Florian Müllner [Thu, 23 Jun 2022 15:02:55 +0000 (17:02 +0200)]
autorunManager: Promisify ContentTypeDiscoverer
Thanks to recent gjs changes - namely the ability to promisify
interface prototypes and promise-based D-Bus wrappers - we can
modernize the ContentTypeDiscoverer API to use an async function
instead of a callback.
Florian Müllner [Thu, 23 Jun 2022 16:13:36 +0000 (18:13 +0200)]
search: Use async functions instead of callbacks
Our search provider API has to be asynchronous to support remote
providers. It doesn't have to be based on callbacks though, now
that async functions provide a nicer alternative.
That is particularly true after gjs's D-Bus wrapper started to
generate promise-based method variants.
Location services aren't expected to change frequently: They are
either disabled globally, or permissions are granted on a per-app
basis.
This is less of a concern while the setting is exposed in a small
submenu, but as we moving to quick settings, it does not deserve
the prominence of a quick toggle. Just the top bar icon and Settings'
privacy panel should be enough.
Other than connecting stream when necessary, the volume menu
only forwards slider events to the indicator, and method calls
from the indicator to the appropriate slider.
Just cut our the middle-man and let the indicator handle the
slider items directly.
There's no good reason for waiting for the proxy to be initialized
to connect signals. In fact, connecting the signal beforehand
ensures that the handler is in place when the proxy fetches the
properties, so we don't have to call the handler explicitly.
That in turn allows us to rely on the signal parameters to only process
changed properties.
To achieve that, construct the proxy manually, and then initialize
it asynchronously in a Promise.
status/nightLight: Use promise to initialize proxy
There's no good reason for waiting for the proxy to be initialized
to connect signals. In fact, connecting the signal beforehand
ensures that the handler is in place when the proxy fetches the
properties, so we don't have to call the handler explicitly.
That in turn will allow us in a follow-up to rely on the signal
parameters to only process changed properties.
To achieve that, construct the proxy manually, and then initialize
it asynchronously in a Promise.
By default, gitlab skips checking out of submodules. So far this
didn't matter to us, as meson will initialize submodules as part
of setup.
It no longer does for the dist command however, and as we run it
from a separate job where the submodule checkout isn't preserved,
the job breaks with recent meson versions.
Fix this by instructing gitlab to checkout the submodule for all
jobs that need it.
extensionSystem: Check if mode is supported when loading
Don't try to initialize and then enable an extension if it is not supported
in the current mode. Otherwise the extension will not be initialized, and the
initial ERROR state is not cleared. Once it is in ERROR state we can't enable
it anymore when we switch mode. Instead, leave the extension in INITIALIZED
state, so that it will be initialized when appropriate. This happens for
extensions that support the unlock-dialog mode but not the user mode.
Emmanuele Bassi [Fri, 20 May 2022 16:40:35 +0000 (17:40 +0100)]
screencast: Use the same format for screencasts and screenshot file names
Screenshots use `%Y-%m-%d %H-%M-%S` for the timestamp, which has the
advantage of allowing proper lexicographical sorting.
The screencast file name pattern, on the other hand, uses
locale-dependent expansions, which break sorting based on file name, and
introduces the chance of potentially invalid characters on different
file systems.
`ShouldShowAirplaneMode` only considers whether airplane mode
should be shown for the form factor, not whether there are
any actual kill switches available.
That's tracked in a separate property, `HasAirplaneMode`.
Take that into account for our `:show-airplane-mode` property,
so that it reflects when airplane mode should and *can* be shown.
Right now we only show airplane mode when it is enabled (and
therefore available), but this will change in the future.
There's no good reason for waiting for the proxy to be initialized
to connect signals. In fact, connecting the signal beforehand
ensures that the handler is in place when the proxy fetches the
properties, so we don't have to call the handler explicitly.
That in turn will allow us in a follow-up to rely on the signal
parameters to only process changed properties.
To achieve that by constructing the proxy manually, and then
initialize it asynchronously in a Promise.
Carlos Garnacho [Mon, 18 Jul 2022 13:11:49 +0000 (15:11 +0200)]
inputMethod: Update to set_preedit_text() API change
Since IBus does not provide this information right away, we
so far cannot do much about providing a truthful anchor position
for the preedit text. But with the Mutter API in place it will
be up to this object to do so in the future.
When gjs overhauled its coding style, it opted for disallowing
spaces around braces in object literals: `{foo: 42}`.
We opted out of that particular rule in commit 09b8e8f, primarily
because it would have been a very intrusive change with the old
style of object literals:
```js
let oldStyle = { foo: 42,
bar: 23 };
```
Alas, all multi-line object literals have been transitioned to
the new style where braces go on separate lines, so they are no
longer subject to the rule one way or the other.
