* lang/qt/tests/t-import.cpp (ImportTest): Add member tempGpgHome and
member function initTestCase.
(keyFpr, keyData): Move to ImportTest::testImportWithKeyOrigin.
--
Use different temporary GNUPGHOME's when running the test. This ensures
that the import in the test is always done with a new keyring.
GnuPG-bug-id: 5733
* lang/qt/src/job.cpp (Job::~Job): Destroy JobPrivate object belonging
to destroyed job.
--
In some tests different Job objects reused the same JobPrivate object
because they were created with the same memory address. Obviously, this
shouldn't happen.
GnuPG-bug-id: 5739
* lang/qt/tests/t-support.cpp, lang/qt/tests/t-support.h
(class QGpgMETest): Add member function doOnlineTests.
* lang/qt/tests/t-wkdlookup.cpp (testWKDLookupAsync, testWKDLookupSync,
testLookupWithNoResultAsync): Skip tests if online tests are not
enabled.
--
GnuPG-bug-id: 5728
* lang/qt/src/Makefile.am (qgpgme_sources): Add qgpgmewkdlookupjob.cpp,
wkdlookupresult.cpp.
(qgpgme_headers): Add wkdlookupjob.h, wkdlookupresult.h.
(camelcase_headers): Add WKDLookupJob, WKDLookupResult.
(private_qgpgme_headers): Add qgpgmewkdlookupjob.h.
(qgpgme_moc_sources): Add qgpgmewkdlookupjob.moc, wkdlookupjob.moc.
* lang/qt/src/job.cpp: Define c'tor and d'tor of WKDLookupJob.
* lang/qt/src/protocol.h (Protocol::wkdLookupJob): New.
* lang/qt/src/protocol_p.h (Protocol::wkdLookupJob): New.
* lang/qt/src/qgpgmewkdlookupjob.cpp, lang/qt/src/qgpgmewkdlookupjob.h,
lang/qt/src/wkdlookupjob.h, lang/qt/src/wkdlookupresult.cpp,
lang/qt/src/wkdlookupresult.h: New.
* lang/qt/tests/Makefile.am (EXTRA_DIST): Add final.test.
(the_tests): New.
(TESTS): Remove all t-*. Add $(the_tests) and final.test.
(moc_files): Add t-wkdlookup.moc.
(t_wkdlookup_SOURCES): New.
(noinst_PROGRAMS): Add t-wkdlookup.
* lang/qt/tests/final.test: New.
* lang/qt/tests/t-support.h (QTest::toString): New template
specialization for std::string.
* lang/qt/tests/t-wkdlookup.cpp: New.
--
The new WKDLookupJob allows running a WKD lookup without implicit
import of the retrieved key. This makes it possible to do WKD lookups
similar to keyserver lookups (aka external keylistings).
The new final.test pseudo test stops a dirmngr that may have been
started by the tests. The toString() template specialization enables
printing of the actual and expected values in case of failed QCOMPARE.
GnuPG-bug-id: 5728
* lang/cpp/tests/Makefile.am (run_wkdlookup_SOURCES, programs_unix):
New.
(noinst_PROGRAMS): Add $(programs_unix).
* lang/cpp/tests/run-wkdlookup.cpp: New.
--
This test runner is a prototype for doing a WKD lookup without implicit
import of the found key(s).
Building this test runner is disabled on Windows because it may not
work there. Feel free to enable it on Windows if the test runner works
there.
GnuPG-bug-id: 5728
lang/cpp/src/context.cpp (Context::engineInfo()): Return engine info
for protocol of context.
--
This change also fixes Context::setEngineFileName() and
Context::setEngineHomeDirectory() which overwrote home dir resp.
file name of the engine info for the protocol of context with the
corresponding value of the "first" engine info (i.e. engine info for
gpg) instead of keeping the current home dir resp. file name.
GnuPG-bug-id: 5722
* 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)
* lang/qt/tests/t-keylist.cpp (KeyListTest::testListAllKeysSync):
Ensure same type for both arguments of QCOMPARE.
* lang/qt/tests/t-various.cpp (TestVarious::testSetExpire):
Ensure same type for both arguments of QCOMPARE.
--
Qt 5.9 does not yet have the generic qCompare helper which supports
arguments of any type that can be compared with ==.
