A "resize shrink width 1px" will cause grow_x to be 0 while grow_width is -1,
incorrectly rejecting the command even though the resize is not a noop. Fix
this by checking width/height instead of x/y.
Prior to this commit, having a layout like T[app1 V[app2]], focusing
app2, and then doing `move left` would result in T[app2 app1]. Now, the
resulting layout is T[app1 app2], which matches i3 behavior.
`container_flatten` updates `container->parent`, meaning that the
existing check would never be true.
`!*rgba` tests if the first byte of rgba isn't `'\0'`.
`hex_to_rgba_hex` returns NULL if `parse_color` fails. There's a null
pointer dereference in that case. The intended behavior is `!rgba`.
If the mouse/cursor/pointer is near the edge of an output when a "move
position to pointer" command is run, then the floating container will be
constrained to fit inside the bounds of the output as much as possible.
This behavior matches what i3 does in this scenario. I also think it is
a better user experience.
Relates to #4906
The logic for the bounds check follows the implementation in i3: 7330778223/src/floating.c (L536)
If moving e.g. `T[app app]` into a new workspace with `workspace_layout
tabbed`, then post-move the tree in that workspace will be `T[T[app
app]]`. This still happens with horizontal or vertical workspace layout,
but is less visible since those containers have no decorations.
Fixes#5426.
This emits frame events for the seat_cmd_cursor subcommands. The
wl_pointer frame event are required to notify clients when to process
the events. It will now be emitted after cursor movement, button press,
button release, and axis events.
Add a separate per-view shortcuts_inhibitor command that can be used
with criteria to override the per-seat defaults. This allows to e.g.
disable shortcuts inhibiting globally but enable it for specific,
known-good virtualization and remote desktop software or, alternatively,
to blacklist that one slightly broken piece of software that just
doesn't seem to get it right but insists on trying.
Add a flag to sway_view and handling logic in the input manager that
respects that flag if configured but falls back to per-seat config
otherwise. Add the actual command but with just enable and disable
subcommands since there's no value in duplicating the per-seat
activate/deactivate/toggle logic here. Split the inhibitor retrieval
helper in two so we can use the backend half in the command to retrieve
inhibitors for a specific surface and not just the currently focused
one. Extend the manual page with documentation of the command and
references to its per-seat sibling and usefulness with criteria.
Signed-off-by: Michael Weiser <michael.weiser@gmx.de>
This allows e.g. triggering one command while a key is held, then
triggering another to undo the change performed by it afterwards. One
use case for this is triggering push-to-talk functionality for VoIP
tools without granting them full access to all input events.
Fixes#3151
* check for workspace command name arg (fix#5131)
For the 'workspace <name> output <output>' command, output_location must
be greater than zero or the attempt to get the workspace name with
join_args will segfault or abort() (depending on the flavor of
sway_assert() in use). This checks and returns an error instead.
* put workspace output error string on one line
To ease grepping as requested
* check for name in workspace gaps command as well
A malformed command here will lead to the same result seen in #5131, so
add a check. Done inside the cmd_workspace_gaps() function itself, to
take advantage of the existing 'Expected...' string.
Remove some unnecessary brackets in an error condition check identified
during review of the shortcuts inhibitor command code (#5021).
Signed-off-by: Michael Weiser <michael.weiser@gmx.de>
Add a command to influence keyboard shortcuts inhibitors. In its current
form it can be used to activate, deactivate or toggle an existing
inhibitor on the surface currently receiving input. This can be used to
define an escape shortcut such as:
bindsym --inhibited $mod+Escape seat - shortcuts_inhibitor deactivate
It also allows the user to configure a per-seat default of whether
keyboard inhibitors are honoured by default (the default) or not. Using
the activate/toggle command they can then enable the lingering inhibitor
at a later time of their choosing.
As a side effect this allows to specifically address a named seat for
actions as well, whatever use-case that might serve.
Signed-off-by: Michael Weiser <michael.weiser@gmx.de>
Adding support for the keyboard shortcuts inhibit protocol allows remote
desktop and virtualisation software to receive all keyboard input in
order to pass it through to their clients so users can fully interact
the their remote/virtual session. The software usually provides its own
key combination to release its "grab" to all keyboard input. The
inhibitor can be deactivated by the user by removing focus from the
surface using another input device such as the pointer.
