Commit message (Collapse) | Author | Age | Files | Lines | ||
---|---|---|---|---|---|---|
... | ||||||
* | added Logger test client (activatable respawnable observer) | Senko Rasic | 2010-04-15 | 1 | -0/+10 | |
| | ||||||
* | add tests for respawning (activatable) observers | Senko Rasic | 2010-04-15 | 3 | -0/+453 | |
| | ||||||
* | update existing tests with slightly different use cases due to new feature | Senko Rasic | 2010-04-15 | 2 | -5/+5 | |
| | ||||||
* | mctest.SimulatedClient: support for optional recovery property for observers | Senko Rasic | 2010-04-15 | 1 | -2/+17 | |
| | ||||||
* | added Logger client service for twisted tests | Senko Rasic | 2010-04-15 | 2 | -1/+4 | |
| | ||||||
* | McdConnection: pass account path _mcd_dispatcher_recover_channel | Senko Rasic | 2010-04-15 | 1 | -1/+2 | |
| | ||||||
* | added internal method _mcd_dispatcher_recover_channel() | Senko Rasic | 2010-04-15 | 2 | -3/+5 | |
| | ||||||
* | McdDispatcher: listen to McdClient's need-recovery signal | Senko Rasic | 2010-04-15 | 1 | -0/+45 | |
| | ||||||
* | pass on the account path to the handler map | Senko Rasic | 2010-04-15 | 1 | -1/+2 | |
| | ||||||
* | added internal _mcd_channel_details_build_from_tp_chan method | Senko Rasic | 2010-04-15 | 2 | -15/+41 | |
| | ||||||
* | added _mcd_client_recover_observer method | Senko Rasic | 2010-04-15 | 2 | -0/+36 | |
| | ||||||
* | McdClient: emit need-recovery signal when an observer needs to be respawned | Senko Rasic | 2010-04-15 | 1 | -0/+21 | |
| | ||||||
* | McdClient: query for and parse Observer.Recovery property | Senko Rasic | 2010-04-15 | 1 | -3/+55 | |
| | ||||||
* | McdHandlerMap: maintain a mapping between handled channels and their account ↵ | Senko Rasic | 2010-04-15 | 2 | -2/+38 | |
| | | | | paths | |||||
* | Set the right environment variable to run tests with a temporary version of ↵ | Simon McVittie | 2010-04-14 | 1 | -1/+1 | |
| | | | | ~/.cache | |||||
* | trivia: .gitignore: ignore GCOV and with-session-bus droppings | Simon McVittie | 2010-04-14 | 1 | -0/+3 | |
| | ||||||
* | Prepare NEWS for 5.4 | Simon McVittie | 2010-04-14 | 1 | -2/+18 | |
| | ||||||
* | Incomplete accounts resulted in account manager load locks never being released. | Vivek Dasmohapatra | 2010-04-14 | 1 | -3/+27 | |
| | | | | | | | This meant that the account manager's dbus name was never claimed, as it sat around forever waiting for the account's manager to become ready, (or fail) but this could never happen as the incomplete account didn't have a manager at all. | |||||
* | Fix error propagation in CreateAccount | Tomeu Vizoso | 2010-04-14 | 2 | -8/+34 | |
| | | | | Fixes fdo#27021 | |||||
* | Remove misleading reference to GConf | Tomeu Vizoso | 2010-04-14 | 1 | -1/+1 | |
| | ||||||
* | Only time out channel requests after many hours | Sjoerd Simons | 2010-04-14 | 1 | -2/+4 | |
| | | | | Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk> | |||||
* | mc-tool: don't crash on invalid account names | Will Thompson | 2010-04-14 | 2 | -0/+13 | |
| | | | | Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk> | |||||
* | mc-wait-for-name: link to GLib properly | Jonny Lamb | 2010-04-14 | 1 | -2/+2 | |
| | | | | | | | | See Debian bug #556488 for more info. Fixes: fd.o#25122 Signed-off-by: Jonny Lamb <jonny.lamb@collabora.co.uk> | |||||
* | Make MC compile with dbus-glib 0.80 by casting const GError* back to GError* | Danielle Madeley | 2009-11-06 | 2 | -2/+2 | |
| | | | | | | | | Ubuntu Karmic ships with dbus-glib 0.80. Which takes a GError* argument to dbus_g_method_return_error(). We could either bump our minimum dbus-glib version to 0.82, or cast the errors so that it works. Reviewed-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk> | |||||
* | Merge branch 'compat' | Simon McVittie | 2009-11-03 | 4 | -2/+23 | |
|\ | | | | | | | Reviewed-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk> | |||||
| * | Update NEWS | Simon McVittie | 2009-11-03 | 1 | -1/+4 | |
| | | ||||||
| * | fd.o #24875: reimplement g_error_new_valist to avoid a GLib 2.22 dependency | Simon McVittie | 2009-11-03 | 3 | -1/+19 | |
|/ | ||||||
* | Nano version | Simon McVittie | 2009-11-02 | 2 | -1/+6 | |
| | ||||||
* | Prepare version 5.3.2upstream/5.3.2telepathy-mission-control-5.3.2 | Simon McVittie | 2009-11-02 | 2 | -2/+4 | |
| | ||||||
* | Update NEWS | Simon McVittie | 2009-11-02 | 1 | -3/+41 | |
| | ||||||
* | Merge branch 'telepathy-mission-control-5.2' | Simon McVittie | 2009-11-02 | 1 | -0/+1 | |
|\ | | | | | | | | | | | | | | | | | | | Conflicts: NEWS configure.ac src/mcd-connection.c src/mcd-dispatcher.c Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk> | |||||
| * | Invoke callbacks in the right order | Alberto Mardegan | 2009-10-23 | 1 | -0/+1 | |
| | | | | | | | | | | | | Callback contexts are added into the list at the beginning (g_slist_prepend), so we must reverse the list before invoking them, to ensure that callbacks are invoked in the same order as they have been installed. | |||||
