libgnome-bluetooth will start to offer a 2.0 library version
depending on GTK4. Given that GNOME Shell already depends on
GTK3, it cannot use this next version of gnome-bluetooth. And
since GJS will automatically try and use the latest version
available of any library, Shell must specify it wants 1.0
explicitly.
Add a required GnomeBluetooth version number when importing it
for the status indicator.
magnifier: Disable desaturation effect when not in use
The desaturation effect was always enabled, regardless of the specified
factor. This was causing the magnifier to always hit offscreen
rendering, even when not necessary. Additionally offscreen rendering is
currently also causing glitches in MetaWindowGroup which this helps to
avoid in some cases.
Carlos Garnacho [Fri, 29 Oct 2021 14:46:18 +0000 (16:46 +0200)]
inputMethod: Do not reset invisible preedit on focus_out
If the preedit is in invisible state, the last preedit string that
was sent around is already null, so there is no need to clear the preedit
in that case.
Carlos Garnacho [Fri, 29 Oct 2021 14:43:15 +0000 (16:43 +0200)]
inputMethod: Do not communicate preedit text change on reset
This is now done in the ClutterInputFocus for us, since
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1940. There
is no need to duplicate this call, and it happens to cause undesired
scrolling to content in some cases when this reset happens during
a change in the IM focused client state.
Florian Müllner [Tue, 26 Oct 2021 17:40:08 +0000 (19:40 +0200)]
dbusServices/extensions: Only allow one dialog at a time
Showing multiple preference dialogs at the same time (for instance
by repeated `gnome-extensions prefs` calls) may or may not work as
expected, depending on whether any of the dialogs is modal or not
(read: opened via the Extensions app).
The easiest way to address this is to disallow more than a single
dialog at the time. It's arguably also the more predictable behavior,
and means extensions don't have to deal with inconsistent state
caused by multiple dialogs.
Florian Müllner [Thu, 28 Oct 2021 17:27:57 +0000 (19:27 +0200)]
dbusServices/extensions: Fix shutdown after showing prefs
GTK4 relies entirely on refcounting for cleanup (that is,
there is no longer a destroy() method that forces a dispose
run regardless of the refcount).
Unfortunately that makes cleanup harder in (some) language
bindings, where an object may be kept alive implicitly by
closures etc.
Address this by releasing the hold count when the window
is closed rather than when it is destroyed.
This isn't the most elegant, but it ensure that the service
doesn't get stuck if an extension doesn't carefully clean
up everything in its prefs widget.
Florian Müllner [Thu, 28 Oct 2021 17:55:23 +0000 (19:55 +0200)]
extensions-tool/prefs: Log D-Bus errors
We currently ignore any error that may occur when calling the
OpenExtensionPrefs D-Bus method. Right now such an error is highly
unlikely, given that we already checked that we are running under
gnome-shell and the extension in question exists and has prefs.
We'll soon make sure that only one dialog is shown at any time,
which is an error that we can realistically expect, so handle that
properly.
Sebastian Keller [Sat, 30 Oct 2021 11:21:17 +0000 (13:21 +0200)]
shell/app: Don't move all workspace windows to their startup workspace
Windows from some applications, such as guake, are created as showing on
all workspaces. When these windows are put on the workspaces via
set_workspace_state() during construction, the first time the window is
added to a workspace in the loop triggers the shell app tracker which
then tries to move the window to its startup workspace. This makes the
window unsticky which triggers another set_workspace_state() which
tries to remove the window from all workspaces, but currently it is only
on the first one and then adds it to the startup workspace. Once that is
finished, the first set_workspace_state() continues adding the window
to the remaining workspaces, despite the window now no longer having
on_all_workspaces set to true.
When the window is now unmanaged, the window according to its internal
state is only found on the startup workspace, so it will only be removed
from that. This causes the assertion to fail that checks that the window
is no longer present on any workspace after this.
Florian Müllner [Thu, 21 Oct 2021 17:50:18 +0000 (19:50 +0200)]
environment: Apply autoReverse/repeatCount to all transitions
Like the old Tweener API, ease() allows to transition multiple
properties at once. If autoReverse or repeatCount are specified,
they should apply to all transitions, but right now we only set
them for the first one.
Ray Strode [Tue, 10 Aug 2021 19:31:00 +0000 (15:31 -0400)]
sessionMode: Allow extensions at the login and unlock screens
Now extensions can specify which session modes they work in,
but specifying the login screen or unlock screen session modes in
an extensions metadata still won't work, because those session
modes disallow extensions.
Ray Strode [Sat, 28 Aug 2021 17:54:39 +0000 (13:54 -0400)]
extensionSystem: Allow extensions to run on the login screen
At the moment it's not realy possible to extend the login screen to do
things it doesn't have built-in support for. This means in order
to support niche use cases, those cases have to change the main
code base. For instance, oVirt and Vmware deployments want to be able
to automaticaly log in guest VMs when a user pre-authenticates through a
console on a management host. To support those use cases, we added
code to the login screen directly, even though most machines will never
be associated with oVirt or Vmware management hosts.