GnuPG-bug-id: 5592
* lang/qt/src/threadedjobmixin.cpp (stringFromGpgOutput): Expect
UTF-8.
--
In recent versions more and more output was changed to be
UTF-8 on Windows. Werner advised to always expect UTF-8
diagnostics.
* lang/qt/tests/t-various.cpp (testSignKeyWithExpiration): Convert
expiration time to uint_least32_t.
--
This fixes the test on 32-bit systems where time_t (the return type of
expirationTime()) is a signed 32-bit integer type.
GnuPG-bug-id: 5522
* lang/qt/src/Makefile.am (qgpgme_sources): Add changeexpiryjob.cpp.
* lang/qt/src/changeexpiryjob.cpp: New.
* lang/qt/src/changeexpiryjob.h (enum ChangeExpiryJob::Option,
typedef ChangeExpiryJob::Options, ChangeExpiryJob::setOptions,
ChangeExpiryJob::options): New.
* lang/qt/src/qgpgmechangeexpiryjob.cpp (change_expiry): Update the
expiration date of primary key and/or some/all subkeys.
(QGpgMEChangeExpiryJob::start): Add options to call of change_expiry.
(ChangeExpiryJob::start): Moved implementation to changeexpiryjob.cpp.
--
Additionally, to the default behavior (update expiration of primary
key if no subkeys are specified resp. update expiration of specified
subkeys) it is now possible to update the expiration of the primary key
_and_ the specified/all subkeys.
GnuPG-bug-id: 4717
* lang/qt/src/job_p.h: New.
* lang/qt/src/job.cpp (typedef JobPrivateHash, d_func, setJobPrivate,
getJobPrivate): New.
--
Because of ABI compatibility requirements we cannot add a d-pointer
to Job. Therefore we store the d-pointers in a global static. This
mechanism will allow Job subclasses to store additional data without
breaking the ABI.
GnuPG-bug-id: 4717
--
This functionality got dropped somewhere after 1.12, as part of the
cleanup of the `Context.decrypt` call signature. Reintroduce it again,
now using an explicit keyword argument `filter_signatures` (which
defaults to hiding signatures by unknown keys).
GnuPG-bug-id: 5292
* lang/qt/src/qgpgmenewcryptoconfig.cpp (parseURL): Only set non-empty
base DN as URL query.
--
This makes it possible to differentiate an unset query from an empty
query.
GnuPG-bug-id: 5465
* lang/qt/src/dataprovider.cpp (QIODeviceDataProvider::write): Call
waitForBytesWritten.
--
The problem here is that QProcess writes into an internal buffer
which is written to stdin of the process triggered by a
signal/slot connection. That connection is broken when we move
the QProcess into our GPGME thread and only restablished when
our Job is finished. This caused Kleopatra to basically keep
everything when decrypting a large archive in memory and
only write it out to the unpack process once the decryption
was finished.
GnuPG-Bug-Id: T5475
* lang/qt/src/qgpgme_debug.h: Make include guard match file name.
* lang/qt/src/qgpgme_debug.h, lang/qt/src/qgpgme_debug.cpp
(QGPGME_CONFIG_LOADING_LOG): New.
* lang/qt/src/qgpgmenewcryptoconfig.cpp
(QGpgMENewCryptoConfig::reloadConfiguration): Use new logging category.
--
Use a separate logging category for the extremely noisy logging of
the result of config loading to make debug logging more useful.
GnuPG-bug-id: 5217
* lang/qt/src/qgpgmenewcryptoconfig.cpp (parseURL): Handle extended
LDAP server option syntax.
(portToString): New.
(splitURL): Append flags to LDAP server option.
--
This adds support for the extended syntax of LDAP server options
introduced in gpg 2.2.18/2.3. The flags are stored as fragment of a
QUrl.
GnuPG-bug-id: 5217
* lang/cpp/src/editinteractor.cpp (EditInteractor::Private): Initialize
members 'state' and 'debug' in-class. Add member 'debugNeedsClosing'.
(EditInteractor::Private::Private): Remove members initializers.
Remember if 'debug' needs to be closed.
(EditInteractor::Private::~Private): Only close 'debug' if it needs to
be closed.
--
This fixes the problem that after destroying an edit interactor all
debug output went to /dev/null instead of stderr (or stdout) if one
enabled debugging of the edit interactors with GPGMEPP_INTERACTOR_DEBUG
set to stderr (or stdout).