| * | Merge branch '52-always-observe-new-channels' into telepathy-mission-control-5.2 | Simon McVittie | 2009-10-13 | 5 | -21/+151 | |
| |\ | | | | | | | | | | Reviewed-by: Alberto Mardegan <alberto.mardegan@nokia.com> | |||||
| | * | McdConnection: correct sense of suppress_handler in on_new_channel | Simon McVittie | 2009-10-12 | 1 | -1/+1 | |
| | | | | | | | | | | | | | | | | | | Channels with suppress_handler=TRUE must be Requested; channels with suppress_handler=FALSE can be assumed to be unRequested. Previously, the opposite logic was used. | |||||
| | * | fd.o#24474: Add a regression test for channels created behind our back | Simon McVittie | 2009-10-12 | 2 | -0/+107 | |
| | | | ||||||
| | * | fd.o#24474: if NewChannels signals Requested channels that we didn't ↵ | Simon McVittie | 2009-10-12 | 3 | -21/+44 | |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | request, observe them but do nothing else Requested channels that we don't know about must have been requested by another process (like Empathy 2.28), by calling Connection methods directly rather than going via MC. As a result, we should not call AddDispatchOperation or HandleChannels. However, we do want to call ObserveChannels, to preserve the invariant that observers are called for *all* channels. This means that observers can do things like present some simple UI for *every* channel, regardless of who the handler is. | |||||
| * | | Nano version | Simon McVittie | 2009-10-13 | 2 | -1/+6 | |
| |/ | ||||||
* | | Makefile.am: re-format SUBDIRS | Simon McVittie | 2009-11-02 | 1 | -3/+10 | |
| | | | | | | | | Reviewed-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk> | |||||
* | | Makefile.am: build tests before docs | Simon McVittie | 2009-11-02 | 1 | -1/+1 | |
| | | | | | | | | | | | | | | This makes it quicker to see whether `make check` has failed - if you see the docs building, it means the tests must have passed :-) Reviewed-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk> | |||||
* | | Merge branch 'wait-for-handlers' | Simon McVittie | 2009-11-02 | 16 | -270/+555 | |
|\ \ | | | | | | | | | | Reviewed-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk> | |||||
| * | | McdDispatchOperation: comment the fact that handle_with_time is global | Simon McVittie | 2009-11-02 | 1 | -0/+9 | |
| | | | | | | | | | | | | | | | | | | | | | | | | In an earlier version of the branch I separated it to be per Approval and used the timestamp from the appropriate approval, but during review Sjoerd considered this to be unnecessarily complicated; it's also undesirable if several HandleWith calls are "compatible", e.g. HandleWith("...Empathy") and HandleWith(""). | |||||
| * | | McdDispatchOperationPrivate: remove now-unused handler member | Simon McVittie | 2009-11-02 | 1 | -1/+0 | |
| | | | ||||||
| * | | _mcd_dispatch_operation_close_as_undispatchable: receive GError from caller | Simon McVittie | 2009-11-02 | 1 | -10/+14 | |
| | | | ||||||
| * | | _mcd_dispatch_operation_set_handler_failed: if all possible handlers have ↵ | Simon McVittie | 2009-11-02 | 2 | -3/+20 | |
| | | | | | | | | | | | | | | | | | | | | | | | | failed, finish the CDO and terminate the channels _mcd_dispatch_operation_close_as_undispatchable will try and fail to call _mcd_dispatch_operation_finish again, which isn't entirely elegant, but works. | |||||
| * | | dispatcher/handle-channels-fails.py: add regression test for retrying ↵ | Simon McVittie | 2009-11-02 | 2 | -15/+15 | |
| | | | | | | | | | | | | HandleWith with the same handler | |||||
| * | | _mcd_dispatch_operation_try_next_handler: fail HandleWith calls if the ↵ | Simon McVittie | 2009-11-02 | 1 | -5/+14 | |
| | | | | | | | | | | | | desired Handler doesn't exist | |||||
| * | | _mcd_dispatch_operation_try_next_handler: remove unnecessary check for approval | Simon McVittie | 2009-11-02 | 1 | -2/+1 | |
| | | | | | | | | | | | | | | | If we're looking at an Approval from the queue, then clearly someone approved :-) | |||||
| * | | _mcd_dispatch_operation_try_next_handler: for HandleWith calls (only), try ↵ | Simon McVittie | 2009-11-02 | 1 | -4/+8 | |
| | | | | | | | | | | | | | | | | | | | | | | | | again even if that handler previously failed This is safe (doesn't cause infinite loops) because on failure, we remove the HandleWith call from the queue of approvals, so we only retry if an approver tries again. | |||||
| * | | McdDispatchOperation: use the approvals queue to choose the preferred handler | Simon McVittie | 2009-11-02 | 1 | -25/+22 | |
| | | | | | | | | | | | | | | | This requires that requests with a preferred handler also put it in the queue. | |||||
| * | | _mcd_dispatch_operation_set_handler_failed: fail any pending HandleWith ↵ | Simon McVittie | 2009-11-02 | 2 | -8/+71 | |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | calls if the nominated handler fails HandleChannels This is a behaviour change: HandleWith with a non-empty argument will no longer count as general approval; so in dispatcher/handle-channels-fails, when the first client fails, the second won't run immediately. |