The DBus specification allows only ASCII characters "[A-Z][a-z][0-9]_".
Dash is illegal and therefore libdbus is complaining about our
interface names not being valid.
Airspace monitor knows all remote aircrafts and which have interim
positions enabled. It passes the list of receivers to INetwork
and INetwork sends it to FSD.
Interim packets are very close to normal position updates. It is
only missing transponder and speed information. On reception
of a interim packet, the information is added to a new
CAircraftSituation. The missing speed information is retrieved
from the last update.
The idea is if you are an addon and running inside the simulator process
you can put his name there. Since swift does (not yet) run as an addon,
set it to "None" instead.
The names were a bit misleading, hence this patch does rename
the methods to something more obvious. It also changes the behaviour
of CNetworkVatlib to automatically answer a custom FSinn query.
In both cases the important details are extracted and signaled to
airspace monitor.
The airspace monitor will send FSinn queries only if the exact model
is not yet known. This would be the case when another client had
sent a FSinn query already.
* Warning flags for elevated warnings as discussed: no /W3, explicit /wd
* fixed include in hotkey samples
* removed outdated QMAKE_CXXFLAGS += -std=c++0x
* Added new "identifier" attribute in JSON metadata files
* "identifier" is now the one that makes plugins unique
* "name" attribute is a human-readable name
refs #404, and discussion https://dev.vatsim-germany.org/issues/404#note-8
* Changed CSimulatorSetup to use CNameVariantPairList as data. The old version was an hack and had to go, CNameVariantPairList would allow complex types in the future and can be already shown in the GUI.
* CNameVariantPairList was improved slightly in the same step
* Functions to get CSimulatorSetup from driver / context
* Removed CSimulatorSetup data from CSimulatorPluginInfo
Qt gets confused by the different ConnectionStatus enums we are using
and complains it cannot queue the argument. It only works when the
fully qualified class name is used.
In the past core was always loading one simulator plugin by default.
This has changed recently with the listeners. So m_simulator is not
guaranteed to be valid.
A proper fix in the future is to not even call them. At the moment they
are hard wired to network.
* Removed obsolete signal from IContextSimulator
* Removed connectToSimulator() method from IContextSimulator as it is no
longer relevant
* CSimulatorComponent updates automatically when simulator disconnects
* FS9 & FSX drivers update (including FSCommon)
To provide maximum flexibility, the simulator info is specified by each
plugin in the JSON metadata file.
* Renamed SCimulatorInfo to CSimulatorPluginInfo
* Removed pre-defined simulator infos (UnspecifiedSim, FSX, etc)
* CContextSimulator keeps track of which plugin is loaded and its info
* QSignalMapper keeps track of which listener emitted the simulatorStared() signal
* CSimulatorPluginInfo contains name, simulator name and description
* TODO: GUI part
* TODO: CTestFsCommon
* The FS9 host is started in the Listener
* When FS9 starts up, it is called via the lobby to join the multiplayer
* Some minor changes in CContextSimulator
* Added ContextSimulator::listenForAllSimulators()
* Added "auto" option in SettingsSimulatorComponent
* Default plugin is UnspecifiedSim, which stands for "auto" option
(!) This in an intermediate step, the feature is not yet fully operational
* max distance for rendering as simulator property
* set max distance / aircraft from GUI
As result of the latest meeting
* changed CCallsignList to CCollection
* fixed some warning with elevated compiler settings (e.g. removed qint values -> int)
* signatures in contexts
* some specialized functions in aircraft list
* context menus in aircraft view
* default "blinking" implementation in driver common base class
The latest vatlib fixes some ambigious methods in the interface and
makes them more clear. Additional it adds a severity level to the
log handlers. Both causing the header to be source incompatible and
therefore need smaller changes in swift.
Before this change signals when users were joining or leaving
the voice room were not properly handled. A new signal in audio
context is emitted now and gui will rebuild the user list when it
is emitted.