Use support for the procotol in wlroots to add support to sway. Extend
the input manager with handlers for inhibitor creation and destruction
and appropriate bookkeeping. Attach the inhibitors to the seats they
apply to to avoid having to search the list of all currently existing
inhibitors on every keystroke and passing the inhibitor manager around.
Add a helper function to retrieve the inhibitor applying to the
currently focused surface of a seat, if one exists.
Extend bindsym with a flag for bindings that should be processed even if
an inhibitor is active. Conversely this disables all normal shortcuts if
an inhibitor is found for the currently focused surface in
keyboard::handle_key_event() since they don't have that flag set. Use
above helper function to determine if an inhibitor exists for the
surface that would eventually receive input.
Signed-off-by: Michael Weiser <michael.weiser@gmx.de>
This enables/disables adaptive synchronization on the output.
For now, the default is disabled because it might cause flickering on
some hardware if clients don't submit frames at regular enough
intervals. In the future an "auto" option will only enable adaptive sync
if a fullscreen client opts-in via a Wayland protocol.
This adds support for input type configs to input_cmd_events. This works
similar to the wildcard handling that existed where configs for the
devices are stored and the type config is reset to INT_MIN so that it
does not override.
This also condenses the toggle_send_events and
toggle_wildcard_send_events functions into a single function to reduce
code duplication.
This patch makes it so users that have configured their screen with a
transform don't have to update their config after the wlroots breaking
change.
References: https://github.com/swaywm/wlroots/pull/2023
The condition in test_mark was checking the negation of list_seq_find.
This works fine for the first mark, but fails for every other mark. This
fixes the condition to check for anything other than -1, which is the
value returned from list_seq_find for index not found.
This makes it so invalid configs will return the exit code 1 when the
validation flag is given. This also reduces the log level to SWAY_ERROR,
which makes it so only the errors are shown. If someone wants more
verbose output, the can use the -V/--verbose or -d/--debug flags.
Additionally, this also makes it so swaybg will not be spawned when
validating the config.
Because the layout code rounds down the dimensions of the windows
resizing would often be off by one pixel. The width/height fraction
would not exactly reflect the final computed width and so the resize
code would end up calculating things wrong.
To fix this first snap the container size fractions to the pixel grid
and only then do the resize. Also use round() instead of floor() during
layout to avoid a slightly too small width. This applies in two cases:
1. For the container we are actually resizing using floor() might result
in being 1px too small.
2. For the other containers it might result in resizing them down by 1px
and then if the container being resized is the last all those extra
pixels would make the resize too large.
Fixes#4391
For i3 compatibility, allow the indicator and child_border colors values
to be optional. The indicator will fallback to sane defaults and
child_border will fallback to the background color for the class.
This is the third commit in a series of commits to refactor color
handling in sway. This removes add_color from commands.c. It was only
being used by bar_cmd_colors. This also changes the functions to use
parse_color which is used to validate rgb(a) colors throughout the code
base and is also what i3bar is using to parse the colors after they are
passed over ipc. After parsing the color and ensuring it is valid, the
rgba hex string is then generated using snprintf. This refactor also
ensures that all the colors for the command are valid before applying
any of them.
This is the second in a series of commits to refactor the color handling
in sway. This removes the duplicated color parsing code in
sway/commands/client.c. Additionally, this combines the parsing of
colors to float arrays with that in sway/config.c and introduces a
color_to_rgba function in commom/util.c.
As an added bonus, this also makes it so non of the colors in a border
color class will be changed unless all of the colors specified are
valid. This ensures that an invalid command does not get partially
applied.
This removes `seat <seat> keyboard_grouping keymap` and replaces it with
`seat <seat> keyboard_grouping smart`. The smart keyboard grouping will
group based on both the keymap and repeat info. The reasoning for this
is that deciding what the repeat info should be for a group is either
arbitrary or non-deterministic when multiple keyboards in the group have
repeat info configured (unless somehow exposed to the user in a
reproducible uniquely identifiable fashion).
This adds seat configuration options which can be used to configure what
events affect the idle behavior of sway.
An example use-case is mobile devices: you would remove touch from the
list of idle_wake events. This allows the phone to stay on while you're
actively using it, but doesn't wake from idle on touch events while it's
sleeping in your pocket.
