* Since ObjectPath and Signature have a user-declared copy-ctor and copy
assignment operator the implicit declaration of corresponding move
operations is disabled. Explicitly add defaulted versions so that
move operations actually move instead of copy.
* See: https://github.com/Kistler-Group/sdbus-cpp/issues/230
fix timeout handling
* Despite what is documented in sd_bus_get_timeout(3), the timeout
returned is actually an absolute time point of Linux's CLOCK_MONOTONIC
clock. Hence, we first have to subtract the current time from the
timeout in order to get a relative time that can be passed to poll.
* For async call timeouts to reliably work, we need a way to notify the
event loop of a connection that is currently blocked waiting in poll.
I.e. assume the event loop thread entered poll with a timeout set to
T1. Afterwards, the main thread starts an async call C with a timeout
T2 < T1. In order for C to be canceled after its timeout T1 has
elapsed, we have to be able to notify the event loop so that it can
update its poll data.
Co-authored-by: Urs Ritzmann <ursritzmann@protonmail.ch>
Co-authored-by: Lukasz Marcul <lukasz.marcul@onemeter.com>
Allow to use human readable chrono literals to specify method call
timeout. The change is backward compatbile - if no unit is provided,
the fallback is "us".
Example:
<annotation name="org.freedesktop.DBus.Method.Timeout" value="500ms"/>
If -DBUILD_TESTS=ON, then creating debian packages fails
because dpkg-shlibdeps obiously can't find the sdbus-c++ dependency
for the tests package.
We could disable CPACK_DEBIAN_PACKAGE_SHLIBDEPS for the tests package.
Nevertheless I don't think we want a tests package at all.
Even though they have the same value, they are something fundamentally different.
Therefore it is extremely confusing if the constant INTERFACE_NAME is passed
where actually a well-known BUS_NAME (destination) should go.
This change addresses conditions where an exception is thrown by the library upon receipt of a malformed message from an external
source, and propagated up to the event loop with no chance of
interception by the application. This issue is only experienced by
proxy convenience APIs, as low-level APIs allow the application to
unpack the message.
Strategy:
1. For malformed signals received by proxies: ignore the signal.
2. For malformed async method responses, translate the unpack
exception into an sdbus::Error, and pass it to the caller as expected.
This internally calls sd_bus_open(), which automatically selects the
system or session bus connection based on the presence and
content of a DBUS_STARTER_BUS_TYPE environment variable and
whether the calling process has root privileges.
This option is very helpful when creating services and clients that will use the system bus in production, but connect to a session
for testing.
Additional changes:
* Removed assertions null-checking make_unique() return values.
make_unique() calls new, and new is expected to throw or abort
on failure, making the assertions unhelpful.
* Corrected a typo in the ClosesAndUnrefsBusWhenDestructed
unit test for the system bus (tested the wrong function).
* Add missing nullptr check in Message::peekType().
* According to its specification, sd_bus_message_peek_type() sets a
given contents parameter to NULL if the next element in a message is
not a container. Since assigning a nullptr to a std::string has
undefined behaviour (typically resulting in an invalid memory access),
Message::peekType() must not assign contentsSig unconditionally.
* fix construct/destruct order of mutex/data in AsnycCalls
Destroying a mutex which might still be owned leads to undefined behaviour.
It wasn't the case here but using the right order is more resistant to future bugs.
* proxy AsnycCalls: don't leave map in unspecified state
* fix 133: use thread-safe container for proxy interfaces list
* proxy callback payload uses concrete signalData
* proxy: revert adding the wrapper class InterfaceContainer
It's no longer required and simplifies the locking logic.
* Proxy: avoid additional lambda wrapper around signal callbacks
As proposed in
https://github.com/Kistler-Group/sdbus-cpp/pull/177#issuecomment-834439707
option 3.
Still TODO: Avoid relying on std::map's non-invalidatable by adding just
std::unique_ptr<Proxy::InterfaceData::SignalData> to the container.
* proxy: add missing underscore prefix
* proxy: put InterfaceData::SignalData into a unique_ptr
This ways, we avoid relying on std::map's non-invalidatable references.
* proxy: code style: get raw pointer directly
* style: fix code style
Co-authored-by: Stanislav Angelovic <stanislav.angelovic@siemens.com>
* Add API to get message path
* Add API to get message destination
* Handle NULL message fields interface and member
Functions sd_bus_message_get_interface() and sd_bus_message_get_member()
can return null in case the message does not use those fields or does
not have them set.
This provides access to the proxy's bus connection so code using
the proxy does not need to store an external reference to it.
A matching function is already available in IObject.
