This patch should not change any behaviors.
This clarifies the semantics of cursor_context returned by
get_cursor_context() as I described in cursor.h; when cursor is on a
subsurface (e.g. xdg/x11/layer/session-lock), the returned ctx.surface
and ctx.node points to the subsurface rather than its parent.
With this change, when a layer-popup has a subsurface and cursor is on
it, wl_pointer.enter/motion events are sent to the subsurface rather
than its parent layer-popup surface. I think this follows wayland spec
more closely.
Tested with my demo: https://github.com/tokyo4j/wayland-demo/tree/layer-popup
Instead, set ctx.type = LAB_NODE_LAYER_SURFACE for both layer-surfaces
and layer-subsurfaces.
This patch preserves the existing behaviors:
- Pressing a subsurface of an on-demand layer-surface gives pointer
focus to the subsurface, but gives keyboard focus to the parent
layer-surface (related: a5fcbfaf).
- Pressing a subsurface of a layer-surface doesn't close a popup
(related: a89bcc3c).
This doesn't change any behaviors.
Attaching LAB_NODE_NONE node-descriptor to ssd->tree looks strange, this
patch uses new LAB_NODE_SSD_ROOT instead. The node-descriptor attached to
ssd->tree is needed for get_cursor_context() to detect cursor hovering on
borders/extents.
I also updated get_cursor_context() to make my intent clearer.
In f347a81, I thought there are only window contents and SSD under
`view->scene_tree` and forgot about the resize-indicator.
I also refactored the logic around it:
- Remove `ret.node->type == WLR_SCENE_NODE_BUFFER` check since it's
already done by `lab_wlr_surface_from_node()`
- Eliminate duplicated call to `lab_wlr_surface_from_node()`
struct ssd_part and struct node_descriptor seem to have essentially the
same purpose: tag a wlr_scene_node with some extra data indicating what
we're using it for.
Also, as with enum ssd_part_type (now lab_node_type), ssd_part is used
for several types of nodes that are not part of SSD.
So instead of the current chaining (node_descriptor -> ssd_part), let's
flatten/unify the two structs.
In detail:
- First, merge node_descriptor_type into lab_node_type.
- Add a separate view pointer in node_descriptor, since in the case of
SSD buttons we need separate view and button data pointers.
- Rename ssd_part_button to simply ssd_button. It no longer contains
an ssd_part as base.
- Add node_try_ssd_button_from_node() which replaces
node_ssd_part_from_node() + button_try_from_ssd_part().
- Factor out ssd_button_free() to be called in node descriptor destroy.
- Finally, get_cursor_context() needs a little reorganization to handle
the unified structs.
Overall, this simplifies the code a bit, and in my opinion makes it
easier to understand. No functional change intended.
ssd_part_type contains several node types that are not actually part of
server-side decorations (ROOT, MENU, OSD, etc.)
Rename it accordingly and move it to a common location, along with some
related conversion/comparison functions.
- Rename `scaled_scene_buffer` to `scaled_buffer`. This makes it clear
that `scaled_{font,img,icon}_buffers` are implementations of it.
- Move the files from `src/common` to `src/scaled-buffer` as
`scaled_icon_buffer` heavily depends on `server` and `view` etc.
Our codebase for ssd scenes has grown with a lot of technical debts:
- We needed to call `ssd_get_part()` everywhere to get the scene node of a
ssd part. We then needed to cast it to `wlr_scene_rect` and
`wlr_scene_buffer`. This bloated our codebase and even blocked
duplicated button types in `<titlebar><layout>`.
- `ssd_get_part_type()` was a dirty hack. It compared parent, grandparent
and grandgrandparent of a node with each subtree in the ssd to get the
part type of the node.
To resolve this issues, this commit changes how ssd scenes are managed:
- Access scene rects and scene buffers just as a member of `struct ssd`.