We also get requests from e.g. government users that need certain features
at the login screen that wouldn't get used much outside of government
deployments. For instance, we've gotten requests that a machine contains
prominently displays that it has "Top Secret" information.
All of these use cases seem like they would better handled via
extensions that could be installed in the specific deployments. The
problem is extensions only run in the user session, and get
disabled at the login screen automatically.
This commit changes that. Now extensions can specify in their metadata
via a new sessionModes property, which modes that want to run in. For
backward compatibility, if an extension doesn't specify which session
modes it works in, its assumed the extension only works in the user
session.
Ray Strode [Tue, 10 Aug 2021 17:25:57 +0000 (13:25 -0400)]
extensionSystem: Get rid of _enabled boolean optimization
At the moment a session mode either allows extensions or it doesn't.
If it allows extensions, then the entire available list of
configured extensions get enabled as soon as the session mode is
entered.
Since enabling or disabling extensions is an all or nothing situation,
the code tracks whether extensions are already enabled when entering
the session mode, and if so, avoids iterating through the extension list
needlessly. It does this using a boolean named _enabled.
In the future, the extensions themselves will be given some say on
whether or not they should be enabled in a given session mode. This
means, the configured extension list may contain extensions that
shouldn't be enabled for a given session mode, and the _enabled boolean
will no longer be appropriated.
This commit drops the _enabled boolean optimization.
Florian Müllner [Thu, 14 Oct 2021 16:53:17 +0000 (18:53 +0200)]
Post-branch mutter API bump
Since the CI pipeline now runs `meson dist` which includes a test
for an up-to-date NEWS entry, post-branch/release version bumps
have become unwieldy.
It's still useful to increase the API version right after branching
though, so return to bumping it manually and do so now.
Florian Müllner [Wed, 6 Oct 2021 16:51:03 +0000 (18:51 +0200)]
theme: Fix high-contrast switches
Commit 9d6fcfdc8506 dropped the different us/intl switches, but
the high-contrast theme was left out at the time (and therefore
ends up with the regular assets now).
Carlos Garnacho [Thu, 16 Sep 2021 08:03:38 +0000 (10:03 +0200)]
ci: Install mutter again on coverity job
This job "needs" the build job and thus gets its artifacts, but
it will not inherit the environment where we already installed
Mutter in /usr. This apparently broke once there was a more
recent mutter-clutter .pc file to look up.
Install Mutter in /usr again in this job, so the new build for
coverity finds all dependencies.
Ray Strode [Tue, 5 Oct 2021 15:01:19 +0000 (11:01 -0400)]
unlockDialog: Properly reset auth prompt when showing it
If a user hits escape twice really fast when coming back to
their machine to unlock it, they made end up getting presented
with a non-functional unlock screen that doesn't show their
user icon and doesn't ask for a password.
This is because showPrompt assumes that if an auth prompt already
exists, it's ready to go. That may not be true, if it's in the
process of getting torn down at the time because it's in the middle
of a cancel animation.
This commit solves the problem by ensuring the auth prompt is always
in a fresh reset state before showing it.
When commit 25bfe99ed53b5b7 replaced the thin libcanberra wrapper
with the (then) new MetaSoundPlayer API, it missed that the latter
expects files as GFile instead of a file path.
Commit dd2cd6286cd3 restricted callers of the screenshot methods to
portal implementations, gnome-settings-daemon and gnome-screenshot.
That restriction does make sense for the actual screenshot methods,
but `PickColor` is actually used by GTK in its color picker (and
therefore may be called from arbitrary applications).
Fix this by unrestricting access to `PickColor` again. Considering that
the method is always interactive, it's not very privacy/security-sensitive
anyway.
st/icon: Ensure icons are updated if theme node is invalid
Icons that are changed while an actor is not mapped may not have a theme
node associated with, and thus we'd end up not updating them at all.
In fact we return early in st_update_icon(), and this was not an issue
until commit 0b1dfbf6f3f, because we'd end up to update the icon anyways
once the style was changed (and so with a valid theme node), but since
such change we might not updating the icon if no theme detail changed.
To prevent this, add a flag to require an icon update when the theme
changed, if no successfully update happened earlier.
Jonas Ådahl [Tue, 21 Sep 2021 14:35:32 +0000 (16:35 +0200)]
Always assume GLSL is supported
The support for GLSL has been advertised as unconditionally supported by
mutter for a while, and has now lost the 'GLSL' feature completely. Lets
remove the checks.
