These traditionally got the various ClutterEvent subtype structs as their
argument, so it was not allowed to use ClutterEvent generic getter methods
in these vfuncs. These methods used direct access to struct fields instead.
This got spoiled with the move to make ClutterEvent opaque types, since
these are no longer public structs so GNOME Shell most silently failed to
fetch the expected values from event fields. But since they are not
ClutterEvents either, the getters could not be used on them.
Mutter is changing so that these vmethods all contain an alias to the
one and only Clutter.Event type, thus lifting those barriers, and making
it possible to use the ClutterEvent methods in these vfuncs.
Closes: https://gitlab.gnome.org/GNOME/mutter/-/issues/2950
Part-of: <https://gitlab.gnome.org/GNOME/gnome-shell/-/merge_requests/2872>
The events that the draggable uses to initiate a drag operation
are the same as the ones a click action consumes for clicks.
It is still possible to combine the two, but it's finicky and far
from being straight-forward. To make this easier, add a dedicated
addClickAction() method to draggables that takes care of the
setup before adding the action to the draggable actor.
In the longer term, we'll want to turn DND into an action, and
have something like GTK's gesture group to allow combining actions
that would otherwise step on each other's toes.
Part-of: <https://gitlab.gnome.org/GNOME/gnome-shell/-/merge_requests/2742>
When a WindowPreview is being destroyed, the class default handler for
the `destroy` signal is responsible for destroying its child actors.
This happens after the emission of the `destroy` signal, i.e. after
`WindowPreview::_onDestroy()` has been run.
The destruction of the WindowPreview's child actors now triggers a
re-pick, but due to WindowPreview having already being marked as
`CLUTTER_IN_DESTRUCTION`, it will not be picked, resulting in a `leave`
event if the cursor was on top of the WindowPreview at the time
`destroy()` was called on it.
So this leads to `WindowPreview::vfunc_leave_event()` being run after
`WindowPreview::_onDestroy()`, which means the idle started by the leave
event handler will not be removed and ends up accessing actors after
they have already been destroyed.
Closes: https://gitlab.gnome.org/GNOME/gnome-shell/-/issues/5512
Closes: https://gitlab.gnome.org/GNOME/gnome-shell/-/issues/6065
Part-of: <https://gitlab.gnome.org/GNOME/gnome-shell/-/merge_requests/2738>
Since the ClutterGrab rework, grabbing properly emits crossing events.
StButtons take a ClutterGrab as soon as they are pressed, so pressing
the close StButton of the WindowPreview takes a grab and causes a
LEAVE+key-focus-out event getting sent to the WindowPreview. This in
turn makes us hide our overlay (which the StButton is part of). We
automatically ungrab ClutterGrabs when hiding actors, so the StButtons
grab now gets released again, key-focus and hover state are updated
again and we emit an ENTER event to the WindowPreview. The preview now
tries show its overlay again and things explode because we re-enter the
mapping machinery.
For the LEAVE event we can break this cycle by detecting the GRAB_NOTIFY
crossing event of our own StButton and not reacting to that.
We should do the same for the key-focus-out event, but these don't pass
context information like a GRAB_NOTIFY flag yet, so just check the
current grab actor here.
https://gitlab.gnome.org/GNOME/gnome-shell/-/issues/3165
Part-of: <https://gitlab.gnome.org/GNOME/gnome-shell/-/merge_requests/2231>
Start using the new methods to simplify signal cleanup. For now,
focus on replacing existing cleanups; in most cases this means
signals connected in the constructor and disconnected on destroy,
but also other cases with a similarly defined lifetime (say: from
show to hide).
This doesn't change signal connections that only exist for a short
time (say: once), handlers that are connected on-demand (say: the
first time a particular method is called), or connections that
aren't tracked (read: disconnected) at all.
We will eventually replace the latter with connectObject() as
well - especially from actor subclasses - but the changeset is
already big enough as-is :-)
Part-of: <https://gitlab.gnome.org/GNOME/gnome-shell/-/merge_requests/1953>
We have made good progress on object literals as well, although there
are still a lot that use the old style, given how ubiquitous object
literals are.
But the needed reindentation isn't overly intrusive, as changes are
limited to the object literals themselves (i.e. they don't affect
surrounding code).
And given that object literals account for quite a bit of the remaining
differences between regular and legacy rules, doing the transition now
is still worthwhile.
