* src/import.c (gpgme_op_import_result): Check fpr for NULL.
(parse_error): New.
(import_status_handler): Handle error status line.
* doc/gpgme.texi (gpgme_import_status_t): Mention that fpr can be NULL.
* tests/gpg/t-import.c (check_result): Check fpr for NULL.
* tests/run-threaded.c (delete_impres): Check fpr for NULL.
--
When importing an encrypted certificate a wrong passphrase may be
entered. In this case gpgsm emits a status line with a bad passphrase
error and an "invalid object" error. To make it possible for callers
to handle a wrong passphrase error more gracefully, an import status
with bad passphrase error is added to the import result for each
status line with bad passphrase error.
GnuPG-bug-id: 5713
* src/context.h (struct gpgme_context): New field import_filter.
* src/engine-backend.h (struct engine_ops): Add arg import_filter to
field 'import'.
* src/engine-gpg.c (gpg_import): Add arg import_filter and pass option
--import-filter with argument value to gpg. Adjust all callers.
* src/engine-gpgsm.c (gpgsm_import): Add dummy arg import_filter.
* src/gpgme.c (gpgme_release): Free 'import_filter'.
(gpgme_set_ctx_flag, gpgme_get_ctx_flag): New flag "import-filter".
* tests/run-import.c (main): Add option --import-filter.
--
This makes the --import-filter option available in the GPGME API for
key imports.
GnuPG-bug-id: 5739
* src/context.h (struct gpgme_context): New field key_origin.
* src/engine-backend.h (struct engine_ops): Add arg key_origin to
field 'import'.
* src/engine-gpg.c (gpg_import): Add arg key_origin and pass option
--key-origin with argument value to gpg. Adjust all callers.
* src/engine-gpgsm.c (gpgsm_import): Add dummy arg key_origin.
* src/gpgme.c (gpgme_release): Free 'key_origin'.
(gpgme_set_ctx_flag, gpgme_get_ctx_flag): New flag "key-origin".
* tests/run-import.c (main): Add option --key-origin.
* tests/gpg/t-import.c (main): Set and verify key origin.
--
This makes the --key-origin option available in the GPGME API for
key imports.
GnuPG-bug-id: 5733
* src/engine-gpgconf.c (gpgconf_parse_option): Skip the string
indicator if it is not the empty string.
--
Some versions of GnuPG (~2.2.28 to 2.2.33) returned a compliance
indicator which should be an int but actually used string as type.
Simply skipping the indicator without checking that this is the
indicator does not allow to work around this type mismatch by the
gpgme users.
* src/dirinfo.c (WANT_SOCKETDIR): New.
(dirinfo): Add field socketdir.
(parse_output): Support "socketdir".
(get_gpgconf_item): Return socketdir.
* tests/t-engine-info.c (main): Add socketdir to the output.
* src/w32-util.c (_gpgme_create_process_utf8): Fix indentation.
--
GnuPG-bug-id: 5613
This is not strictly necessary because we could deduce this from one
of the other socket info items but it is more clean to have a
dedicated info item. The socketdir item is available for 5 years now
(GnuPG commit 8e3fa5a4b205c534de2142e5d071712f957cf06a)
* src/gpgme.pc.in (Require): Move gpg-error, libassuan to...
(Requires.private): ... here.
(Cflags): Add -I${includedir}.
(Libs): Add -L${libdir}.
--
This will result better library dependency for applications using
gpgme. It will break build of an application which directly uses
gpg-error or libassuan without specifying them (which means, depending
on redundant/wrong gpgme.pc).
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
src/debug.h (TRACE_SYSRES_OFF_T, _trace_sysres_off_t,
TRACE_SYSRES_SSIZE_T, _trace_sysres_ssize_t): New.
src/data.c (gpgme_data_read, gpgme_data_write, gpgme_data_seek): Use
appropriate new tracing macros instead of casting the results to int.
--
This change adds tracing macros for results of system functions of
type __off_t and __ssize_t.
GnuPG-bug-id: 5481
* src/context.h (struct gpgme_context): Add 'cert_expire'.
* src/engine-gpg.c (append_args_from_cert_expire): New.
(gpg_edit): Set option according to the new flag.
* src/gpgme.c (gpgme_release): Free 'cert_expire'.
