As with all D-Bus signal subscriptions, it’s possible for a signal
callback to be invoked in one thread (T1) while another thread (T2) is
unsubscribing from that signal. In this case, T1 is the main thread, and
T2 is the D-Bus connection worker thread which is unsubscribing all
signals as it’s in the process of closing.
Due to this possibility, all `user_data` for signal callbacks needs to
be referenced outside the lifecycle of the code which
subscribes/unsubscribes the signal. In other words, it’s not safe to
subscribe to a signal, store the subscription ID in a struct,
unsubscribe from the signal when freeing the struct, and dereference the
struct in the signal callback. The data passed to the signal callback
has to have its own strong reference.
Instead, it’s safe to subscribe to a signal and add a strong reference
to the struct, store the subscription ID in that struct, and unsubscribe
from the signal when the last external reference to your struct is
dropped. That unsubscription should break the refcount cycle between the
signal connection and the struct, and allow the struct to be completely
freed. Only with that approach is it safe to dereference the struct in
the signal callback, if there’s any possibility that the signal might be
unsubscribed from a separate thread.
The tests need specific additional main loop cycles to completely emit
the NameLost signal callback. Ideally they need refactoring, but this
will do (1000 test cycles passed).
Signed-off-by: Philip Withnall <withnall@endlessm.com>
Fixes: #978
Calling g_variant_get (parameters, "(&s)") when parameters has a
signature other than (s) is considered to be a programming error.
In practice the message bus (dbus-daemon or a reimplementation) should
always send the expected type, but be defensive.
(Modified by Philip Withnall to improve type check.)
Signed-off-by: Simon McVittie <smcv@collabora.com>
Signed-off-by: Philip Withnall <withnall@endlessm.com>
https://bugzilla.gnome.org/show_bug.cgi?id=784392
If we have an input parameter (or return value) we need to use (nullable).
However, if it is an (inout) or (out) parameter, (optional) is sufficient.
It looks like (nullable) could be used for everything according to the
Annotation documentation, but (optional) is more specific.
The default GNOME 3.10 login process right now has gdm spawn a session
for the login screen, retaining the X server, but closing the session
bus. Right now in this scenario many GNOME components such as
gnome-settings-daemon attempt to "clean up" on shutdown by releasing
their owned names.
But they're shutting down because the session bus went away, so
releasing the name is pointless, and presently spews an error into the
journal.
This patch avoids that error spew, which helps system administrators
find *real* problems.
https://bugzilla.gnome.org/show_bug.cgi?id=704567
Add g_main_context_ref_thread_default(), which always returns a
reffed GMainContext, rather than sometimes returning a (non-reffed)
GMainContext, and sometimes returning NULL. This simplifies the
bookkeeping in any code that needs to keep a reference to the
thread-default context for a while.
https://bugzilla.gnome.org/show_bug.cgi?id=660994
To help cross compilation, don't use glib-genmarshal in our
build. This is easy now that we have g_cclosure_marshal_generic().
In gobject/, add gmarshal.[ch] to git (making the existing entry
points stubs).
In gio/, simply switch to using g_cclosure_marshal_generic().
https://bugzilla.gnome.org/show_bug.cgi?id=652168
The GClosure API is a bit funky (and badly documented), and requires
you to set a marshaller on the closure, and the marshaller has an
implicit 'this' argument, and the caller is reponsible for unsetting
the values after invoking the closure.
I've added some calls of the _with_closures variants to the
gdbus-names test now.
This is currently unused but will probably be useful in the
future. For example, we could have a _ARG0_IS_PATH to specify that
arg0 should be used for arg0path.
This commit breaks API and ABI. Users of
g_dbus_connection_signal_subscribe() will need to port to this new
version.
Signed-off-by: David Zeuthen <davidz@redhat.com>
There's a couple of places in GDBus where it's a programming error
(e.g. we'll assert or spew via e.g. g_warning()) to use the API on a
closed connection. This approach can never work since a
GDBusConnection can be closed at any point in time outside of
programmer control.
Just change the code to return a run-time error (e.g. return
G_IO_ERROR_CLOSED when sending messages, invoking methods) or silently
accept the request (e.g. exporting objects, registering for signals)
without doing anything.
Note that a GDBusConnection object is always useless after being
closed - e.g. there's no way to "reopen" a connection - the user will
have to create a new object and use that instead.
https://bugzilla.gnome.org/show_bug.cgi?id=623143
Signed-off-by: David Zeuthen <davidz@redhat.com>
accept more than one callback.
g_bus_own_name_with_closures
g_bus_own_name_on_connection_with_closures
g_bus_watch_name_with_closures
g_bus_watch_name_on_connection_with_closures
g_bus_watch_proxy_with_closures
g_bus_watch_proxy_on_connection_with_closures
https://bugzilla.gnome.org/show_bug.cgi?id=621092
This allows the caller to specify the reply type that they are expecting
for this call. If the reply comes back with the wrong type, GDBus will
generate an appropriate error internally.
- add a GVariantType * argument to g_dbus_connection_call() and
_call_sync().
- move the internal API for computing message types from introspection
data to be based on GVariantType instead of strings. Update users
of this code.
- have GDBusProxy pass this calculated GVariantType into
g_dbus_connection_call(). Remove the checks done in GDBusProxy.
- Update other users of the code (test cases, gdbus-tool, GSettings
tool, etc). In some cases, remove redundant checks; in some other
cases, we are fixing bugs because no checking was done where it
should have been.
Closes bug #619391.
Lots of people been suggesting this. We still use MethodInvocation /
method_invocation for handling incoming method calls so use call()
instead of invoke_method() helps to separate the client and server
facilities. Which is a good thing(tm).
Things compile and the test-suite passes. Still need to hook up
gio.symbols and docs. There are still a bunch of TODOs left in the
sources that needs to be addressed.
Signed-off-by: David Zeuthen <davidz@redhat.com>
