Until now, the solution to ensure that even large messages are fully sent out has been to flush the connection queues after each sending of a message, which is likely an unnecessary call (with unnecessary cost) in vast majority of cases, and which may block the connection from doing other work until the large message is fully sent out. This was a rather quick, hacky workaround.
Now, after the sending the message we check whether it has been sent out fully or not. If not (outbound queues are non-empty), then we send a wake-up signal to the connection event loop. The event loop thread then fetches new sd-bus timeouts and events and will see that there are pending outbound messages to process, and will process them together with any other prospective pending events, until there is nothing to process (i.e., the outbound message has been fully dispatched).
This reorganizes the layers of abstraction in the sense how `Message` depends on `Connection` and vice versa. Now, `Message` has a link to the `Connection`. This replaces the shortcut link to the low-level `SdBus` interface that the `Message` kept. The interactions from `Message` now go through `Connection` which forwards them to `SdBus`. `Connection` is now a sole owner of the low-level `SdBus` interface. This allows for future changes around `SdBus` (e.g. a change from virtual functions back to non-virtual functions) without affecting the rest of the library. `Proxy`s and `Object`s can now send messages directly without having to go through `Connection`. The `Connection` no more depends on `Message` business-logic methods; it serves only as a factory for messages.
The flow for creating messages: `Proxy`/`Object` -> `Connection` -> `SdBus`
The flow for sending messages: (`Proxy`/`Object` ->) `Message` -> `Connection` -> `SdBus`
This also better reflects how dependencies are managed in the underlying sd-bus library.
Additionally, `getSdBusInterface()` methods are removed which was anyway planned, and improves the design by "Tell, don't ask" principle.
This refactoring is the necessary enabler for other upcoming improvements (regarding sending long messages, or creds refactoring, for example).
Fix inconsistencies in documentation and CMake files regarding minimum required libsystemd version (which is v238 for sdbus-c++ v2.0). Also fix compiler error bug in sdbus-c++ code when compiled against older libsystemd versions. Plus some extra related cleanups.
This fixes the template specialization ambiguity error when a signature_of applied upon a const (and/or volatile) enum type would render the const specialization and the enum specialization as two equally-ranked candidates.
This makes that you can use these macros in headers as well as
source files.
Else you will get a warning about multiple definitions of the same
function.
This introduces SDBUSCPP_REGISTER_STRUCT macro that helps clients conveniently, in one line, teach sdbus-c++ to recognize and accept their custom C++ struct types wherever D-Bus structs are expected in a D-Bus API.
The macro saves some boilerplate that would otherwise be needed on client side for every registered struct.
Make all the API consistent by using return_slot_t-based overloads for returning the slots to clients, and overloads without that tag for "floating" slots. floating_slot_t tag was previously added for API backwards compatibility reasons, but now is a (counter-)duplicate to the return_slot_t.
Moving adaptor or proxy instances changes their `this` pointer. But `this` is captured by value in closures used by those instances, and this remains unchanged on move, leading to accessing an invalid instance when a lambda expression executes. Supporting move semantics would require unsubscribing/unregistering vtable, handlers, etc. and re-subscribing and re-registering all that, which is too complicated and may have side effects. Hence it has been decided that these classes are not moveable. One may use an indirection with e.g. `std::unique_ptr` to get move semantics.
Since sdbus-c++ knows types of D-Bus call/signal/property input/output parameters at compile time, why not assemble the D-Bus signature strings at compile time, too, instead of assembling them at run time as std::string with likely cost of (unnecessary) heap allocations...
std::array is well supported constexpr type in C++20, so we harness it to build the D-Bus signature string and store it into the binary at compile time.
Having explicit conversion operator is a good practice according to the C++ core guidelines, as it makes the code safer and better follows the principle of least astonishment. Also, it is consistent with the standard library style, where wrappers like std::variant, std::any, std::optional... also do not provide an implicit conversion to the underlying type. Last but not least, it paves the way for the upcoming std::variant <-> sdbus::Variant implicit conversions without surprising behavior in some edge cases.
This introduces strong types for `std::string`-based D-Bus types. This facilitates safer, less error-prone and more expressive API.
What previously was `auto proxy = createProxy("org.sdbuscpp.concatenator", "/org/sdbuscpp/concatenator");` is now written like `auto proxy = createProxy(ServiceName{"org.sdbuscpp.concatenator"}, ObjectPath{"/org/sdbuscpp/concatenator"});`.
These types are:
* `ObjectPath` type for the object path (the type has been around already but now is also used consistently in sdbus-c++ API for object path strings),
* `InterfaceName` type for D-Bus interface names,
* `BusName` (and its aliases `ServiceName` and `ConnectionName`) type for bus/service/connection names,
* `MemberName` (and its aliases `MethodName`, `SignalName` and `PropertyName`) type for D-Bus method, signal and property names,
* `Signature` type for the D-Bus signature (the type has been around already but now is also used consistently in sdbus-c++ API for signature strings),
* `Error::Name` type for D-Bus error names.
As of now sdbus-c++ supports C++20 but does not require it, and the used C++20 features are conditionally compiled depending on whether they are available or not.
A new CMake configuration variable SDBUSCPP_SDBUS_LIB has been introduced that enables clients to specify which sd-bus implementation library shall be used by sdbus-c++.
This PR makes things around connection factories a little more consistent and more intuitive:
* createConnection() has been removed. One shall call more expressive createSystemConnection() instead to get a connection to the system bus.
* createDefaultBusConnection() has been renamed to createBusConnection(), so as not to be confused with libsystemd's default_bus, which is a different thing (a reusable thread-local bus).
Proxies still by default call createBusConnection() to get a connection when the connection is not provided explicitly by the caller, but now createBusConnection() does a different thing, so now the proxies connect to either session bus or system bus depending on the context (as opposed to always to system bus like before).
The integration tests were modified to use createBusConnection().
In sdbus-c++ v1, new virtual functions (e.g. overloads of existing virtual functions) were always placed at the end of the class to keep backwards ABI compatibility. Now, with v2, these functions are reordered and functions forming a logical group are together.
This improves usability of sdbus-c++ in downstream CMake projects. CMake options now have `SDBUSCPP_` prefix so potential conflicts with downstream options/variables are minimized. Also, all configurable options and variables are placed in a single place in the root CMake file.
This switches from a raw pointer to std::optional type to pass prospective call errors to the client (using std::optional was not possible years back when sdbus-c++ was based on C++14). This makes the API a little clearer, safer, idiomatically more expressive, and removes potential confusion associated with raw pointers (like ownership, lifetime questions, etc.).
This makes D-Bus proxy signal registration more flexible, more dynamic, and less error-prone since no `finishRegistration()` call is needed. A proxy can register to a signal at any time during its lifetime, and can unregister freely by simply destroying the associated slot.
This improves the D-Bus object API registration/unregistration by making it more flexible, more dynamic, closer to sd-bus API design but still on high abstraction level, and -- most importantly -- less error-prone since no `finishRegistration()` call is needed anymore.
BrianandStanislav Angelovič
Karl Ljungkvist
Laurence Crestani Tasca
Marcel Hellwig
Stanislav Angelovic