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.
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.
There are currently only three values: Keep in Outbox, Send immediately
and Send after 5 minutes. It is partly related with the "flush-outbox"
option, but as that is used for filtering, I rather kept it untouched.
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.
CamelFolder holds a weak reference to its parent CamelStore, thus
if the store is freed before the folder, then the folder cannot
access it, which can lead to crashes.
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.
I considered replacing the "session" property with a "registry"
property, but that just complicates application startup even more.
Fact is, if we have a CamelStore then we can get the CamelSession
and even the ESourceRegistry from it. Kinda dirty, but works.
It goes a little something like this...
camel_service = CAMEL_SERVICE (camel_store);
camel_session = camel_service_get_session (camel_service);
mail_session = E_MAIL_SESSION (camel_session);
registry = e_mail_session_get_registry (mail_session);
Removed functions:
mail_folder_cache_get_session()
When the SASL mechanism reports that it can try an "empty password", give
it the opportunity to do so before we demand that the user provide one.
(Patch mostly by mbarnes; thanks)
This now uses e_source_get_oauth2_access_token_sync() instead of talking
directly to GNOME Online Accounts and therefore no longer has to live in
the "online-accounts" module.
The "online-accounts" module will be removed in Evolution 3.9.1, when we
can require GNOME Online Accounts 3.8. Evolution 3.8 will still support
older versions of GNOME Online Accounts, which use OAuth 1.0a tokens for
Google authentication.
Evolution consists of entirely too many small utility libraries, which
increases linking and loading time, places a burden on higher layers of
the application (e.g. modules) which has to remember to link to all the
small in-tree utility libraries, and makes it difficult to generate API
documentation for these utility libraries in one Gtk-Doc module.
Merge the following utility libraries under the umbrella of libeutil,
and enforce a single-include policy on libeutil so we can reorganize
the files as desired without disrupting its pseudo-public API.
libemail-utils/libemail-utils.la
libevolution-utils/libevolution-utils.la
filter/libfilter.la
widgets/e-timezone-dialog/libetimezonedialog.la
widgets/menus/libmenus.la
widgets/misc/libemiscwidgets.la
widgets/table/libetable.la
widgets/text/libetext.la
This also merges libedataserverui from the Evolution-Data-Server module,
since Evolution is its only consumer nowadays, and I'd like to make some
improvements to those APIs without concern for backward-compatibility.
And finally, start a Gtk-Doc module for libeutil. It's going to be a
project just getting all the symbols _listed_ much less _documented_.
But the skeletal structure is in place and I'm off to a good start.
Because we now check for junk filtering software in configure, junk
filtering modules can now safely assume the required junk filtering
software is installed. No more having the module installed but not
the required software. If this invariant is broken, an error will
be shown for each new message received.
camel_service_get_settings() is now camel_service_ref_settings()
and it returns a new CamelSettings reference which the caller must
release with g_object_unref().
These functions now return new references:
camel_session_add_service()
camel_session_list_services()
These functions have been renamed and also return new references:
camel_session_get_service() -> camel_session_ref_service()
camel_session_get_service_by_url() -> camel_session_ref_service_by_url()
Disabling the mail part of an online account through the Control Center
panel will now remove the CamelService from the EMailSession in addition
to disabling the account/identity/transport ESources, causing it to be
delisted from the account list in Preferences.
Furthermore, hide the Enabled check box for accounts linked to GOA in
Preferences. The collection ESource for these accounts can no longer
be disabled through Evolution; all such account manipulation must be
done through the Control Center panel.
Lastly, display an icon next to accounts linked to GOA in Preferences.
* Might be nice to show the actual provider icon instead of the generic
Online Accounts icon from the Control Center, but need to think about
how best to do that. Don't want a GOA dependency in core Evolution.
Maybe ESourceCollection should grow a GIcon property for the online-
accounts module in the registry service to set?
We're leaking CamelService references when we remove a CamelService
from a CamelSession. I don't yet know where or how.
If we remove a CamelService without finalizing the corresponding ESource,
and then add a new CamelService with the same UID, the ESource will have
a bidirectional "display-name" binding to multiple CamelService instances.
This creates an endless cascade of "notify" signals as soon as any of the
bound "display-name" properties change.
Until I can fix the leaking CamelService references, make the binding
one-way: ESource -> CamelService. This means the ESource's display name
is authoritative, and camel_service_set_display_name() MUST NOT be called
explicitly or else it will become out-of-sync with the ESource.
Use e_source_camel_generate_subtype() to register an ESourceCamelNone
class based on the newly-registered "none" CamelProvider.
e_source_camel_register_types() won't pick up the new provider if it was
already called prior to the provider being registered. That might be a
bug... I'm not sure yet. The Camel/ESource integration is kinda messy.
