There could happen a busy loop on a Month view select, when the week
starts on Sunday was set. The date range clamp made the first day Sunday,
but the Month view itself corrected it to Saturday (it doesn't split
weekends), this was noticed and the time range was recalculated
again, by six days back. Then the round repeated down to 1970, till
a runtime critical warning wasn't hit.
Since this change the client is responsible to provide credentials
to use to authenticate backends (through ESource-s, to be more precise),
unless the credentials are already saved.
There is a fake change notify on timezone when either a corresponding
file changes or the setting for "use system timezone" changes,
to propagate the change in the code properly. This notify requires
proper key identification for GSettings bindings, because without it
the binding crashes on a runtime check.
A simple Evolution run and move between all views means creation of
more than 100 GSettings objects, with only a bit more than 10 schemas.
Reusing the objects should have a positive impact on a performance too.
This is related to an ESourceSelector, which checks ESource's
selected state when it is removed or disabled, but the corresponding
row in the selector could be already gone, which could produce
a runtime warning about invalid GtkTreeRowReference.
This was a remaining thing from the 'Make calendar views non-UI-blocking'
work, to show progress of views in UI. This is currently done by a spinner
beside source's name in the ESourceSelector and a tooltip above that row.
The Calendar, Memos and Tasks views use to do D-Bus calls to
the backends on the main (UI) thread, which could result in UI
freezes, until the operation was done on the backend (and server)
side. This commit fixes that by invoking the operations in
a dedicated thread. It has few additional advantages too:
- operations can be cancelled
- proper error reporting to a user
- less code duplication between the views for common operations
There had been fixed some performance issues when selecting/unselecting
sources in the source selector as well.
The icons at the search bar, to search forward, backward and stop
searching were too large, which didn't look good. This makes them
smaller, though even here can be seen a little gap around
the images in the buttons which might not be there ideally.
Similar to GObject::notify, the GSettings::changed can be emitted
even if a key didn't change. It's up to the user (aka evolution)
to test for real changes, thus let's do it. It may have certain
performance positive impact too.
This is related to bug 698275, which did not cover all cases.
The problem here is that the dconf can in certain situation claim
that everything changed (path "/" changed), which GSettingsBinding
propagates to a GObject property unconditionally and GObject's
property setter (g_object_set_property()) also notifies about
the property change unconditionally, despite the real descendant
property setter properly checks for the value change. After all
these false notifications a callback on "notify" signal is called
and possibly an expensive operation is run.
Checking whether the value really changed helps in performance, for
which were added new e-util functions:
e_signal_connect_notify()
e_signal_connect_notify_after()
e_signal_connect_notify_swapped()
e_signal_connect_notify_object()
which have the same prototype as their GLib counterparts, but they allow
only "notify::..." signals and they test whether the value really changed
before they call the registered callback.
This ports the following two function calls throughout Evolution:
• e_categories_get_list() to e_categories_dup_list()
• e_categories_get_icon_file_for() to e_categories_dup_icon_file_for()
It necessarily changes some internal e-util API:
• e_util_get_searchable_categories() to
e_util_dup_searchable_categories()
This bumps the EDS requirement to 3.13.1.
https://bugzilla.gnome.org/show_bug.cgi?id=727221
This makes the code free of Coverity scan issues.
It is sometimes quite pedantic and expects/suggests some
coding habits, thus certain changes may look weird, but for a good
thing, I hope. The code is also tagged with Coverity scan
suppressions, to keep the code as is and hide the warning too.
Also note that Coverity treats g_return_if_fail(), g_assert() and
similar macros as unreliable, and it's true these can be disabled
during the compile time, thus it brings in other set of 'weird'
changes.
EShellView no longer needs help from subclasses other than getting
the needed GalView subclasses registered.
A nice side-effect of this is EShellView subclasses can now use the
G_DEFINE_DYNAMIC_TYPE macro.
No longer needed. Instead, use g_type_ensure() to ensure the necessary
GalView subclasses are registered in the GType system before loading a
GalViewCollection. Best place to ensure types is from GClassInitFunc.
