We're storing in the texture cache images and scaled images appending
the scaling factor to the key. When a file changes the cache key
corresponding to that file is removed, but not the keys for the scaled
ones so that images in the cache are never reloaded.
This patch removes all keys from the cache related to the file that
changes, including those with the scaling factor.
A new set (hash table) was added to keep track of scale used to be able
to remove all possible images in the cache.
When the KEY is removed from the cache, we can look now in the scale set
for and each scale we also remove the key "KEY1.000000", "KEY2.000000",
etc.
Assuming that the number of used scales is small (I would typically
expect one or two), the overhead should be negligible.
https://gitlab.gnome.org/GNOME/gnome-shell/merge_requests/567
If the GNOME shell crashes, we run a service that may disable
extensions. This is important so that users will not be locked out of
their own session in case an extension is causing crashes.
As this is a very agressive action, we tried to only do this in the
first two minutes of the session. Unfortunately, the logic was broken
and would result in an unclean session shutdown.
Fix this by using the newly introduced gnome-shell-disable-extensions
file. This is created by the extension subsystem for a period of time to
indicate the extensions may be the cause of a gnome-shell failure.
See
https://gitlab.gnome.org/GNOME/gnome-session/issues/43
for a log of the bug happening and the gnome-session part to fix this.
https://gitlab.gnome.org/GNOME/gnome-shell/merge_requests/858
When the extension system is loaded, create the
gnome-shell-disable-extensions file in the users runtime directory. This
file is automatically removed 60s later. The sole purpose of this file
is to be consumed by the systemd units. If the file exists, the systemd
units will disable extensions when the gnome-shell fails.
https://gitlab.gnome.org/GNOME/gnome-shell/merge_requests/858
When removing a string from a settings list, we iterate over all
existing entries and copy all strings except the one that's being
removed to a new list, which is then written to GSettings.
However we currently always increment the index, so we end up with
a NULL entry in place of the removed entry, which is then interpreted
as the end of the list. In other words, we also remove all entries
that follow the removed string.
Fix this by looping over the list entries instead of the index, and
only increment the index for entries we copy.
https://gitlab.gnome.org/GNOME/gnome-shell/issues/1946
The whitelist is a list of well-known D-Bus names, which we then search
for the unique name we get from the method invocation - unsuccesfully.
Fix this by watching the bus for any name in the whitelist in order
to maintain a map from wel-known to unique name that we can use for
matching.
https://gitlab.gnome.org/GNOME/gnome-shell/issues/1916
(cherry picked from commit eee1ab4890)
Since the notification message close button had no border, or mouse
over effect, there was no way to determine whether the mouse cursor
were over the button.
Improve this by adding a message-close-button class for the close
button, and a styling for its hovered state, based on media control
button styling.
https://gitlab.gnome.org/GNOME/gnome-shell/merge_requests/855
Increases contrast between normal and hovered states in
message-media-control buttons. Previously there was very little
difference between the two states, making it hard to distinguish
whether the mouse cursor was over the button.
https://gitlab.gnome.org/GNOME/gnome-shell/merge_requests/855
Since commit mutter/a2a8f0cda we force the focus surface of the
meta-wayland-pointer to NULL while the pointer is hidden. This
introduced an issue with the magnifier, where we use
`set_pointer_visible` to hide the real cursor and show our own cursor at
the correct position: Because the meta-wayland-pointer is still used to
communicate with Wayland clients, the UI of the windows will not respond
to mouse movement anymore as soon as the real cursor is hidden.
To fix this, use the newly added API of MetaCursorTracker to switch back
to the old behavior while the magnifier is hiding the system cursor.
In the future and as a more correct fix, we might want to rewrite the
magnifier so it doesn't have to hide the cursor and can simply show the
default one (eg. by scaling the actual view instead of a clone of the
view).
https://gitlab.gnome.org/GNOME/gnome-shell/merge_requests/754
When a user has no password and a polkit authentication is started,
instead of blindly initiating the admin session, show the regular
"Authentication Requested" dialog (but without the password entry). This
makes sure that the user's admin session is only effectively started
after the user chooses to proceed with the authentication, which
provides an extra confirmation step that can be vital for critical
tasks.