- `ssd_part` is now a attachment to a scene node that can be accessed via
node_descriptor->data, with a new node-descriptor type
`LAB_NODE_DESC_SSD_PART`. `LAB_NODE_DESC_SSD_BUTTON` is unified into it.
Now the scene graph under ssd->tree looks like below. The parentheses
indicate the type of ssd_part attached to the node:
ssd->tree (LAB_SSD_NONE)
+--titlebar (LAB_SSD_PART_TITLEBAR)
| +--inactive
| | +--background bar
| | +--left corner
| | +--right corner
| | +--title (LAB_SSD_PART_TITLE)
| | +--iconify button (LAB_SSD_BUTTON_ICONIFY)
| | | +--normal close icon image
| | | +--hovered close icon image
| | | +--...
| | +--window icon (LAB_SSD_BUTTON_WINDOW_ICON)
| | | +--window icon image
| | +--...
| +--active
| +--...
+--border
| +--inactive
| | +--top
| | +--...
| +--active
| +--top
| +--...
+--shadow
| +--inactive
| | +--top
| | +--...
| +--active
| +--top
| +--...
+--extents
+--top
+--...
When hovering on SSD, `get_cursor_context()` traverses this scene node
from the leaf. If it finds a `ssd_part` attached to the node, it returns
`ssd_part_type` that represents the resizing direction, button types or
`Title`/`Titlebar`.
In 9e3785f8cd, a heuristic was added to assume that NORMAL and DIALOG
window types were always focusable. (This was before we had the "offer
focus" mechanism in place.)
However, we should still call wlr_xwayland_surface_offer_focus() for
these views, in case they actually don't want focus. (This is uncommon
but has recently been seen with WeChat popups, which have both NORMAL
and UTILITY type.)
To make this possible, refine view_wants_focus() to return either
LIKELY or UNLIKELY for Globally Active input windows. This decouples
the question of "should we try to focus this view" from the actual
mechanism used to do so.
For longer cases, factor out the logic to new functions.
For very short cases, just move the declaration before the switch.
v2: in one case, replace the switch with if/else.
Offer focus by sending WM_TAKE_FOCUS to a client window supporting it.
The client may accept or ignore the offer. If it accepts, the surface will
emit a focus_in signal notifying the compositor that it has received focus.
The compositor should then call wlr_xwayland_surface_activate(surface, true).
This is a more compatible method of giving focus to windows using the
Globally Active input model (see wlr_xwayland_icccm_input_model()) than
calling wlr_xwayland_surface_activate() unconditionally, since there is no
reliable way to know in advance whether these windows want to be focused.
v2: add caution not to use view_offer_focus() directly
v3: remove obsolete comment
This commit moves the check against server->input_mode from the callers
of desktop_focus_view() into desktop_focus_view() itself. This
eliminates code duplications and makes it harder to mess up the window
stacking order while window switching.
I also added the same check in view_minimize() so that minimize requests
from panels never messes up the window stacking order (I think only this
should be described in the release note).
Eliminate corner extents and instead use cursor position to map SSD
borders and extents to corner contexts, with a size configurable by the
<resize><cornerRange> parameter. This simplifies extent handling,
eliminates bugs in the detection of corner context, and allows users to
expand corner targets if they wish.
Co-authored-by: Andrew J. Hesford <ajh@sideband.org>
Before this commit, `top` layers were hidden whenever there is a
fullscreen window in the corresponding output.
With this commit, `top` layers are hidden only when there is a fullscreen
window without other windows above it in the corresponding output.
A caveat is that `bottom` layer is still always hidden by a fullscreen
window.
Add custom field with subset of printf style formatting
to replace the original field formats.
Example:
<windowSwitcher preview="no" outlines="no" allWorkspaces="yes">
<fields>
<field content="custom" format="foobar %b %3s %-10o %-20W %-10i%t" width="100%" />
</fields>
</windowSwitcher>
Mono space font recommended. May need OSD width adjusted
Co-authored-by: @Consolatis (based on work done by them)
...to give keyboard focus to layer-shell clients if exclusive or on-demand
interactivity is set, so that menu popups can be navigated with the
keyboard. This still only works if the client is in top (or overlay)
layers. Support for bottom and background to be done as a separate patch
set.
