This doesn't (or shouldn't) change the visual appearance of the fade
effect, but does do all the testing math inside the shader, rather
than on the CPU. This will make fading the offset much easier in
the future.
https://bugzilla.gnome.org/show_bug.cgi?id=689249
GLSL 1.20 is a better language, and we'll rely on it in future updates.
This doesn't have any additional constraints, since GLSL 1.20 was
standardized before GLSL-supporting drivers came out.
https://bugzilla.gnome.org/show_bug.cgi?id=689249
Theme nodes are interned and shared with other widgets, so they cannot
be disposed, otherwise we blow useful resources, and in particular we
break the parent-child chain.
https://bugzilla.gnome.org/show_bug.cgi?id=689029
We where not setting a duration for the frames, which caused the resulting videos to
have a broken header.
Fix that by making the source a live source and setting the correct frame duration.
https://bugzilla.gnome.org/show_bug.cgi?id=688487
The AnimatedIcon does not have an API for controlling the animation but
relies on the :visible property changes to start and stop a timeout used
to update the frame.
This has the inconvenient of having a side effect when visible is set to
true multiple times, and is not really the API expected from such
component. Also, there is a race if it is displayed before the images
are loaded: there is no child yet and thus we get this._frame = NaN
which leads to a crash.
Switch to a play/stop API instead, and add a load event callback to the
TextureCache.load_slice_image to exactly know when we can start using
the images.
https://bugzilla.gnome.org/show_bug.cgi?id=687583
We are currently using a hack to allow a select set of keybindings
in the overview. Implement the new MetaPlugin keybinding_filter
hook, which provides a cleaner way to achieve the same.
https://bugzilla.gnome.org/show_bug.cgi?id=688202
It appears to be somewhat common for st_widget_style_changed() to be
called when no style-relevant attributes have, in fact, changed. Now that
we cache theme nodes, we're likely to get the same theme node back from
the cache. If we do, we don't need to waste time asking whether its
geometry and painting are equal to itself: we can just note that nothing
really changed and get on with our lives.
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=687465
Reviewed-by: Jasper St. Pierre <jstpierre@mecheye.net>
If you copy a theme node's paint state into itself, it should be an
inexpensive no-op. What actually happened was that we destroyed the
old paint state, re-initialized to blank, then copied the blank state
back into itself. In the process, we lost (for instance) the textures
for rounded corners.
Until I introduced the texture cache, this never actually happened,
because when st_widget_recompute_style() calls st_widget_get_theme_node(),
we'd always get a fresh theme node. Now, we get a theme node T back
from the cache, notice that paint_equal(T, T) is true, short-circuit
slightly by copying its drawing state into itself, and destroy drawing
state that we still needed.
I'm going to fix this in recompute_style() too, but as a general
principle, self-assignment ought to be harmless.
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=687465
Reviewed-by: Jasper St. Pierre <jstpierre@mecheye.net>
Because we calculate and cache CSS properties once per StThemeNode,
and only a certain set of attributes can affect the CSS properties,
it's advantageous for as many widgets as possible to share a single
StThemeNode. Similarly, if a widget changes state and then changes back
(e.g. gaining and losing the :hover pseudo-class), it should ideally
get its original StThemeNode back again when it returns to the old
state.
Here, I'm using the StThemeContext as the location for a cache.
StThemeNodes are currently never freed: this seems OK for Shell's usage
(a finite number of IDs, classes, pseudo-classes and types).
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=687465
Reviewed-by: Jasper St. Pierre <jstpierre@mecheye.net>
In my testing this cuts the longest time to dispatch(), when showing the
calendar menu for the first time, from 604 to 442 milliseconds,
while reducing additional_selector_matches_style() from 32% to 13% of
CPU time used.
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=687465
Reviewed-by: Jasper St. Pierre <jstpierre@mecheye.net>
The descriptions for keybindings shared with Metacity are now
duplicated in Mutter, so only use the latter to avoid duplicate
entries in System Settings.
https://bugzilla.gnome.org/show_bug.cgi?id=687672
The screen grabber was a workaround for an extremely slow path in Mesa
when reading back pixel data from the frame buffer. It was using pixel
buffer objects by directly calling into GL to hit a fast blit path in
Intel's driver. This should no longer be necessary with the latest
Mesa because the normal read pixels path now has a fast path to just
memcpy the data. Using PBOs in that case just adds an extra
indirection because the data is read into an intermediate buffer and
then copied back out again.
We want to be able to remove the dependency on linking against libGL
directly from Gnome Shell because that will not work if Cogl is
actually using GLES. Also libGL includes GLX which means gnome-shell
ends up with a hard dependency on Xlib which hinders the goal of
getting Gnome Shell to be a Wayland compositor.
https://bugs.freedesktop.org/show_bug.cgi?id=46631https://bugzilla.gnome.org/show_bug.cgi?id=685915
shell_mobile_providers_parse() was returning the country information split
into a hash table with providers and a hash table with country names. This
patch merges both outputs into a single per-country object, so the parse()
method now returns a GHashTable with the following element-type:
(element-type utf8 ShellCountryMobileProvider>)
This also avoids more complex setups like returning lists inside of hash tables,
which was actually breaking either g-i or gtk-doc.
shell_mobile_providers_parse() was also modified to allow inputting the paths
of the country codes and provider list files to use. If paths are not given, the
default ones will be used. This helps us to provide test files during unit
tests.
Both the findProviderForMCCMNC() and findProviderForSid() methods are exported
out of the GSM and CDMA specific classes, and new unit tests for them are
implemented. Tests can be run manually with:
$> ./tests/run-test.sh tests/unit/mobileProviders.js
https://bugzilla.gnome.org/show_bug.cgi?id=687356.
Rather than using a complicated set of function calls across
library boundaries and our own scanning logic, use strtok(),
which glibc already provides, and is probably much more optimized.
https://bugzilla.gnome.org/show_bug.cgi?id=687465
* Fix wrong parameter name to on_name_appeared callbacks
* optparse doesn't just leave extra command line arguments, it
errors out, so don't try to pass through extra arguments -
instead add explicit passthrough for '--replace'
* Fix usage of Gio.DBusProxy
* Add a default value for --perf so that if it's not supplied
things don't die with a mysterious error message. (This wasn't
needed when --perf enabled perf-mode)
https://bugzilla.gnome.org/show_bug.cgi?id=687287
The libedataserverui dependency is a relic of the old E-D-S API.
