|
88cb4962
|
2020-12-09T06:42:31
|
|
Fixed bug 5291 - SDL_SetRenderTarget unnecessarily changes target when current target is the native texture of the passed in texture
|
|
|
f2fff217
|
2020-12-09T06:24:40
|
|
Fixed bug 5374 - WGI: Use fast-pass strings.
Joel Linn
Eliminate additional heap allocation for short-lived HSTRINGs.
Uses `WindowsCreateStringReference()` to disable reference counting and memory management by the Window Runtime.
|
|
|
42c5b4ac
|
2020-12-09T06:17:55
|
|
Fixed bug 5366 - cmake build doesn't detect Metal on macOS
Tom Seddon
2nd time lucky, perhaps. patch 2 applies to current HEAD at time of writing - 4eb049c9bb1ca94efe3c40b57beda3169984d0cb from https://github.com/SDL-mirror/SDL.
This basically goes back to what was there originally, but now manually adding "-x objective-c" to the clang command line rather than "-ObjC". clang is then invoked without the "-x c" that was causing the problem, the snippet builds, and Metal is detected. (I had a quick trawl through the cmake code, but I couldn't see where this is handled.)
I was moved to try this after finding SDL's own CHECK_OBJC_SOURCE_COMPILES macro, and noting what it does: https://github.com/SDL-mirror/SDL/blob/4eb049c9bb1ca94efe3c40b57beda3169984d0cb/cmake/macros.cmake#L67
An alternative fix of course would be to use CHECK_OBJC_SOURCE_COMPILES instead of cmake's check_objc_source_compiles - but that had the same problem of getting confused by "return 0;". (Maybe that's because it's a macro? I'll defer to a cmake expert on this one.)
I decided in the end to err on the side of leaving things looking basically the same as they were before my first patch.
|
|
475405e0
|
2020-12-09T12:03:24
|
|
CMakeLists.txt: sync DYLIB_CURRENT_VERSION to Xcode project
|
|
|
250a0557
|
2020-12-09T12:01:10
|
|
fix bug #5384 -- define DLL_EXPORT in DLL builds and adjust begin_code.h
|
|
|
50203d58
|
2020-12-08T22:00:06
|
|
Fixed bug 5329 - SDL_SetWindowGrab(SDL_FALSE) fails to unlock cursor if window is partially offscreen
Ivan Mogilko
With SDL 2.0.12 under MS Windows, if the window is partially offscreen calling SDL_SetWindowGrab(w, SDL_TRUE) works, but subsequent call to SDL_SetWindowGrab(w, SDL_FALSE) does not work.
I tested this in both real program, and a small test app, where unlocking cursor worked perfectly while window is fully in desktop bounds, but did not work if it was at least few pixels outside.
For the reference, following code is enough to reproduce the issue:
#include <windows.h>
#include <SDL.h>
int WinMain(
HINSTANCE hInstance,
HINSTANCE hPrevInstance,
LPSTR lpCmdLine,
int nShowCmd)
{
SDL_Init(SDL_INIT_VIDEO);
SDL_Window* w = SDL_CreateWindow("", SDL_WINDOWPOS_CENTERED, SDL_WINDOWPOS_CENTERED, 640, 400, 0);
bool grabbed = false;
bool want_quit = false;
while (!want_quit)
{
SDL_Event event;
while (SDL_PollEvent(&event))
{
switch (event.type)
{
case SDL_QUIT: want_quit = true; break;
case SDL_KEYDOWN:
if (event.key.keysym.scancode == SDL_SCANCODE_SPACE)
{
SDL_SetWindowGrab(w, static_cast<SDL_bool>(!grabbed));
grabbed = !grabbed;
}
}
}
}
SDL_DestroyWindow(w);
SDL_Quit();
return 0;
}
|
|
|
43aad966
|
2020-12-08T19:03:50
|
|
Fixed bug 5222 - Crash when running with -DHIDAPI=ON
Mathieu Eyraud
SDL dynamically loads libusb but does not check the return value of 'SDL_LoadFunction'.
Also libusb is loaded and initialized several time because 'SDL_hidapi_wasinit' is never set to true.