* lang/qt/src/qgpgmesignkeyjob.h: Remove unneeded includes. Include
<memory>.
(QGpgMESignKeyJob): Remove all member variables. Add pimpl pointer.
* lang/qt/src/qgpgmesignkeyjob.cpp: Include <QString>. Don't include
<memory>.
(QGpgMESignKeyJob::Private): New.
(QGpgMESignKeyJob::QGpgMESignKeyJob): Remove initialization of removed
members. Initialize d.
(QGpgMESignKeyJob::start, QGpgMESignKeyJob::setUserIDsToSign,
QGpgMESignKeyJob::setCheckLevel, QGpgMESignKeyJob::setExportable,
QGpgMESignKeyJob::setSigningKey, QGpgMESignKeyJob::setNonRevocable,
QGpgMESignKeyJob::setRemark, QGpgMESignKeyJob::setDupeOk): Adapt to move
of member variables to pimpl.
--
GnuPG-bug-id: 5245, 5421
* lang/qt/src/cryptoconfig.h, lang/qt/src/cryptoconfig.cpp
(CryptoConfig::entry): Move implementation to cpp. Add overload not
requiring a group name.
--
The group name is not needed for identifying a config entry because the
groups only provide a logical grouping of config entry for user
interfaces. To improve usability entries are sometimes moved to a
different group, but this shouldn't break existing applications trying
to access those entries. The new group-agnostic overload makes
applications robust against regrouping of config entries.
GnuPG-bug-id: 5217
* lang/qt/src/encryptjob.h, lang/qt/src/signjob.h (start):
Undeprecate QByteArray based functions.
--
While the QIODevice access might avoid a copy it is more
difficult to manage the data and lifetime of the
QIODevices in calling code. The QByteArray calls are
convienient for small data objects like mails where
an in memory copy is not really expensive anymore.
This reverts commit 88294023c1.
--
This commit was too early and i needed to test more
this is breaking more then it helps so for now
revert it before we can do a proper solution.
* configure.ac: Configure cpp data.h.in
* lang/cpp/src/Makefile.am: Generate data.h
* lang/cpp/src/data.cpp, lang/cpp/src/data.h: Use portable
types.
* lang/qt/src/Makefile.am: Include build dir.
* lang/qt/tests/makefile.am: Include build dir.
--
These kind of patches have been around for a while, IMO this
should not create an ABI incompatbility for cases where
it already works because the types should be the same
so I think this is not an interface break.
GnuPG-Bug-Id: T3996
* lang/qt/src/qgpgmesignkeyjob.cpp (sign_key): Check remark for being
an empty string instead of a null QString.
--
GnuPG-bug-id: 5142
Co-authored-by: Daniel Kahn Gillmor
* lang/python/tests/t-quick-subkey-creation.py: Specify RSA.
--
Using GnuPG 2.3, a key with default (ed25519/cv25519) cannot have some
combination of capabilities (e.g., "encr auth").
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* lang/qt/src/debug.h: Add export macro.
--
This is public API so it should be exported. This did not
resolve a link issue for Windows that I've experienced
when building. But this might be to some uncleanlyness
in my build environment.
lang/qt/src/quickjob.h (QuickJob::result): Make parameters
auditLogAsHtml and audigLogError optional.
--
This aligns the signature of QuickJob's result signal with the result
signals of the other Job subclasses.
GnuPG-bug-id: 5094
lang/cpp/src/key.h, lang/cpp/src/key.cpp (UserID::Signature::operator<):
New.
lang/cpp/src/key.cpp (signature_index): New.
--
operator< defines a canonical sort order for signatures belonging to
the same user ID. It is based on the sort order defined by
cmp_signodes() in g10/keylist.c of gnupg. In particular, the ordering
of signatures made with the same key by creation time allows to see
easily whether the most recent signature is a certification or a
revocation.
GnuPG-bug-id: 5094
recent
* lang/qt/src/qgpgmequickjob.h: (GpgME::Key, QDateTime, QString):
Remove superfluous forward declarations.
(QGpgMEQuickJob): Remove superfluous second template argument. It is
identical to the default.
(QGpgMEQuickJob::result): Remove. It is already defined in the base
class.