As of 3.6.0, E-D-S now centralizes authentication prompts so clients
don't have to display their own. This also allows trading the GTK+
main loop for a plain GMainLoop in gnome-shell-calendar-server.c.
https://bugzilla.gnome.org/show_bug.cgi?id=687189
Often the first thing a user wants to do after making a recording
is post it somewhere.
This commit adds the video to recently used items, so that it shows
up prominently in open file choosers.
https://bugzilla.gnome.org/show_bug.cgi?id=680647
Recording continues for some time after the recorder object
is closed, since closing isn't a synchronous operation.
This commit defers quiting the test-recorder application until
the recording is finished.
https://bugzilla.gnome.org/show_bug.cgi?id=680647
We want to make sure the recorder isn't finalized until the
saved recording hits disk. This means the pipeline object needs
a hard reference on the recorder.
https://bugzilla.gnome.org/show_bug.cgi?id=680647
The stage is a floating object. We don't own a reference
to it, so we shouldn't unref it.
This commit removes the erroneous unref call and makes sure
we call clutter_actor_destroy on the stage when we're done
with it.
https://bugzilla.gnome.org/show_bug.cgi?id=680647
While looking at how the plymouth implementation was built, I was so
short-sighted and focused on the string "_XROOTPMAP_ID" that I didn't
realize it was the name of the standard background on the root window.
Remove our own implementation, and switch to using a standard mutter
MetaBackgroundActor.
https://bugzilla.gnome.org/show_bug.cgi?id=682428
Currently we miss changes to a file referenced in background-image
or border-image.
Connect to the StTextureCache::texture-file-changed signal to keep
up with file changes and update the drawing state if necessary.
https://bugzilla.gnome.org/show_bug.cgi?id=679268
For textures loaded from files, the cache might hide image changes
by keeping the data of a previous version around indefinitely. For
instance AccountsService will notify of avatar changes, but as new
image is copied over the old one, we will continue to use the old
image data.
Install a file monitor for each file resource we load and clear
the corresponding data from the cache on changes, emitting the
new StTextureCache::texture-file-changed signal.
https://bugzilla.gnome.org/show_bug.cgi?id=679268
ClutterText will only queue a relayout after font changes if it has
any contents other than the empty string. As a result, its height
request may change after the first character has been entered. To
avoid this visual glitch, force a relayout on actual font changes.
https://bugzilla.gnome.org/show_bug.cgi?id=685534
The actor's GtkIMContext is freed in dispose and reset in unrealize - as
ClutterActor's dispose will unrealize the actor if necessary, chaining
up to the parent after clearing the im context will result in warnings
if the actor is still realized, so chain up first.
https://bugzilla.gnome.org/show_bug.cgi?id=686016
The application menu is currently unusable with non-maximized
windows when using focus-follows-mouse mode. Override mutter's
focus-change-on-pointer-rest preference, so that the actual
focus change is delayed until the pointer stops moving.
https://bugzilla.gnome.org/show_bug.cgi?id=678169
The signals for the action group were being disconnected when the action
group was explicitly removed from the GActionMuxer but the same was not
being done when it was finalized.
This means that a change in the state of an action group that used to be
associated with a finalized GActionMuxer would result in a crash. This
would happen for stateful application actions after closing a window.
https://bugzilla.gnome.org/show_bug.cgi?id=681399
For performance reasons, resources required to paint a widget are
aggressively cached; we know of at least one case where our caching
prevents updating the used background-image correctly, so add explicit
API to clear all associated cache data.
https://bugzilla.gnome.org/show_bug.cgi?id=679268
StThemeNode caches its resources aggressively to keep the required
work on paint to a minimum - right now, resources are only recreated
on allocation changes.
In order to update the background-image property correctly when the
underlying file changes, resources need to be recreated without a
size change, so add an explicit method for that.
https://bugzilla.gnome.org/show_bug.cgi?id=679268
The current API assumes that image data loaded from files remains
valid during the life time of the shell. This assumption is mostly
valid for image files we provide ourselves (with the exception being
designers working on those files), but not necessarily for "external"
files - provide API to explicitly remove cached data associated with
a URI for those cases.
https://bugzilla.gnome.org/show_bug.cgi?id=679268
The screenshield requires gdm 3.5, which can be problematic in
jhbuild configurations, or distributions that don't use GDM as the display
manager. Allow transparent fallback to gnome-screensaver in that case.
https://bugzilla.gnome.org/show_bug.cgi?id=683060
The speed and quality properties have been removed in favor of properties
closer to the upstream library.
Removing the properies from the pipeline would result into a huge
slowdown so we have to map the old values to the new ones.
According to the source code of the old vp8enc element quality maps to
(int)(63 - quality * 6.2) for min_quantizer and max_quantizer, while
speed maps to cpu-used = speed == 0 ? 0 : (speed - 1).
So set min_quantizer and min_quantizer to 13, and cpu-used to 5 based on
the above formulas.
https://bugzilla.gnome.org/show_bug.cgi?id=684206
Remove the PlacesManager, its search provider and all associated code.
Places search is now provided by nautilus using the external search
provider API.
https://bugzilla.gnome.org/show_bug.cgi?id=683506
Rather than unconditionally removing a focus root in remove_group(),
decrement a counter that add_group() increments, and only actually
remove a focus root when the counter drops to 0.
https://bugzilla.gnome.org/show_bug.cgi?id=682243
When using an input method like IBus, the IM is expected to process
key events before anything else. Currently this doesn't always work
as expected, as the event filtering is done in the default handlers
of the key-press and key-release events, e.g. only after other
handlers have been run.
To allow the IM to filter events earlier, move the code to a
captured-event handler instead.
https://bugzilla.gnome.org/show_bug.cgi?id=658325
If an actors is not mapped (visible and all parents visible), then don't
allow navigating focus to it.
This fixes a regression in the keyboard navigation of the panel with
invisibile items.
https://bugzilla.gnome.org/show_bug.cgi?id=683529
st_texture_cache_load_from_raw() enforces a square ClutterTexture,
resulting in the texture being stretched if the passed in image
data has a different width:height ratio.