To do this, we show the dialog inside `_onUserChanged()` right after the
dialog was created instead of calling `performAuthentication()` from
`_onInitiate()`. The bug mentioned in `_onInitiate()` is no longer an
issue since we show the dialog in all cases now anyway.
Ideally we should use a different wording than "authentication" when the
user has no password set, and use "confirmation" instead. However polkit
already sends the requests with such messages (e.g. "Authentication is
required to configure software repositories"), and it's important to
show those to the user, so this patch keeps the regular wording.
https://gitlab.gnome.org/GNOME/gnome-shell/merge_requests/829
Currently when dragging an icon to the space above or below the appGrid
to switch pages, we do so very quickly without checking when the last
page-switch happened. This makes it hard to move icons to pages which
are not the first or the last one, since the other pages are skipped
very quickly.
To fix this, add a timeout of 1 second that blocks switching pages after
a page-switch using drag overshoot occured.
Fixes https://gitlab.gnome.org/GNOME/gnome-shell/issues/1693
We currently always switch app pages when a dragged app icon
moves outside the grid boundaries, regardless of any previous
page switches. This makes it too easy to switch multiple pages
accidentally, so add a small threshold that the icon has to
move back towards the grid before allowing another page switch.
https://gitlab.gnome.org/GNOME/gnome-shell/issues/1693
There's no need for a `inhibitEventBlocker` interface. Since we connect
to "open-state-changed" of our folders in the AllView anyway, we can
just make the event blocker visible while a folder is opened, and hide
the event blocker during DnD.
This allows keeping the eventBlocker reactive at all times and fixes an
issue where DnD to create a new folder is impossible if no folders are
present because the eventBlocker would not get inhibited.
Fixes https://gitlab.gnome.org/GNOME/gnome-shell/issues/1652
While we support a reasonable list of layouts nowadays, we don't
include many variants like `fr+oss`. Instead of directly falling
back to the `us` layout, try stripping the variant first, as the
base layout is likely closer to the expectation than `us`.
https://gitlab.gnome.org/GNOME/gnome-shell/issues/1907
(cherry picked from commit c1ec7b2ffa)
Commit c899453800 lifted the requirement of switcher keybindings
to contain a modifier, however it is currently only possible to
finish it by letting it time out.
Improve that by also accepting space/enter key presses to confirm the
selection immediately.
https://gitlab.gnome.org/GNOME/gnome-shell/issues/1883
The noModsTimeout obviously finishes inside a timeout callback, which
means `global.get_current_time()` might return Clutter.CURRENT_TIME (ie.
0) when called inside it, because it's not called while handling an
event. This means when switching apps or activating a window, the
timestamp passed to `activate_window` may be 0, which is the reason why
the altTab switcher is currently broken when using modifier-less
keybindings.
Fix that by using `meta_display_get_current_time_roundtrip`, which
always return a valid timestamp, instead of
`shell_global_get_current_time`.
https://gitlab.gnome.org/GNOME/gnome-shell/merge_requests/847
g_desktop_app_info_get_categories() may return null. In that case, the
previous code would fail to create a folder when dragging an app with
no categories onto another app. Instead, simply continue with the next
app info.
https://gitlab.gnome.org/GNOME/gnome-shell/merge_requests/782
We currently only update the windows section when either the focus
app changes, or when the app's windows change (that is, a window is
opened or closed). This allows the menu item labels to become stale
if the window title changes after one of those events (for example
when switching tabs).
Fix this by updating menu items when the corresponding window
title changes.
https://gitlab.gnome.org/GNOME/gnome-shell/issues/1830
I've observed that UPower can occasionally report a charge level of 100%
while the state is still "charging". This usually doesn't last very long
but it is noticeable because the power icon changes to a "missing icon"
icon. This will handle that rare case correctly.
https://gitlab.gnome.org/GNOME/gnome-shell/merge_requests/814
The close dialog is added as a child to MetaWindowActor, and, in Wayland
sessions, since commit [1] MetaWindowActor applies a transformation
matrix which scales all it's children using the geometry scale factor.
