This fixes the serial number for Nintendo Switch Pro, which is queried from the hardware in device initialization, and was later clobbered by the USB string which isn't correct.
Hiding a window and then clearing the minimized, maximized, fullscreen, or grabbed flags wouldn't result in the new state being applied to the window when shown, as the pending flags would always be zero, resulting in the flag application function never being entered. Calling SDL_RestoreWindow() didn't result in the minimized or maximized state being cleared on the hidden window either.
Save the current window flags to the pending flags when hiding a window, and universally apply the pending flags when the window is shown again to ensure that all state that was set or cleared when the window was hidden is applied. Any pending flags that match the existing window flags will be automatically deduplicated by the corresponding functions.
Alt-Enter will go from the current state to fullscreen desktop, or if already there, will leave fullscreen mode.
Ctrl-Enter will go from the current state to exclusive fullscreen mode, or if already there, will leave fullscreen mode.
- Removed the blanking window used in SDL_WINDOW_FULLSCREEN. This allows CMD + Tab to work.
- Changed the fullscreen NSWindow level so this works properly.
Fixes issue #7776
Raw input will not send game controller events while Remote Desktop is active, so dynamically switch between XInput and raw input when Remote Desktop state changes.
Fixes https://github.com/libsdl-org/SDL/issues/7759
1. Send display disconnected events for all displays no longer connected
2. Send display connected events for all displays that have been connected
3. Send window display events for any windows that have changed displays
Add aspect-correct output of scaled video modes and a hint to control this behavior (aspect, stretch, or none).
The Wayland spec states that fullscreen surfaces that do not cover the entire output shall be centered with the borders masked by the compositor, so no additional work is required aside from calculating the proper window dimensions.
The default is still 'stretch' mode, as some window managers as of this time (KDE and older versions of GNOME still found in LTS distros) don't behave according to the spec and present an unmasked window that is not centered, so it's not yet safe to change the default.
- Fix places working with window coordinates that need to call SDL_RelativeToGlobalForWindow or SDL_GlobalToRelativeForWindow
- Remove NSScreen param from ConvertNSRect(). Reflecting the Y coordinate is done relative to the main screen height (which ConvertNSRect
was already doing) so the explicit screen isn't needed.
- Refactor NSScreen lookups for point/rect and fix getting the screen for Cocoa_SetWindowPosition() to get the screen for the new position and
not the window's current screen (which may not exist if the window is off-screen).
- Fix re-associating the popup and parent window when the child window is shown. Hiding a child window removes it from the window hierarchy
and so must be added when the window is shown again.
- Allow popup windows that are not tooltips to gain key focus.
If SDL_HINT_APP_ID is set, pass it as the application.id to pipewire.
This gives any pipewire-based tools a hint to find an associated
.desktop file for icons, etc.
Apparently when using the Xbox One Wireless Adapter, using XInput at the same time as raw input will cause the controller to turn off immediately after connecting. This appears to be a bug in the Windows 11 driver stack, but since WGI provides all the extended functionality we need, this can be turned off for now.
Fixes https://github.com/libsdl-org/SDL/issues/3468
(cherry picked from commit b2e88ecfeb5e4d7db021e43c1b9bc4c9d14f615c)
In case something reports "Device" when we expected "device", etc.
Reference Issue #6835.
(cherry picked from commit df9d0fb332ea65c3fc47f72574851c91da2c912b)
Consolidate the X11_WMCLASS and WAYLAND_WMCLASS envvars into one SDL_HINT_APP_ID hint. This hint serves the same purpose on both windowing systems to allow desktop compositors to identify and group windows together, as well as associate applications with their desktop settings and icons.
The common code for retrieving the value is now consolidated under core/unix/SDL_appid.c as it's common to *nix platforms, and the value is now retrieved at window creation time instead of being cached by the video driver at startup so that changes to the hint after video initialization and before window creation will be seen, as well as to accommodate cases where applications want to use different values for different windows.
