popupMenu: Adjust submenu arrows to RTL icon changes
Since commit e04e507659027, we will already get the right icon for the
submenu arrow, so we must not mirror it again. However we do need to
take the text direction into account for the rotation now (but that's
not actually too bad - the resulting code gets quite a bit easier).
appDisplay: Only allow ctrl-click for apps that can open new windows
Trust the heuristics in shell_app_can_open_new_window() to get it right
more often than not, and add an appropriate check in activate(). This
makes the behavior consistent with the dash, e.g. we will try to open
a new window (and show the corresponding animation) for apps that don't
have a "New window" item in their dash context menu.
Carlos Soriano [Sun, 7 Sep 2014 08:22:26 +0000 (10:22 +0200)]
iconGrid: Make actors reactive on animating out as well
Commit 9e5704b introduced a regression due to a sloppy mistake not
making actors reactive on animation out.
Fix that making actors reactive when animating out as well.
extensionPrefs: Improve handling of OutOfDate extensions
Our current UI is horrible in dealing with outdated extensions - we
don't give users any indication at all that something is not working as
expected (unless hiding the prefs button as we did until the last commit
counts).
To fix this, make the enable switch insensitive to indicate OutOfDate
extensions.
extensionPrefs: Remove version check from extensionAvailable()
It is confusing to treat outdated extensions as if they didn't exist
in some places, but like any other extensions elsewhere. In particular
gnome-tweak-tool still allows to launch prefs for them while making
it clear to users that the extension will not work - opening the
list of installed extensions in that case is unexpected and confusing.
Just remove the version check for now, we will soon follow tweak-tool
and use it to disable the extension switch instead.
extensionSystem: Fix reloading on version-validation changes
The current code only works for enabled extensions, which means
that extensions that were marked OUT_OF_DATE cannot be enabled
without a restart when disabling the version check.
Fix this by reloading all extensions while making sure to only
enable any extensions when we're supposed to.
Carlos Soriano [Sat, 6 Sep 2014 10:47:33 +0000 (12:47 +0200)]
iconGrid: Fix last item check on spring animation
Currently we consider the last item animating being the last actor of
the list, but that's wrong since the last item animating depends on the
distance between the source actor and the original position of the
animated actor.
Take that into account to check the last item of the animation.
Owen W. Taylor [Thu, 4 Sep 2014 14:01:52 +0000 (10:01 -0400)]
Fix problem when background settings change multiple times in rapid succession
The code that cleaned up when a background actor was destroyed tried to access
the Background as backgroundActor.background._delegate but when the destroy
signal is emitted, the MetaBackgroundActor was already disposed and
backgroundActor.background NULL.
Owen W. Taylor [Mon, 11 Aug 2014 16:15:45 +0000 (18:15 +0200)]
Adapt to Mutter background changes
The rewrite of Mutter's background code (see bug 735637) requires
corresponding changes here - we no longer need to layer multiple
MetaBackgroundActors together.
The general strategy is that a BackgroundSource object is created
per GSettings schema, and keeps either one Background/MetaBackground pair,
or, for animation, a Background/Metabackground pair for each monitor.
popupMenu: Force submenu arrow rotation angles to multiples of 90
When you open or close a PopupSubMenu the arrow icon rotates, but
the code previously assumed that the starting angle was a multiple
of 90. If you click the submenu fast enough the rotation could overlap
with the previous rotation and you'd get something that wasn't a multiple
of 90.
Now we ensure that the ending arrowRotation angle is always a multiple
of 90 regardless of what the starting angle is.
Carlos Soriano [Tue, 2 Sep 2014 20:49:10 +0000 (22:49 +0200)]
iconGrid: Make actors non reactive while animating
Although the actor are with opacity 0, they were reactive, so the if the
user quickly make a click on the grid while animating it can launch an
app accidentally.
To avoid that make the actors non reactive while animating.
David King [Tue, 23 Jul 2013 15:23:10 +0000 (16:23 +0100)]
keyboard: Monitor the ShowOSK property of g-s-d
Show the on-screen keyboard if gnome-settings-daemon requests that it
should be shown, such as when using a touchscreen input device. Do this
in addition to monitoring the accessibility setting for the OSK.
Florian Müllner [Sun, 31 Aug 2014 00:20:43 +0000 (02:20 +0200)]
overviewControls: Use correct translation in dragBegin
Commit 12845f0eefa reset translationX to 0 to fix the slide-in
animation of the dash when dragging a search result, but 0 is the
wrong value for the thumbnail slider - the next time it will be
shown, it will just pop up instead of sliding in due to translationX
already being at its target value.
Fix this by making _updateTranslation take the actual visibility into
account and call that to set the correct translation in dragBegin.