Add padding in those cases as we already do when loading from pixbufs.
https://bugzilla.gnome.org/show_bug.cgi?id=683483
Clutter will try to unmap during a dispose if we have a parent, so if we
set our own actor to NULL before the chain up, we're going to attempt to
unmap our own NULL actor. Fix that by swapping the order in which we
chain up.
https://bugzilla.gnome.org/show_bug.cgi?id=672790
The default filename isn't localized and isn't the same one that
the shell sets. Just remove the fallback mechanism, and abort
recording if somebody didn't set the filename
https://bugzilla.gnome.org/show_bug.cgi?id=677434
Components are pieces of the shell code that can be added/removed
at runtime, like extension, but are tied more directly to a session
mode. The session polkit agent, the network agent, autorun/automount,
are all components, keyring, recorder and telepathy client are all
now copmonents.
https://bugzilla.gnome.org/show_bug.cgi?id=683156
The supposed reason for launching the calendar server in a peculiar
way was so that the process would be killed when the Shell was killed,
but that didn't actually work. Launch the calendar server through auto-start,
and persist all throughout the session.
https://bugzilla.gnome.org/show_bug.cgi?id=683156
Introduce a StShadowHelper to manage drop shadows from JS (which
cannot use Cogl directly), and use it in a new StWidget-derived
JS class to draw the arrow.
https://bugzilla.gnome.org/show_bug.cgi?id=682285
Reroute setting those properties to a GIcon. API users are expected
to create GIcon directly now.
The advantage is that from a StIcon you can now create a similar one
by accessing :gicon.
https://bugzilla.gnome.org/show_bug.cgi?id=682540
The original design for the overview had buttons for searching for
Wikipedia and Google, but in practice this is a bad idea. The buttons
are the default activations, meaning that using the overview as a
fluent motion of launching something - "firefxo<Enter>", will launch
Google/Wikipedia.
https://bugzilla.gnome.org/show_bug.cgi?id=670168
If the user is already active when the notification pops up, we
won't get an idle watcher because there's no transition from
active to idle or vice versa. Correct this by initializing the
state correctly from XSync.
StScrollBar was intercepting motion events by using captured-event on
the stage, which required additional dirty tricks, which required
additional hacks. Simplify it by just using clutter_grab_pointer()
instead.
https://bugzilla.gnome.org/show_bug.cgi?id=671001
It doesn't make sense to have multiple ShellIdleMonitors, since
each has its own GDK filter function, but they all get the same
events. In preparation for having it accessed from other places
than the message tray, make it a singleton.
https://bugzilla.gnome.org/show_bug.cgi?id=682041
Add support for the CSS "background-repeat" property. Currently, this
only supports on/off, rather than allowing tiling in each individual
dimension. It is supported for both the cogl and cairo rendering paths.
https://bugzilla.gnome.org/show_bug.cgi?id=680801
gnome-session and gnome-settings-daemon rely on the screensaver
interface to know the locked state. Since gnome-screensaver is no
longer running, it's up to gnome-shell to provide it.
https://bugzilla.gnome.org/show_bug.cgi?id=619955
To allow more than one summary icon actor for a source we split
the model of the source icon (which is iconName, if the default
implementation is used, or a GIcon otherwise) and replace
createNotificationIcon() with a generic createIcon(size). Also,
the actual source actor is split into a separate class, that handles
the notification counter automatically.
https://bugzilla.gnome.org/show_bug.cgi?id=619955
The :reactive property is used on StButton to like the :sensitive
property on GtkWidgets, that is, to indicate that the user is not
(yet) expected to click the button, and therefore should affect
styling too.
This allows to remove some code at the JS layer.
https://bugzilla.gnome.org/show_bug.cgi?id=619955
While modal dialogs were attached to the parent's titlebar, it
made sense to leave the top of the parent window at full color.
With the new position of modal dialogs, it makes more sense to dim
the entire parent window, so we can use a combination of Clutter's
BrightnessContrast- and DesaturateEffect instead of our own custom
shader.
https://bugzilla.gnome.org/show_bug.cgi?id=674499
We used to use realize/unrealize instead of map/unmap in ShellEmbeddedWindow
because there originally was no map/unmap. The days of this are long gone...
https://bugzilla.gnome.org/show_bug.cgi?id=672790
This currently causes the shell to freeze very often in a thread
deadlock, and the gjs garbage collector behavior is currently getting
fixed at the right level in gjs itself.
https://bugzilla.gnome.org/show_bug.cgi?id=679832
When selecting "Open Calendar" in the date menu, the configured
application is launched via command line, so we don't get any
startup notification. In order to fix the issue at least for our
default calendar, add a hidden .desktop file for evolution's
calendar component.
https://bugzilla.gnome.org/show_bug.cgi?id=677907
get_app_for_window() is (transfer full), but shell_app_system_lookup_wmclass()
is (transfer none), so we must reference the result, or crash
occur.
https://bugzilla.gnome.org/show_bug.cgi?id=678992
When running with Valgrind, this helps us ensure that we're managing
memory correctly. We need to GC twice as finalizing an object in the
sweep can unroot objects which were already marked. Technically, it
could be that we'll need to GC more than twice, but GCing twice should
hopefully last us for now.
https://bugzilla.gnome.org/show_bug.cgi?id=678737
clutter_actor_get_children returns a newly allocated GList and it was
not freed.
However, as there's no reason to copy the children list, switch to
iterator api.
https://bugzilla.gnome.org/show_bug.cgi?id=678406
The dependency chain spirals out from folks->zeitgeist->xapian...and
I'm really not interested in pulling in all of that into the core
shell.
There is work on splitting out contact search into gnome-contacts; I'd
add a bug link but Bugzilla is down.
Currently we assume that GTK_UNIQUE_BUS_NAME is shared between all
windows of an application. This assumption does not hold true for
applications that specify G_APPLICATION_NON_UNIQUE, so make sure
to update the menu as necessary.
https://bugzilla.gnome.org/show_bug.cgi?id=676238
There is a race if a channel is invalidated during its preparation: the
'invalidated' signal is already emitted so the Shell will never notice.
We fix this by simply checking if the channel is already invalidated when
receiving it from telepathy-glib.