I made a patch if you want to test:
- check that 'hid_init' is called once and only once,
- check return value of 'hid_init',
- check return value of 'SDL_LoadFunction',
- check return value of 'SDL_malloc',
- add some debug logging.
|
|
|
a2098a47
|
2020-12-08T18:56:06
|
|
Updated SDL to 2.0.14 in preparation for release candidate
|
|
|
2a9591a9
|
2020-12-08T09:16:34
|
|
Accepted patch https://github.com/microsoft/vcpkg/blob/master/ports/sdl2/fix-space-in-path.patch
|
|
|
7f1c6e82
|
2020-12-08T09:13:08
|
|
Accepted patch https://github.com/microsoft/vcpkg/blob/master/ports/sdl2/enable-winrt-cmake.patch
|
|
|
7665f887
|
2020-12-08T09:07:21
|
|
Accepted patch https://github.com/microsoft/vcpkg/blob/master/ports/sdl2/disable-wcslcpy-and-wcslcat-for-windows.patch
|
|
|
695499ae
|
2020-12-08T09:04:28
|
|
Accepted patch https://github.com/microsoft/vcpkg/blob/master/ports/sdl2/disable-hidapi-for-uwp.patch
|
|
|
5610e5ae
|
2020-12-08T00:11:10
|
|
Makefile.os2: disable W303 (add -wcd=303 to CFLAGS)
newer OpenWatcom versions enable W303 by default. without this, we get
multiple "Parameter '%s' has been defined, but not referenced" warnings.
|
|
|
c9723c40
|
2020-12-07T09:38:21
|
|
Fixed potential hang in joystick close if the rumble thread is blocked for some reason
It's still possible to hang when shutting down, if the rumble thread is still hung, but it won't block indefinitely at runtime.
|
|
|
3c68051e
|
2020-12-04T15:47:28
|
|
Allow background input when testing game controllers
|
|
|
09909d02
|
2020-12-03T19:44:47
|
|
Fixed handling of BACK button on newer Xbox One S controllers
|
|
|
54e5136b
|
2020-12-03T18:17:04
|
|
Refactored Xbox One Bluetooth protocol and verified Xbox One S, Xbox Series X, and Xbox One Elite Series 2 controllers
|
|
|
1031231b
|
2020-12-03T18:17:03
|
|
Fixed duplicating a device between XInput and HIDAPI
|
|
|
59f28b7f
|
2020-12-03T18:17:01
|
|
Fixed whitespace
|
|
|
f487d63a
|
2020-12-03T18:16:56
|
|
Fixed crash when printing NULL wide character string
|
|
754286c6
|
2020-12-02T13:45:24
|
|
SDL Renderer: specify the correct flag when recreating the window
|
|
|
035f8f23
|
2020-12-02T13:37:59
|
|
SDL_RecreateWindow: allow clearing VULKAN when recreating the window
|
|
|
93fbab0f
|
2020-12-02T11:04:53
|
|
SDL_ReCreateWindow: allow to unload METAL window and switch back to OpenGL.
On older mac, where METAL Renderer METAL fails to create, it allows to switch back to OpenGL SDL_Renderer
by re-creating the window (METAL flags was previously persistent).
|
|
|
7854f43b
|
2020-12-01T14:43:15
|
|
Disable SDL_JOYSTICK_HIDAPI on iOS and tvOS by default
It's only needed for Steam Controller support, and introduces a Bluetooth framework dependency which requires additional permissions on the App Store.
|
|
|
ebf315e0
|
2020-12-01T14:01:12
|
|
Fixed bug 5369 - iOS static library build copies public headers files into xcode archive
Dominik Reichardt
Trying to integrate the latest SDL2 changes into our iOS project of Exult I've stumbled over the fact that when I added the static iOS library the public header files were copied to the archive of our project when you let Xcode build the archive.
This makes the archive invalid for upload to the AppStore Connect.
To fix this you need to delete the public headers from the build phase:
Open the xcode project, select the target "Static Library-ios", got to build phases, and in "headers" delete all the headers in the "public" group. This is safe to do as this actually just copies the public headers for some odd counterintuitive reason.