(gpgme_set_ctx_flag, gpgme_get_ctx_flag): Add "cert-expire".
* tests/gpg/Makefile.am (c_tests): Add new test.
(gpg.conf): Write "allow-weak-key-signatures" to gpg.conf.
* tests/gpg/t-edit-sign.c: New.
--
The new context flag "cert-expire" allows setting the expiration date
for key signatures created with gpgme_op_interact.
GnuPG-bug-id: 5336, 5505
* src/data.c (_gpgme_data_release): Free buffers.
(gpgme_data_seek): Adjust from renamed fields.
(gpgme_data_set_flag): Implement new flags.
(_gpgme_data_inbound_handler): Allow the use of a malloced buffer.
(_gpgme_data_outbound_handler): Ditto.
* src/data.h (BUFFER_SIZE): Move out of the struct definition.
(struct gpgme_data): Remove pending filed and introduce inbound and
outbound fields.
* src/conversion.c (_gpgme_wipememory): New. Taken from GnuPG.
* src/cJSON.c (wipememory): Use this here too.
* tests/run-decrypt.c (main): Add options "--large-buffers" and
"--sensitive".
--
GnuPG-bug-id: 5478
Signed-off-by: Werner Koch <wk@gnupg.org>
* src/data.h (BUFFER_SIZE): Increase to 4096 for Windows.
--
This brings it in line to the PIPE_BUF size on desktop Linux
systems. This should increase performance when working with
large files on Windows a bit.
GnuPG-Bug-Id: T5478
* src/Makefile.am (AM_CFLAGS): Add GPG_ERROR_CFLAGS
--
This fixes the include directory for libgpg error if
it is installed in a different prefix then libassuan. Previously
libassuan provided the include directory also implicitly.
* src/keylist.c (keylist_colon_handler): Check for correct number of
fields.
--
This prevents NULL pointer dereferencing with older versions of gpg
that may output less fields.
* src/gpgme.h.in (struct _gpgme_key_sig): Move trust_scope to the end
of the struct.
--
Fixes-commit: 276187f6b6
Note that we assume that splitting a reserved 28 bit field into 3
fields does not change the ABI.
* src/gpgme.h.in (struct _gpgme_key_sig): Add fields 'trust_depth',
'trust_value', and 'trust_scope'.
* src/key.c (gpgme_key_unref): Free trust_scope.
* src/keylist.c (keylist_colon_handler): Set the fields.
* tests/run-keylist.c (main): Print the fields.
--
The trust signature values are part of gpg's output since the year 2003.
This now adds support to gpgme.
GnuPG-bug-id: 5245, 5419
* src/export.c (export_start): Remove GPGME_EXPORT_MODE_NOUID check.
* src/engine-gpg.c (export_common): Ditto.
--
Note that this constant was never announed and used only for internal
tests. It made it into some release only by accident.
The constants is kept in gpgme.h to not break any code but it is
non-fucntionals. We do not consider this an ABI break.
GnuPG-bug-id: 5284
* src/export.c (export_start): Set command handler for passphrase
interaction for the case when it's secret keys.
(export_ext_start): Likewise.
--
GnuPG-bug-id: 5046
Co-authored-by: Louis Dupré Bertoni
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* configure.ac (NEED_GPG_ERROR_VERSION): Require 1.36.
* src/cJSON.c: Remove code for older version.
* src/engine.c (gpgme_get_engine_info): Ditto.
* src/gpgme-json.c: Ditto.
* src/op-support.c: Ditto.
* src/util.h: Ditto.
--
Libgpg-error 1.36 has been released more than 18 months ago so it is
time to avoid hacks and require this verion. This will for example
help Kleopatra to support PIV cards and improves the gpgme-json.
Signed-off-by: Werner Koch <wk@gnupg.org>
* src/engine-assuan.c (inquire_cb): Implement returning data.
* tests/opassuan/t-command.c (inq_cb): Send some test data.
--
The old code only allowed to send an empty response which was good
enough for scdaemon's KNOWNCARDP inquire but not to send actual data.