St's theme test now requires the libmutter-test library, which is only
built when tests are enabled. Instead of mandating a particular build
configuration in mutter, add a corresponding option in gnome-shell as
well.
build: Check sassc errors when generating stylesheets
We currently just call sassc, but don't check its return value. That
means as long as sassc is available, the script (and therefore the
newly added dist CI job) will succeed.
Florian Müllner [Tue, 31 Aug 2021 11:06:22 +0000 (13:06 +0200)]
ci: Add dist job
So far, releases are done locally by invoking `meson dist`.
We can do better and leverage the existing CI infrastructure, to get
to the following release workflow:
- bump version in meson.build, update NEWS etc.
- open merge request for the release
- merge when the pipeline (including dist check) succeeds
- tag the release
- wait for the tag pipeline to spit out the tarball artifact
Robert Mader [Sun, 6 Jun 2021 12:29:47 +0000 (14:29 +0200)]
windowManager: Use MetaWindowActor.paint_to_content() for window effects
Previously we used `get_image()`, which returned `cairo_surface`,
in order to create a `ClutterContent` with a screenshot of the
`MetaWindowActor`. This added a roundtrip from GPU to CPU memory.
Instead, use the new `paint_to_content()` API which lets us use a
`CoglTexture` directly as source of our `ClutterContent`.
Carlos Garnacho [Tue, 27 Jul 2021 14:52:32 +0000 (16:52 +0200)]
inputMethod: Use update-preedit-with-mode signal
This signal announces the preferred commit mode of the preedit text when
the input context is reset. Keep this mode around, and ensure to honor this
mode (e.g. maybe commit the preedit string) whenever the input method would
be reset.
This is delegated to the internal layers, so propagate this mode via
clutter_input_method_set_preedit_text().
Florian Müllner [Thu, 17 Jun 2021 13:29:42 +0000 (15:29 +0200)]
shellDBus: Restrict callers
The org.gnome.Shell interface provides a private API to other core
components to implement desktop functionalities like Settings or
global keybindings. It is not meant as a public API, so limit it
to a set of expected callers.
If we throw an error when the invocation isn't allowed instead of
returning false, we can simply return that error instead of duplicating
the error handling.
Florian Müllner [Wed, 16 Jun 2021 23:50:50 +0000 (01:50 +0200)]
shellDBus: Use MetaContext:unsafe-mode to restrict Eval()
The Eval() method is unarguably the most sensitive D-Bus method
we expose, since it allows running arbitrary code in the compositor.
It is currently tied to the `development-tools` settings that is
enabled by default. As users have become accustomed to the built-in
commands that are enabled by the same setting (restart, lg, ...),
that default cannot easily be changed.
In order to restrict the method without affecting the rather harmless
commands, guard it by the new MetaContext:unsafe-mode property instead
of the setting.
panel: Show warning indicator when unsafe-mode is on
MetaContext added an unsafe-mode property, which we will use to restrict
a number of privileged operations unless it is enabled. It is meant to
only be enabled temporarily for development/debugging purposes, so add
a scary icon to the top bar as a reminder to turn it off again.
Since touchpad smooth scroll events with source finger are handled by
the swipeTracker, the workspace scroll handler can focus on discrete
events.
Thanks to Mutter emulating discrete scroll events, see
meta_seat_impl_notify_scroll_continuous_in_impl in meta-seat-impl.c,
it is safe to ignore smooth scroll in the workspace scroll handler
and handle exclusively discrete events.
In addition, once high-resolution scroll events land in Mutter [1], a
mouse will be able generate non emulated smooth scroll events that
should be ignored in favour of the discrete scroll events.
Otherwise, a single mouse wheel click will scroll through multiple
workspaces at once.
Florian Müllner [Thu, 26 Aug 2021 19:54:15 +0000 (21:54 +0200)]
layout: Exclude hidden actors from struts
Currently we only exclude hidden actors from the input region,
but not from struts. This is an oversight that went unnoticed
until now, probably because we didn't consider the work area
in the overview until recently.
Sebastian Keller [Tue, 17 Aug 2021 22:45:16 +0000 (00:45 +0200)]
keyboard: Ensure chrome input region is updated after slide animation
The keyboard is placed outside of the screen when shown and then slides
in via a transition that changes the translate-y property. This
translation does not affect the allocation of the Keyboard actor and as
such does not trigger any of the signals LayoutManager is connected to
to update the input region. This means the input region remains at the
original position of the actor outside of the screen and as a result on
X11 clicks will go through to the underlying window.
There was a workaround for this by queuing a relayout at the end of the
transition, but this stopped working due to optimizations avoiding
unnecessary allocation changes.
This updates that workaround to toggle the visibility of the actor
instead, which is the other signal that LayoutManager reacts to. Once
ClutterActor provides better ways to react to transforms this can
hopefully be removed entirely.
projects / gnome-shell.git / log