I think this needs to be done for all the library build targets but likely not for the framework targets.
|
|
|
4b35a18d
|
2020-12-01T13:50:42
|
|
Fixed bug 5366 - cmake build doesn't detect Metal on macOS
Tom Seddon
This is as of commit 50d804ea729accf9e3a9ce83238d0a2976a17545 from https://github.com/SDL-mirror/SDL, which is HEAD as I write (apologies, not confident with Mercurial)
# Config
macOS: 10.14.6 (18G6042)
cmake --version: cmake version 3.16.20200101-g23e782c
clang --version: Apple clang version 11.0.0 (clang-1100.0.33.17)
Xcode version: Version 11.3.1 (11C504)
# Repro steps
Run the following commands in the shell.
cd /tmp/
git clone https://github.com/SDL-mirror/SDL
mkdir build.SDL
cd build.SDL
cmake -G ../SDL/
Examine cmake output.
# Expected result
Metal is detected.
# Actual result
It appears that Metal is not detected! Note this line in the summary:
-- RENDER_METAL (Wanted: 0): OFF
# Fix
Change check_c_source_compiles to check_objc_source_compiles. The cmake script tries to add -ObjC to the clang command line, but, for whatever reason, this doesn't seem to work.
Change the test source to have an empty main. The "return 0;" line seems to confuse cmake somehow, causing it to crap out with an error about HAVE_FRAMEWORK_METAL being an unknown argument. (Maybe I'm just dense, but it's not obvious to me what the problem is here.)
With these two changes:
-- RENDER_METAL (Wanted: ON): ON
Patch attached.
|
|
|
798b2288
|
2020-12-01T13:46:34
|
|
Fixed bug 5367 - SDL_OpenUrl disabled on iOS due to unified Xcode project file
Dominik Reichardt
When you unified the Xcode project file you forgot to add the iOS file (/src/misc/ios/SDL_sysurl.m) to the project file, so the dummy in that folder takes over and the call does not work.
Relevant commits:
http://hg.libsdl.org/SDL/rev/c86bbf75f55e (SDL_OpenUrl enabled for macOS/iOS)
http://hg.libsdl.org/SDL/rev/282c3dc1cf65 (and following: unified Xcode project files)
Adding an ios group in the misc group of the project file and then adding /src/misc/ios/SDL_sysurl.m to it fixes the problem.
|
|
|
c78ca2d1
|
2020-12-01T13:38:42
|
|
Fixed bug 5371 - Rawinput: Fix truncating cast of string length.
Joel Linn
Fixes an implicit truncation of a string length on 64bit systems.
|
|
|
a3ccf9ad
|
2020-12-01T13:36:41
|
|
Fixed bug 5373 - [PATCH] Rawinput: Get correlated XInput battery info
Joel Linn
Currently the rawinput driver always reports a device as "wired". This changes that to "unknown" and updates it once the device is correlated with xinput.
|
|
|
e3966e25
|
2020-11-30T13:04:30
|
|
Use the correct internal API for updating the battery level for PS5 controllers
|
|
|
f4ed07de
|
2020-11-30T13:02:34
|
|
We don't know whether the PS5 controller is Bluetooth or not when we open it
|
|
|
04f0fd13
|
2020-11-27T18:57:42
|
|
Remember to close the game controller when we're done with it
|
|
|
1f2f536b
|
2020-11-27T18:57:40
|
|
Fixed XInput correlation for raw input controllers after hotplug events
|
|
|
012471e9
|
2020-11-27T18:57:38
|
|
Open and test all connected controllers
|
|
|
1c865c46
|
2020-11-27T18:57:36
|
|
Load the raw input device list at init time so it's available when DirectInput is doing device detection
|
|
|
a0c5bfa3
|
2020-11-27T13:08:40
|
|
Moved raw input event processing from the main thread to the joystick thread
This allows fast joystick event delivery regardless of what the main thread is doing.
|
|
|
4fbefbe2
|
2020-11-27T11:33:53
|
|
Sort the raw input axes by usage, so X comes before Y, etc.
