Sat, 17 Feb 2018 21:23:58 +0100 ENGINE-382 - fixing update bug
Fri, 16 Feb 2018 08:35:04 +0100 ENGINE-382: changed alter table with non-constant default for timestamp to copy-insert-drop for version 8
Thu, 15 Feb 2018 00:15:05 +0100 Added prioritisation of own ids and forgotten prioritisation of recent records, necessary for update_identity to work properly; message to self now gets green from outgoing_message_rating
Wed, 14 Feb 2018 18:49:09 +0100 Other SQL fixes for, for example, constraint errors on person.main_key_id. This is like pulling your own teeth without novacaine...
Wed, 14 Feb 2018 16:02:08 +0100 ENGINE-379 and probably other problems: was calling set_identity from within set_as_pep_user, which was called by get_trust, which is called from set_identity. So... yeah. Bad.
Tue, 13 Feb 2018 16:08:48 +0100 Damnit, shut off trusted own key import again. Fixed.
Tue, 13 Feb 2018 00:13:18 +0100 ADD_TO_LOG and GOTO are now simply removed everywhere.
Mon, 12 Feb 2018 12:45:27 +0100 Removing log_event errors - if it fails, it fails.
Mon, 12 Feb 2018 12:17:03 +0100 Removing logging from the log_event function when logging fails
Sun, 11 Feb 2018 18:31:35 +0100 Argh, accidentally committed import key disable. Sorry.
Fri, 09 Feb 2018 09:03:11 +0100 Added set_as_pep_user to set trust
Fri, 09 Feb 2018 08:14:27 +0100 ENGINE-365: is_pep_user() is now exposed to the outside. See pep_engine.h documentation
Thu, 08 Feb 2018 20:05:10 +0100 Removed Roker test that broke when run outside of the testing environment (this isn't a case of avoiding a bug, but rather the case would fail because the keys tested are actual keys which some of us have present in our real keyrings, and test presumes we do not.
Wed, 07 Feb 2018 19:08:11 +0100 Argh... if true...
Wed, 07 Feb 2018 18:46:29 +0100 Removed inadverent commit of sqlite3 header include
Wed, 07 Feb 2018 16:53:36 +0100 Turned automatic first-time import of ultimately trusted keys back on
Wed, 07 Feb 2018 16:42:56 +0100 ENGINE-250: foreign keys are now on and tests pass. Be aware that with the amount of time we may have been running with them off, there may be other execution paths that cause trouble. File bugs against any DB errors.
Wed, 07 Feb 2018 16:05:04 +0100 ENGINE-250: address fixes and more; lower() still does something really goofy on the FIRST run with make test, but only then, and even if all databases and keyrings are removed, case_and_dot_address_test still passes for every time after until the next time you make clean *with the whole engine* and rebuild. BIZARRE.
Wed, 07 Feb 2018 10:22:23 +0100 ENGINE-250: Pretty sure the current bugs are now actual bugs and not sql-related. But will check. insert-or-replace for identity-based set functions is now replaced with a slightly convoluted callback mechanism which does a check for existence and then updates or inserts appropriately, with optional transaction guards (to be set to false if they're called in what's already a transaction, e.g. set_identity)
Wed, 07 Feb 2018 09:23:51 +0100 ENGINE-250: Marginally working, but changes() doesn't seem to be a sufficient way of differentiating between update and insert. COuld be my bug though. Some tests still fail, but we need a commit here.
Tue, 06 Feb 2018 18:48:45 +0100 ENGINE-250: fixed some things, am screaming at others. Have overcome the 'insert or update' issue at least, probably.
Mon, 05 Feb 2018 10:58:03 +0100 Intermittent commit
Sun, 04 Feb 2018 03:52:27 +0100 Merged in ENGINE-352
Sun, 04 Feb 2018 03:36:08 +0100 A whole horde of trust fixes (mistrust list added, tests fixed, and sneaky null string in elect key removed)
Fri, 02 Feb 2018 11:03:20 +0100 ENGINE-352: put in the mechanism to attach an is_pep mechanism for the user; tacking pEp messages is complicated, however, as it seems we've also always fouled up when we checked on that (we don't check what key it's encrypted with before we declare whatever key gets retrieved by update_identity is the same, and then set the comm_type on that, which could be the wrong key.) So that's the next fix.
Thu, 01 Feb 2018 13:57:04 +0100 shelving mistrust investigation and fixes to fry bigger fish
Mon, 29 Jan 2018 12:36:02 +0100 Fixing bad merges still.
Mon, 29 Jan 2018 12:15:51 +0100 Urgh. Fixing bad merge (???)