For the validation pass of reloading, there is no need to touch swaybg,
swaynag, inputs, outputs, or seats. This drastically improves the speed
of a reload by skipping over the expensive I/O configuration and
handling of wayland clients. As long as the syntax is valid, the
CMD_FAILURE's can be relayed during the actual reload.
A wlr_keyboard_group allows for multiple keyboard devices to be
combined into one logical keyboard. This is useful for keyboards that
are split into multiple input devices despite appearing as one physical
keyboard in the user's mind.
This adds support for wlr_keyboard_groups to sway. There are two
keyboard groupings currently supported, which can be set on a per-seat
basis. The first keyboard grouping is none, which disables all grouping
and provides no functional change. The second is keymap, which groups
the keyboard devices in the seat by their keymap. With this grouping,
the effective layout and repeat info is also synced across keyboard
devices in the seat. Device specific bindings will still be executed as
normal, but everything else related to key and modifier events will be
handled by the keyboard group's keyboard.
This adds complete support for the barconfig_update ipc event. This also
changes the bar command and subcommand handlers to correctly emit the
event. This makes it so all bar subcommands other than id and
swaybar_command are dynamically changeable at runtime. sway-bar.5 has
been updated accordingly
In cmd_mode, the mode is currently being reset to the default mode after
a mode subcommand handler is executed. This stores and restores the mode
instead
When moving to a scratchpad hidden container (using `move
[window|container] [to] mark <mark>`), this moves the container to the
scratchpad (equivalent to `move [window|container] [to] scratchpad`).
Previously, this would crash since the destination did not have a
workspace.
This just adds a force option to cmd_xwayland that allows for xwayland
to be immediately launched instead of lazily launched. This is useful
for slower machines so it can be part of the startup time instead of
when the user is actively trying to use it
This enhances the opacity command to support relative assignment as well
as the currently implemented absolute assignment. The syntax is copied
from the same format that gaps uses for relative and absolute setting.
An example usage in a sway config looks like:
// relative change (this feature)
bindsym button4 opacity plus .1
bindsym button5 opacity minus .1
// absolute change (this feature)
bindsym button4 opacity set 1
bindsym button5 opacity set .3
// old way, still supported
bindsym button4 opacity 1
bindsym button5 opacity .3
This adds support for specifying a binding for a specific group. Any
binding without a group listed will be available in all groups. The
priority for matching bindings is as follows: input device, group, and
locked state.
For full compatibility with i3, this also adds Mode_switch as an alias
for Group2. Since i3 only supports this for backwards compatibility
with older versions of i3, it is implemented here, but not documented.
Sibling amounts were being calculated after the original fraction had
been altered. This led to broken resize amounts. Fix that by calculating
things upfront before adjusting values which also makes the code
cleaner.
For sanity checks also calculate the sibling amount with the
ceiling so we never go below the sanity check even by one pixel.
Fixes#4386
Similar to seat command, this provides an alias for the current output.
Instead of the output name or identifier, `-` can be used to operate on
the focused output by name and `--` can be used to operate on the
focused output by its identifier. This will prevent operating on the
no-op output when using either alias.
Adds a new commend "xkb_file", which constructs the internal
xkb_keymap from a xkb file rather than an RMLVO configuration.
This allows greater flexibility when specifying xkb configurations.
An xkb file can be dumped with the xkbcomp program.
Instead of tracking gaps per child apply gaps in two logical places:
1. In tiled containers use the layout code to add the gaps between
windows. This is much simpler and guarantees that the sizing of children
is correct.
2. In the workspace itself apply all the gaps around the edge. Here
we're in the correct position to size inner and outer gaps correctly and
decide on smart gaps in a single location.
Fixes#4296
Instead of using container->width/height as both the input and output
of the layout calculation have container->width_fraction/height_fraction
as the share of the parent this container occupies and calculate the
layout based on that. That way the container arrangement can always be
recalculated even if width/height have been altered by things like
fullscreen.
To do this several parts are reworked:
- The vertical and horizontal arrangement code is ajusted to work with
fractions instead of directly with width/height
- The resize code is then changed to manipulate the fractions when
working on tiled containers.
- Finally the places that manipulated width/height are adjusted to
match. The adjusted parts are container split, swap, and the input
seat code.
It's possible that some parts of the code are now adjusting width and
height only for those to be immediately recalculated. That's harmless
and since non-tiled containers are still sized with width/height
directly it may avoid breaking other corner cases.