|
|
|
4ddac485
|
2020-11-27T11:33:51
|
|
Backed out minor optimization that prevented correlation_id from being set
|
|
|
8973a258
|
2020-11-27T10:44:56
|
|
Enable dispatching of WM_INPUT_DEVICE_CHANGE events directly, in case the application hasn't created a window with the normal message loop
|
|
|
e8adc648
|
2020-11-27T10:44:55
|
|
Enable dispatching of WM_INPUT events directly, in case the application hasn't created a window with the normal message loop
|
|
|
0252235e
|
2020-11-27T10:44:53
|
|
Recheck devices if another API queries raw input for a new device
|
|
|
a7dede7e
|
2020-11-27T10:44:51
|
|
Re-enable axis correlation for raw input controllers, for twin stick shooters that don't need face buttons
|
|
|
ce77966d
|
2020-11-27T10:44:49
|
|
Fixed RAWINPUT_IsDevicePresent() not returning TRUE for Xbox One controllers
|
|
|
8a449de2
|
2020-11-27T10:44:47
|
|
Fixed Xbox 360 wireless controller being picked up by WGI when it's being managed by RAWINPUT
|
|
|
37c9e4af
|
2020-11-27T06:03:15
|
|
Fixed processing WM_INPUT_DEVICE_CHANGE at startup
|
|
|
248fc75b
|
2020-11-27T05:53:56
|
|
Correlate just based on buttons, joystick axes are not as precise and could potentially cause incorrect uncorrelation.
It's okay if the triggers aren't precise until someone presses a button on their controller.
|
|
|
e7e615de
|
2020-11-27T05:53:54
|
|
Removed extraneous windows message pumping
Most of the raw input events are dispatched in the main windows message loop. We only dispatch device change messages separately when we need them to be completely up to date.
|
|
|
2931eccd
|
2020-11-27T05:53:52
|
|
Fixed detecting Bluetooth raw input devices, which have device names longer than 128 characters
|
|
|
849ce803
|
2020-11-27T05:53:50
|
|
Renamed SDL_JOYSTICK_RAWINPUT_GAMING_INPUT to SDL_JOYSTICK_RAWINPUT_WGI
|
|
|
219a28dd
|
2020-11-27T03:45:05
|
|
Fixed D-pad uncorrelating raw input controllers (thanks Jimbly!)
|
|
|
2c079a2f
|
2020-11-27T09:42:14
|
|
SDL_FillRects: prevent empty SDL_surface from raising an error message
It's legitimate to have a surface with 0 width or height (null 'pixels' pointer).
But calling SDL_FillRects would wrongly set the error "You must lock the surface".
|
|
|
3fbff2a4
|
2020-11-26T10:47:33
|
|
SDL_rawinputjoystick.c: fix ambiguous 'else' warning
src/joystick/windows/SDL_rawinputjoystick.c: In function 'RAWINPUT_HandleStatePacket':
src/joystick/windows/SDL_rawinputjoystick.c:1343:9: warning: suggest explicit braces to avoid ambiguous 'else'
|
|
|
845b9033
|
2020-11-25T16:46:42
|
|
Implemented trigger rumble for raw input controllers
|
|
|
517be808
|
2020-11-25T16:05:19
|
|
Implemented battery status for Bluetooth Xbox One controllers
Also switched the rumble loop count to 0xEB (one hour) to match Windows driver
|
|
|
24cc0012
|
2020-11-25T16:05:15
|
|
Some controllers don't always reset their triggers to zero when they are released (e.g. Xbox One S in Bluetooth mode), so only trigger rumble if the trigger is pulled halfway or more.
|
|
|
6a7f29cd
|
2020-11-25T11:31:17
|
|
Fixed compiling on tvOS
|
|
|
9ec2cf56
|
2020-11-25T10:31:59
|
|
Fixed building with Windows.Gaming.Input enabled
|
|
|
3c07dd12
|
2020-11-25T10:28:48
|
|
Strip an extra '.' off of the HORI manufacturer name
|
|
|
e5783e11
|
2020-11-25T14:51:56
|
|
cmake: add missing checks for wcscasecmp, _wcsicmp, wcsncasecmp, _wcsnicmp
|
|
|
7c18088f
|
2020-11-25T14:51:56
|
|
SDL_config_os2.h: define HAVE__WCSICMP and HAVE__WCSNICMP
|
|
|
159d1b3d
|
2020-11-25T01:35:45
|
|
Don't set the serial number after the controller is opened
We'll use the USB value instead so it's available as soon as it's opened.