Revert 06b19f0 to process layer-shell subsurfaces in
`cursor_button_press()`, but only when their parent layer-shell surface
has keyboard interactivity.
Fix bug in `get_cursor_context()` which resulted in layer-surfaces not
being detected correctly.
Background:
Commit 06b19f0 (issue #1131) disabled processing of layer-shell
subsurfaces in cursor_button_press() because when pressing a task in
Waybar (Gtk panel using layer-shell subsurfaces) the foreign-toplevel
minimize-raise action did not work correctly as the action logic relied on
the recipient window being activated and by clicking on the panel, the
panel itself was both surface-focusd and activated (and thus the window
de-activated).
The un-intended consequence was that by not responding to layer-subsurface
cursor buttons presses, layer-shell clients (such as panels) were not
given keyboard focus if they indeed wanted it by setting exclusive or
on-demand keyboard interactivity.
The good news is that that following @jlindgren90's refactoring (various)
the only place where we call `view_set_actived()` is in
`focus_change_notify()` in `seat.c` and we now only do it for views
(bb8f0bc).
Another side-effect (positive) of 06b19f0 was that a Waybar dnd bug was
fixed (pointer-serial-number validation failure).
Have tested with sfwbar, waybar and tint (test-panel) the following
results:
- Minimize-raise works even when on-demand keyboard interactivity is set
- Keyboard interactivity is given popup-menus (sfwbar and tint) when the
panels are in the top layer (support for bottom will be as a separate
patch set)
- Waybar dnd still works (even when hard-coding keyboard-interactivity)
References:
- bb8f0bc960
- 40ce95a68c/src/seat.c (L481-L483)
- 40ce95a68c/src/dnd.c (L24)
- https://github.com/johanmalm/tintFixes: #1572
Also share common config option (rc.window_switcher.criteria) in osd.c and
desktop.c to make sure the window lists are always consistent.
Configure with `<windowSwitcher allWorkspaces="yes|no">`
This is undesired when a focusable popup (e.g. applications menu) of an
unfocusable view (e.g. XWayland panel) is closed.
This reverts commit f6e3527767.
We already allow some xwayland-unmanaged surfaces to take focus on map,
if indicated by wlr_xwayland_or_surface_wants_focus(). But once these
surfaces lose focus, they never regain it again.
Add desktop_focus_view_or_surface() and call it in the appropriate
places to allow these views to regain focus in the usual ways (e.g.
clicking on them or focus-follows-mouse).
Commit 7e72bf975f changed behavior to not automatically focus xwayland
views using the "Globally Active" input model (WM_HINTS.inputs = false
but WM_TAKE_FOCUS listed in WM_PROTOCOLS).
One undesired side effect of this change is that when a dialog is
closed, the parent window is not re-focused if "Globally Active". This
issue is seen for example with JDownloader. It can be solved taking a
similar approach to what is done for unmanaged xwayland views: allow
automatic re-focus between views sharing the same PID.
Note that it's difficult to completely solve all of the focus issues
with Globally Active views without proper WM_TAKE_FOCUS support.
Implementing proper support is difficult since it requires wlroots
changes and would also mean waiting for a message round-trip in
desktop_focus_topmost_view().
Fixes (partially): 7e72bf975f
("view/xwayland: avoid focusing views that don't want focus")
We were checking for a locked session in desktop_focus_view(), but there
are several other call sites of seat_focus_surface() which were missing
such a check. Any one of those could cause the lock screen to lose focus
(making the session impossible to unlock) or another surface to gain it
(breaching the session lock).
To fix the issue, make any call to seat_focus_surface() no-op when the
session is locked. Add a specific seat_focus_lock_surface() function
which is the only way to bypass the check and is called only from
session-lock.c.