In the approving case, we reject the full ChannelDispatchOperation as we only
support approving one channel at the time.
https://bugzilla.gnome.org/show_bug.cgi?id=677457
Commit de8a66d4ce removed our own DPI handling for the one found
it Clutter, but broke resolution updates at runtime (for instance
when setting the "Large Text" option in Universal Access).
https://bugzilla.gnome.org/show_bug.cgi?id=677975
We explicitly include NoDisplay applications in the ShellAppSystem because
we want app tracking for them, but we explicitly filter NoDisplay applications
out when showing them to the user because we don't want to show them to the
user. We also based our "All" apps view on a flattened list of apps. While
we did check for NoDisplay on the app item itself, we didn't check against
its parents. Refactor the app display view to not use a separate flat list
of applications, but instead a concatenation of all the applications in all
the loaded categories.
https://bugzilla.gnome.org/show_bug.cgi?id=658176
Instead of falling back to a set of default values or crashing the
window manager when an invalid mode is specified, check the value
of the ShellGlobal:session-mode property before taking over as WM
and make a clean exit if it cannot be resolved to an existent mode.
https://bugzilla.gnome.org/show_bug.cgi?id=676156
Add a session-mode property on ShellGlobal which corresponds to the
new --mode switch. Make the existing ShellGlobal:session-type property
readonly and base it on ShellGlobal:session-mode to avoid conflicts.
https://bugzilla.gnome.org/show_bug.cgi?id=676156
ShellApp keeps track of windows sorted by most recently used first
which means that when raising windows we need to start from the end of
the list to finish with the correct stacking order.
This patch just makes the code agree with the comment which was
already there.
https://bugzilla.gnome.org/show_bug.cgi?id=676371
Add a dummy call to shell_js_add_extension_importer() to ensure that the
link to libgnome-shell-js.so is not removed when using -Wl,--as-needed,
which is the default on many distros.
https://bugzilla.gnome.org/show_bug.cgi?id=670477
Depending on the exact linker options and versions, rpath's written
into the binary may, or may not, be honored by dlopen(). Work around
this by simply linking the gnome-shell binary to gnome-shell-js, since
then dlopen() doesn't need to search paths.
https://bugzilla.gnome.org/show_bug.cgi?id=670477
It doesn't make sense to have multiple requests for the same
connection/setting combination at the same time, since we would be
asking the user twice for the same password. Instead, report cancellation
to NetworkManager if this happens.
Note that does make sense to have multiple requests in sequence though
(they could have different flags), but this is not affected.
https://bugzilla.gnome.org/show_bug.cgi?id=674961
Lots of things here were unused. First, we don't need to calculate
the js_dir when not running from the source tree, which makes the
branch mostly empty, meaning we can merge the two branches.
This allows us running uninstalled. While we're at it, though, remove
JHBUILD_TYPELIBDIR. jhbuild shell should add its own stuff to GI_TYPELIB_PATH,
and we don't want to half reimplement jhbuild. The wrapper script should be
solely for the case of running from the source directory, and not care about
jhbuild at all.
If start_shell throws an exception for whatever reason, the finally
handler will run before shell has been set, leading to a "referencing
an undefined local variable" error.
If a request isn't found in shell_network_agent_set_password() or
shell_network_agent_respond(), then g_return_if_fail() rather than
crashing. OTOH, if a request is not found in
shell_network_agent_cancel_get_secrets(), then just ignore it
silently, since that could legitimately happen.
https://bugzilla.gnome.org/show_bug.cgi?id=674961
Right now, we are hard-depending on the presence of Evolution by
using its settings schemas. This is likely to be unpopular, and
also causes instability if someone happens not to have Evolution
installed, so install a schema that has the same data path as
the Evolution schema, but a different name and install that
for the keys we need.
To avoid a string-freeze break, we rely on the translations in
Evolution - if Evolution isn't installed, the key descriptions
will be untranslated in dconf-editor.
https://bugzilla.gnome.org/show_bug.cgi?id=674424
This swaps a use of GLfloat for a regular float. Cogl might stop
including a GL header in its public headers soon so this would fix a
compilation error.
https://bugzilla.gnome.org/show_bug.cgi?id=672711
If run under jhbuild, gnome-shell is in PATH, so trying to restore
is just running it again, likely failing in the same way (with the
additional problem of detaching the shell from the terminal). Also,
this is using deprecated GConf keys, and deprecated pygobject
bindings.
The jhbuild wrapper is expected to be run through jhbuild run, or
in the uninstalled case inside a jhbuild shell, so there should
be no need to set XDG_DATA_DIRS and friends.
dconf-daemon is available since 2.32, so there is no need to
start it manually (and in fact, it should not be running normally).
Similarly, notify-osd and notification-daemon should not be
dbus activated, but rather autostarted by gnome-session, and
therefore not running in a gnome-shell session.
Introduce a new gnome-shell-perf-tool, which can be used instead
of the old gnome-shell-jhbuild wrapper to gather data about gnome-shell
performance and submit to shell-perf.gnome.org. This runs the
shell with no extra setup beyond the WM_CLASS filter, so it can
be used for a jhbuild setup or for an installed shell.
Currently the scroll event code only handles scroll events if the
adjustment's value is within the "lower" and "upper" limits. The
likely intent was to pass events to a parent scroll view when
reaching the bounds (uh, nested scroll views!), but apparently
we never made use of this, as the upper bound is actually wrong
(an adjustment's maximum value is upper - page_size, not upper).
Just handle all scroll events unconditionally and rely on the
bound checks in StAdjustment.
https://bugzilla.gnome.org/show_bug.cgi?id=672413
Evolution now stores its selected calendars and tasks in GSettings, not
in GConf. If we don't look at the new location, then we'll not pick up
newly added and enabled calendars, making the calendar effectively not
work for new installs.
https://bugzilla.gnome.org/show_bug.cgi?id=673610
If evolution-data-server needs to prompt for a password, it will try
to pop up a GTK+ dialog. When GTK+ is not initialized, the result is
a crash. So, initialize GTK+ and run a main loop.
See https://bugzilla.redhat.com/show_bug.cgi?id=809681
The result is ugly since we have a Gnome-shell-calendar-server fallback
application, but I don't think it's worth installing a desktop file
and having a string break, since this is pretty uncommon (only for
manually added calendars without the password stored in gnome-keyring),
and apparently this is being rewritten for 3.5 to have the dialogs come
the e-d-s daemon rather than from the individual application.
https://bugzilla.gnome.org/show_bug.cgi?id=673608
gnome-shell-extension-prefs uses format(), but can't pull in Shell
(which is a dependency for the module), since that in turn would pull in
Meta. Fix this by moving the introspected int format function to ShellJS
instead.
https://bugzilla.gnome.org/show_bug.cgi?id=673106
Icon theme change signals aren't noticed immediately, they're usually
noticed when trying to load an icon. Since icon theme changes cause a
style change, and most icon widgets try to re-load their texture during
a style change, this means that we get a stack like this:
st_texture_cache_load_icon
gtk_icon_theme_lookup_icon
gtk_icon_theme_changed
st_widget_style_changed
st_texture_cache_load_icon
Rather than making every place that uses StTextureCache re-entrant,
punt the notifying of icon theme changes to an idle handler instead.
https://bugzilla.gnome.org/show_bug.cgi?id=673512
We've been dangling on the edge of unsafety, unnoticed, for a little while
about the reference count safety of ShellScreenshot. GJS owns the entire
reference count, so as soon as it goes out of scope it could die, causing
GJS to try and fetch the corresponding wrapper object for a stale pointer.
