Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | fd.o#22705: turn on both MC and tp-glib debug if MC_DEBUG=all | Simon McVittie | 2009-09-10 | 1 | -2/+64 |
| | | | | | Also put in basic support for MC having Gabble-style debug categories in future. | ||||
* | mcd_debug_[un]ref: deprecate, and stop overriding g_object_[un]ref | Simon McVittie | 2009-09-10 | 3 | -43/+12 |
| | | | | | | | | | | | | The output from these turns out not to be as useful as you might think: about half the refs and unrefs happen somewhere in the bowels of GObject (e.g. g_value_dup_object creates a ref without calling mcd_debug_ref), so the debug log, while verbose, doesn't actually provide enough information to locate leaks. refdbg is what I actually ended up using to find leaks during 5.1 development; it works considerably better in practice, and doesn't require source code modification (or namespace pollution). | ||||
* | mc-server: use a TpDebugSender, to log everything over D-Bus if enabled | Simon McVittie | 2009-09-10 | 1 | -0/+21 |
| | | | | | | This isn't in the library, so other things that link libmissioncontrol-server (mainly osso-mission-control on Maemo) won't automatically have a TpDebugSender. | ||||
* | with-session-bus.sh: fix bashism (&> doesn't work in dash) | Simon McVittie | 2009-09-10 | 1 | -1/+1 |
| | |||||
* | src: set a proper G_LOG_DOMAIN (mcd) | Simon McVittie | 2009-09-10 | 1 | -0/+1 |
| | | | | | | This will lead to log messages like: (process:12345): mcd-DEBUG: telepathy-mission-control version 5.2.3+ | ||||
* | Merge branch 'telepathy-mission-control-5.2' | Simon McVittie | 2009-09-10 | 1 | -0/+19 |
|\ | | | | | | | | | Conflicts: configure.ac | ||||
| * | Post-release version bump | Simon McVittie | 2009-09-10 | 2 | -1/+6 |
| | | |||||
| * | Prepare new bugfix releasetelepathy-mission-control-5.2.3 | Simon McVittie | 2009-09-10 | 2 | -1/+15 |
| | | |||||
* | | Merge branch '52-try-more-handlers' into master | Simon McVittie | 2009-09-10 | 6 | -254/+336 |
|\ \ | | | | | | | | | | | | | | | | Originally targeted for 5.2, but probably too risky there. Reviewed-by: Alberto Mardegan <alberto.mardegan@nokia.com> | ||||
| * | | dispatcher/handle-channels-fails.py: assert that we try more than one handler | Simon McVittie | 2009-09-09 | 1 | -2/+32 |
| | | | |||||
| * | | If HandleChannels fails, remember that the handler is broken and move on to ↵ | Simon McVittie | 2009-09-09 | 1 | -33/+43 |
| | | | | | | | | | | | | | | | | | | | | | | | | the next one If there are no handlers we haven't tried, mcd_dispatcher_run_handlers will close the channels anyway, so there's no longer any need to close them in the HandleChannels callback. | ||||
| * | | McdDispatcher: rename skip_approval to approved, and set it from more places | Simon McVittie | 2009-09-09 | 1 | -13/+35 |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Instead of implicitly approving channels in the following situations, we now set the flag explicitly: - all approvers reject the channels (or, as a trivial case, no approver is run at all) - a handler exists that bypasses approval (this is not quite right, fd.o #23687, but preserves existing behaviour) - a CDO finishes - the channel has Requested=TRUE to start with | ||||
| * | | _mcd_dispatcher_enter_state_machine: don't short-cut to running clients if ↵ | Simon McVittie | 2009-09-09 | 1 | -14/+4 |
| | | | | | | | | | | | | | | | | | | | | | there are no filters If there are no filters, mcd_dispatcher_context_proceed will run the clients anyway, so we might as well reduce the number of code paths. | ||||
