kmx git

Commit	Date	Message
3cfd907d	2019-01-11T15:33:02	Android: Audio thread is already setup for the JVM In 'src/thread/pthread/SDL_systhread.c' RunThread() calls first 'Android_JNI_SetupThread()'
9d82f4e9	2019-01-11T15:27:53	Android: use pthread_once for creating thread key 'mThreadKey'
9a98e5af	2019-01-11T14:50:43	Android: don't call Android_JNI_ThreadDestroyed() for Java SDLThread SDLThread is a Java Thread, it's not needed to call 'Detach' from the JVM. Clear mThreadKey, so that the pthread_create destructor is not called for this thread.
b44a7aea	2019-01-10T21:49:00	Android: fix prototype of Android_JNI_InitTouch
7a1d1bae	2019-01-10T21:40:57	Android: add name for Touch devices and simplification, from bug 3958
d23c2f07	2019-01-10T18:05:56	Fixed bug 3930 - Android, set thread priorities and names SDLActivity thread priority is unchanged, by default -10 (THREAD_PRIORITY_VIDEO). SDLAudio thread priority was -4 (SDL_SetThreadPriority was ignored) and is now -16 (THREAD_PRIORITY_AUDIO). SDLThread thread priority was 0 (THREAD_PRIORITY_DEFAULT) and is -4 (THREAD_PRIORITY_DISPLAY).
66fbfe1d	2019-01-10T15:43:07	Android: nativeQuit for SDLActivity thread - destroy Android_ActivityMutex - display any SDL error message that may have occured in this thread, since SDL_GetError() is thread specific, and user has no access to it.
dad81611	2019-01-10T15:35:46	Android: only send Quit event to SDLThread if it's not already terminated And it avoids reporting errors using Android_Pause/ResumeSem that are NULL.
8dd91550	2019-01-09T23:19:26	Android: prevent a dummy error message sending SDL_DISPLAYEVENT_ORIENTATION In the usual case, first call to onNativeOrientationChanged() is done before SDL has been initialised and would just set an error message "Video subsystem has not been initialized" without sending the event.
02f292eb	2019-01-09T22:49:49	Android: add some SetError for Android_SetWindowFullscreen First error could happen if Android_SetWindowFullscreen somehow gets called between SurfaceDestroyed() and SurfaceCreated() Second error should not happen has native_window validity is guaranteed. (It would happens previously with error -19)
68c0e69f	2019-01-09T22:41:52	Android: native_window validity is guaranteed between surfaceCreated and Destroyed It's currently still available after surfaceDestroyed(). And available (but invalid) between surfaceCreated() and surfaceChanged(). Which means ANativewindow_getWidth/Height/Format() fail in those cases. https://developer.android.com/reference/android/view/SurfaceHolder.html#getSurface()
1803944b	2019-01-09T15:18:41	Android: concurrency issue for Android_SetWindowFullscreen() It accesses data->native_window, which can be changed by onNativeSurfacedChanged(). Currently, Android_SetWindowFullscreen() may access data->native_window after it has been released, and before a new reference is acquired. (can be reproduced by adding some SDL_Delay() in onNativeSurfacedChanged and Android_SetWindowFullscreen() ).
cfe2924d	2019-01-07T11:35:31	Android: some robustness when quitting application from onDestroy() Make sure there is not pending Pause accumulated, so the the application doesn't remain paused and stucked in onDestroy(). Can be tested by adding: SDLActivity.nativePause(); SDLActivity.nativePause(); mSingleton.finish();
462e62e1	2019-01-06T20:25:54	Android: better fix for bug 3186. Run those commands from SDL thread.
9f23d181	2019-01-06T17:35:42	Android: allow multiple calls to nativeResume() Doesn't seem to happen manually, but symetrical to the pause handling. Can be tested by adding: mNextNativeState = SDLActivity.NativeState.PAUSED; handleNativeState(); mNextNativeState = SDLActivity.NativeState.RESUMED; handleNativeState(); mNextNativeState = SDLActivity.NativeState.PAUSED; handleNativeState(); mNextNativeState = SDLActivity.NativeState.RESUMED; handleNativeState(); Before, it ends in 'paused' state. Now, it ends in 'resumed' state.