We haven't seen any crashes because of luck -- SpiderMonkey tries to group
together deallocations to limit GC pauses, and there isn't really a lot
of GC pressure in the duration that a screenshot happens, so we tend to
be mostly stable. But in the case that you create a lot of objects while
a screenshot is going on, by hammering the "Print Screen" button, for
example, you can destroy the GObject before the callback finishes.
https://bugzilla.gnome.org/show_bug.cgi?id=672775
Currently compilation fails with -Werror, as we don't handle the
(newly introduced) smooth scroll events in switch statements; add
some basic support, which should make the compiler happy.
https://bugzilla.gnome.org/show_bug.cgi?id=672413
More than one outstanding request to the same URI should now be
deduplicated, and the framework is there if we want to cache async loaded
URIs as well
https://bugzilla.gnome.org/show_bug.cgi?id=672273
In the case that we don't have an icon corresponding to the gicon, it's
more than likely that the code calling load_gicon will know more about
what it wants as a fallback than the texture-cache itself. In fact -
we had a whole lot of dead code that would try to fall back, but never
did because we always returned a valid actor.
This was causing certain applications with invalid icons to not get the
fallback icon because an icon couldn't be found. Fix up the one place
where we don't have an explicit fallback icon codepath, and then stop
doing what we were doing.
https://bugzilla.gnome.org/show_bug.cgi?id=671656
This assignment was shadowed by the giant switch above. Since the
switch has a comment or two explaining the logic inside of it,
keep that instead of the assignment.
Since the invoker for navigate_focus has an extra parameter, annotations
from the invoker aren't applied on the vfunc itself. Fix that by annotating
the vfunc separately.
In the workspace-collecting code we add a check to avoid collecting a
workspace if any startup sequence is running there. Since the sequence
can take some time to load, an helper function is also added which keeps
the (empty) workspace around for a very short time, while waiting for the
sequence to start.
https://bugzilla.gnome.org/show_bug.cgi?id=664202
If the user was inactive while a notification was shown, we show the summary
when the user becomes active again. This ensures that we inform the user of
the existance of new notifications that the user might have missed.
When the user comes back from away, the summary is now only shown if it has
new notifications.
https://bugzilla.gnome.org/show_bug.cgi?id=643014
ECal is deprecated and replaced by ECalClient. This has the
advantage of using e_utils to handle authentication, and should
fix NotOpened errors (that affect in particular webcal calendars
prior to evolution running)
https://bugzilla.gnome.org/show_bug.cgi?id=671177
This allows us to do directional keyboard navigation when there's no
actor inside the horizontal or vertical strip extending from the
origin actor but there are other actors to the sides of that strip
that could still be used as targets even if that means the focus would
move diagonally.
https://bugzilla.gnome.org/show_bug.cgi?id=663901
For arrow keys navigation, when moving from a widget which isn't a
descendant of the widget we are going to, it's unexpected that focus
moves to the target's first descendant instead of the closest to the
source widget.
This requires us to use absolute coordinates to compare widgets since
we no longer have the guarantee that the widgets we are comparing are
siblings.
https://bugzilla.gnome.org/show_bug.cgi?id=663901
Our DBus API (and mostly every other API in existence) define an
area as the top-left corner and width/height; glReadPixels on the
other hand uses the bottom-left corner, so we have to transform the
coordinates before passing them to GL.
https://bugzilla.gnome.org/show_bug.cgi?id=670979
When porting from st_container_get_children_list to this code, I accidentally
broke StTable by a copy/paste error. This broke notifications that updated,
leaving them with what they think were 0 columns and 0 rows.
https://bugzilla.gnome.org/show_bug.cgi?id=670640
* Add a keyring prompter based on GcrSystemPrompter
* Adds dependency on gcr version 3.3.5 or higher
* Not yet using unmerged support for non-pageable memory
https://bugzilla.gnome.org/show_bug.cgi?id=652459
The very similar clutter_actor_allocate_align_fill is close enough
that this is just needless duplication. Additionally, allocate_fill
already inverts the align if the text direction is RTL, so we don't
need to do that here.
https://bugzilla.gnome.org/show_bug.cgi?id=670034
Now that StWidget is a group of sorts, it needs to account for its children
in its paint volume. Unfortunately, this causes havoc for StBoxLayout, so it
needs fixing - it's unknown why it worked when chaining up to near-identical
code in StContainer.
https://bugzilla.gnome.org/show_bug.cgi?id=670034
The get_focus_chain() implementation in StWidget just returns all
children, it should filter for visible children instead. This
breaks keyboard navigation in various places since commit 72dad591
removed the correct implementation in StContainer.
https://bugzilla.gnome.org/show_bug.cgi?id=670904
Now that ClutterActor has a ClutterContainer implementation, we
can start removing StContainer. To help make this a bit more
understandable, instead of converting everything at once, make
StContainer a compatible API wrapper around the ClutterActor
implementation, and then we'll remove those wrappers in later
commits.
https://bugzilla.gnome.org/show_bug.cgi?id=670034
Since an StWidget now has children, it needs to allocate those children
properly. Defer to the currently installed layout manager, like Clutter
does.
Now that we have something that allocates children in St, to prevent
double allocations, we use clutter_actor_set_allocation rather than
chaining up to StWidget::allocate.
https://bugzilla.gnome.org/show_bug.cgi?id=670034
Now that StWidget is concrete and instantiable, we need to do something
other than return an adjusted 0 for width and height. Just chain up
to ClutterActor's default implementation, which uses the layout manager.
https://bugzilla.gnome.org/show_bug.cgi?id=670034
Since we want to paint children by default in StWidget, we need to
provide a way for custom subclasses to paint their CSS backgrounds
without painting children... introducing st_widget_paint_background.
Additionally, remove any custom paint/pick handlers added by subclasses
of StWidget that just painted their children. This will cause double
painting if left alone.