|
|
|
7d92b14f
|
2020-11-25T01:18:18
|
|
Simplified Xbox One controller initialization state, and don't query for the serial number.
|
|
|
9fc4a4c9
|
2020-11-24T22:25:26
|
|
Revamped Xbox One HIDAPI init sequence
Added support for querying the controller serial number on newer firmware
|
|
|
e16afa79
|
2020-11-24T16:38:49
|
|
Automatically switch to testing a new controller when it's plugged in
|
|
|
46a84478
|
2020-11-24T12:43:01
|
|
Added SDL_wcscasecmp() and SDL_wcsncasecmp()
|
|
|
e4b7d9a2
|
2020-11-24T07:56:59
|
|
Removed usage of TARGET_OS_OSX for building with older SDKs
|
|
|
16d789fb
|
2020-11-24T07:30:39
|
|
Disable compiler warning when using -fobjc-weak building C code
|
|
|
a5cde4cb
|
2020-11-24T07:29:55
|
|
Fixed trying to use @available() on older SDK
|
|
|
b4338dda
|
2020-11-24T07:22:29
|
|
Enable -fobjc-weak when building MFI controller code
|
|
|
97782e6c
|
2020-11-24T07:11:49
|
|
Don't try to call IOS_SupportedHIDDevice() if it's not defined
|
|
|
a063b9fb
|
2020-11-24T07:06:31
|
|
Fixed compiler warning
|
|
|
c63bbb06
|
2020-11-24T06:55:33
|
|
Including SDL_config_iphoneos.h enables MFI controller code
|
|
|
eba069ea
|
2020-11-24T06:53:48
|
|
It looks like the Objective C support needed to build MFI controller code is in 10.8
|
|
|
7a05dbf4
|
2020-11-24T06:42:53
|
|
Fixed building on FreeBSD
Alex S
Looks like we have a collision with https://hg.libsdl.org/SDL/rev/cd774daff9f6. (Again, the headers in the base system are intended for drivers and should not be used for compiling non-base applications. At least that's the policy for now: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=240964#c19.)
|
|
|
feab9d42
|
2020-11-24T06:40:13
|
|
5363 - Memory leak of joystick->sensors in SDL_JoystickClose
Mathieu Eyraud
Joystick->sensors is never freed.
|
|
|
23ad4f4b
|
2020-11-24T14:10:30
|
|
configure.ac: check GameController framework support when targeting Darwin
disables SDL_JOYSTICK_MFI for i386 or if MAC_OS_X_VERSION_MIN_REQUIRED < 1090
--disable-joystick-mfi disables it unconditionally.
|
|
|
711d4090
|
2020-11-24T14:10:30
|
|
CMakeLists.txt: add src/joystick/iphoneos/*.m to Darwin joystick sources
.. so that there won't be missing symbols.
TODO: add checks for SDL_JOYSTICK_MFI ???
|
|
|
34bea84a
|
2020-11-23T23:03:55
|
|
Fixed bug 5335 - Patch: enable joystick/haptic/evdev support by default on FreeBSD
Alex S
Ah, that's not quite enough. You need to:
1. rename src/joystick/bsd/SDL_sysjoystick.c to something;
2. regenerate configure.
|
|
|
1e943e2a
|
2020-11-23T22:59:22
|
|
Fixed building with an older SDK and macOS target
|
|
|
1fc5ca64
|
2020-11-23T22:24:54
|
|
Fixed building with an older SDK and macOS target
|
|
|
4121ab62
|
2020-11-23T21:25:07
|
|
Fixed bug 5354 - DirectFB: Add Vulkan support
Nicolas Caramelli
This patch adds Vulkan support for the DirectFB video driver.
