This restores the intended behavior of keybinds set by `<default />`
to be overwritten by manually configured keybinds which come later in
the config.
In `src/keyboard.c`, `handle_keybinding()` is going backwards through
the list of keybindings and breaks after the first match.
`wl_list_insert(&list_node, item)` will insert the new item *after* the
list_node so if its called multiple times with the same list_node as
fist argument the result will be a reversed list. Using `list_node.prev`
instead will result in a non-reversed list.
Backport of 39cdba36a8
Allow underscore in environment variable names.
Closes issue #439
Helped-by: @Consolatis
...who both found the bug and told us how to fix it :)
Backport of fd7ec0ab95
Use wlr_output_test() instead of output_ensure_buffer() which is not a
public function and only compiled when linking statically to wlroots.
Also wlr_output_rollback() on failed test.
Fixes issue #273
Backport of 5fd5024ca6
Some xwayland clients leave unmapped child views around, typically when a
dialog window is closed. Although handle_destroy() is not called for
these, we have to call foreign-toplevel-destroy to avoid panels and the
like incorrecly showing them.
Backport of 62ae87bbd9
This is needed to allow X11 applications to create surfaces as
non-override_redirect and then change them to override_redirect later
Without this gitk-menus and rofi are treated as xwayland-views with
associated server-side-decoration and forced positioning.
Backport of 81f216b6a2
Note: view->padding was deleted in commit b279550 as the wlroots
scene-graph xdg-surface commit handler offsets the CSD invisible border.
If view->padding still existed, we could have used that, but considering
that this is probably the only place where this offset will now be
needed (because we're generating surface local coordinates), it's simpler
to just do it locally.
Related to issue #340
Written-by: @Consolatis
Backport of 81810db3b6
Instead of using the stored view for comparison use the actual surface.
Before this patch, there were situations where the branch intended for
out-of-window text selection / scrollbar dragging was used even though
the cursor never left the surface.
Partly fixes#340Fixes#347
Backport of 6dd290afc9
...in both resistance_resize_apply() and resistance_move_apply()
Failure to initialize these variables sometimes causes windows to stick to
edges during move/resize.
Issues #331 and #309 may be related to this
Commit ec68806354 ("xwayland: Honor size increments from
WM_SIZE_HINTS") adjusted only the window width/height according
to the size hints. If resizing from the top or left edge of the
window, we also need to adjust the window position to keep the
bottom or right edge from jumping around.
...and call it from desktop_move_to_front() in order force an enter event
on the surface below the cursor when cycling views.
Fixes#162 and #225
Inspired by PR #164 - just restructured it a bit.
Suggested-by: @bi4k8
Co-authored-by: Consolatis <35009135+Consolatis@users.noreply.github.com>
Prevents a single action like ToggleDecorations + ToggleMaximize to
position the view somewhere with negative coordinates when unmaximizing.
It may still position the view on negative coordinates but later commit
events will fix the position. This issue only exists on xwayland because
there are no configure serials which we could use to ignore all
repositioning until we are at the latest desired state.
This is necessary for menus in X11 apps to work properly.
Otherwise, any region of the menu that extended out beyond the main
application window was not receiving any mouse input.
Adapted from sway's code.
Add the 'cycleViewPreview.core' option to rc.xml to enable previews of
the selected view when cycling between windows. Default this option to
be disabled to match current behavior.
This will still prevent touchscreens or other absolute pointers to
function correctly after switching the TTY when using a multi
screen setup. But that is still better than segfaulting.
...in preference to views and other layer surfaces (except those in the
OVERLAY layer which always go first)
Fixes issue #207 where a BOTTOM layer popup lost mouse focus when a
terminal was behind it.