This also removes the hacky things that some subclasses of StBin did
to prevent their one child to be painted by StBin.
https://bugzilla.gnome.org/show_bug.cgi?id=670034
Clutter now provides two new properties on ClutterActor - first-child
and last-child, so we have notifiers on when they change. Unfortunately,
it still doesn't help us too much - we need to keep track of the previous
values of the properties so we can remove their pseudoclasses.
https://bugzilla.gnome.org/show_bug.cgi?id=670034
We can't get rid of the implementations in StContainer just yet,
as StContainer still keeps its own child list. But this should
lower the amount of code that has to be moved around when we
remove StContainer.
https://bugzilla.gnome.org/show_bug.cgi?id=670034
clutter_container_foreach is deprecated, so let's replace that
with some ClutterActorIter usage. Additionally, remove the checks
for ClutterContainer, as all ClutterActors are now ClutterContainers.
https://bugzilla.gnome.org/show_bug.cgi?id=670034
Although not all "Finding and reminding" applications are ready
yet, the integration with gnome-documents' search results overlaps
enough with the "Recent Items" provider to justify its removal.
https://bugzilla.gnome.org/show_bug.cgi?id=670150
Empathy uses to support 2 D-Bus API for calls:
- StreamedMedia: legacy API
- Call.DRAFT: experimental version of the new API
Since 3.3.90, Empathy only supports Call1, the first stable version of the new
API, so the Shell should do the same.
https://bugzilla.gnome.org/show_bug.cgi?id=667694
The function may return FALSE without setting the GError, so don't
assume it is set to prevent a crash in that case. While at it, free
the GError we were leaking before.
https://bugzilla.gnome.org/show_bug.cgi?id=670418
When using systemd, polkit doesn't set the error even when returning false in g_initable_init
Signed-off-by: Marc-Antoine Perennou <Marc-Antoine@Perennou.com>
https://bugzilla.gnome.org/show_bug.cgi?id=670319
The handle was a child of the trough, but it was allocated and painted
like it was a child of the bar. This will wreak havoc when we port over
to the new Clutter API, so let's just make it a child of the bar.
https://bugzilla.gnome.org/show_bug.cgi?id=670034
When the operation is cancelled by NetworkManager, the request is
cancelled immediately. Later when gnome-keyring invokes the callback
notifying the error we must therefore not access its memory.
Previously the callback would mistakenly treat "cancelled" (which
indicates a programmatic cancel) as "denied" (which means the user
clicked "Cancel" on the keyring prompt)
https://bugzilla.gnome.org/show_bug.cgi?id=658484
VPN secrets are stored by the plugins, that provide separate
helpers for authentication. This commit adds the support for invoking
the binaries and pass them connection details.
For plugins that support it (as exposed by their keyfile), we invoke
them in "external-ui-mode" and expect a set of metadata about the
secrets which is used to build a shell styled dialog.
https://bugzilla.gnome.org/show_bug.cgi?id=658484
ConsoleKit is being obsoleted by systemd. Accordingly port the CK logic
in the gnome-shell automount manager to systemd-logind APIs.
This makes use of systemd-logind's native C APIs which are much easier
to use than the D-Bus APIs in this case, and much faster too (since they
are synchronous and directly query the kernel for the information we
need). The dependency is compile time optional, and in order to be nice
to the Debian folks g-s compiled with this enabled fill automatically
fall back to CK support on systems lacking systemd.
A new tool, 'gnome-shell-extension-prefs' can load a new entry point from
extensions, 'prefs.js', which has an entry point to return a GTK+ widget.
This allows extensions to have their own preferences dialog, without each
extension needing to ship its own Python script and .desktop file.
https://bugzilla.gnome.org/show_bug.cgi?id=668429
Clutter, since version 1.8, does The Right Thing™ inside the default
implementation of ClutterActor::map and ClutterActor::unmap, even for
non-container actors: it will iterate over the list of children and
map, or unmap, each one of them, respectively.
This means that the requirement to override map and unmap for composite
actors to map and unmap the internal children has been dropped.
https://bugzilla.gnome.org/show_bug.cgi?id=669239
Since our implementation of background-size is now CSS-compliant, we
do not need this subtexture hack that clips a "leak". The comment here
is also incorrect.
https://bugzilla.gnome.org/show_bug.cgi?id=633462
It seems that accidentally, two variables were swapped in one code path
of the background-size implementation, causing interesting but wrong
images for some elements.
https://bugzilla.gnome.org/show_bug.cgi?id=633462
Recent mutter changes made MetaShapedTexture not a ClutterTexture,
but instead a special ClutterActor subclass that implemented the texture-y
bits itself. Use recently introduced API in MetaShapedTexture so that we can
get the raw texture data and spit it out as a PNG.
Use the new meta_shaped_texture_get_image() to get a window's texture data.
meta_shaped_texture_get_image() flattens the image against any mask it may
have, so a screenshot of it should look exactly as it does on the display.
https://bugzilla.gnome.org/show_bug.cgi?id=662486
Instead of adding every rendered frame into the recording, drop frames
and only buffer and record enough frames to match the target framerate.
Increase the default frame rate from 15 to 30, since now that we're
actually enforcing framerate, it's noticeable that 15fps is not smooth.
https://bugzilla.gnome.org/show_bug.cgi?id=669066
The default pipeline was fairly aggressive about quality, and could
be too expensive for some computers. Decrease the quality setting for
the vp8 codec from 10 to 8, and increase the speed setting from 2 to 6.
(Basically, quality affects the visual fidelity of the end result, while
speed affects how much CPU the encoder uses to get a high compression
ratio at that quality level.)
Remove videorate from the pipeline, since the GStreamer VP8 encoder can
handle variable-framerate streams. This means that we won't spend CPU
encoding duplicate frames added by videorate.
https://bugzilla.gnome.org/show_bug.cgi?id=669066
We need to indicate that our GStreamer source produces timestamped
frames, instead of the default, which is to produce a stream of bytes.
This is needed for correctness, and to avoid warnings for some
pipelines.
https://bugzilla.gnome.org/show_bug.cgi?id=669066
For the Intel drivers, using glReadPixels() to read into client-memory
directly from the frame buffer is much slower than creating a pixel
buffer, copying into that, and then mapping that for reading. On other
drivers, the two approaches are likely to be similar in speed. Create
a ShellScreenGrabber abstraction that uses pixel buffers if available.
Use that for screenshots and screen recording.
https://bugzilla.gnome.org/show_bug.cgi?id=669065
Use the correct clip offsets when taking the screenshot of a window, to
exclude possible invisible borders and to include the case where the
window doesn't have any frame itself.
