- Add `active` argument for consistency with `ssd_set_active()`
- `assert()` that `ssd_create()` is not called twice without an
`ssd_destroy()` in between
Gather related logic from `reload_config_and_theme()` in `server.c` and
`ssd_reload()` in `ssd.c` into a new function, `view_reload_ssd()`.
Also drop the `view->mapped` check since we want to update any view that
has SSD nodes created, mapped or not.
Add xdg_surface_from_view() + xwayland_surface_from_view() accessors
that assert() the view is of the expected type before returning.
Fix a real bug in xdg.c parent_of() that dereferenced
`view->xdg_surface->toplevel` without first checking `view->type`.
The goal of the new accessors is to catch similar bugs in future.
IMHO it encourages better design (by making dependencies more obvious)
to have source file/header file pairs like view.c/view.h, rather than a
monolithic header like labwc.h with everything in it.
I don't think we need to break up all of labwc.h at once, but maybe we
can start pulling it apart bit by bit as it's convenient.
Also:
- Move "struct border" to ssd.h so that view.h can use it without pulling
in all of labwc.h.
- Add a missing required #include within scaled_font_buffer.h (forward
declaration of "struct font" is not enough).
Currently, snapping to a screen edge and then snapping to maximize
results in both the natural_geometry and tiled state of the view
getting messed up. After unmaximize, the view ends up in a weird
state (tiled location but natural/untiled size).
There are also a couple of sketchy things going on in the code:
- interactive_begin() pokes its own values into view->natural_geometry
to force view_maximize() to set a particular geometry.
- interactive_end() "fixes" view->natural_geometry after calling
view_maximize() to save the original geometry from the start of the
interactive move/resize.
To fix all this:
- Adjust/expand the API of view.c so that the interactive.c can
avoid this "back door" of overwriting view->natural_geometry
directly.
- Save the natural geometry and the tiled state of the view in
interactive_begin() when starting to move the view. When done,
interactive_end() will update the tiled state if appropriate but
*not* overwrite the natural geometry.
It seems that every Wayland client is expected to implement its own
key-repeat logic, rather than doing it server-side as in X11. This
means that labwc also has to implement its own key-repeat logic for
compositor keybindings.
This is a very simplistic timer-based implementation. It doesn't
attempt to synthesize accurate timestamps, and may lag depending
on system load, but it appears to get the job done.
v2: Use server->wl_event_loop
v3: Comments and formatting
...so that it can be determined what wlr_keyboard events come from.
This is required to manage virtual keyboards alongside the keyboard_group
of physical keyboards.
Key events associated with keybindings (both pressed and released) are not
sent to clients. When using wlr_seat_keyboard_notify_enter() it it
therefore important not to send the keycodes of _all_ pressed keys, but
only those that were actually _sent_ to clients (that is, those that were
not bound).
This approach is consistent with sway's implementation in input/seat.c
cffb006feb/sway/input/seat.c (L173-L175)
Fixes issue #510
If a user lets go of the modifier (e.g. alt) before the 'normal' key (e.g.
tab) when window-cycling, we do not end the cycling until both keys have
been released. If we end the window-cycling on release of the modifier
only, some XWayland clients such as hexchat realise that tab is pressed
(even though we did not forward the event) and because we absorb the
equivalent release event it gets stuck on repeat.
Just to clarify the position here: Issue #176 describes a behaviour
whereby dmenu gets stuck on repeat after being launched with a keybind.
This patch does not resolve that issue but reflects that in Wayland, the
client is responsible for implementing "key repeat".
Changing the key repeat rate/delay in (labwc/labwc@e62bb51) was dirty fix
that need should never have been made.
This fixes an issue with the wlr_output_cursor not being properly
initialized on new outputs, because wlr_output_layout.events.change
is triggered too soon, before the wlr_output_cursor exists.
Attempting to open a GTK3 menu and activate a menu item in it,
using a single mouse motion (press-move-release), was broken due
to GTK apparently expecting to receive leave/enter events when the
cursor enters the menu (XDG popup).
To fix the issue, allow leave/enter events when the cursor is
moved between an XDG toplevel and popups of the same.
v2:
- Use (struct view *) as proxy for toplevel in comparisons
- Update seat->pressed.surface when entering/leaving popups
v3:
- Go back to using get_toplevel() rather than (struct view *)
Fix a couple of glitches seen when exiting interactive move/resize:
- Cursor briefly set to left_ptr rather than the correct cursor image
- Cursor not updated if the view being moved/resized is destroyed
Also make sure to exit interactive mode if the view is going fullscreen
(labwc gets very confused otherwise).
Code changes in detail:
- Factor out set_server_cursor() which will set the correct cursor
image for non-client areas (either XCURSOR_DEFAULT or one of the
resize cursors).
- Unify the logic from cursor_rebase() and process_cursor_motion by
factoring out cursor_update_common(). This corrects some logic
discrepancies between the two, which should be a good thing(TM).
- Remove the extra cursor_set(XCURSOR_DEFAULT) from interactive_end()
and instead rely on cursor_update_focus() to do the right thing.
- Simplify cursor_button() by just calling interactive_end() when we
want to exit interactive mode.
- Call cursor_update_focus() from view_destroy() if the view had mouse
focus or was being interactively moved/resized.
v2: Eliminate force_reenter parameters and figure out automatically
when we need to re-enter the surface.
v3: Rename wlseat -> wlr_seat.
v4: Simplify client/server cursor logic.
