GitLab

Closes GNOME/evolution-data-server#65

Without it the process cannot own the D-Bus name, thus it doesn't start.

Closes GNOME/evolution-data-server#64

Closes GNOME/evolution-data-server!6

Closes GNOME/evolution-data-server#63

That might be the same as having set NULL authentication method.

Related to GNOME/evolution-ews#27

Closes GNOME/evolution-data-server#61

Couple changes in the test:
1) add also a component into the calendar, thus the returned
   Free/Busy component is not empty
2) let the sync variant wait for the D-Bus signal with Free/Busy
   data again
3) wait for the D-Bus signal in case it's not received in time
   of the method execution finish.

Especially the 3) is important, because the D-Bus signal can be delivered
after the method invocation is received, which could break the test.

The test could fail when creating the instance, not when opening it,
thus retry also when the instance creation fails. Left the code
for the open() too, just in case.

As the passwords are saved in the host system, the eds-origin key cannot
use the same value as the host system, because otherwise the secret-monitor
module of evolution-source-registry would delete the password, because
the ESource is unknown to it. Similarly the other way around, the inner
(in Flatpak) source registry would delete system passwords, because those
ESource-s are unknown to it.

This runtime warning had been shown when running Evolution in Flatpak
with GNOME 3.30 runtime. It had been due to libical not finding
tzdata file in the expected directory. Let's relax the runtime
warning, because it can happen the time zone detection fails, even
in a corner case.

This is needed, because some API functions use these structures directly, which is
a problem when those are used through the introspection code, because the boxed type
made a copy of the structure, instead of using the original pointer. This ref/unref
detail is not exposed to the C API.

Closes GNOME/evolution-data-server#62

Even the compiler thought the 'text' can be used uninitialized,
it wasn't the case, but initialize it anyway, to make it happy.

Related to GNOME/evolution-data-server#56

The option is still meant to be used in Flatpak only, to prefix
the D-Bus services with certain prefix (usually the same as the Flatpak
application ID), thus the services do not clash with those installed
in the host system. It fixes many issues in evolution-dbus-session.

Related to GNOME/evolution#165

Closes https://bugzilla.gnome.org/show_bug.cgi?id=651197

When a filter rule contains action to copy message, this copy cannot be
batched as message move, because the copy doesn't stop further processing,
thus the to-be-copied message can be further modified or even removed,
thus the copy would not be correct.

This had been reported downstream:
https://bugzilla.redhat.com/show_bug.cgi?id=1650665

Missed filter's search code, which didn't know about the new
compare-date and make-time functions, causing errors int he UI
when filtering.

Related to GNOME/evolution-data-server#5
and GNOME/evolution#245

When a component/contact is loaded from the server when it's not part
of the local cache (which can happen for example with meeting invitations),
then the fetched object was stored without the href in the local cache,
which made it impossible to edit the object later (it resulted
in an "Invalid Object" error).

There could happen a deadlock when two threads change for example
message info flags, where one threads holds message info's summary lock
and is waiting for the info's property lock, which the other thread
hold's info's property lock and is waiting for the summary lock.
It's safer to hold summary lock before acquiring the message info's
lock when anything will be changed in the message info, which didn't
happen here.

This had been reported downstream at:
https://bugzilla.redhat.com/show_bug.cgi?id=1649373

Closes GNOME/evolution-data-server#56

Closes GNOME/evolution-data-server#55

Handle also cases where the fdlist is only a return value, not an input
argument.

Related to GNOME/evolution#169

This is required to have properly propagated calls like OpenFile,
which use it.

Related to GNOME/evolution#169

...and address the newly shown warnings as well.

The test had been mostly working, except cases when the delivery
of a D-Bus property change signal had been delayed. Then the value
on the client side was not the same as the value on the server side.
There's a very similar test for EBookClient, which does slightly more
things, which gives D-Bus enough time to deliver the property change
to the client, thus this change adds the component removal to address it.

Closes GNOME/evolution-data-server#54

Thus it can be used to influence where the needed libraries are
looked for from the outside.

The only two parts using it were GDBus objects for address book and
calendar view, which could be easily transformed to use gdbus-codegen,
the same as the rest of the GDBus objects defined in src/private/.

Related to GNOME/evolution#113

Closes GNOME/evolution-data-server#52

Closes GNOME/evolution-data-server#41

Closes GNOME/evolution-data-server#51

(cherry picked from commit 37e582ad)

Similar to GNOME/evolution#196
when intltool-merge is called in parallel, it could either rewrite
the ongoing attempt to build it or use an incomplete data, which
results in broken output files (.desktop, .metainfo and so on).
This change ensures the intltool-merge cache is created only once
and any other requests which would use it will wait until it's created.