Part-of: <https://gitlab.gnome.org/GNOME/gnome-shell/-/merge_requests/2200>
Move the get_preferred_width/height() and allocate() vfunc
implementations of WindowPreview to C, subclassing the C GObject from
JS.
This gets us another significant performance gain, allocating a
workspace with 20 windows now only takes 1.2 ms.
Part-of: <https://gitlab.gnome.org/GNOME/gnome-shell/-/merge_requests/1743>
This layout manager is used quite often and the time we spend calling
it's allocate and get_preferred_width/heigth functions increases with
every open window.
We can save a lot of precious time during the layout cycle by moving
this layout manager from JS to C, thus avoiding the overhead of
trampolining between C and JS land.
In a measurement where the average time spent in vfunc_allocate() of the
Workspace actor was measured while opening an overview with 20 windows,
the average time spent went down from 3.1 ms to 2.3 ms.
Part-of: <https://gitlab.gnome.org/GNOME/gnome-shell/-/merge_requests/1743>
This is used to detect whether a click was short enough to select a
window and activate it or long enough to start a drag. Usually when
clicking on a window and selecting it, this leaves the overview, but
when clicking on a window on a neighboring workspace, the overview is
kept open, but selected is not unset in this case. So all attempts at
dragging the window after using it to switch workspaces will fail.
Fixes https://gitlab.gnome.org/GNOME/gnome-shell/-/issues/3783
Part-of: <https://gitlab.gnome.org/GNOME/gnome-shell/-/merge_requests/1717>
The preview was getting scaled up by a factor based on what is needed to
increase the width by activeExtraSize pixels. With windows that are
wider than than they are tall, this means that the size of the window
will not increase any more than activeExtraSize in any direction, but
for windows that are taller than they are wide, the vertical scaling
can exceed this. This would break some of the assumptions in the
reported size for the preview chrome and could for very narrow windows
result in a rather large scale.
To fix this, calculate the scaling factor based on whatever is larger,
the height or the width.
Part-of: <https://gitlab.gnome.org/GNOME/gnome-shell/-/merge_requests/1707>
Right now the rowSpacing and columSpacing of the layout strategy is
calculated by looking at the overlapping sizes of the close button and
the app icon of the WindowPreview, plus a constant spacing read from CSS
by the WorkspaceLayout that's added to that. We're not factoring in the
extra size of the scaled-up WindowPreviews here and instead depend on
the constant spacing being large enough. If we don't want to depend on
the spacing here, we should add the scaled-up extra size to the sizes
returned by chromeWidths() and chromeHeights().
Since the last commits all previews scale up by the same amount of
pixels, so we can now just add that size to the values returned by
chromeWidths() and chromeHeights().
Part-of: <https://gitlab.gnome.org/GNOME/gnome-shell/-/merge_requests/1692>
Scaling differently sized WindowPreviews by a constant factor will
result in smaller windows getting enlarged by a smaller amount of pixels
than larger windows (1000*1.02=1020 but 100*1.02=102, one will grow by
20 pixels and the other one by 2), this can look a bit weird because
smaller windows don't scale up as much as larger windows.
So introduce a constant extra size to use when scaling windows up, we
set only the half size there because we want to ensure that the size
added on both sides is not fractional and we remain aligned to the pixel
grid.
Part-of: <https://gitlab.gnome.org/GNOME/gnome-shell/-/merge_requests/1692>
Right now the spacing between icon and title works using a little trick
that doesn't really seem intended: The title is offset by
(icon-height * ICON_OVERLAP), when the icon is actually overlapping the
preview by ICON_OVERLAP, and *overflowing* the preview by
(1 - ICON_OVERLAP).
So correct that and offset the title by
(icon-height * (1 - ICON_OVERLAP)), and since now there's no spacing
anymore, add a proper ICON_TITLE_SPACING to that offset.
Also add the new ICON_TITLE_SPACING to the overlapHeight, where the
spacing was ignored so far.
Part-of: <https://gitlab.gnome.org/GNOME/gnome-shell/-/merge_requests/1692>
The icon and close button might be overlapping the window actor but
were not considered in has_overlaps() which gets used to decide whether
to offscreen the actor for transparency. Since currently the icon is
visible when the preview is dragged and the whole actor is turned
transparent, the opacity will not be applied to everything as a whole
but the child actors individually. This leads to the window becoming
visible behind the icon.