| * | | _mcd_dispatcher_enter_state_machine: use mcd_dispatcher_context_proceed ↵ | Simon McVittie | 2009-09-09 | 1 | -1/+1 |
| | | | | | | | | | | | | | | | | | | | | | instead of the equivalent call to mcd_dispatcher_context_process _process's semantics are less obvious and it will be deprecated eventually. | ||||
| * | | Correct a comment | Simon McVittie | 2009-09-09 | 1 | -1/+1 |
| | | | |||||
| * | | McdDispatchOperation: change _mcd_dispatch_operation_handle_with to ↵ | Simon McVittie | 2009-09-08 | 3 | -27/+26 |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | _mcd_dispatch_operation_approve It's only called from outside McdDispatchOperation in order to say "handling with any handler is acceptable" (because a request constitutes approval), never to say "handle with this specific handler". It also doesn't need to return the error, because the only possible error is NotYours, which the caller ignores anyway. | ||||
| * | | Remove McdDispatchOperation from public API/ABI | Simon McVittie | 2009-09-08 | 5 | -110/+79 |
| | | | |||||
| * | | McdDispatchOperation: factor out mcd_dispatch_operation_check_handle_with ↵ | Simon McVittie | 2009-09-08 | 1 | -13/+36 |
| | | | | | | | | | | | | (the checks that can be done synchronously) | ||||
| * | | McdDispatcherContext: always release all client locks, and use a separate ↵ | Simon McVittie | 2009-09-08 | 1 | -22/+17 |
| | | | | | | | | | | | | | | | | | | flag to suppress handlers This makes the handling of CTXREF14 more obviously correct. | ||||
| * | | Replace McdDispatcherContext->client_locks with a boolean | Simon McVittie | 2009-09-08 | 1 | -19/+9 |
| | | | | | | | | | | | | It only has one use, now. | ||||
| * | | McdDispatcherContext: replace the troublesome CTXREF13 with a ref held ↵ | Simon McVittie | 2009-09-08 | 1 | -12/+18 |
| | | | | | | | | | | | | | | | | | | | | | exactly when we're awaiting approval This is more obviously correct; previously, it wasn't at all clear that the refcounting worked correctly. | ||||
| * | | McdDispatcherContext: ref the context for the duration of on_operation_finished | Simon McVittie | 2009-09-08 | 1 | -1/+8 |
| | | | |||||
| * | | Improve comments that explain how client locks work | Simon McVittie | 2009-09-08 | 1 | -6/+12 |
| | | | |||||
| * | | McdDispatcherContext: separate pending approvers out from client_locks into ↵ | Simon McVittie | 2009-09-08 | 1 | -25/+41 |
| | | | | | | | | | | | | awaiting_approval and approvers_pending | ||||
| * | | McdDispatcherContext: separate pending observers from other "client locks" | Simon McVittie | 2009-09-08 | 1 | -12/+31 |
| | | | |||||
* | | | Merge branch 'telepathy-mission-control-5.2' | Simon McVittie | 2009-09-10 | 7 | -82/+13 |
|\ \ \ | | |/ | |/| | | | | | | | Conflicts: configure.ac | ||||
| * | | Merge branch '52-trivia' into telepathy-mission-control-5.2 | Simon McVittie | 2009-09-10 | 7 | -82/+13 |
| |\ \ | | |/ | |/| | | | | Reviewed-by: Alberto Mardegan <alberto.mardegan@nokia.com> | ||||
| | * | Don't distribute generated sources in tarball | Simon McVittie | 2009-09-09 | 1 | -10/+6 |
| | | | |||||
| | * | Distribute assorted files that are present in git but not in tarballs | Simon McVittie | 2009-09-09 | 2 | -0/+3 |
| | | | |||||
| | * | Remove src/mcd-main.c | Simon McVittie | 2009-09-09 | 2 | -52/+0 |
| | | | | | | | | | | | | | | | It's not compiled (server/mc-server.c is the real daemon), and not distributed in tarballs. | ||||
| | * | Remove unused mcd-filtering-int.h from git | Simon McVittie | 2009-09-09 | 2 | -19/+1 |
| | | | | | | | | | | | | It's not distributed in tarballs, or compiled. | ||||