e4f558a5	2019-01-05T22:46:52	Android: un-needed check of "isPausing" and minor typos
35722b64	2019-01-05T22:27:25	Android: fix wrong state after immediate sequence pause() / resume() / pause() It may happen to have the sequence pause()/resume()/pause(), before polling any events. Before, it ends in 'resumed' state because as the check is greedy. Now, always increase the Pause semaphore, and stop at each pause. It ends in 'paused' state. Related to bug 3250: set up a reconfiguration of SurfaceView holder. Turn the screen off manually before the app starts (repro rate is not 100%..)
d4c0f498	2019-01-04T22:09:38	Fixed bug 4255 - SDL_GetGlobalMouseState() returns incorrect Y on secondary display Julian Raschke I use an open Mac laptop with an additional external monitor. The coordinate spaces from SDL_GetGlobalMouseState() and SDL_GetWindowPosition() match on the primary display, but not on the secondary display. Cocoa window coordinates are vertically flipped in relation to the primary display: https://github.com/spurious/SDL-mirror/blob/release-2.0.8/src/video/cocoa/SDL_cocoawindow.m#L219-L222 However, Cocoa_GetGlobalMouseState inverts the cursor Y coordinate per-display: https://github.com/spurious/SDL-mirror/blob/release-2.0.8/src/video/cocoa/SDL_cocoamouse.m#L320-L323 Suggested fix: Replace the for-loop with this simpler calculation: x = (int) cocoaLocation.x; y = (int) (CGDisplayPixelsHigh(kCGDirectMainDisplay) - cocoaLocation.y);
5e13087b	2019-01-04T22:01:14	Updated copyright for 2019
2a885eb0	2019-01-04T23:39:27	Android: fixed immediate transition to pause and resume. "Pause" transition will add events: SDL_WINDOWEVENT_ENTER SDL_WINDOWEVENT_FOCUS_LOST SDL_WINDOWEVENT_MINIMIZED SDL_APP_WILL ENTER BACKGROUND SDL_APP_DID ENTER BACKGROUND "Resume" transition will add events: SDL_APP_WILL ENTER FOREGROUND SDL_APP_DID ENTER FOREGROUND SDL_WINDOWEVENT_FOCUS_GAINED SDL_WINDOWEVENT_RESTORED If Android application doesn't empty the event loop in between, it enters in "paused" state when SDL_WINDOWEVENT_RESTORED is fetched. See bug 3250 for pratical case.
cf122344	2019-01-04T23:11:21	Android: make Android_PumpEvents() more readable No behavior change in this commit.
ca184ac3	2019-01-03T23:22:50	Android: concurrency issue with egl_surface EGL_BAD_SURFACE - (bug 4142) Occurs when application goes to background: - Java activity is destroying SurfaceView holder and "egl_surface" (in onNativeSurfaceDestroyed()) - While native thread is in Android_GLES_SwapWindow(), prepared to call SDL_EGL_SwapBuffers() The error is "call to eglSwapBuffers failed, reporting an error of EGL_BAD_SURFACE" It an be reproduced easily by adding a SDL_Delay(100) at the begining of SDL_EGL_SwapBuffers(), and putting the application into background.
2e19343d	2019-01-03T20:18:29	Android: use Mutex instead of Semphore for bug 4142
23478642	2019-01-03T16:22:33	Android: prevent the error message from SDL_EGL_CreateSurface() to be masked.
cc8f1136	2019-01-03T14:18:06	Fixed bug 4142 - Concurrency issues in Android backend Use a semaphore to prevent concurrency issues between Java Activity and Native thread code, when using 'Android_Window'. (Eg. Java sending Touch events, while native code is destroying the main SDL_Window. )
d11f7615	2019-01-03T13:38:33	Android: minor preparation for bug 4142 (concurrency issues)
5dc25fef	2019-01-03T13:14:16	Android: preparation bug 4142, reduce usage of global variable Android_Window
a95f91bc	2019-01-02T18:06:33	Fixed bug 3250 - Wrong backbuffer pixel format on Android, keep getting RGB_565 Use the egl format to reconfigure java SurfaceView holder format. If there is a change, it triggers a surfaceDestroyed/Created/Change sequence.
03b0e1de	2018-12-30T22:44:25	Android: on rare occasion, prevent Android_JNI_GetNativeWindow() from crashing If Java getNativeSurface() returns null, then ANativeWindow_fromSurface() would crash().
