Florian Müllner [Thu, 27 Feb 2020 21:46:44 +0000 (13:46 -0800)]
environment: reduce calls to g_time_zone_new_local()
Creating a new GTimeZone for the local timezone can be quite expensive if
done repeatedly. It requires an open(), mmap(), and parsing of
/etc/localtime.
This patch was provided by Florian, and I've tested it as far back as
3.28.4 to ensure that we are really reducing the number of open() calls
on the compositor thread.
global: force fsync() to worker thread when saving state
The g_file_replace_contents_async() API can potentially call fsync() from
the thread calling into it upon completion. This can have disasterous
effects when run from the compositor main thread such as complete stalls.
This is a followup to 86a00b6872375a266449beee1ea6d5e94f1ebbcb which
assumed (like the rest of us) that the fsync() would be performed on the
thread that was doing the I/O operations.
You can verify this with an strace -e fsync and cause terminal to display
a command completed notification (eg: from a backdrop window).
This also fixes a lifecycle bug for the variant, as
g_file_replace_contents_async() does not copy the data during the operation
as that is the responsibility of the caller. Instead, we just use a GBytes
variant and reference the variant there.
Florian Müllner [Tue, 18 Feb 2020 18:05:47 +0000 (19:05 +0100)]
ci: Check that files with translatable strings are listed in POTFILES
Also known as "Piotr Drąg Bot".
We will soon make sure that files processed by xgettext don't use template
strings. To make that check as adequate as possible, ensure that no source
code files are missing from POTFILES.
Jonas Dreßler [Mon, 27 Jan 2020 22:42:28 +0000 (23:42 +0100)]
altTab: Don't fade out thumbnails on destroy
The thumbnails actor `this._thumbnails` has already been destroyed when
calling `_destroyThumbnails()` from the `destroy` signal handler because
it is a child actor of the AppSwitcherPopup. So stop destroying the
thumbnails separately (fading them out inside a destroy handler wouldn't
make sense anyway).
Jonas Dreßler [Wed, 15 Jan 2020 16:04:07 +0000 (17:04 +0100)]
switcherPopup: Use correct scroll-direction property
The correct property for the scroll-direction with scrolling events is
`direction`, no `scroll_direction`. This fixes scrolling in the alt-tab
popup, which broke with the actorization changes.
Jonas Dreßler [Thu, 6 Dec 2018 17:05:39 +0000 (18:05 +0100)]
switcherPopup: Always return true when the popup finished early
When the switcherPopup was initiated successfully, we return true,
otherwise the WindowManager will try to destroy it. Since an early
release of the keystroke will also switch to another application and
close the switcher just fine, we should return true to indicate success
here.
Jonas Dreßler [Wed, 15 Jan 2020 15:19:41 +0000 (16:19 +0100)]
switcherPopup: Lookup index of items when hovering or clicking
Right now the index that gets selected on click and motion events is set
when connecting the event inside the addItem function. If items are
added or removed (for example when an application is closed by pressing
"q"), this index isn't valid anymore and has to be updated.
To fix this, use the items themselves instead of the index as arguments
for the event handlers and lookup the index of the item inside the event
handler.
Jonas Dreßler [Thu, 26 Apr 2018 09:19:40 +0000 (11:19 +0200)]
switcherPopup: Implement functions to add and remove accessible states
Set the accessible states of the switcherList items by calling a
function instead of manipulating class-internal variables from outside
the class, which is considered bad practice.
The check whether the item at `_selectedIndex` exists can also be
removed since we always select a new index after an item was removed
(i.e. an app was closed) and destroy the alt-tab switcher right away if
no more items exist (see `SwitcherPopup._itemRemovedHandler`).
Jonas Dreßler [Wed, 25 Apr 2018 21:56:29 +0000 (23:56 +0200)]
switcherPopup: Swap Math.max and Math.min in _scrollToLeft
Limit the minimum and maximum value to scroll to inside the box to 0 and
the upper limit, for some reason this was done right in _scrollToRight,
but not in _scrollToLeft.