| | * | fd.o #23807: distribute mc-account-convert.1 in tarballs even if gconf was ↵ | Simon McVittie | 2009-09-09 | 1 | -1/+3 |
| |/ | | | | | | | | | | | | | not available at configure time Otherwise, people who have gconf headers can't compile tarballs produced by people who don't have gconf headers. | ||||
| * | Post-release version bump | Simon McVittie | 2009-09-07 | 1 | -1/+1 |
| | | |||||
* | | Merge commit 'telepathy-mission-control-5.2.2' | Simon McVittie | 2009-09-07 | 15 | -47/+323 |
|\ \ | |/ | | | | | | | Conflicts: configure.ac | ||||
| * | Prepare stable release 5.2.2telepathy-mission-control-5.2.2 | Simon McVittie | 2009-09-07 | 2 | -1/+29 |
| | | |||||
| * | Merge branches '52-permissions' and '52-tubes' into ↵ | Simon McVittie | 2009-09-07 | 7 | -18/+85 |
| |\ | | | | | | | | | | | | | | | | telepathy-mission-control-5.2 Reviewed-by: Will Thompson <will.thompson@collabora.co.uk> | ||||
| | * | Factor out identical "should I close it?" logic from _mcd_channel_close, ↵ | Simon McVittie | 2009-09-07 | 1 | -32/+24 |
| | | | | | | | | | | | | _mcd_channel_undispatchable | ||||
| | * | _mcd_channel_close: don't close old Tubes channels | Simon McVittie | 2009-09-07 | 1 | -2/+10 |
| | | | |||||
| | * | Don't close Tubes channels if undispatchable | Simon McVittie | 2009-09-07 | 1 | -0/+6 |
| | | | | | | | | | | | | This implicitly closes every Tube in the channel too, which isn't useful. | ||||
| | * | McdChannel: improve debug when not closing an undispatchable channel | Simon McVittie | 2009-09-07 | 1 | -6/+17 |
| | | | |||||
| | * | fd.o#23778: fix permissions of any existing account storage with ↵ | Simon McVittie | 2009-09-07 | 2 | -2/+5 |
| | | | | | | | | | | | | _mcd_chmod_private, and don't use 0777 permissions to start with | ||||
| | * | mcd-misc: add _mcd_chmod_private() | Simon McVittie | 2009-09-07 | 2 | -0/+33 |
| | | | |||||
| | * | McdMaster: if umask is available, create dirs/files with 0700/0600 permissions | Simon McVittie | 2009-09-07 | 1 | -0/+12 |
| | | | |||||
| | * | configure.ac: check for umask and its headers | Simon McVittie | 2009-09-07 | 1 | -0/+2 |
| | | | |||||
| * | | Merge branch '52-pre-presence' into telepathy-mission-control-5.2 | Simon McVittie | 2009-09-07 | 7 | -29/+210 |
| |\ \ | | |/ | |/| | | | | Reviewed-by: Alberto Mardegan <alberto.mardegan@nokia.com> | ||||
| | * | auto-connect: be less strict about ordering of events | Simon McVittie | 2009-09-03 | 1 | -6/+5 |
| | | | | | | | | | | | | | | | | | | Now that we "pre-load" presence, CurrentPresence can change sooner; so, don't expect the signal (it can happen at basically any time), but just do a Get for the property at the end. | ||||
| | * | dispatcher/already-has-*.py: succeed despite early GetInterfaces call | Simon McVittie | 2009-09-03 | 3 | -0/+12 |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | These tests make assumptions about the order of method calls in order to set up unusual situations, so the early call to GetInterfaces made them deadlock. Also take the opportunity to exercise the code path in which GetInterfaces fails before the connection goes CONNECTED, emulating a telepathy-glib 0.7.35 (or older) CM. | ||||
| | * | account-manager/presence.py: assert that we try to set presence before Connect() | Simon McVittie | 2009-09-03 | 1 | -9/+39 |
| | | | |||||
| | * | mctest: allow expecting events before Connect() | Simon McVittie | 2009-09-03 | 1 | -2/+7 |
| | | | |||||
| | * | mctest: simulate SimplePresence more accurately | Simon McVittie | 2009-09-03 | 1 | -7/+18 |
| | | |