Part-of: <https://gitlab.gnome.org/GNOME/gnome-shell/-/merge_requests/1684>
Scaling the icons all the way from/to 0 is a relatively big transition,
which is fairly distracting when playing simultaneously for multiple
previews after reaching the WINDOW_PICKER state.
Instead, tie the scale to the overview state itself, so that the animations
runs in parallel.
Part-of: <https://gitlab.gnome.org/GNOME/gnome-shell/-/merge_requests/1654>
Resource icons are added to the fallback icon theme, so they won't
get used if a matching icon is found in the configured theme.
That includes fallback names, so Adwaita's "window-close-symbolic"
takes precedence over "window-close-24-symbolic" in hicolor.
Fix this by using a custom name for a custom icon.
Part-of: <https://gitlab.gnome.org/GNOME/gnome-shell/-/merge_requests/1640>
Accessing GObject properties from JS has proven to be quite slow because
of the JS->C->JS roundtrip involved. With the WindowPreview this
actually has an impact since we're accessing those properties very often
while creating new layouts.
So cache the boundingBox and the windowCenter properties of the
WindowPreview using a this._cachedBoundingBox JS object. This might
speed up opening the overview with lots of open windows significantly.
Part-of: <https://gitlab.gnome.org/GNOME/gnome-shell/-/merge_requests/1617>
We sometimes add dialogs multiple times to the WindowPreview, for
example for modal dialogs we receive both the "window-added" and the
"window-entered-monitor" signal, which means we call
WindowPreview.addDialog() twice.
We handle that fine already in the WindowPreviewLayout and return NULL
in case the window already was added, so simply handle that NULL return
value and bail out of WindowPreview.addDialog() in this case.
https://gitlab.gnome.org/GNOME/gnome-shell/-/merge_requests/1482
WindowPreviews now contain and manage overlaid elements like close
button or title label themselves. That's generally better, but right now
the only way to disable those overlays (for example during transitions)
is to prevent any hover or focus events from getting to the preview.
Instead, add some explicit API for enabling or disabling overlay support.
https://gitlab.gnome.org/GNOME/gnome-shell/-/merge_requests/1345
For the windowPreview we need to ensure the style information of the
border and title is up-to-date when chromeWidths() or chromeHeights() is
called. Since the introduction of the WorkspaceLayout those functions
may be called during an allocation cycle, which means we should avoid
calling queuing relayouts inside them. Calling StWidgets ensure_style()
method will queue a relayout though in case the newly generated theme
node has a different geometry.
So avoid queueing a relayout during allocation cycles (and the warning
Clutter logs because of that) by ensuring the style of the border and
title earlier, as soon as the WindowPreview is attached to a stage.
https://gitlab.gnome.org/GNOME/gnome-shell/-/merge_requests/1305
Switch to the new WorkspaceLayout layout manager to allocate the window
clones of the overview properly using Clutters layouting mechanisms.
Since we now no longer make use of the fullGeometry, we can remove the
setFullGeometry() function from the Workspace class. Also we can stop
setting the actualGeometry on the Workspaces and WorkspaceViews and
instead just set the fixed position and size of the views to their
full or actual geometry. This also has the benefit that we no longer
have to set a custom clip, but can simply enable clip_to_allocation.
The geometry needs to be set inside a BEFORE_REDRAW later because
_updateWorkspacesActualGeometry() is called from a notify::allocation
handler.
This isn't doing any animations when showing/hiding the overview yet,
we'll add that in the next commit.
https://gitlab.gnome.org/GNOME/gnome-shell/-/merge_requests/1305
When the bounding box size is 0 during allocation (which happens right
after creating a window for example), we're doing a division by zero and
end up with a NaN scale. This ends up making the childBox NaN, which
triggers an error in Clutters allocation machinery.
So fix that and fall back to a scale of 1 in case the bounding box is
empty.
https://gitlab.gnome.org/GNOME/gnome-shell/-/merge_requests/1320
The workspace.js file is quite large and is a bit confusing when it
comes to the term "window" in there, because it can either refer to a
WindowPreview of a complete window or to an individual window like an
attached dialog.
So try to avoid that confusion and split the new WindowPreview class and
its WindowPreviewLayout layout manager out into a new windowPreview.js
file.
https://gitlab.gnome.org/GNOME/gnome-shell/-/merge_requests/1307