A screenshot with SDL tests running on DirectFB is available on the HiGFXback project:
https://github.com/caramelli/higfxback/wiki/DirectFB#sdl
|
|
|
e44bf8a0
|
2020-11-23T21:18:37
|
|
Fixed bug 5359 - Incorrect sensor data from DualShock4
multiply gyro values by sensitivity
When the hardware calibration fails, values read from sensors need to be multiplied by default sensitivity (16 for gyro, 1 for accelerometer).
|
|
|
c8c818d7
|
2020-11-23T21:14:37
|
|
Fixed bug 5360 - non-libudev joystick detection doesn't see controllers that were already connected
Simon McVittie
When watching for hotplug events we can poll the inotify fd, but we
still need to scan /dev/input once per process, otherwise we'll fail
to detect devices that were already connected.
|
|
8e2746cf
|
2020-11-23T21:10:48
|
|
joystick: Don't use udev in Flatpak or pressure-vessel container
Flatpak[1] and pressure-vessel[2] are known to use user namespaces,
therefore udev event notification via netlink won't work reliably.
Both frameworks provide a filesystem API that libraries can use to
detect them. Do that, and automatically fall back from udev-based
device discovery to the inotify-based fallback introduced in Bug #5337.
[1] <https://flatpak.org/>
[2] <https://gitlab.steamos.cloud/steamrt/steam-runtime-tools/-/tree/master/pressure-vessel>
Signed-off-by: Simon McVittie <smcv@collabora.com>
|
|
|
e9869e07
|
2020-11-23T21:08:19
|
|
Fixed bug 5335 - enable joystick/haptic/evdev support by default on FreeBSD
Alex S
Evdev headers aren't actually included in the base system (well, it has a private copy), they are available through the devel/evdev-proto port instead. We also have devel/libinotify and devel/libudev-devd shims, I didn't verify whether they work with SDL.
|
|
|
38ab8bf7
|
2020-11-23T21:03:43
|
|
Fixed bug 5362 - Mac OS ARM doesn't build Metal/Vulkan back-end/renderers
C.W. Betts
As it is, SDL2's built-in config on macOS for Metal excludes Apple Silicon. This is due to thinking that the 64-bit Mac platform would always be x86_64. My patch fixes this by using the catch-all of 64-bit platforms.
|
|
|
62e39b5f
|
2020-11-23T20:57:14
|
|
Fixed building with an older SDK and macOS target
|
|
|
268aa456
|
2020-11-23T20:26:28
|
|
Fixed compiling with SDL_JOYSTICK_RAWINPUT disabled
|
|
eaa53a19
|
2020-11-23T22:16:07
|
|
joystick: On Linux, don't try to close an invalid inotify file descriptor.
|
|
|
5c957747
|
2020-11-23T22:14:22
|
|
joystick: Fix up Linux joystick code to (mostly) compile on FreeBSD.
|
|
|
179bd665
|
2020-11-23T18:27:43
|
|
Don't enable PS5 controller support if DS4Windows is running - they conflict.
|
|
|
5b3616c3
|
2020-11-23T18:24:05
|
|
Generalized the raw input controller driver and moved XInput/WGI detection into it for XInput devices
This fixes bad report parsing for various newer Xbox controllers, and this driver is now preferred over XInput, since it handles more than 4 controllers.
|
|
|
0ff5d55a
|
2020-11-23T21:07:28
|
|
x11: Don't try to grab the pointer on an unmapped window (thanks, Lee!)
Fixes Bugzilla #5352.
|
|
|
04b45b55
|
2020-11-23T14:28:30
|
|
Allow testing a specific controller
|
|
|
ff4bc138
|
2020-11-23T14:28:16
|
|
Fixed mapping controllers after adding the touchpad button
|
|
|
4c96faee
|
2020-11-23T20:37:10
|
|
remove non-existing tslib support from autofoo and cmake
|
|
|
a94fe009
|
2020-11-21T18:40:27
|
|
Moved comment with the code it covers
|
|
|
49f7be6d
|
2020-11-21T18:37:56
|
|
Fixed compile warning C4127: conditional expression is constant
|
|
|
63b52e8e
|
2020-11-21T18:34:32
|
|
Remove SDL dependency in Windows code
|
|
|
d4954eb0
|
2020-11-21T18:26:43
|
|
Removed debug code
|