Fixes#3547Fixes#4297
Previously, `layout toggle` and `layout toggle split` would set L_VERT
when layout was L_HORIZ, otherwise it would set L_HORIZ. This meant
that when the layout was L_TABBED or L_STACKED, it would always be
L_HORIZ. This extends #4315 (which corrects the handling when multiple
layouts are given) to try prev_split_layout,
config->default_orientation, and then falling back to L_VERT when the
output is taller than wide and L_HORIZ when wider than tall.
This fixes the logic of split for layout toggle when the default
layout is L_TABBED or L_STACKED. When the default layout is L_TABBED
or L_STACKED, the container/workspace may not have a prev_split_layout.
This was causing L_NONE to be returned by get_layout_toggle, which was
being handled as a syntax error. This adds logic to try
config->default_orientation when prev_split_layout is L_NONE. If that
is also L_NONE, then L_VERT is used when the output is taller than
wide, otherwise, L_HORIZ is used.
This adds a --reload flag to cmd_bindswitch that allows for the binding
to be executed on reload. One possible use case for this is to allow
users to disable outputs when the lid closes and enable them when the
lid opens without having to open and re-close the lid after a reload.
This allows for modes to be created, bindings to be added to modes, and
bindings to be removed from modes at runtime. Additionally, this also
allows for `mode <mode>` to be deferred in the config to set an initial
mode.
This separates the logic for seat subcommand handlers that only perform
actions on the seat and handlers that alter the seat config. The former
group can immediately free the seat config after running the command as
it is only used by the subcommand to find the name of the seat to
operate on. The latter group alters the seat config so it will need to
go through the storage and application stage (assuming success).
Since xwayland can only be enabled/disabled at launch, the xwayland
status should be retained on reload. Having `xwayland enabled|disabled`
in the config, should not cause `config->xwayland` to be invalid on
reload. This also returns `CMD_FAILURE` with a message that xwayland
can only be enabled/disabled on launch when trying to set the invalid
status on reload. This allows swaynag to notify the user that the
change will not take effect until sway is restarted.
Currently container_replace removes the container from the scratchpad
and re-adds it afterwards. For the split commands this results in the
container being send to the scratchpad, which results in a NULL segfault
if the same container should be shown.
Pass an optional workspace to root_scratchpad_add_container, if the
workspace is passed the window will continue to show on the workspace.
If NULL is passed it is sent to the scratchpad.
This was an issue if no other window except the scratchpad container was
on the workspace.
Fixes#4240
This adds the logic to defer binding execution while sway is still
initializing. Without this, the binding command would be executed, but
the command handler would return CMD_DEFER, which was being treated as
a failure to run. To avoid partial executions, this will defer all
bindings while config->active is false.
This allows for an optional validation stage when storing an input
config. Currently, only the xkb keymap is validated. If storing the
delta input config will result in any invalid xkb keymaps, the input
config will not be stored and error will be populated with the first
line of the xkbcommon log.
Before the delta input config is stored, this attempts to compile a
keymap with it. If the keymap fails to compile, then the first line of
the xkbcommon log entry will be included with a `CMD_FAILURE`, the
entire xkbcommon log entry will be included in the sway error log, and
the delta will not be stored.
This only handles basic issues such as a layouts not existing. This
will NOT catch more complex issues such as when a variant does
exist, but not for the given layout (ex: `azerty` is a valid variant,
but the `us` layout does not have a `azerty` variant).
New 'seat <name> xcursor_theme <theme> [<size>]' command that
configures the default xcursor theme.
The default seat's xcursor theme is also propagated to XWayland, and
exported through the XCURSOR_THEME and XCURSOR_SIZE environment
variables. This is done every time the default seat's configuration is
changed.
This adds the missing argument count check after the --i3 flag
processing in cmd_hide_edge_borders. Without the check,
`hide_edge_borders --i3` would result in a SIGSEGV instead of a syntax
error. There are some minor adjustments to make it so nothing gets
altered if this check fails
This changes the behavior of bindings to make the `BINDING_LOCKED` flag
conflicting, which will allow for both unlocked and locked bindings.
If there are two matching bindings and one has `--locked` and the other
does not, the one with `--locked` will be preferred when locked and
the one without will be preferred when unlocked.