Fri, 26 Jan 2018 16:30:39 +0100 ENGINE-329: gpg implementation of first-time import of ultimately trusted keys is in. Apparently, netpgp bluffs trust completely, so a stub is in here for them instead. Too bad :(
Wed, 24 Jan 2018 18:23:21 +0100 merged in default
Wed, 24 Jan 2018 16:06:50 +0100 why not on the phones?
Wed, 24 Jan 2018 11:57:10 +0100 ENGINE-340: removed extra mistrust check - problem was actually a missing set in key_mistrust when there is not yet a set identity
Wed, 24 Jan 2018 10:54:04 +0100 ENGINE-340: fixed persistence issues caused by removing mistrusted keys as defaults in ENGINE-332 branch
Mon, 22 Jan 2018 17:40:44 +0100 ENGINE-332: key_reset_trust and key mistrust fixed
Mon, 22 Jan 2018 11:11:38 +0100 ENGINE-332: rewrite complete, but needs checking and testing.
Wed, 17 Jan 2018 12:23:44 +0100 ENGINE-289: added tests for myself() default user_id aliases, fixed consts in signature, fixed blacklist test issue
Wed, 17 Jan 2018 01:01:10 +0100 ENGINE-289: added getter and setter methods for user id aliases, plus error codes for failure (PEP_CANNOT_SET/FIND_ALIAS)
Wed, 17 Jan 2018 00:25:36 +0100 ENGINE-289: added user_id alias queries
Wed, 17 Jan 2018 00:09:58 +0100 ENGINE-289: added alternate user id mapping tables
Mon, 15 Jan 2018 17:12:49 +0100 ENGINE-289: myself() changes and related fallout. Committing to merge in default.
Wed, 10 Jan 2018 16:19:07 +0100 ENGINE-289: All old tests now pass!
Wed, 10 Jan 2018 09:19:44 +0100 ENGINE-289: empty fpr on identity input now ok. removed superfluous (and incorrect) check.
Mon, 08 Jan 2018 12:19:34 +0100 Merged in default AND fixed SQL problem (oops)
Mon, 25 Dec 2017 06:22:37 +0100 ENGINE-289: Argh.
Tue, 19 Dec 2017 00:00:35 +0100 ENGINE-289: Ok, really. Someone tell me what's wrong with 'update person set id = ?1 where id = ?2 ;', because honestly... anyway, some of the sql is fixed. I should give it up.
Mon, 18 Dec 2017 23:00:56 +0100 ENGINE-289: intermittent commit - almost compiling cleanly (yes, it's that far away from tested. It's complicated.
Mon, 18 Dec 2017 13:27:42 +0100 ENGINE-289: Intermittent commit - need to refactor some logic big-time.
Fri, 15 Dec 2017 12:49:03 +0100 ENGINE-320: intermittent commit
Wed, 13 Dec 2017 11:42:23 +0100 ENGINE-289: shelving to work on another bug
Sat, 09 Dec 2017 13:04:40 +0100 ENGINE-293: initialise_own_identities (cheap myself() only keygen/retrieval) + additional internal fprless get_identity function added so that temporary identities w/o fpr can be retrieved
Thu, 07 Dec 2017 01:13:36 +0100 ENGINE_293: some sql issues
Tue, 05 Dec 2017 16:35:23 +0100 ENGINE-293: Compiles and runs SOME tests. SQL and myself() fixes.
Tue, 05 Dec 2017 10:00:48 +0100 ENGINE-293: compiling again. Need to fix test cases and see what I have wrought with the PEP_OWN_USERID removals. After that, next step is to really fix myself()
Mon, 04 Dec 2017 18:23:14 +0100 ENGINE-293: Intermittent commit. Probably not compiling yet, but PEP_OWN_USERID is out from src where it needs to be.
Mon, 27 Nov 2017 16:23:13 +0100 Committing to move to proper branch
Wed, 22 Nov 2017 09:43:14 +0100 Shelving to check for changes. Next step here is to continue dealing with PEP_OWN_USERID as the only marker all over the DB
Tue, 21 Nov 2017 17:19:25 +0100 shelving changes to context switch - finish DB change after this for .me
Mon, 20 Nov 2017 17:44:50 +0100 ENGINE-289: reverted most of the identity->me removal. While tests pass, this is not in its final form.
Thu, 02 Nov 2017 17:30:01 +0100 Removed all 2-party keysync components from engine.
Thu, 19 Oct 2017 11:03:28 +0200 Forcing normal locking on the DB, though this is the default anyway