| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
| |
We can now use this to switch input sources so there's no reason to
not convert it.
|
|
|
|
|
|
|
| |
The layout we want is always in the first XKB group index so we should
enforce it to make sure we never end up with the wrong one.
https://bugzilla.gnome.org/show_bug.cgi?id=696996
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The only way for external components to activate an input source is by
setting the gsettings key. That's not optimal since external
components then have no way of knowing when exactly the switch is
completed.
This commit introduces a DBus method to set an input source and we
make sure to only return the method after all the changes have been
made, i.e. the XKB keyboard description has been changed and the IBus
engine (if any) has been activated.
E.g. gnome-shell can use this method and freeze keyboard events in the
X server until it hears back from g-s-d to ensure that events won't be
misinterpreted after an input source switch.
https://bugzilla.gnome.org/show_bug.cgi?id=696996
|
|
|
|
|
|
|
| |
For now we'll just claim the name and export an empty interface. We'll
grow the interface as needed.
https://bugzilla.gnome.org/show_bug.cgi?id=696996
|
|
|
|
|
|
| |
Prevents potential uses after free.
https://bugzilla.gnome.org/show_bug.cgi?id=696996
|
| |
|
| |
|
|
|
|
|
|
| |
For both vertical and horizontal scrolls.
https://bugzilla.gnome.org/show_bug.cgi?id=651134
|
| |
|
|
|
|
|
|
| |
We need stdlib.h to get the definition of NULL.
https://bugzilla.gnome.org/show_bug.cgi?id=696515
|
| |
|
|
|
|
|
|
|
|
| |
Hide deprecation warnings by definition of
_PPD_DEPRECATED="" because of missing replacement for
deprecated function ppdLocalizeIPPReason().
https://bugzilla.gnome.org/show_bug.cgi?id=696571
|
|
|
|
|
|
|
|
|
|
|
|
| |
The intent of the virtual machine detection was to avoid changing
power states; however, while we avoided some parts of the code, we
still ended up flipping the state variable to which other parts of the
code reacted.
With this patch, in a VM, we remain GSD_POWER_IDLE_MODE_NORMAL
permanently.
https://bugzilla.gnome.org/show_bug.cgi?id=696500
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
gnome_idle_monitor_new_for_device() can now return NULL on
errors, such as when per-device idletime counters aren't available.
Do error checking, on startup, for the creation of per-device
idletime monitors. Note that we don't do error checking later
in the run-time as this should be the only the only possible
failure, bar system problems.
https://mail.gnome.org/archives/release-team/2013-March/msg00289.html
https://bugzilla.gnome.org/show_bug.cgi?id=696118
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
When the shell crashes while using the screensaver (or it needs to be
killed because it won't remove a black screen), we checked for the
state of the screensaver, but didn't do anything with it.
Now we use the same codepath and make sure we disable the screensaver's
aggressive timeout if gnome-shell isn't in screensaver mode anymore.
https://bugzilla.gnome.org/show_bug.cgi?id=696290
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
NOTE: This is just the default value, and our tools only ever read the first
Unicode entry. When applications are modified to display the correct translation
for the locale, it'll be important to match the correct language and country values.
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
| |
Switching the gtk+ IM module when switching input source adds delays
and races with the input events themselves. Thus it's better to always
use the ibus module as long as the user has at least one ibus input
source.
https://bugzilla.gnome.org/show_bug.cgi?id=696141
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We set XMODIFIERS=@im=ibus globally at session start so that XIM
clients can immediately benefit from IBus input sources when/if the
user adds one. Unfortunately, these clients then won't work correctly
in case we don't actually activate the ibus daemon. E.g. dead keys
don't work in this case.
Also, previously we'd try to set an "xkb:..." ibus engine according to
the XKB layout. It happens that this wouldn't work for most layouts
since there isn't an ibus engine for each layout and it's actually not
needed since all of those "xkb:..." engines work the same way despite
having different names.
As such we'll always try to activate the daemon even if none of the
user's current input sources need it and set one of the trivial "echo"
engines so that XIM clients work correctly.
https://bugzilla.gnome.org/show_bug.cgi?id=695867
|
|
|
|
|
|
| |
Otherwise we fail to send method calls.
https://bugzilla.gnome.org/show_bug.cgi?id=696114
|
| |
|
| |
|
|
|
|
| |
Signed-off-by: Gheyret Kenji <gheyret@gmail.com>
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
Some users might have XKB options configured in their xorg.conf. In
case the current user's setting is empty or when running under GDM we
should honor the system options.
https://bugzilla.gnome.org/show_bug.cgi?id=694877
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We used to get the current X server's XKB layouts and variants from
the X property when populating empty settings or running under
GDM. This means that the g-s-d instance running under GDM would then
change the X server's configuration (as a side-effect of applying the
input source) and thus, after a freshly created user account logged
in, the user session's g-s-d instance would no longer see the
system's xorg.conf layouts and variants but instead the ones that were
last active in GDM.
Using localed to get layouts and variants will allow us to always get
what we actually want i.e. the system settings.
https://bugzilla.gnome.org/show_bug.cgi?id=694877
|
| |
|