Nowadays the rule mostly affects destructuring, and there it
is a bit odd to apply different spacing rules to arrays and
objects:
This is now the main divergence from the canonical gnome style,
and the only one where we directly contradict it.
It would be good to transition away from that, and as the rule
now only affect destructuring and single-line object literals,
we can do that on a case-by-case bases by moving the override
to the legacy configuration.
As desctructuring imports makes up a fair share of the affected
code, and those will change when moving to ES6 modules, this
seems like a good moment to start that transition.
Florian Müllner [Sun, 12 Jun 2022 16:27:34 +0000 (18:27 +0200)]
notificationDaemon: Stop auto-dismissing on focus changes
This behavior dates back all the way to the original gnome-shell
release, and it did make sense at the time:
- we were first to make notifications persistent instead of
closing them after a timeout; apps were therefore used to
treat notifications as fire-and-forget instead of closing
them when no longer relevant
- neither web notifications not portals did exist yet, so the
D-Bus API was only used directly, instead of as a backend
for other APIs; as a result, focusing the app was more likely
than not to put the source of the event that the user was
notified about into view
Nowadays both persistent notifications and web notifications
are wide-spread, so the original reasons no longer apply.
This change helps web notifications in particular, as it reduces
the differences between XDG and portal-based implementations.
Matthew Jakeman [Fri, 1 Jul 2022 11:24:08 +0000 (23:24 +1200)]
extensions: Allow updating with only extension-manager installed
GNOME Shell will only update extensions if the org.gnome.Extensions
app is present. This commit adds alternative support for
com.mattjakeman.ExtensionManager to the extension updating mechanism.
It enables updates to occur when extension-manager is the sole installed
tool on the system. When both applications are installed,
org.gnome.Extensions is preferred at all times.
Meson now warns that the default will change in a future version,
so set it explicitly to maintain the existing behavior (and shut
up the meson warning).
Maksym Hazevych [Mon, 20 Jun 2022 14:43:09 +0000 (17:43 +0300)]
dateMenu: Remove custom aligning of time strings in the World Clocks
Since now Glib uses tabular space for aligning numbers (merged in
https://gitlab.gnome.org/GNOME/glib/-/merge_requests/2698), there is no
more need for custom aligning.
The iterator was pointing to NULL when going out of scope, leading to
autofree not clearing the list.
==300183== 32 bytes in 2 blocks are definitely lost in loss record 14,798 of 38,939
==300183== at 0x484586F: malloc (vg_replace_malloc.c:381)
==300183== by 0x4D7D980: g_malloc (gmem.c:127)
==300183== by 0x4D95AB3: g_slice_alloc (gslice.c:1074)
==300183== by 0x4D96D96: g_slist_prepend (gslist.c:282)
==300183== by 0x487854F: shell_app_get_windows (shell-app.c:794)
==300183== by 0x48791A1: shell_app_get_pids (shell-app.c:1201)
==300183== by 0x488B293: shell_window_tracker_get_app_from_pid (shell-window-tracker.c:702)
==300183== by 0x488B632: get_app_from_window_pid (shell-window-tracker.c:370)
==300183== by 0x488B632: get_app_for_window (shell-window-tracker.c:436)
==300183== by 0x488B632: track_window (shell-window-tracker.c:549)
==300183== by 0x4CDBB75: g_cclosure_marshal_VOID__OBJECTv (gmarshal.c:1910)
==300183== by 0x4CD8BE9: _g_closure_invoke_va (gclosure.c:893)
==300183== by 0x4CF2A28: g_signal_emit_valist (gsignal.c:3406)
==300183== by 0x4CF2C2C: g_signal_emit (gsignal.c:3553)
environment: Remove inline import of extensionUtils
The logging function cannot be asynchronous, so move the override
into main.js where ExtensionUtils can be imported at the top level.
Importing ExtensionUtils in environment.js at the top level is not
possible because it would import Main prematurely.
extensionUtils: Add utility for setting the current extension
ES modules do not allow exports to be overriden, in anticipation
of the ESM port add a `setCurrentExtension` utility which will
throw if used in the shell. This is tested using a conditional
import of Main.
main: Define a custom log writer to dump on structured messages
Even though GNOME Shell is not explicitly using structured logging via
G_LOG_USE_STRUCTURED, GLib uses it as default since 2016 [1], and so
we're de facto using it.
As per this, if backtrace on warnings is enabled, it is ignored since the
log handler isn't used anymore, and no dump is printed.
Thus, replace the default log handlers with writer functions instead, honoring
backtrace-warnings debug string.
signalTracker: Correctly guard against invalid GObject.ConnectFlags
We considered any ConnectFlag value major than SWAPPED as invalid, while
it's technically not fully true as we need to ensure that the passed
value is respecting the whole flags mask.