This fixes the behavior of scrolling to the left: Before, scrolling one
item to the left moved the view to the first element of the list (this
can make the selected element invisible in large lists). Instead, scroll
one item to the left, just like scrolling to the right works.
Jonas Dreßler [Wed, 25 Apr 2018 21:24:05 +0000 (23:24 +0200)]
switcherPopup: Use local variable for index in scrolling functions
Make sure the index that's being scrolled to doesn't change while the
scrolling animation is running by using an argument instead of the
this._highlighed class scope variable.
This fixes wrongly shown arrows when selecting new elements faster than
the scrolling animation takes for one index. The check run to disable
the arrow might be checking against a newer index than the one at the
start of the animation which results in the arrow not getting hidden
even if no more scrolling is possible.
Jonas Dreßler [Wed, 25 Apr 2018 20:14:57 +0000 (22:14 +0200)]
switcherPopup: Use this._highlighted to check for reentrancy
Since this._highlighted is always set to the currently highlighted
index, there's no need to save the index to a separate variable. This
obviously depends on getting the new item highlighted as a result of the
item-entered event.
This fixes bugs in situations where the highlighted element changes
after an event that is not calling _onItemEnter, for example after
scrolling or pressing a key. In those cases the _currentItemEntered
variable wouldn't be updated and the old item couldn't be entered
anymore without entering another one before.
Daniel van Vugt [Wed, 26 Feb 2020 03:53:40 +0000 (11:53 +0800)]
Revert "js/ui: Use Clutter.OffscreenRedirect.ON_IDLE"
This reverts commit c0c027c608d0aa1c547a2a82bd3ce6b84bb83b83. Because for
some reason animating external opacity and position is still incurring
internal repaints, which disables offscreening and makes fading of
overlapping actors look wrong. `ON_IDLE` should be fixed in mutter before
it is used (in boxpointer at least) again.
Jonas Dreßler [Mon, 24 Feb 2020 10:19:28 +0000 (11:19 +0100)]
overview: Hide the overview on session mode hasOverview changes
If the sessionMode does not allow to show the overview, we should also
hide an already visible overview.
This fixes a bug where, if the lockscreen was shown while the overview
was visible, the Ctrl+Alt+Tab popup would allow navigating inside the
overview because the overview actor is still mapped.
Florian Müllner [Sat, 22 Feb 2020 15:35:32 +0000 (16:35 +0100)]
environment: Handle reversed transition with 0 duration
If a transition is reversed, the final property values will be the
same as before the transition. However this currently only works
correctly when we actually use a transition; to fix this with a
duration of 0, simply skip the set() call when the transition is
reversed.
Florian Müllner [Thu, 20 Feb 2020 15:10:36 +0000 (16:10 +0100)]
theme: Fix horizontal default user widget
Commit 6c6c89c63470b5 added a pill around the default avatar, but
assumed the sizes from the vertical widget used on the lock screen.
In order to fix the horizontal widget on the login screen, move the
size-specific bits to the corresponding .horizontal and .vertical
sections, and half the sizes for the former (which corelates with
the icon sizes).
Jonas Dreßler [Thu, 20 Feb 2020 18:57:20 +0000 (19:57 +0100)]
dateMenu: Sync initial state of the message indicator
While the unread messages indicator is updated when starting a new
session because we call _onSourceAdded() on existing sources, we should
also update the do-not-disturb setting which might still be enabled.
Implement ClutterActorClass.has_accessible() to ensure that CallyActor does not
recreate accessibles during the removal/destruction of an actor. This relies
on GNOME/mutter!1083 for the ClutterActorClass.has_accessible virtual function.
Running GNOME Shell for about 30 seconds results in a difference between
the two runs.
Florian Müllner [Wed, 19 Feb 2020 17:06:38 +0000 (18:06 +0100)]
unlockDialog: Set accessible name of icon-only buttons
We turned both the auth prompt's cancel button and the switch user
button into icon buttons now, which means they are completely cryptic
when using a screen reader.
Just use the previously used labels as accessible names, which has the
nice side effect of lowering the impact of the string freeze break.
