gpgme/lang, branch gpgme-1.24.3

cpp: Ensure correct expiration time on 32-bit arch with 64-bit time_t

2025-05-05T14:14:55Z

* lang/cpp/src/key.cpp (Subkey::expirationTime): Cast away the signedness of _gpgme_subkey.expires before casting to time_t. -- _gpgme_subkey.expires stores the expiration as `long int` although the expiration is always an unsigned value. Casting the value to unsigned long int before casting it to time_t ensures that we get the correct expiration value for 64-bit time_t even on 32-bit systems. With signed 32-bit time_t we still get a negative value as before. GnuPG-bug-id: 7627

cpp: Add missing transition, remove two ignored (and wrong) transitions

2025-04-15T07:20:25Z

* lang/cpp/src/gpgsignkeyeditinteractor.cpp (makeTable): Remove two transitions. Add one transition. * lang/cpp/src/gpgsignkeyeditinteractor.cpp (makeTable) : Add one more transition. -- This adds a missing transition for the CONFIRM2 state which ensures that the certification is properly completed in case two confirmations were requested. And it removes two useless (and wrong) transitions which are overwritten by two later correct transitions in the map. GnuPG-bug-id: 7600 This is gpgmepp commit 5b77f4072d03342b8719bbe93fc77641fffb476a The second patch fixed a bug seen before we could test Ingo's patch. - wk

cpp: Validate the transition map

2025-04-11T08:28:20Z

* lang/cpp/src/gpgsignkeyeditinteractor.cpp (makeTable): Assert that all transitions go from one state to a different state. -- The state machine gets stuck if a transition doesn't change the state. Make sure that this cannot happen again by mistake. GnuPG-bug-id: 7600 Taken from gpgmepp commit 6f2e91d4d25afa6934ceaf1563a4d826a904d644

cpp: Ensure that all transitions go from one state to a different state

2025-04-11T08:26:31Z

* src/gpgsignkeyeditinteractor.cpp (makeTable): Replace transitions from CONFIRM to CONFIRM with transitions from CONFIRM to CONFIRM2 and vice versa. Add transitions from CONFIRM2 to some other state for all transitions from CONFIRM to some other state. -- The state machine gets stuck if a transition doesn't change the state but an action is required. To avoid this situation the CONFIRM2 state was introduced, but it was not used correctly. GnuPG-bug-id: 7600 Taken from gpgmepp commit aee2b30482406e677dbddf1a68b2a11a5dc70adf

build,python: Fix uninstalling in distcheck

2024-12-04T16:02:40Z

* lang/python/Makefile.am (uninstall-local): Get platlib path relative to data path. -- For unknown reasons (at least on my system) Python 3.1x's sysconfig returns a platlib (== directory for site-specific, platform-specific files) starting with /usr/local although the prefix config variable is /usr. This broke uninstallation because a wrong folder was (tried to) remove. Python 2.7 and 3.9 return a platlib starting with just /usr. In order to always get the correct relative path we use the data path instead of the prefix config variable as anchor directory.

build,python: Add Python 3.13 and remove Python 3.8

2024-12-04T16:02:40Z

* configure.ac: Look for Python 3.13. Don't look for Python 3.8 anymore. * lang/python/setup.py.in: Update Python versions in classifiers. * m4/python.m4: Add python3.13 as valid interpreter. Remove python3.8. --

cpp: Add Kyber algorithm

2024-11-14T15:22:36Z

* lang/cpp/src/key.h (enum Subkey::PubkeyAlgo): Add AlgoKyber with same value as GPGME_PK_KYBER. -- GnuPG-bug-id: 7397

qt: Fix typo

2024-10-23T08:13:25Z

* lang/qt/src/Makefile.am (camelcase_headers): Fix typo in header name. --

qt: Put attributes before declarations

2024-10-23T08:12:23Z

* lang/qt/src/downloadjob.h (DownloadJob::start): Move QGPGME_DEPRECATED before the whole function declaration. --

cpp: Put attributes before declarations

2024-10-23T08:12:08Z

* lang/cpp/src/key.h (UserID::Signature::Notation): Move GPGMEPP_DEPRECATED before the whole typedef declaration. * lang/cpp/src/verificationresult.h (Signature::Notation): Ditto. --