.desktop files have been designed for browsing, so the existing
fields often produce insufficient results when used for search.
gnome-control-center used X-GNOME-Keywords for that purpose, which
has now been standardized as Keywords. It makes sense for us to
support it in gnome-shell as well (and encourage its use outside
of settings panels).
https://bugzilla.gnome.org/show_bug.cgi?id=609702
It seems that Debian has their own prefixes in something like
debian-xterm.desktop. To properly do application matching in these cases,
we need to strip the debian- prefix.
https://bugzilla.gnome.org/show_bug.cgi?id=665647
Writting the screenshot to a file can take a relativly long time
in which we block the compositor, so do that part in a separate
thread to avoid the hang.
https://bugzilla.gnome.org/show_bug.cgi?id=652952
A new texture has undefined contents - when we're creating a shadow,
we need to clear the contents of the texture before drawing the border
and background into it.
https://bugzilla.gnome.org/show_bug.cgi?id=668048
g_dbus_method_invocation_return_value() adopts a floating reference,
so we don't also need to unreference it; fix by replacing the code
using a more compact form using the ^ convenience character in
GVariant type specifications. (Thanks to Ryan Lortie for the
suggestion.)
https://bugzilla.gnome.org/show_bug.cgi?id=667378
According to the GIO docs, sections can have labels too. We support
them by inserting a non reactive menu item at the beginning of the
section. This item is specially flagged to be ignored while processing
changed signals from the model (since it does not correspond to any
model item)
https://bugzilla.gnome.org/show_bug.cgi?id=666681
shell_embedded_window_hide() can be called during widget destruction,
after the associated ClutterActor has been already cleared out.
Fix a crash in that case.
https://bugzilla.gnome.org/show_bug.cgi?id=633028
The application proxy is created asynchrously after the dbus name
is registed. This means that when tracking the first window (and
therefore creating the first window GActionGroup) there is no
app proxy yet.
https://bugzilla.gnome.org/show_bug.cgi?id=633028
Implement the background-size CSS property, specified by the CSS
Backgrounds and Borders Module Level 3, including the keywords
"contain", "cover", and fixed-size backgrounds.
https://bugzilla.gnome.org/show_bug.cgi?id=633462
It's not guaranteed that the application DBus proxy appears before
we receive the first focus event from the toplevel window.
Ensure that the first method to access the action muxer creates it if
hasn't been created yet.
We now have GRemoteActionGroup interface with the needed API,
implemented by GDBusActionGroup.
With the new API we could theoretically turn GActionMuxer itself into a
GRemoteActionGroup and expose the _full API to the shell so that the
timestamps could be passed from there.
Pass the current timestamp as platform data when activating
an action. This is implemented slightly hacky, since we use
clutter_get_current_event_time() to get at the timestamp, but
the alternative is to expose g_action_muxer_activate_action_full
to js, which would be quite a bit more involved.
GTK+ also exports window-specific actions, by putting the object path
for the exported action group in the _DBUS_OBJECT_PATH X property.
We add this action group to the app's muxer with a 'win' prefix,
since that is what the exported menu expects. Whenever the focus
window changes, we update the window-specific actions of its
application, and emit notify::action-group to cause the app
menu to be updated.
GDBusActionGroup api has changed again, adapt to that.
Also, use a GActionMuxer to add the 'app.' prefix to actions,
instead of manually stripping it out of the action names.
In the future, the muxer will also contain per-window actions
with a 'win.' prefix.
GMenuProxy has been replaced by GDBusMenuModel, and the object path
has been moved (now needs to be retrieved from the AppMenu GApplication
property).
Update the test to prefix each action with "app." as documented,
and use a GtkApplicationWindow instead of a plain GtkWindow.
Use the new GApplication support in ShellApp to create the application
menu. Supports plain (no state), boolean and double actions.
Includes a test application (as no other application uses GApplication
for actions)
https://bugzilla.gnome.org/show_bug.cgi?id=621203
GDBusActionGroup and GMenuProxy are new objects in GIO 2.32 that
help with accessing menus and actions of remote applications.
This patch makes it possible for the shell to associate an
application with a dbus name and from that a GMenu, that will
be shown as the application menu.
https://bugzilla.gnome.org/show_bug.cgi?id=621203
grid_width and grid_height were inverted, which caused a crash
in GdkPixbuf code. This was never noticed because the animation
in the panel is a square.
https://bugzilla.gnome.org/show_bug.cgi?id=666606
commit 5350302b09 dropped the possibility
to make a dbus name acquisition failure non-fatal.
Btw, it has also overriden the name in the error message.
Signed-off-by: Marc-Antoine Perennou <Marc-Antoine@Perennou.com>
Signed-off-by: Colin Walters <walters@verbum.org>
https://bugzilla.gnome.org/show_bug.cgi?id=663941
Rewrite code acquiring dbus names so that it uses GDBus, and rewrite
ShellDBus so that it is exposed on the GDBus connection. Ports of
the other objects will follow.
https://bugzilla.gnome.org/show_bug.cgi?id=648651
Not all desktop files tracked by the shell have
Exec lines. This could be because they're actually
run by another process, for instance, and the desktop
file is merely there to provide metadata. For example,
nautilus-pastebin provides a desktop file without an
Exec line.
The shell currently crashes if one of these partial
desktop files is installed and the user attempts to
search from the overview.
commit 37726a4cb6 fixed
a similar crasher.
This commit fixes the next one lower in the code.
https://bugzilla.gnome.org/show_bug.cgi?id=663815
Instead of saving the last_used_time per-app, grab the maximum time for all
windows. The logic is less hard to keep track of, and it solves some edge
case issues where windows that no longer exist update the user time, even
if none of the other windows have been used recently.
https://bugzilla.gnome.org/show_bug.cgi?id=660650
Allow push_modal to optionally only work with a keyboard only grab and
use that in altTab as a fallback to allow switching windows while a pointer grab
is in effect (like during DND operations).
https://bugzilla.gnome.org/show_bug.cgi?id=660457
For some reason, the texture cache decides to make a request and then look up
an icon in the icon theme. If it's valid, it just returns, fine, but if it
doesn't add the icon, it tries to undo the request, leaking an
AsyncTextureLoadData that isn't freed in the process.
https://bugzilla.gnome.org/show_bug.cgi?id=660968
Rather than have five or six structs allocated duplicating data,
just keep one and simplify the code considerably.
Again, part of my ongoing quest to merge St and Mx.
https://bugzilla.gnome.org/show_bug.cgi?id=660968
shell_global_get_memory_info tries to zero initialize the output
parameter with memset, but it passes the wrong size (because of
a missing *). There's no reason to do the memset, though. In the
normal case all members of the struct gets initialized before the
function returns anyway.