In fact, per se SWAPPED|AFTER (> SWAPPED) is a valid value (even if we
don't support the AFTER value).
signalTracker: Avoid getting the same owner object proto multiple times
While untracking an object we used to compute it's proto for each signal
we were disconnecting from, while this is not needed when we're just
iterating over all the same owner signals, so let's add few more
functions to compute an object prototype, and repeat the disconnections
in the simplest way we can.
signalTracker: Avoid creating a temporary keys array when clearing
We used to create a temporary array of signal tracker keys and then to
iterate through them in order to untrack the objects, but the Map's can
be iterated directly so let's just use their native forEach.
signalTracker: Avoid doing two maps lookups once they're initialized
Avoid checking if a map contains a value, given that in the case it's
not there, get() will just return an undefined value. And only in such
case we need to set the value.
Adopt EventEmitter class instead of injecting Signal methods
Introduce a new class, EventEmitter, which implements signal
handling for pure JavaScript classes. EventEmitter still
utilizes GJS' addSignalMethods internally.
EventEmitter allows static typechecking to understand the
structure of event-emitting JS classes and makes creating
child classes simpler.
The name 'EventEmitter' mirrors a common name for this pattern
in Node and in JS libraries.
Florian Müllner [Fri, 6 May 2022 14:45:20 +0000 (16:45 +0200)]
ci: Try falling back to stable branch before HEAD
For non-MR pipelines, we check for a matching branch or tag name
in mutter. As we use the same naming policy for stable branches
and branch at about the same time, this works generally fine for
branches.
However for tags, it is less unlikely that there is no matching
tag in mutter, in particularly late in a stable cycle when most
changes are translation updates (which are much rare in mutter
than gnome-shell).
Before falling back to main (which is doomed to fail in that case),
try to guess the correct stable branch based on the tag name.
ibusCandidatePopup: Prevent it from covering screenshot UI
If we only raise screenshot UI to top, ibusCandidatePopup can still
cover screenshot UI because each time ibusCandidatePopup updates it's
visibility it will raise it to top and this also happens if we open
screenshot UI via keybinding. This commit fixes it by only raising it
above keyboardBox, because keyboardBox is above all entries in modal
dialogs.
ibusCandidatePopup actually works as a top chrome because it always
raises itself above top_window_group when updating visibility, so just
initing it as a top chrome instead of initing it below top_window_group.
Florian Müllner [Thu, 30 Jun 2022 13:36:47 +0000 (15:36 +0200)]
layout: Force size of overviewGroup
The overviewGroup's size is currently determined by the coverPane
actor. That actor is only shown during transitions, so we rely on
ClutterFixedLayout including hidden children in its size request.
That odd behavior is about to change, so we need to make sure the
overview still requests the correct size.
Florian Müllner [Thu, 30 Jun 2022 13:29:55 +0000 (15:29 +0200)]
layout: Use constraint to resize screenShieldGroup
We want the actor to always fill the entire uiGroup. It seems
better and simpler to express that with a contraint than resizing
it explicitly on every monitor change.
oreo639 [Mon, 27 Jun 2022 00:56:48 +0000 (17:56 -0700)]
ibusManager: Pass on envrionment variables to ibus-daemon
Previously only DISPLAY was being passed on to ibus, which results in issues
with missing the WAYLAND_DISPLAY or the DBUS_SESSION_BUS_ADDRESS environemnt
variables for example.
GNOME_SETUP_DISPLAY is only necessary when using xwayland on demand,
which is only supported on systemd, in which case you should consider
using the user service provided by ibus.
When a window with a modal dialog gets minimized and at the same time
the dialog is closed, the WindowDimmer undim animation starts and gets
cancelled when the minimize animation is done, because that unmaps the
window actor.
In this case we want ensure the dimming effect still goes into a
proper state instead of being stuck mid-animation, so listen to
onStopped instead of onComplete for syncing state of the window dimmer.
While at it, clean things up a little and move the check for the
attach-modal-dialogs pref inside the _syncEnabled() function.
Carlos Garnacho [Wed, 29 Jun 2022 11:06:07 +0000 (13:06 +0200)]
ibusCandidatePopup: Set OSK candidates list visibility correctly
Right now, we don't pay much attention to visibility hints from IBus
about the candidates list for the OSK suggestions bar. Since some
IMs rely on this visibility being honored, do that.
To fix this, the visibility hints for the lookup table are now
propagated to the keyboard, so the Suggestions actor hides its
internal candidate buttons. Since the Suggestions actor gets its
minimum height from CSS, this does not result in OSK size jumps.
projects / gnome-shell.git / log