If there are two matching bindings and one has both a matching
`--input-device=<input>` and `--locked` and the other has neither, the
former will be preferred for both unlocked and locked.
This also refactors `get_active_binding` in `sway/input/keyboard.c`
to make it easier to read.
For compatibility with i3, `bar mode` and `bar hidden_state` do not
require bar-ids (in the normal location) at runtime since they follow
the alternative syntax: `bar mode|hidden_state <option> [<bar-id>]`
This removes the incorrect error that the bar-id is missing for those
two bar subcommands
Running a command like this produced a confusing error message:
$ swaymsg bar bar-0 colors background #ff0000
Error: Unknown/invalid command 'bar-0'
This patch makes the error message use argv[1] instead of argv[0] (from
config_subcommand's implementation), so it actually uses the name of the
command, rather than the id of the bar.
Prior to this patch, if I ran something like this, sway would crash:
swaymsg bar height 50
or
swaymsg bar not-a-bar-id color bg #ff0000
This was in contrast to other bar subcommands, like status_command,
which would exit with a "No bar defined" message.
The difference between the subcommands that crashed and the ones that
exited was that some subcommands had a check to see if a bar was
specified, while others just assumed that it had been and carried on
until they segfaulted.
Because this check was identical in every subcommand it was present in,
and I couldn't think of a case where it would be valid to run a bar
subcommand without specifying which bar to apply it to, I moved this
check from individual subcommands into the bar command, which is already
responsible for actually setting the specified bar. This reduced code
duplication, and fixed the crash for the subcommands that were missing
this check.
This adds support for the following commands for i3 compatibility:
- `move [window|container] [to] output current`
- `move workspace to [output] current`
- `move workspace [to] output current`
The above commands are only useful when used with criteria.
* `bindsym --to-code` enables keysym to keycode translation.
* If there are no `xkb_layout` commands in the config file, the translation
uses the XKB_DEFAULT_LAYOUT value.
* It there is one or more `xkb_layout` command, the translation uses
the first one.
* If the translation is unsuccessful, a message is logged and the binding
is stored as BINDING_KEYSYM.
* The binding keysyms are stored and re-translated when a change in the input
configuration may affect the translated bindings.
This modifies cmd_move to allow for the syntax options allowed by i3.
The following syntaxes are supported:
- `move left|right|up|down [<amount> [px]]`
- `move [--no-auto-back-and-forth] [window|container] [to] workspace
<name>|next|prev|next_on_output|prev_on_output|current|number <num>`
- `move [window|container] [to] output <name/id>|left|right|up|down`
- `move [window|container] [to] mark <mark>`
- `move workspace to [output] <name/id>|left|right|up|down`
- `move [window|container] [to] [absolute] position <x> [px] <y> [px]`
- `move [window|container] [to] [absolute] position center`
- `move [window|container] [to] position mouse|cursor|pointer`
This also allows retains the following syntax option that is not
supported by i3, but is supported in sway 1.0:
- `move workspace [to] output <name/id>|left|right|up|down`
The changes are:
- `window` and `container` are now optional
- `output` is now optional for `move workspace` when `to` is given
There is also stricter command checking now. If `absolute` or
`--no-auto-back-and-forth` are given for commands that do not support
them, it will be treated as invalid syntax instead of being silently
ignored.
Disable the i3-compatible behavior if the option '--i3' is not given.
Previously it was only possible to disable it by changing the config
file. Now it also works via swaymsg.
When setting fullscreen on a hidden scratchpad container, there was a
check to see if there was an existing fullscreen container on the
workspace so it could be fullscreen disabled first. Since the workspace
is NULL, it would cause a SIGSEGV. This adds a NULL check to avoid the
crash.
This also changes the behavior of how fullscreen is handled when adding
a container to the scratchpad or changing visibility of a scratchpad
container to match i3's. The behavior is as follows:
- When adding a container to the scratchpad or hiding a container back
into the scratchpad, there is an implicit fullscreen disable
- When setting fullscreen on a container that is hidden in the
scratchpad, it will be fullscreen when shown (and fullscreen disabled
when hidden as stated above)
- When setting fullscreen global on a container that is hidden in the
scratchpad, it will be shown immediately as fullscreen global. The
container is not moved to a workspace and remains in the
scratchpad. The container will be visible until fullscreen disabled
or killed. Since the container is in the scratchpad, running
`scratchpad show` or `move container to scratchpad` will have no
effect
This also changes `container_replace` to transfer fullscreen and
scratchpad status.
