Commit Graph

9173 Commits

Author SHA1 Message Date
Matthias Clasen
5ce05a8fd0 events: Stop doing elaborate virtual modifier handling
These are going away in the GdkModifierType cleanup.
Just compare the modifiers we got.
2020-04-06 15:13:54 -04:00
Matthias Clasen
be2a0971c2 broadway: Stop using MOD2..MOD5
These are going away in the GdkModifierType cleanup.
2020-04-06 15:13:53 -04:00
Matthias Clasen
25e9a54902 win32: Stop using MODx modifiers
The win32 backend is using GDK_MOD2_MASK for AltGr,
so define GDK_MOD2_MASK locally to keep this working,
but remove any mention of GDK_MOD3_MASK,...,GDK_MOD5_MASK.
2020-04-06 15:13:00 -04:00
Matthias Clasen
7fa103717d wayland: Stop setting MOD2..MOD5
These are going away in the GdkModifierType cleanup.
2020-04-06 01:40:49 -04:00
Matthias Clasen
f8646b9733 gdk: Remove reserved bit from GdkModifierType
This really has no purpose and obscures the header.
2020-04-06 01:40:49 -04:00
Matthias Clasen
b02db72e17 Rename GDK_MOD1_MASK to GDK_ALT_MASK
We've hardcoded Mod1 = Alt for a long time, there is
no need to keep the confusing naming around anymore.
2020-04-06 01:40:49 -04:00
Matthias Clasen
13213c710f x11: Notify new device properties 2020-04-06 01:40:49 -04:00
Matthias Clasen
e165267924 wayland: Notify new device properties 2020-04-06 01:40:49 -04:00
Matthias Clasen
00b25d5594 Move keymap mapping functions to GdkDisplay
We want to stop exposing GdkKeymap, so this
functionality needs a new home.
2020-04-06 01:40:40 -04:00
Matthias Clasen
31fb5c287a Move GdkKeymapKey to gdktypes.h
We are going to use this type in a different header.
2020-04-05 23:43:41 -04:00
Matthias Clasen
416ec580dc gdk: Move modifier mask api to GdkDisplay
GdkKeymap is on the way out.
2020-04-05 23:43:41 -04:00
Matthias Clasen
ee14be8ac1 gdk: Add keymap properties to GdkDevice
GdkKeymap is on the way out.
2020-04-05 23:43:41 -04:00
Matthias Clasen
c9aef3fc4d Add gdk_event_get_match
This is a counterpart to gdk_event_matches() that can
be used to obtain a shortcut matching an event.
2020-04-05 23:37:13 -04:00
Matthias Clasen
e8330c5eec Add gdk_event_matches
Move the elaborate key event matching code from
GtkShortcutTrigger to GdkEvent, which greatly reduces
the amount of keymap api use outside of GDK.
2020-04-05 23:37:13 -04:00
Emmanuele Bassi
de481a7b5e Merge branch 'gir-fixes' into 'master'
Annotation fixes

See merge request GNOME/gtk!1616
2020-04-03 15:21:59 +00:00
Emmanuele Bassi
4e264a74be Annotation fixes
Match the arguments name in the declaration and the definition, and make
sure we have the proper transfer annotations.
2020-04-03 15:30:56 +01:00
Olivier Fourdan
4f49d83e43 gdk: Move GdkGrabStatus and GdkGrabOwnership decl.
With the removal of `gdk_seat_grab` we do not need to keep the
definitions of `GdkGrabStatus` and `GdkGrabOwnership` public.

Move those definitions to become internal only.
2020-04-03 15:23:01 +02:00
Matthias Clasen
cffa45d5ff wayland: Add a comment about keyboard state handling
Clarify a point that took me a few hours to fully track
down, so lets preserve what I found for the next poor
sould coming this way.
2020-04-02 17:43:14 -04:00
Jan Alexander Steffens (heftig)
7444b15d47 gdk/wayland: Clear modifiers when we lose keyboard focus
When we `Alt+Tab` away from a GTK application, it loses keyboard focus.
If we don't clear the modifiers, events from other devices that we
receive while unfocused will assume `Alt` is still pressed. This results
in e.g. Firefox navigating through the history instead of scrolling the
page when using the mouse wheel on it.