Florian Müllner [Wed, 19 Feb 2020 16:32:22 +0000 (17:32 +0100)]
authPrompt: Don't focus hidden cancelButton
The button is hidden on the lock screen, so it shouldn't be allowed to
activate it, be it via click or keyboard. The latter is still possible
by keynaving to the button and hitting space/enter. Fix that by making
the button unfocusable when we make it unreactive.
Florian Müllner [Wed, 19 Feb 2020 16:30:15 +0000 (17:30 +0100)]
shell/stack: Ignore hidden children for focus navigation
ShellStack implements custom focus navigation, and will only ever
navigate into its top-most child. That kind of makes sense as long
as that child is actually visible, but not when it is hidden.
Descend into the stack to look for a focusable child instead.
Jonas Dreßler [Mon, 7 Oct 2019 13:02:17 +0000 (15:02 +0200)]
magnifier: Use inhibit-unfocus API to keep wayland focus while hidden
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 clutter_seat_inhibit_unfocus() API to
temporarily disable unsetting the focus-surface while the magnifier is
hiding the system cursor.
Florian Müllner [Fri, 21 Feb 2020 18:38:53 +0000 (19:38 +0100)]
shellEntry: Restore natural-height-set instead of forcing it
If we are transitioning the label from 0 to its natural height, we
must set natural-height-set again after querying the preferred height,
otherwise Clutter would skip the transition.
However when transitioning in the opposite direction, setting the
property to true can go horribly wrong:
If the actor hasn't been allocated before, it will store a fixed
natural height of 0. But as there is no fixed min-height, we can
end up with min-height > natural-height, which is a fatal error.
(This isn't an issue when *actually* setting a fixed height, as
that will set both natural and minimum height)
So instead of always setting natural-height-set to true, restore
its previous value to fix the issue.
Florian Müllner [Sat, 22 Feb 2020 13:25:37 +0000 (14:25 +0100)]
util: Don't wiggle when animations are disabled
repeatCount and autoReverse don't play well with animations disabled:
They cause password entries to wiggle themselves off-screen (by ending
up with some off-scale translation-x value).
While we should handle this more gracefully in the transition helpers,
it also makes sense to handle the case directly in wiggle(): As it
uses a chain of three transitions, we would still end up with a crude
one-frame-per-transition wiggle "animation".
Instead, do no animation at all as you would expect when animations are
disabled.
Jonas Dreßler [Thu, 20 Feb 2020 13:23:01 +0000 (14:23 +0100)]
layout: Show and hide keyboard using translation_y
ClutterActors anchor-y property is deprecated, which means we get a
warning when animating it, so use the translation property instead to
show or hide the osk.
Florian Müllner [Thu, 13 Feb 2020 22:16:38 +0000 (23:16 +0100)]
unlockDialog: Use "normal" background
Now that we apply a strong blur effect to the background, it doesn't
make too much sense to use a separate lock-screen background: It will
be mostly unrecognizable anyway.
The alternative would be to turn off the blur effect if a different
background is used (or have a hidden setting for that), but that would
then imply that we must keep the contents readable without blur.
Let's avoid that rabbit hole and just re-use the regular background.
Daniel van Vugt [Wed, 19 Feb 2020 11:09:07 +0000 (19:09 +0800)]
js/ui: Use Clutter.OffscreenRedirect.ON_IDLE
To bypass offscreening in cases where continuous animation is happening.
Offscreening is slower in such cases so this reduces the render time of
animations within offscreenable actors.
On an i7-7700 this reduces the render time of boxpointers for example by
25-30%.
Daniel van Vugt [Thu, 20 Feb 2020 08:02:20 +0000 (16:02 +0800)]
slider: Calculate handle position in whole pixel units
`-slider-handle-radius` is a floating point value and even in the default
theme it's not a whole number. Regardless of the fractional part that's
still going to occupy a whole extra pixel with antialiasing. So make room
for it. Otherwise it looks clipped, which it is by the Cairo context of
its `StDrawingArea`.
Jonas Ådahl [Tue, 1 Oct 2019 10:07:03 +0000 (12:07 +0200)]
introspect: Add AnimationsEnabled property
While the gsetting is available for all who needs it, the Shell might
override it given various hueristics. Expose the decision made by the
Shell via a new property.
