be nice to also selectively hide recipients (that is gpg
--hidden-recipient) but our API does not ye allow this because it is
based on key objects. A possible way to implement that would be a API
to set processing flags into a key but this is complicated due to the
reference counting and thus the possibility that a key object is used
by different context.
Signed-off-by: Werner Koch <wk@gnupg.org>
* tests/start-stop-agent: Do not create 'gpg-agent.conf' if it does
not exist.
Fixes-commit: 16b202d999
Signed-off-by: Justus Winter <justus@g10code.com>
* lang/python/tests/Makefile.am (gpg-agent.conf): Do not hard-code the
option. This breaks gpg-agent from GnuPG 2.0.
* tests/start-stop-agent: Rather, check if the option is supported and
add it to the configuration if it is.
GnuPG-bug-id: 3008
Fixes-commit: bbf19124bb
Signed-off-by: Justus Winter <justus@g10code.com>
* tests/gpg/t-gpgconf.c (main): Reduce iterations to 10.
* tests/gpg/t-thread-keylist-verify.c,
tests/gpg/t-thread-keylist.c (THREAD_COUNT): Reduce to 10.
--
While these tests tested for race conditions a smaller number
of iteration should still show problems if they are run on
multiple systems and regulary. While the 100 Thread count in
the t-thread tests could lead to resource problems.
* tests/start-stop-agent: Don't autostart agent on --stop and
running check. Use --debug-quick-random when starting.
--
This should speed up the tests especially on low entropy systems.
Possibly fixing a hang on pythons op_genkey test in the Launchpad
build enviorment (see launchpad issue 1655298)