We don't get any information about modifiers while we are missing
keyboard focus, so assuming no modifiers are active is the best we can
do.

The shell sends us a modifier update immediately before we regain
keyboard focus, so the state shouldn't get out of sync.

Fixes https://gitlab.gnome.org/GNOME/gtk/-/issues/2112
2020-04-02 16:18:18 -04:00
Olivier Fourdan
c3ff175e3f win32: Add inhibit shortcut properties
Win32 backend doesn't have support for inhibit shortcuts, yet it needs
support the standard set of GdkToplevel properties.

Add support for the "inhibit-list" object property to GdkToplevel on
win32.
2020-03-30 18:25:36 +02:00
Olivier Fourdan
b22c7b8b77 broadway: Add inhibit shortcut properties
Broadway doesn't have support for inhibit shortcuts, yet it needs to
support the standard set of GdkToplevel properties.

Add support for the "inhibit-list" object property to GdkToplevel on
Broadway.
2020-03-30 18:25:36 +02:00
Olivier Fourdan
83027c68f1 x11: Implement inhibit_system_shortcuts API
On X11, there is no such equivalent to the inhibit shortcut protocol
found on Wayland.

To implement the inhibit_system_shortcuts API on X11, we emulate the
same behavior using grabs on the keyboard.

To avoid keeping active grabs on the keyboard that would affect other
X11 applications even when the surface isn't focused, the X11
implementation takes care of releasing the grabs as soon as the toplevel
loses focus.
2020-03-30 18:25:36 +02:00
Olivier Fourdan
44931a66df x11: Use gdk_x11_* rather than gdk_wayland_*
Rename the local functions for the x11 backend as gdk_x11_* rather than
gdk_wayland_*
2020-03-30 18:25:36 +02:00
Olivier Fourdan
36d986f33c wayland: Implement inhibit_system_shortcuts API
On Wayland, this is a direct mapping with the inhibit shortcut protocol
that GDK already supports.

Simply wire the new API to the existing entries.
2020-03-30 18:25:36 +02:00
Olivier Fourdan
5ac73be976 wayland: Use a GdkSeat as key for our internal hash table
With the current implementation, we use a `wl_seat` as the key for our
internal has table where we store the Wayland shortcuts inhibitors.

There is however no technical reason for this, and we could use a
GdkSeat instead, which will ease the implementation of the GdkToplevel
shortcut inhibition API.
2020-03-30 18:25:36 +02:00
Olivier Fourdan
a07a69e24e gdk: add internal gdk_surface_get_seat_from_event()
This can come handy to get the seat from a surface/event.
2020-03-30 18:25:36 +02:00
Olivier Fourdan
3e1f59af61 gdk: Add gdk_toplevel_inhibit_system_shortcuts API
With the removal of grabs from the public API, we need a replacement API
to let applications bypass system keyboard shortcuts.

A typical use case for this API is remote desktop or virtual machine
viewers which need to inhibit the default system keyboard shortcuts so
that the remote session or virtual host gets those instead of the local
environment.

Close: https://gitlab.gnome.org/GNOME/gtk/issues/982
2020-03-30 18:25:36 +02:00
Matthias Clasen
9ccd514682 x11: Don't use g_object_unref on events
Events are not object anymore.
2020-03-28 14:28:36 -04:00
Emmanuele Bassi
1ddd8586dd Add definition macro for internal types
GTK defines various types that are meant to be derivable only within GTK
itself, and "final" from the perspective of consumers of the GTK API.
The existing macros defined by GObject, such as G_DECLARE_FINAL_TYPE and
G_DECLARE_DERIVABLE_TYPE, lack this functionality.

While we wait for GObject to get this kind of macro, we should define
our own.
2020-03-25 23:14:45 -04:00
Benjamin Otte
2244eeb727 gdk: Remove GDK_RELEASE_MASK
It was only ever supported by keybindings and those are gone now.
2020-03-25 23:14:27 -04:00
Matthias Clasen
abeabc39e8 x11: Fix key event state translation
In the event refactoring, a bit was lost.