Intended to be used by gsd-xsettings as well as xdg-desktop-portal-gtk.
This also add a version property to the API, so that semi external
services (xdg-desktop-portal-gtk) can detect what API is expected to be
present.
Jonas Ådahl [Tue, 1 Oct 2019 09:56:34 +0000 (11:56 +0200)]
st/settings: Add API to inhibit animations
There may be situations where we shouldn't enable animations. Make it
possible for the Shell to decide when there are such situations and in
when needed inhibit animations.
Daniel van Vugt [Fri, 14 Feb 2020 10:24:58 +0000 (18:24 +0800)]
iconGrid: Create icon clones in a separate loop
For reasons not yet fully understood, `Main.uiGroup.add_actor` takes around
10 milliseconds to complete.
Because of this, each `actor.opacity = 0` has a good chance of falling
on a different frame. And when it does, `_opacityChangedId` also lands
on multiple different frames each incurring a separate relayout cycle.
It is this excessive number of relayouts that causes stuttering in the
icon grid animation (#2065). But it is the slowness of `uiGroup.add_actor`
that causes the number to be excessive when it should be one.
By creating the clones and adding them to `uiGroup` early, we then enable
the existing loop starting the animation to complete within a single frame.
And by completing within a single frame all the opacity changes land within
the same frame interval, thus incurring only a single relayout instead of
many.
This issue went unnoticed until 004a5e1042 (!704), after which the slow
emissions of `notify::opacity` became a more visible performance problem.
Jonas Dreßler [Wed, 19 Feb 2020 08:49:36 +0000 (09:49 +0100)]
theme: Make the focus state of buttons a bit more visible
This changed with the theme refactor was probably not intended, so make
the border on key-focused buttons a bit less transparent to make it
easier to see.
Now that the bind constraint changes the preferred size of
the actor, a major flaw in the AppDisplay code was exposed:
the folder dialog depends on the preferred size of a parent,
and the parent depends on the preferred size of the folder
dialog.
While we know this is not actually true, we shouldn't rely
on broken behavior to achieve this result. What's interesting
is that the bind constraint used by the folder dialog is a
relic of the development phase; we now control the position
and size of the dialog with a combination of CSS, and alignment.
Jonas Dreßler [Sun, 16 Feb 2020 20:57:34 +0000 (21:57 +0100)]
dateMenu: Hide overlay scrollbar in the notification popup
Since the design of the notification popup changed with the theme
refactor and there are now boxes around the world-clock and weather
sections, the overlay scrollbar that is shown above them looks rather
bad. So simply hide that scrollbar, we still have the vfade effect to
indicate the container is scrollable and we also depend on that in the
new popup app-folders.
Feichtmeier [Wed, 19 Feb 2020 09:36:23 +0000 (10:36 +0100)]
Style looking-glass $variant independant
- use osd colors
- also fix the toolbar buttons jumping on click
- adding better unchecked/hover/checkedifferences
- use $link_color for links
- remove the strong text-shadow
Jonas Dreßler [Wed, 19 Feb 2020 10:53:05 +0000 (11:53 +0100)]
lookingGlass: Also handle null objects in objectToString
Whoops, while ff4623454 fixed the handling of `undefined` objects,
`null` objects obviously still don't have a toString() method. So also
handle those and return 'null' in case the object is null.
Jonas Dreßler [Tue, 18 Feb 2020 22:21:52 +0000 (23:21 +0100)]
workspace: Ensure style of window-chrome titles before requesting width
Make sure the stylesheet properties of the window-chrome title are
updated before requesting the preferred width of the title to prevent
size changes of the title after we animated the width.
Jonas Dreßler [Sun, 16 Feb 2020 22:03:14 +0000 (23:03 +0100)]
userWidget: Simplify icon size setting a bit
Remove the default icon size of -1 and always set the container StBin to
a real size. This fixes an error where the "width" and "height"
properties get set to -2 (which is -1 * scaleFactor) in the `_init`
function.
projects / gnome-shell.git / log