Now because the dialog actor is not a window (i.e. a MetaSurfaceActor),
but a subclass of StWidget, the scale factor is also applied to the
properties of the dialog by StThemeNode, so we end up applying the
geometry scale twice to the close dialog.
Fix this by applying the inverted scale to the dialog, which leaves the
scaling only to MetaWindowActor. This means we also can't apply a pivot
point other than 0 to the dialog actor, so apply the 0.5-pivot point to
the `_dialog` child of the Dialog class (the actual visible dialog box)
and also perform scaling animations on this child.
[1] https://gitlab.gnome.org/GNOME/mutter/commit/fb9e8768https://gitlab.gnome.org/GNOME/gnome-shell/merge_requests/783
It may happen that the app icon is destroyed with a drag
monitor still around, in which case, a load of warnings
will be shown.
Make sure to remove any pending drag monitor on destroy.
https://gitlab.gnome.org/GNOME/gnome-shell/merge_requests/841
When removing the last icon of a folder, FolderView first removes
the folder from org.gnome.desktop.app-folders.folder-children, then
proceeds to reset all its keys, which removes the relocatable schema.
That order of operations turns out to be problematic. Removing the
folder from 'folder-children' destroys the folder icon, which in turn
destroys the folder view, which throws a load of warnings in the
journal.
Fix that by removing the folder after resetting the schema keys. In
fact, what we're doing here is not using 'this' anymore.
https://gitlab.gnome.org/GNOME/gnome-shell/merge_requests/841
The variable that holds the list of application icons is
called 'newApps', but that technically was never true,
since we only create new app icons when necessary.
Rename it to 'appIcons'.
https://gitlab.gnome.org/GNOME/gnome-shell/merge_requests/841
The views (AllView and FrequentView) build a list of all applications
they contain. BaseView then diffs between what's currently added, and
what needs to be added, and removed.
This approach has a problem though: creating an AppIcon or a FolderIcon
connects to various signals, and we confuse the garbage collector.
When building the list of applications, instead of always creating new
icons, try to use already existing icons first.
Fixes https://gitlab.gnome.org/GNOME/gnome-shell/issues/1610
Fixes https://gitlab.gnome.org/GNOME/gnome-shell/issues/1694
Resizing effects are more finicky as other effects, as the actual
animation is delayed until we receive the ::size-changed signal.
However that signal may never be emitted if the window is destroyed
just after starting the size-change effect, in which case the effect
is never completed, blocking mutter from destroying the corresponding
window actor.
Address this by tracking when a resize effect is pending, and complete
the effect when appropriate.
https://gitlab.gnome.org/GNOME/mutter/issues/655https://gitlab.gnome.org/GNOME/gnome-shell/merge_requests/815
For stable branches, we currently only check out the correct mutter
branch for merge requests. For the regular pipeline, our code to
determine the current shell branch fails because CI runs on a
temporary "pipeline/12345" branch that doesn't exist for mutter.
Switching to the correct gitlab environment variable fixes that.
https://gitlab.gnome.org/GNOME/gnome-shell/merge_requests/811
(cherry picked from commit fd5989e99a)
In the case of bugs in a drag target's acceptDrop() function, it may
throw an exception. In the previous code, this would break out of the
loop entirely and never cancel the drag, so the mouse button release
event would be ignored and you would have to press Esc to get out of the
drag.
In this change, if acceptDrop() throws an exception, we log it and move
on to the next parent target instead.
https://gitlab.gnome.org/GNOME/gnome-shell/merge_requests/777
(cherry picked from commit 055c007ac2)
When accessing properties on ClutterActor for size and position there is
a notable access time overhead. This overhead adds considerable user lag
when opening the overview if many windows are open.
This is primarily due to these properties being accessed while sorting
WindowClone instances by their window's center for placement in the
overview. By pre-computing this center value only once when
initializing WindowClone, the induced lag can be significantly reduced.
https://gitlab.gnome.org/GNOME/gnome-shell/merge_requests/763
(cherry picked from commit d91927674d)