XRandR supports applying transformations to an output's picture
including changes to scale. Such scaling is used by some desktop
environments under feature names such as "fractional scaling" to
accomodate HiDPI devices. Alternatively, such scaling can be enabled by
a command such as the following:
xrandr --output DP1 --scale 0.5x0.5 --filter nearest
XRandR scaling has no "HiDPI awareness" or other way for an application
to signal that it wants to work with physical display pixels, and so all
we do is scale SDL's returned display modes so that applications receive
the number of usable pixels that they expect.
By default SDL will only enumerate controllers, to reduce risk of hanging or crashing on devices with bad drivers and avoiding macOS keyboard capture permission prompts.
The hidapi method of storing the error on the device is not thread-safe, and not only could it result in a double free if multiple threads were setting the error at the same time, but SDL could be trying to use the error message and have it be freed out from under it by another thread.
Use SDL's error functions since they already use thread-local storage.
This reverts commit 2b386b6c80.
This isn't the right approach. Even if the string itself isn't double-freed, it can be returned to the application and then freed while the application is trying to use it. This really needs to be in thread-local storage to be completely safe.
In SDL we already have a global thread-local error string, so I'm going to make an SDL-specific change to handle the error strings safely.
The error string is not protected by a mutex, and can be set from multiple threads at the same time. Without this change, it can be double-freed. It can still be double-allocated, leading to a memory leak, but at least it won't crash now.
Signed-off-by: Sam Lantinga <slouken@libsdl.org>
They can vanish for UP TO EIGHT SECONDS...!
This is for devices that connect to HDMI/DisplayPort/etc, where it
presumably has to wait for a display to get up and running before it
can play audio through it, so one can see the audio device fail when
changing display modes, or the system returning from sleep. Since this
can be triggered by a game changing video resolutions at startup (either
before or after opening the audio device!), it's important to deal with.
In normal conditions, it shouldn't take this long to open or recover an
audio device, but this is better than unexpectedly losing the device
in this situation.
Fixes#7044.
Fixes#5571.
(cherry picked from commit 48e71ae87be425f117dece3735b148fbc5f2606e)
This is unsafe because the event is auto-reset, therefore the call to
WaitForSingleObject() resets the event which GetOverlappedResult() will
try to wait on.
Even though the overlapped operation is guaranteed to be completed at
the point we call GetOverlappedResult(), it will still wait on the event
handle for a short time to trigger the reset for auto-reset events. This
amounts to roughly a 100 ms sleep each time GetOverlappedResult() is called
for a completed I/O with a non-signalled event.
In the context of HIDAPI, this extra sleep means that callers that loop
on hid_read_timeout() with timeout=0 will loop forever, since the 100 ms
sleep each iteration ensures ReadFile() will always have new data.
Signed-off-by: Cameron Gutman <aicommander@gmail.com>
Signed-off-by: Sam Lantinga <slouken@libsdl.org>
- hidapi already called CancelIo on hid_close but that only cancels pending IO for the current thread. Controller read/writes originate from multiple threads (serialized, but on a different thread nonetheless) but device destruction was always done on the main device thread which left any pending overlapped reads still running after hidapi's internal read buffer is deallocated leading to intermittent free list corruption.
Signed-off-by: Sam Lantinga <slouken@libsdl.org>
Touching HID devices with keyboard usages will trigger a keyboard capture
permission prompt on macOS 11+. See #4887
Like the IOKit joystick backend, we accept HID devices that have joystick,
gamepad, or multi-axis controller usages. We also allow the Valve VID for
the Steam Controller, just like the Windows HIDAPI implementation does.
Signed-off-by: Cameron Gutman <aicommander@gmail.com>
Signed-off-by: Sam Lantinga <slouken@libsdl.org>
The get_usb_string call is rather expensive on some USB devices, so we
cache the vendor/product strings for future lookups (e.g. when
hid_enumerate is invoked again later).
This way, we only need to ask libusb for strings for devices we haven't
seen since before we started.
Signed-off-by: Steven Noonan <steven@valvesoftware.com>
Signed-off-by: Sam Lantinga <slouken@libsdl.org>
We weren't meant to have multiple contexts and mainloops, but we had one
for each opened device and the hotplug detection thread. Instead, use
pa_threaded_mainloop, which can be shared between threads and objects, and
a single context (which, according to the PulseAudio documentation, is
usually meant to be a singleton that represents a global server connection,
possibly with multiple streams hung on it).