1e22fc15	2018-12-30T15:41:28	Android: fixed comments and spaces
2a412eb9	2018-12-30T15:39:37	Fixed bug 3186 - Android SW keyboard not restored when app becomes foreground.
9cdd0dd9	2018-12-29T17:59:34	PSP renderer: use colors from 'draw' union (very likely, but un-tested)
ebd9efb3	2018-12-29T16:37:44	opengles 1: same fix as in bug #4433
87b76369	2018-12-29T16:34:50	opengles 1: use color from 'draw' union in SetDrawState()
dc344360	2018-12-21T20:53:31	render: Fix internal state getting out of sync when destroying a texture that was just rendered and then creating a new one, in the GL and GLES2 backends. Fixes bug #4433.
0a705901	2018-12-19T18:10:02	render: Prefer the Metal renderer over OpenGL. This is the best option for macOS and iOS, the only platforms with Metal. Pre-Metal versions of these platforms will fall back to OpenGL (ES), as appropriate. Huge thanks to Alexander Szpakowski, who worked incredibly hard to get the Metal renderer to such a high-quality state!
c7db6ade	2018-12-19T18:27:21	metal: Implement SDL_LockTexture for YUV formats.
ce8c716a	2018-12-18T14:23:05	metal: Implement SDL_LockTexture for non-YUV textures.
c0c8f2d7	2018-12-16T11:15:21	Gesture: remove warnings when ENABLE_DOLLAR is undefined.
13869f19	2018-12-16T01:03:17	cocoa: Implement OpenGL swap interval support with CVDisplayLink. Not only does this fix macOS 10.14 ("Mojave")'s broken NSOpenGLCPSwapInterval support, it also lets us implement "adaptive vsync" on macOS! CVDisplayLink is supported back to macOS 10.4 ("Tiger"), so we just use it universally without version checks and dump NSOpenGLCPSwapInterval, Mojave or not.
1ed60219	2018-12-15T16:21:24	Fixed bug 4426 - allows re-creation of software renderer Switching between renderers "software -> opengl -> opengles2 -> software" fails. "opengl -> opengles2" calls SDL_RecreateWindow() and frees "window->surface" without marking it as "surface_invalid".
18296927	2018-12-15T09:57:05	emscripten: SDL_PrivateJoystickAdded() wants an instance id, not device index.
e5476c65	2018-12-15T14:50:12	Fixed bug 4425 - promote to alpha format, palette surface with alpha values. SDL_CreateTextureFromSurface() forgets to choose a texture format with alpha for surfaces that have palettes with alpha values.
bd08d72d	2018-12-11T20:04:10	Fixed building with the 10.10 SDK
98829d37	2018-12-11T19:50:58	Handle both "Sony Interactive Intertainment" and "Sony Computer Entertainment" when ignoring motion sensors
1c9595b1	2018-12-08T11:06:40	Fixed bug 4415 - SDL menu bar is nonstandard on Mac foo.null I'm on macOS 10.14 and I think I'm using or around SDL 2.0.9. This is about the menu bar that SDL sets up which looks like: <App Name> <Window> <View> 1. View menu never proceeds after the Window menu in any Mac application (it is always before). 2. For SDL, the only purpose of the View menu is for a single fullscreen menu item, which is not justifiable enough to reserve space for a menu. The View menu should thus be removed, and the full screen menu item should be added at the end inside of Window's menu. See built in apps like Dictionary, Chess, App Store (on 10.14) that do this. 3. SDL should add a "Close" menu item to the Window's submenu, and it should be the first item. Its key equivalent should map to command w. Without this, you cannot close the game window via this shortcut, and you cannot close the app's About window via this shortcut. 4. Apps typically use "Enter Full Screen" or "Exit Full Screen" depending on context, not "Toggle Full Screen" which is less user friendly -- I personally care about this point the least.