This commit drops the memset call in favor of one explicit 0 assignment
that only gets executed on on atypical platforms.
https://bugzilla.gnome.org/show_bug.cgi?id=662236
Without the desktop, even Nautilus hardcodes "Home" in all places
nowadays (except for the desktop itself). I think a run-time dependency
on nautilus being installed (GSettings will abort if the schema is not
found) is not worth it to keep compatibility with the desktop.
(Also, nautilus itself could probably hardcode "Home" for the desktop as
well and remove the preference).
https://bugzilla.gnome.org/show_bug.cgi?id=659895
If we add a 0-sized actor with a border-radius, we will crash as we try to
allocate a 0-sized texture in Cogl. Bail out early instead of doing that.
https://bugzilla.gnome.org/show_bug.cgi?id=661617
We originally OR'ed search terms and favored results which matched
multiple times to get more relevant results. When changing search
to AND search terms, the semantics of "multiple matches" were
changed to refer to a single term matching multiple criteria (name,
executable), which seemed like a good idea at the time.
However in practice this just results in applications whose
user-visible name matches the executable name on disk being
favored over applications using a more generic name, which
isn't too useful (in particular when taking usage frequency
into account).
https://bugzilla.gnome.org/show_bug.cgi?id=623372
Currently we use a very strict definition of "prefix", where the
search term has to match at the very beginning of the searched
criteria (application name, executable name). Use a more liberal
definition by including matches where the preceding character is
a space (application name) or hyphen (executable name) as well;
as many applications use a prefix, this should improve the quality
of results.
https://bugzilla.gnome.org/show_bug.cgi?id=623372
Application search results are internally categorized in four sets,
multiple and single prefix matches and multiple and single substring
matches. Each set is currently sorted alphabetically by application
name when concatenating the sets to the final result.
Change the last step to sort each set by usage frequency instead,
which is more likely to favor the most relevant match than
"arbitrary" alphabetic order.
https://bugzilla.gnome.org/show_bug.cgi?id=623372
Require that all terms match. This is the expected behaviour
and matches what gnome-contacts does. Keep the prefix/infix
weights in place for now.
https://bugzilla.gnome.org/show_bug.cgi?id=660912
Folks uses collection/set objects from libgee to store email addresses
associated with an individual. Unfortunately to extract addresses, parts
of libgee which are unusable from (introspected) bindings have to be
used[0], so add a helper method.
[0] in particular gee_iterator_get(), which is annotated as
"return: (transfer full): gpointer"
https://bugzilla.gnome.org/show_bug.cgi?id=660580
Match folks' name and nick fields, in addition to alias,
and look at email addresses in addition to im addresses.
This is more in line with what gnome-contacts does.
To match this new usage, rename the ALIAS_..._WEIGHT and
IM_..._WEIGHT constants to NAME_ and ADDR_, respectively.
https://bugzilla.gnome.org/show_bug.cgi?id=660580
The translate coordinates are calculated as the offset after the scale, so it
needs to be applied after the scale as well. This fixes random centering issues
in the UI.
https://bugzilla.gnome.org/show_bug.cgi?id=660674
For GIcons we use g_icon_to_string() in the key, but the function
will return NULL if the icon cannot be serialized. As a result,
all non-serializable GIcons of the same size end up with the same
cache key - an example for this are contacts with avatars, which
currently all end up with the same image.
To fix, opt out of caching for GIcons which cannot be serialized.
https://bugzilla.gnome.org/show_bug.cgi?id=660585
Setting up the framebuffers for transitions may fail, in which case
the material used for drawing is left uninitialized, so trying to
access it results in a crash.
Instead bail out in this case, which means that we won't paint
anything during the transition - still, drawing errors are better
than crashes ...
https://bugzilla.gnome.org/show_bug.cgi?id=659676
While we allow for arbitrary modifiers in keybindings, both the
alt-tab and ctrl-alt-tab popups close when ALT is not present in
the modifier mask, resulting in ALT being de-facto hardcoded.
Instead, pass the actual modifier mask when invoking the popups.
https://bugzilla.gnome.org/show_bug.cgi?id=645200
VPN secrets are currently unhandled by the UI code. To avoid
lengthy timeouts, bail out early with an error, so NetworkManager
falls back to the nm-applet agent directly.
https://bugzilla.gnome.org/show_bug.cgi?id=658484
Commit 0af108211c introduced a
regression where applications that appear in multiple categories were
duplicated in the "All Apps" list, because we switched from
uniquifying on desktop file ID to the GMenuTreeEntry.
Switch back to keeping the set of apps based on ID. To flesh this
out, we keep the ShellApp instance for a given ID around forever, and
when we're loading new contents, we replace the GMenuTreeEntry inside
the app. That means callers still get new data.
We still keep around the running app list, though we could just
recompute it from the app list now.
https://bugzilla.gnome.org/show_bug.cgi?id=659351
The apps and settings loading code duplicated the part to traverse a
GMenuTree. Unify this by adding a new function to return a flattened
set.
This will also be useful for a future change to how we store apps -
this way we can look at both the current set of apps and the new set.
https://bugzilla.gnome.org/show_bug.cgi?id=659351
Instead of doing complex computations in the shader just pass in the correct
fade area (taking padding, scrollbars and rtl into account) and just work
with that in the shader.
That fixes a bug where we would fade the scrollbar when padding is present.
https://bugzilla.gnome.org/show_bug.cgi?id=659159
The way the window dimmer shader is applied will cause rendering errors with
the rounded corners, invisible borders or shaped textures since it doesn't deal
well with the multitexturing used by the MetaShapedTexture. Use an off-screen
buffer to flatten the texture before being applied.
https://bugzilla.gnome.org/show_bug.cgi?id=659302
While I've been trying to make the GC kick in more often, I've decided
it's a better tradeoff to aggressively GC at "leisure", for multiple
reasons.
We can and should revisit this at a later time, but basically:
* The shell doesn't generate *that* much JS data - garbage collection
is very fast here.
* Long periods without GC mean we're not calling free() when we
could, which in turn makes heap fragmentation much worse.
* Ensuring the GC runs at idle makes it much less likely we'll take
a random large GC hit in the middle of an animation.
https://bugzilla.gnome.org/show_bug.cgi?id=659254
It was always taking the first .service.in file to create all .service
files: org.gnome.Shell.CalendarServer.service.in was the only one being
used.
https://bugzilla.gnome.org/show_bug.cgi?id=659194