Florian Müllner [Sun, 31 Aug 2014 00:18:51 +0000 (02:18 +0200)]
overviewControls: Fix initial slide-in animation
Controls are slided in by animating translationX from the actor's
width to 0; however as _updateTranslation() will skip the animation
when the property is already at its target value and 0 happens to be
the initial value of translationX, the initial animation is skipped.
Fix this by initializing translationX to undefined, which will always
differ from a valid target value.
Florian Müllner [Sat, 30 Aug 2014 23:16:00 +0000 (01:16 +0200)]
overviewControls: Don't update translation on slideIn()
The comment is right, updating the translation should be deferred
to pageEmpty, or else overview controls will just pop up instead
of sliding in. So revert that part of commit 6a7fa52879d53.
Florian Müllner [Thu, 21 Aug 2014 10:17:46 +0000 (12:17 +0200)]
search: Remove DND code from GridSearchResult
Providers that need drag-and-drop behavior can implement this via the
createResultObject() hook (as the app search provider already does), no
need to duplicate that code in the generic result objects
(ListSearchResult already does not implement DND).
Carlos Soriano [Tue, 17 Jun 2014 19:31:53 +0000 (21:31 +0200)]
appDisplay: Animate appIcon for new window of apps
Following design mockups, animate the icons on AllView, FrequentView,
Dash and Search to zoom out when opening a new window of the app or when
the app is not running and the user execute it.
Carlos Soriano [Tue, 17 Jun 2014 17:10:54 +0000 (19:10 +0200)]
appDisplay: Animate AllView and FrequentView
Following design decision, we want to animate AllView and FrequentView
when opening and closing with a swarm spring form.
This involves a few changes needed to allow that, since from some time
now, we are animating page changes in viewSelector, using only a fade
transition. However now we want to let appDisplay and iconGrid apply its
own animation.
For that we special case the change to and from apps page on
viewSelector to let appDisplay to animate its own items, using and API
on appDisplay which at the same time uses an API on iconGrid.
Carlos Soriano [Tue, 17 Jun 2014 10:47:00 +0000 (12:47 +0200)]
appDisplay: Animate folder view items
Add a new animation to folder view based on designers mockups that
emulates pulsating icons.
The code on iconGrid is though to work well for the upcoming patches to
animate AllView and FrequentView.
If you turn on natural scrolling, instead of swiping up, you need to
swipe down, which is bizarre. Just accept any type of scroll and have
them contribute to the delta.
This commit didn't make the OSK fully operational yet on Wayland, and
caused the double emission of key events on X11 due to the OSK keys
receiving first touch events from the passive touch grab, and then
emulated pointer events from event selection after the touch sequence
was rejected in the grab.
When we make a better effort at handling touch events just once on X11,
this commit can be reapplied and remaining wayland OSK support resumed
from there. In the mean time, this patch is better reverted.
Florian Müllner [Thu, 28 Aug 2014 17:43:05 +0000 (19:43 +0200)]
grabHelper: Adjust to property name change
Commit bbfa616f275cdd41c052 renamed _ignoreRelease to _ignoreUntilRelease
in some places, but not others, which broke GrabHelper.ignoreRelease().
Complete the name change to fix the fallout (e.g. app launcher menus
closing on button release).
Florian Müllner [Thu, 28 Aug 2014 15:27:00 +0000 (17:27 +0200)]
popupMenu: Use an appropriate key to move focus into the menu
The code from PanelMenu.Button assumed menus would open below their
source actor, making KEY_Down a good choice; however with the new
generic code, we should base the key used on the actual menu position.
Florian Müllner [Thu, 28 Aug 2014 15:04:23 +0000 (17:04 +0200)]
popupMenu: Handle key-press events on sourceActor
The behavior of opening/closing/navigating a menu from its source
actor is generic enough to not limit it to PanelMenu.Buttons, so
move the code into PopupMenu itself.
Florian Müllner [Thu, 28 Aug 2014 14:22:49 +0000 (16:22 +0200)]
workspace: Always remove removed windows
Since mutter commit 527c53a2a0582eba, MetaWorkspace::window-removed
is emitted *before* MetaWindow:workspace is updated, so the test
whether the removed window should still be on the workspace in
question will always return true.
Assume the test is no longer necessary nowadays to fix this very
obvious regression.
Carlos Soriano [Tue, 26 Aug 2014 17:10:02 +0000 (19:10 +0200)]
workspace: Remove tweens of overlay.title before animate again
Currently we are removing tweens of the button and border, but not from
the title. That causes the title to be in wrong positions sometimes when
dragging windows on the overview, given that the slider is moving and
therefore the windows + overlay are moving too.
To avoid that, remove tweens of the title as well.