A quick test using a test smartcard might be
./t-command 'scd setattr --inquire PRIVATE-DO-1'
and then reading the data back using
gpg-connect-agent 'scd getattr PRIVATE-DO-1' /bye
Signed-off-by: Werner Koch <wk@gnupg.org>
* src/genkey.c (setexpire, gpgme_op_setexpire_start,
gpgme_op_setexpire): Move to ...
* src/setexpire.c: New.
* src/Makefile.am (main_sources): Add that file.
* src/context.h (ctx_op_data_id_t): Add OPDATA_SETEXPIRE.
* lang/qt/tests/t-various.cpp (testSetExpire): Test error handling.
--
Errors (and failures) emitted via status-fd need to be handled
explicitly, i.e. we need to provide an appropriate status handler with
corresponding op_data_t. Additionally, we need to set a passphrase
command handler if a passphrase callback is set in the context, e.g.
during tests.
GnuPG-bug-id: 4395
* src/gpgme.h.in (gpgme_op_setexpire_start, gpgme_op_setexpire): New.
* src/libgpgme.vers, src/gpgme.def: Add new functions.
* src/genkey.c (setexpire): New.
(gpgme_op_setexpire_start, gpgme_op_setexpire): New.
* src/engine.h, src/engine.c: (_gpgme_engine_op_setexpire): New.
* src/engine-backend.h (engine_ops): Add 'setexpire' and adjust all
engine initializers.
* src/engine-gpg.c (gpg_setexpire): New.
(_gpgme_engine_ops_gpg): Set setexpire to gpg_setexpire.
* doc/gpgme.texi: Document new functions.
* tests/run-genkey.c: Add option --setexpire.
--
This extends GPGME to support the --quick-set-expire command
added by GnuPG 2.1.22. This allows changing subkeys expiry
date without going through the editinteractor interface.
Co-authored-by: Andre Heinecke <aheinecke@gnupg.org>
GnuPG-bug-id: 4999
* src/export.c (export_start): Allow that mode.
--
The test tools uses gpgme_op-export_ext and thus did not caught it.
Signed-off-by: Werner Koch <wk@gnupg.org>
* 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>
* src/verify.c (parse_trust): Fix detection of "chain" keyword.
--
Fixes-commit: da6f3dc0c5
from 2007 shortly after introducing this. I doubt that this info has
ever been used (it is for qualified signatures, which are only
supported using the legacy German RegTP rules for them which were soon
overturned by the commercial CAs).
Signed-off-by: Werner Koch <wk@gnupg.org>
* src/Makefile.am (gpgme_w32_extra_libs): New.
(libgpgme_la_LIBADD, libgpgme_glib_la_LIBADD): Use it.
--
Since w32-io.c directly uses functions from ws2_32 it should
not rely on libgpg-error to pull in this dependency.
* src/engine-gpg.c (gpg_keylist_build_options): Send --with-keygrip
if --with-secret is used.
--
Since GnuPG 2.2.19 the keygrip was not anymore send if only
--with-secret was used in a public key listing. Given that
--with-secret requires computation of the keygrip anyway, there is no
point in not requesting the keyrip in this case.
GnuPG-bug-id: 4820
* src/Makefile.am (gpgme_tool_LDADD): Add @GPG_ERROR_LIBS@.
--
Since gpgme-tool uses gpg_strerror, it should be linked
to -lgpg-error.
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* src/posix-util.c (find_executable): New.
(walk_path_str): Factored out from walk_path.
(walk_path): Replaced by find_executable.
(_gpgme_get_gpg_path, _gpgme_get_gpgconf_path): Use find_executable.
--
This should help to locate GnuPG on MacOS systems where
it is not part of the PATH environment variable and
should reduce the need to have fixed path known
at GPGME compile time.
mailvelope/issue699
* 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>
* src/engine-gpg.c (gpg_decrypt): only send --no-keyring when we are
not verifying.
--
Without this change, the signature verification would fail. This
problem was introduced in bded8ebc59 in
an attempt to avoid an error when *not* verifying. Clearly more test
suite coverage is needed to avoid introducing this sort of problem in
the future.
GnuPG-bug-id: 3464
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
* src/keylist.c (keylist_colon_handler): Do not clear
out the tmp_uid when a subpacket line is encountered.
--
When the uid is cleared no more signatures following
the subpacket are parsed, so we now no longer clear
it when a subpacket is encountered.