* lang/cpp/src/context.cpp, lang/cpp/src/context.h
(Context::revokeSignature, Context::startRevokeSignature): New.
* lang/cpp/src/context.cpp
(getLFSeparatedListOfStrings, getLFSeparatedListOfUserIds): New.
(getLFSeparatedListOfFingerprintsFromSubkeys): Extracted second part
of function to getLFSeparatedListOfStrings.
--
GnuPG-bug-id: 5094
* lang/cpp/src/context.cpp, lang/cpp/context.h
(Context::cancelPendingOperationImmediately): New.
* NEWS: Mention added API
--
This allows immediate canceling of running operations, e.g. when a
thread with a long running operation (like SCD DEVINFO --watch) is
going to be terminated (on application shutdown).
GnuPG-bug-id: 5066
* lang/cpp/src/Makefile.am: Add new files.
* lang/cpp/src/interfaces/statusconsumer.h,
lang/cpp/src/statusconsumerassuantransaction.cpp,
lang/cpp/src/statusconsumerassuantransaction.h: New.
* NEWS: Mention new API.
--
This Assuan transaction is useful for long running Assuan commands.
Classes implementing the StatusConsumer interface can process received
status lines while the Assuan command is still running.
GnuPG-bug-id: 5066
* lang/qt/src/Makefile.am: Add new files.
* lang/qt/src/debug.h, lang/qt/src/debug.cpp: New.
* NEWS: Mention it.
--
This was previously implemented in Kleopatra, but it makes sense for
all users of qgpgme.
* lang/cpp/src/engineinfo.h
(EngineInfo::Version::operator>(const Version &)): Fix logic.
(EngineInfo::Version::operator>(const char *)): Use Version-overload of
operator>.
* lang/qt/tests/t-various.cpp: Add test.
--
This fixes a logic error that 2.0.0 > 2.0.0 would return true.
* lang/cpp/src/key.cpp (Key::update): Call Context::key() only once
with KeyListMode::WithSecret.
--
With gpg >= 2.1, get the key with --with-secret instead of first trying
to get the secret key and, if that fails, getting the public key.
GnuPG-bug-id: 4794
* lang/qt/src/qgpgmelistallkeysjob.cpp
(do_list_keys): Rename to do_list_keys_legacy and put into unnamed
namespace.
(merge_keys): Put into unnamed namespace.
(list_keys): Rename to list_keys_legacy and put into unnamed namespace.
(do_list_keys, list_keys): New.
* lang/qt/tests/t-keylist.cpp (testListAllKeysSync): New.
--
With gpg >= 2.1, list keys once with --with-secret instead of listing
public keys and secret keys and then merging (part of) the information
about the keys.
GnuPG-bug-id: 4794
* lang/cpp/src/util.h
(add_to_gpgme_keylist_mode_t): Sort modes as in enum definition and
add missing modes to check.
(convert_from_gpgme_keylist_mode_t): Add missing handling of
GPGME_KEYLIST_MODE_WITH_TOFU and GPGME_KEYLIST_MODE_WITH_KEYGRIP.
* lang/cpp/src/key.cpp (Key::mergeWith): Also merge secret flag and
keygrip
--
The keygrip was lost if the result of a public key listing was merged
with the result of a secret key listing as done by QGpgMEListAllKeysJob.
GnuPG-bug-id: 4794
* lang/qt/src/dn.cpp (parse_dn_part): Fix parser.
--
This could in theory result in reading bytes after a after Nul in a
string and thus possible segv on unallocated memory or reading other
parts of the memory. However, it is harmless because the rfc2253
strings have been received from GnuPG which is expected to emit
correct syntax.
GnuPG-bug-id: 5037
* 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
* lang/qt/src/changeexpiryjob.h (ChangeExpiryJob::start): New overload
that accepts subkeys (with empty implementation).
* lang/qt/src/qgpgmechangeexpiryjob.h,
lang/qt/src/qgpgmechangeexpiryjob.cpp (QGpgMEChangeExpiryJob::start):
New overload that accepts subkeys.
--
This adds the possibility to change the expiry of subkeys.