This makes it so there will only be one swaybg instance running
instead of one per output. swaybg's cli has been changed to a xrandr
like interface, where you select an output and then change properties
for that output and then select another output and repeat. This also
makes it so swaybg is only killed and respawned when a background
changes or when reloading.
This implements the following command to set/unset a user idle
inhibitor for a view:
`inhibit_idle focus|fullscreen|open|none|visible`
The modes are as follows:
- focus: inhibited when the view is focused by any seat
- fullscreen: inhibited when the view is fullscreen (or a descendant of
a fullscreen container) and is visible on any output
- open: inhibited until the view is closed or the inhibitor is unset or
changed
- none: unsets any user set idle inhibitors for the view
- visible: inhibited when the view is visible on any output
This should have no effect on idle inhibitors set by the applications
themselves and those should still work as intended.
Since this operates on the view in the handler context, it is possible
to set it on the currently focused view, on any existing view with
criteria, or for any future view with for_window.
Many laptop screens report unknown subpixel order. Allow users to manually set subpixel hinting to work around this.
Addresses https://github.com/swaywm/sway/issues/3163
This change adds support for renaming a workspace when `exec` command
is being processed by keeping sway_workspace and pid_workspace names in
sync.
The change can be verified by running following command:
swaymsg exec <application>; swaymsg rename workspace number 1 to 5
Fixes: #3952
This kind of worked before in that focus would change, but it wasn't
intentionally supported and had side effects such as not raising
the container, and being unable to cycle through all floaters depending
on the direction used.
This commit makes it properly supported. The new focus is chosen based
on the distance to the center point of each floating container in the
workspace, and the container is raised.
In a multi output setup, if both visible workspaces have floating
containers, focus will NOT cross into the other output. It is assumed
the user will use a workspace binding in this case.
If two floating containers occupy the exact same center point and you
try to focus in a direction, the behaviour is undefined.
This commit adds support for laptop lid and tablet
mode switches as provided by evdev/libinput and
handled by wlroots.
Adds a new bindswitch command with syntax:
bindswitch <switch>:<state> <command>
Where <switch> is one of:
tablet for WLR_SWITCH_TYPE_TABLET_MODE
lid for WLR_SWITCH_TYPE_LID
<state> is one of:
on for WLR_SWITCH_STATE_ON
off for WLR_SWITCH_STATE_OFF
toggle for WLR_SWITCH_STATE_TOGGLE
(Note that WLR_SWITCH_STATE_TOGGLE doesn't map to
libinput and will trigger at both on and off events)
This changes the way zero (which is the default) is interpreted for both
the width and height of `floating_maximum_size`. It now refers to the
width and height of the entire output layout, which matches i3's
behavior.
This also removes duplicated code to calculate the floating constraints
in three files. Before this, `container_init_floating` used two-thirds
of the workspace width/height as the max and the entire workspace
width/height was used everywhere else. Now, all callers use a single
function `floating_calculate_constraints`.
Enables i3-compatible behavior regarding hiding the title bar on tabbed and
stacked containers with one child.
Related issues and merge requests: #3031, #3002, #2912, #2987.
This makes it so running `move [to] scratchpad` on a container already
in the scratchpad does not return an error. To match i3's behavior, a
visible scratchpad container will be hidden and a hidden scratchpad
container will be treated as a noop.
This modifies the places where output_get_active_workspace is called to
handle a NULL result. Some places already handled it and did not need a
change, some just have guard off code blocks, others return errors, and
some have sway_asserts since the case should never happen. A lot of this
is probably just safety precautions since they probably will never be
called when `output_get_active_workspace` is not fully configured with a
workspace.
This fixes a few misc memory leaks reported by asan:
- Items of `config->config_chain` are now freed instead of just the list
itself
- `bar->swaybar_command` is now freed
- The result returned by a seat subcommand is now returned instead of
leaked
If the bar was set to "invisible" and subsequently "toggle" was send twice, the
new mode was never set and the bar->mode was double freed.
Fix this by not requiring the bar->mode to be "hide" and instead show it
unconditionally, because it was either hidden or invisible.
Fixes#3637