Owen W. Taylor [Mon, 25 Aug 2014 13:48:19 +0000 (09:48 -0400)]
perf: fix wallpaper style for overridden background
The code that loads SHELL_BACKGROUND_IMAGE, which is used to load the
performance background was loading it in WALLPAPER mode, not ZOOM
mode. Zoom mode is what we use for the actual GNOME defaultiwallpaper
and what we want to test: the background will be scaled except when
the resolution matches the 2560x1440 default backgrounds.
location: Only show menu when geolocation is in use
Users can now toggle geolocation off/on from privacy panel of
gnome-control-center so we don't need to clutter the menu with a
settings that most users won't touch most of the time.
Florian Müllner [Sun, 17 Aug 2014 17:44:52 +0000 (19:44 +0200)]
iconGrid: Move code around a bit
The scale passed to _updateChildrenScale is exclusively computed from
properties, so it can be moved into the function itself - as the scale
now becomes a mere detail, rename to a more appropriate _updateIconSizes
at the same time.
Florian Müllner [Tue, 19 Aug 2014 16:57:36 +0000 (18:57 +0200)]
search: Fix custom drag actor source
The existing code broke when commit 792b963bda8c changed the custom
result actor hook to return an object instead of an actor - stop
trying to go through a _delegate to make it work again.
Florian Müllner [Thu, 16 Jan 2014 12:41:46 +0000 (07:41 -0500)]
dateMenu: Try to use the default calendar application
Commit 14ceb10555699 changed the "Open Calendar" item to open the
"recommended" calendar application rather than the default one to
avoid problems with MIME subclassing (namely falling back to the
default text editor when no calendar app is installed).
With this change however, the application launched does no longer
necessarily match the one configured in Settings, which is unexpected.
To avoid both problems, use the default calendar application again,
but only if it is in the list of recommended applications.
We need to put the actual actors in the history, not just the labels,
otherwise all emptyLine (which are not messages but are not empty
either) and all lines with a timestamp will get stuck in the scrollback.
PortalHelper: fix portals that don't redirect properly
We assumed that either a portal would give us a target URI, or we
would eventually reach the URI of our choice (which is http://www.gnome.org)
This is not always the case: even after a successful login we may
stay in a confirmation page, so make sure that every time we see a
redirect we trigger a connectivity check (or queue one for when
the user closes the window)
search: add support for default disabled search providers
Search providers that should be disabled by default come with
a DefaultDisabled=true key in their keyfile, and are enabled
with the "enabled" whitelist, not with the "disabled" blacklist.
Owen W. Taylor [Mon, 11 Aug 2014 12:35:52 +0000 (14:35 +0200)]
Add a special background to use for performance testing
Performance testing was producing inconsistent values at different
times in the day since the GNOME default background is animated and
sometimes has a single layer, and sometimes two blended layers.
So we have consistent numbers, install a simple animated background
with GNOME Shell that has 40-year long transition ending in 2030,a
and set an environment variable in gnome-shell-perf-tool so that the
background is override with that background. (The background depends
on files installed by gnome-backgrounds; we assume that the person
running performance tests is doing so within the scope of a full
GNOME install.)
Remove 'backwards' argument from SwitcherPopup:_keyPressHandler
All derived classes are already checking explicitly for action names
(FOO and FOO_BACKWARDS). mutter used to have a META_KEY_BINDING_REVERSES
flag for keybindings which required special handling of "shift"+FOO as
FOO_BACKWARDS, but this has been removed now, so this special handling
is no longer necessary.
Stop using Meta.KeyBindingFlags.REVERSES for IM switch keybinding
When this flag is set on a MetaKeyBinding, mutter will know that
the keybinding has an associated reverse keybinding triggered with
the shift modifier. However, an undesirable side-effect is that
gnome-control-center keyboard panel does not know that this 'shift'
is reserved for these reverse keybindings and cannot detect
conflicting bindings in this case.
This 'reverse' logic can now be handled at a higher level (in gcc keyboard
panel) so this commit removes it from gnome-shell so that they do not
conflict.
Now that mutter gives a way to check if a MetaKeyBinding was marked as
'reversed' or not, gnome-shell does not have to hardcode that a
MetaKeyBinding using a shift modifier is reversed, it can directly check
if the appropriate flag is set.
Having the on/off setting be backed by a boolean in dconf makes sense
anyway but this is mainly to be able to remember the max accuracy set
before user disabled geolocation so that when they enable it next time,
we have the max accuracy level on same value as before.
There hasn't been a real need for this but now we are about to add
geolocation settings in control center and it'll be easiser for
control-center to simply toggle a boolean property rather than to have
to know about and deal with accuracy levels.
Later we might also want to add accuracy level settings to privacy panel
so keeping the accuracy level setting around still. However we no longer
support 'off' accuracy level as the new boolean property covers that.
This also implies that we no longer track available accuracy level,
which made the code a bit hard to follow/maintain.
projects / gnome-shell.git / log