GnuPG-bug-id: 4717
* lang/cpp/src/context.h, lang/cpp/src/context.cpp
(Context::getLFSeparatedListOfFingerprintsFromSubkeys): Remove
from Context; make it a static function instead
--
GnuPG-bug-id: 5003
* lang/qt/src/qgpgmechangeexpiryjob.cpp (change_expiry): Use new
setExpire() instead of edit() with GpgSetExpiryTimeEditInteractor
--
Using the new setExpire() (which uses --quick-set-expire) is a lot less
complex than using gpg's edit interface. It also feels faster. Next,
I'll add support for changing the expiration time of subkeys.
GnuPG-bug-id: 5003
* lang/cpp/src/key.cpp (Subkey): Add ostream operator.
* lang/cpp/src/key.h: Update accordingly.
--
This is helpful for debugging / showing the state of a subkey.
* lang/cpp/context.cpp, lang/cpp/context.h
(Context::startPublicKeyExport, Context::exportPublicKeys): Extend
with flags paramenter.
(Context::ExportMode): New.
* lang/qt/src/exportjob.h (ExportJob::setExportMode): New.
* lang/qt/src/qgpgmeexportjob.cpp, lang/qt/src/qgpgmeexportjob.h:
Update accordingly.
--
This adds the C++ and Qt API for export modes.
* 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>
* lang/qt/src/*: Change logging category to macro to QGPGME_LOG.
--
The old logging category macro had a typo and this way we
are more consistent with other logging rules. For example
you could write gpg.* in the logging conf.
* lang/qt/src/Makefile.am: Add new files.
* lang/qt/src/job.cpp (GpgCardJob): Add impl stuff.
* lang/qt/src/protocol.h (gpgCardJob): Get one.
* lang/qt/src/qgpgmebackend.cpp,
lang/qt/src/qgpgmebackend.h: Add helpers to get the job.
* lang/qt/src/qgpgmegpgcardjob.cpp,
lang/qt/src/gpgcardjob.h,
lang/qt/src/qgpgmegpgcardjob.h: New.
--
This is annoyingly complex to add a simple new job.
In the future we should implement something like this
without the threadedjobmixin stuff. But the idea was
to follow the usual job pattern.
GnuPG-Bug-Id: T4794
* lang/cpp/src/gpgsignkeyeditinteractor.cpp (action):
Use uidhash instead of number.
(GpgSignKeyEditInteractor::setKey): New.
* lang/cpp/src/gpgsignkeyeditinteractor.h: Update accordingly.
* lang/cpp/src/key.h, lang/cpp/src/key.cpp: Wrap uidhash.
* lang/qt/src/qgpgmesignkeyjob.cpp: Set the key.
--
Using the uidhash avoids problems when the user ids
on --edit-key are different ones then the uids
captured by gpgme when listing keys. Or if
they are in a different order. This can happen
with cached keys or keys with user attributes.
* lang/qt/tests/t-remarks.cpp (initTestCase): Supply
allow-weak-key-signatures flag for GnuPG 2.3, which
normally rejects use of SHA1 digest.
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* lang/qt/tests/t-encrypt.cpp (decryptSupported): Moved to
t-support as loopbackSupported.
* lang/qt/tests/t-remarks.cpp: Check for loopbackSupported.
* lang/qt/tests/t-support.cpp, lang/qt/tests/t-support.h
(loopbackSupported): New.
--
This ensures that the tests do not fail with GnuPG 2.0.x
* lang/qt/test/t-remarks.cpp (testRemarkReplaceSingleUIDExportable):
Use QVERIFY instead of assert.
--
An assert causes the temp directory not to be cleaned up. I
use this to analyze test failures.
If you want to see the tests home directory after a failure you
can just change a failing verify to an assert.
* lang/cpp/src/key.cpp, lang/cpp/src/key.h (UserID::remarks): New.
* NEWS: Mention this.
--
This can be useful if we want to show remarks made by others, too.
For:
GnuPG-Bug-Id: T4734
* lang/qt/tests/t-various.cpp (testRemarks): Move out.
* lang/qt/tests/t-remarks.cpp: New.
* lang/qt/tests/Makefile.am: Update accordingly.
--
While t-various is nice to add a quick test for more
extensive stuff it is better to move it in its own test.
This also tests with a key that has only a single uid.
* lang/cpp/src/key.cpp (UserID::remark): Use C-API.
--
The other parts of that function already use the C API
so we can also avoid function calls for the keyListMode.