Now instead of polling in a loop, threads will block until the
threaded_mainloop runs a callback, and the callback will fire a signal to
unblock the thread.
Prior to this, the code upset ThreadSanitizer, as Pulse has some unprotected
global resource that each mainloop/context would touch.
Reference Issue #7427.
SDL mutexes are always recursive in modern times, so no need to check this,
plus the test triggers a false-positive on ThreadSanitizer.
Reference Issue #7427.
There are many toolkit specific ways to set a font DPI in X11 desktop
environments. The primary approach of reading a Gnome specific setting
from the portal is ok, it will work on Gnome and on Plasma most the
time.
The current fallback GDK_SCALE is less great; it's an internal GTK
setting relating to the mapping of logical pixels to device pixels
within the toolkit, it's a developer setting for GTK devs. We were
instructed within Plasma to not set this as it caused issues.
Xft.dpi in xresources is a good universal fallback, it's very dated to
the point that it works in clients like xterm, Qt on X11 uses it in our
font DPI path.
Re-writes clipboard data handling in X11 to an on demand approach where
data can be produced on request instead of storing it in X11 properties
on the window.
Primary selection has been changed to mimic this behavior even though
it's only possible to use it for text as of now.
Ensure that incoming touch events originate from valid surfaces owned by SDL and have proper window data before forwarding them to the touch subsystem, or the window focus pointer that is sent with the event may not be a pointer to an SDL window.
Additionally, ensure that allocated touch events are always cleaned up on exit.
- SDL_MaximizeWindow, MinimizeWindow, SetFullScreenWindow, etc are not meant to show the window if it isn't currently visible, but on some platforms
(e.g. Windows) it isn't possible to set this state without also showing the window so cache the flags in a pending_flags field until SDL_ShowWindow is
called. Then replay the pending flags through ApplyPendingFlags (hoisted out from SDL_FinishWindowCreation).
If the text-scaling-factor setting is available via D-Bus, add a listener and update the content scale values for the displays if the value is changed during runtime.
Factors out the D-Bus message pump from the system theme detection code to the general D-Bus code, as it's now used for more purposes than just the system theme.
Expose the text scaling factor (aka the global scale factor on KDE) as the display content scale value so applications can scale themselves as necessary.
If D-Bus is unavailable or retrieving the setting fails, fall back to trying the legacy GDK_SCALE envvar before reverting to the default 1.0 value.
Change 8067023 by mikela:
Add SDL_HINT_WINDOW_ACTIVATE_WHEN_RAISED to SDL_RaiseWindow
- When set to false, this allows SDL_RaiseWindow to bring a chosen window to the top of the stack but not force input focus to it
Change 8067041 by mikela:
Rename SDL_HINT_WINDOW_NO_ACTIVATION_WHEN_SHOWN to SDL_HINT_WINDOW_ACTIVATE_WHEN_SHOWN
In theory this is illegal, but legit wavefiles in the field do it, and
it's easy to bump it to 1 for general purposes.
Formats with more specific alignment requirements already check for them
separately.
Fixes#7714.
First one to load and have the necessary symbol is the one we accept,
if any. Once we accept one, we won't try loading others.
Fixes#7613.
(cherry picked from commit 32999798e0232cdeda777a32ce2fe6b78ec2bb4e)
We have gotten feedback that abstracting the coordinate system based on the display scale is unexpected and it is difficult to adapt existing applications to the proposed API.
The new approach is to provide the coordinate systems that people expect, but provide additional information that will help applications properly handle high DPI situations.
The concepts needed for high DPI support are documented in README-highdpi.md. An example of automatically adapting the content to display scale changes can be found in SDL_test_common.c, where auto_scale_content is checked.
Also, the SDL_WINDOW_ALLOW_HIGHDPI window flag has been replaced by the SDL_HINT_VIDEO_ENABLE_HIGH_PIXEL_DENSITY hint.