This makes Ctrl-Shift-i work again to open
the inspector.
2020-03-21 10:26:11 -04:00
Matthias Clasen
31db615885 Revert "Merge branch 'disable-window-test' into 'master'"
This reverts commit 3ac4c76b18, reversing
changes made to 6ec96d2e98.
2020-03-19 18:03:16 -04:00
Benjamin Otte
8c23ccd86e gdk: Remove GDK_RELEASE_MASK
It was only ever supported by keybindings and those are gone now.
2020-03-18 23:00:50 -04:00
Matthias Clasen
1576cced0f Drop GdkToplevel:accept-focus/:focus-on-map
These properties were only ever implemented in the
X11 backend. If you want to use them, just use the
X properties directly.
2020-03-14 19:28:00 -04:00
Matthias Clasen
32b9dd64db Drop GdkToplevel:keep-above/below
These properties were only ever implemented in the X11
backend. If you want to keep windows above or below
on X11, just use the X properties.
2020-03-14 15:46:55 -04:00
Matthias Clasen
60ea97187a Drop GdkToplevel:sticky
This was only ever implemented on X11. If you
want to use it there, just use the X properties
yourself.
2020-03-14 15:16:18 -04:00
Matthias Clasen
0d3df4fb1d Document gdk_toplevel_set_modal 2020-03-14 01:23:09 -04:00
Matthias Clasen
4be675b0c9 Small documentation fixes 2020-03-14 01:22:31 -04:00
Matthias Clasen
60a9e9d342 Merge branch 'x11-resize' into 'master'
x11: Update surface size for popups too

See merge request GNOME/gtk!1525
2020-03-14 03:35:31 +00:00
Matthias Clasen
8010bc4596 x11: Update surface size for popups too
Without this, the back buffers of the wrong size
keep being used, causing flickery misdraws, as
seen when expanding the expander in the popover
in widget-factory.
2020-03-13 21:33:23 -04:00
Matthias Clasen
331069f378 wayland: Stop capping cursor scales
We don't load entire themes anymore, so we can
easily accomodate larger cursor scales now.
2020-03-13 11:13:02 -04:00
Alexander Larsson
9e2829756f broadway: Inherit frame clock from parent surface 2020-03-13 15:26:22 +01:00
Alexander Larsson
b74a26b275 broadway: Keep popups above their parent
We set the parent as a "transient parent" for popups, which means
they will not be stacked below the parent.
2020-03-13 15:06:02 +01:00
Alexander Larsson
03d6d272f7 broadway: Track surface position correctly
surface->x/y (and various x,y arguments) should be in the parent
coordinates, so treat it as such. We also keep track of the root coords
as these are needed for popup positioning.

Also, drop the isTemp property server side and the weird initial
placement at (100, 100) in the daemon. We now fully control window
placement from the client instead. If this is not we want we should do
a serious design for that but until then lets do the simplest thing.
2020-03-13 15:05:59 +01:00
Alexander Larsson
1a763c440f broadway: Correctly track surface visibility in client
We forgot to unset visible when hiding surfaces
2020-03-13 15:00:32 +01:00
Alexander Larsson
43598fc5f2 broadway: Ensure images are decoded, not only loaded
In firefox, onload will trigger when the image is loaded, but at
that point it may not be decoded yet so showing it will sometimers
trigger flashes. We use the new decode() feature instead which ensures
both that the image is loaded *and* decoded, thus fixing the flashes.
2020-03-13 15:00:24 +01:00
Alexander Larsson
a4b718da4e broadway: Fix image updates for firefox
For whatever reason we need to assign the function to a variable
to be able to call it.
2020-03-13 15:00:18 +01:00
Alexander Larsson
086a577fd1 broadway: Handle browser hidpi scale factor 2020-03-13 14:59:37 +01:00
Matthias Clasen
9d98707f66 Merge branch 'present-toplevel-2' into 'master'
surface api splitup

See merge request GNOME/gtk!1511
2020-03-13 13:34:57 +00:00
Matthias Clasen
19e2ab76ce win32: Build fixes
These fixes were done blindly, to make the ci pass,
and will need review by somebody with access to an
actual win32 system to make sure the surface subtypes
are implemented properly.
2020-03-13 07:47:48 -04:00