Fixes https://github.com/libsdl-org/SDL/issues/7709
sdl12-compat can get into a state where a color-keyed surface is
marked for blending, but wants to blend with full alpha (which
is the same as _not_ blending), so rather than fail to find a
blitter in that case, it just selects the colorkey blitter.
Reference https://github.com/libsdl-org/sdl12-compat/issues/233
(cherry picked from commit 0eea92c8fcc4b5c572a0dd9848ca8886978ee0a2)
It turns out that screen coordinates were confusing people, thinking that meant pixels, when instead they are virtual coordinates; device independent units defined as pixels scaled by the display scale. We'll use the term "points" for this going forward, to reduce confusion.
This was only including the resampling buffer needs if it was larger
the other allocation needs, but it needs to be included unconditionally.
For safety's sake, we also make sure the pre-resample buffer doesn't risk
overflow, too, but this might not be necessary in practice.
Re-writes the clipboard data handling in wayland to an on demand
solution where callbacks are provided to generate/provide the clipboard
data when requested by the OS.
Wayland doesn't support getting the true global cursor position, but it can be faked well enough for what most applications use it for: querying the global cursor coordinates and transforming them to the window-relative coordinates manually.
The global position is derived by taking the cursor position relative to the toplevel window, and offsetting it by the origin of the output the window is currently considered to be on. The cursor position and button state when the cursor is outside an application window are unknown, but this gives 'correct' coordinates when the window has focus, which is good enough for most applications.
Before, as ConvertAudio might have expanded data in-place temporarily during
its work, this could blow up. Now if there's a chance of that, it'll
work out of an internal buffer and copy the final results to the output
buffer.
If the output format can handle the temporary expansion, we write directly
to the output buffer without the extra copy.
Fixes#7668.
Otherwise, a CoreAudio thread lingers forever, and coreaudiod eats CPU
until the SDL process terminates.
Fixes#7689.
(cherry picked from commit 86786ed5447fe32ea2e48f12f0598816a76721c2)
If the popup is positioned such that it requires correction on both the x and y axes, it will be aligned with parent only at the window corners, which is neither overlapping nor adjacent. In this case, nudge the window plus or minus one screen unit on the x axis so it is properly adjacent to the parent and within spec guidelines.
An in-place swizzle mutation was erroneously inside of a loop, which
caused each consecutive 4-pixel vector to alternate between correct and
incorrect endianness.
The bug was introduced in 715e070d29.
Thanks to RobbieAB for reporting the bug.
Fixes https://github.com/libsdl-org/SDL/issues/3428
The window geometry will be updated when in underlying shell surface config handler, before the config is ack-ed, so no need to do it in the popup config handler.
Validate and reposition popups in any case where the position or size may have changed. In particular, this fixes cases where the position parameters were adjusted while the window was hidden, as the new values weren't being applied in all cases.
Popups beyond the right and bottom borders of the window must be width/height minus one in order to be considered adjacent and not be instantly closed or cause a protocol error.
In this case we want the display mode pixel to screen coordinates to be 1:1 ... but we lose information about the UI scaling of the display - is that okay?
As child windows can be recursively destroyed when their parents are destroyed, emit an event to notify the application when a window is being or has been implicitly destroyed so that it can appropriately clean up any associated resources.
If the application has registered an event watch, the destroy message will be received when the window handle is still valid, so the application can retrieve and release any userdata associated with the window. If the message is processed at any time after that, the window handle is already invalid and the ID is only useful for application-side bookkeeping purposes.
The docs say you should, if not statically initializing an SRWLOCK--which
we aren't--but in practice this is probably just being pedantic.
Still, better (thread) safe than sorry!
- SDL_AudioCVT is gone, even internally.
- libsamplerate is gone (I suspect our resampler is finally Good Enough).
- Cleanups and improvements to audio conversion interfaces.
- SDL_AudioStream can change its input/output format/rate/channels on the fly!
c1b9d2ad98
Properly handle the close of run loop on macOS
(https://github.com/libusb/hidapi/pull/522)
- as per documentation `kCFRunLoopRunStopped` should be handled once the runloop is closed via `CFRunLoopStop`;
- if it is not handled - a race condition/crash may happen on the latest macOS when a device gets disconnected while being open;
Insert new displays at the end of the list instead of the front so that the initial ordering, as exposed by the compositor, is preserved. This is particularly important when the compositor exposes the xdg-output instance after the wl_output instance, as xdg-output must be attached to the outputs in the same order that wl_output exposed them, or they can be added to the SDL output list in reverse order.
Destroy any proxy wrappers and callbacks before the associated event queues to silence libwayland warnings about destroying the queues while proxies are still attached.
wl_compositor v6 introduces the preferred buffer scale event, which serves a similar function to the fractional scale protocol, but deals in integer scale factors. Listen to this event when the wl_compositor version is >= 6 and the fractional scale protocol is not present to set the scale factor for surfaces.
- Composited windows seem to need to actually paint (or appear to paint) through calls to Begin/EndPaint to properly validate
their update region. If not done they will continue to receive WM_PAINT messages for the same region.
When X11_UpdateWindowPosition() was called and the position didn't update
we would fire an SDL_EVENT_WINDOW_MOVED event with the global x,y for
the pop-up instead of the relative position for the pop-up.
This change ensures we always have a relative position for pop-ups before sending
the SDL_EVENT_WINDOW_MOVED event.
xdg-output names and descriptions are deprecated in favor of those provided by the core wl_output protocol as of v4. Bump wl_output support to v4 and ignore the xdg-output description in favor of the wl_output description when appropriate.
This allows clients to be aware of "natural scrolling" settings in the
compostior and adjust accordingly to optionally keep the same wheel and
trackpad behaviour accross platforms.
Testing using the "testmouse" example, behaviour in the demo is now the
same regardless of the compositor setting for scrolling.
./src/video/SDL_video.c:1734:13: warning: no previous prototype for function 'SDL_CreateWindowInternal' [-Wmissing-prototypes]
SDL_Window *SDL_CreateWindowInternal(const char *title, int x, int y, int w, int h, SDL_Window *parent, Uint32 flags)
^
./src/video/SDL_video.c:1734:1: note: declare 'static' if the function is not intended to be used outside of this translation unit
SDL_Window *SDL_CreateWindowInternal(const char *title, int x, int y, int w, int h, SDL_Window *parent, Uint32 flags)
^
static
Track the serial numbers of key events, mouse button presses, touch down events, tablet tool down events, and tablet button presses, and pass the serial, along with the seat, to the xdg-activation protocol when raising windows to increase the chances of it succeeding.
Don't perform a round trip when setting the minimized and maximized window states during initial window creation, as it will be done later in the window creation process.
When changing the fullscreen state and performing a roundtrip, more than one configure event can be sent before getting the event that changes the fullscreen state. Don't change the higher-level window state and avoid recursive calls back into the video system during these events.
According to the xdg-shell spec, xdg-toplevel min/max values are just hints, which compositors are free to ignore, and they often do so in several cases (usually when tiling, maximizing, or entering fullscreen). Even with the hints set, manually constraining the surface dimensions for non-maximized, non-fullscreen surfaces is required for consistent behavior.
In the case of maximized windows, the dimensions as specified in the configure event are non-negotiable and must be respected by the client, as not doing so is a protocol violation.
This also eliminates some redundant calls to set the libdecor resizable state, which can cause unnecessary commits internally.
It turns out there's a race condition on X11 where the window could be placed by the window manager while being placed by the application, so we need to have the initial position available at window creation.
Ozkan Sezer
Sam Lantinga
Sylvain
Frank Praznik
Ryan C. Gordon
Rusty Moyher
Alibek Omarov
Nikita Krapivin
David Gow
Jeffrey Knockel
Anonymous Maarten
David Edmundson
Linus Probert
Caleb Cornett
M. P. Halpin
valid-ptr
Vid Tadel
Semphris
Jeremy Rand
Semphriss
Brick
Mingjie Shen
Cyril Dubet
Ethan Lee