70ce0f2e	2018-12-07T12:02:08	Added the hint SDL_HINT_GAMECONTROLLERCONFIG_FILE to specify a file to load at initialization containing SDL game controller mappings
0b6e24f7	2018-12-06T14:26:18	Linux Haptic: Fix periodic.magnitude value
898644d1	2018-12-06T09:09:05	Made it more clear that the values being compared are floats
f9192ab8	2018-12-06T10:39:33	wayland: Do not try to lock on an invalid pointer This happens if you try to lock the pointer and (caps & WL_SEAT_CAPABILITY_POINTER) is false Leading to input->pointer being NULL which ends up bringing the wayland client down (at lease on weston)
7468d1e0	2018-12-06T15:46:40	Fix warnings detected on Android build
39ec1699	2018-12-06T10:24:44	opengles2: fix prototype of glDeleteBuffers
252dc85e	2018-12-06T09:22:00	Fix warnings detected on Android build
a7563bcd	2018-12-05T19:03:15	joystick: Removed unused variable.
c7b71371	2018-12-05T17:53:38	Whoops, forgot to add a new source file. :/
6ed76ae1	2018-12-05T14:46:03	Fixed the ROCCAT Tyon mouse showing up as a joystick on Windows
fbead635	2018-12-05T16:55:59	joystick: Added controller config for IMS Passenger Control Unit Devices.
2878d4f8	2018-12-05T16:53:15	egl: Don't force X11 support when testing for EGL. Fixes building Wayland support on embedded systems without X11. (TODO: maybe move the EGL test out of the X11 tests at some point, too.)
1689e9f9	2018-12-05T16:51:22	linux: Move SDL_LinuxSetThreadPriority() elsewhere to fix build. Fixes Bugzilla #4393.
bd3ee07c	2018-12-05T16:49:38	wayland: Send SDL_TOUCH_MOUSEID mouse events for touches.
6259a726	2018-12-05T16:13:12	Warnings: fix a documentation warning and missing prototypes
db320e46	2018-12-04T14:21:29	Fixed the PS4 motion controls showing up as a separate game controller on Linux
09b46204	2018-12-04T16:50:31	Windows: NEON detection and intrinsic includes on Visual Studio Visual Studio doesn't define __ARM_ARCH nor _ARM_NEON, but _M_ARM and _M_ARM64, so SDL_HasNEON() was bypassed. PF_ARM_NEON_INSTRUCTIONS_AVAILABLE doesn't see to be defined (but still works when defined as 19).
aea7e56a	2018-12-04T12:34:45	android: use __ARM_NEON instead of __ARM_NEON__ to include <arm_neon.h> Only __ARM_NEON is defined with Android NDK and arm64-v8a Tested on ndk-r18, ndk-r13 and also Xcode. (Visual Studio needs a different fix). Fixes Bugzilla #4409.
cca9d24c	2018-12-03T09:26:05	direct3d: be more aggressive about resetting invalidated cached state. Fixes Bugzilla #4402.
939bf1c4	2018-12-03T02:06:17	render: fix some static analysis warnings.
33f78eb1	2018-12-03T01:58:23	direct3d: Make sure streaming textures update before being used for drawing. Fixes Bugzilla #4402.
b744108a	2018-12-02T21:57:33	Patched to compile on C89 compilers.
3c936150	2018-12-02T20:55:57	direct3d: Release and NULL out vertex buffers on reset. Otherwise they are irretrievably lost on window resize, etc, which makes rendering freeze and other disasters. Fixes Bugzilla #4358.
65a7c98c	2018-12-02T02:33:06	directfb: Updated render backend to new internal API. Totally untested, beyond it now compiles again. Probably needs some fixes. Fixes Bugzilla #4405.
583f61b2	2018-12-01T16:31:56	kmsdrm: Check for resources when validating KMSDRM device in check_modesetting. Fixes Bugzilla #4403.
b3ac87d5	2018-12-01T13:09:00	kmsdrm: uninitialized KMSDRM fixes * Search for valid encoder in connector when it doesn't have an encoder ID set. * Search for valid CRTC in resources and encoder's possible CRTCs when encoder doesn't have one set. * Select default mode if CRTC doesn't have a valid one. * Pick current_mode's W/H/Refresh Rate basing on current and valid CRTC mode, not the saved one.
3323b355	2018-12-01T12:19:11	android: use cpufeatures to support SDL_HasNEON() (thanks, Sylvain!). Fixes Bugzilla #4406.
fac40f8e	2018-12-01T11:14:20	Patched to compile on Linux with threads enabled. (whoops!)
c4bc59a5	2018-12-01T10:36:26	Patched to compile on Linux with --disable-threads. Fixes Bugzilla #4393.
f205f3a8	2018-11-27T15:10:26	Added support for the Razer Raiju Mobile
a95291c9	2018-11-27T11:16:04	Fixed the hotspot for cursors on Raspberry Pi
aea48357	2018-11-27T10:20:29	Fixed bug changing cursors on Raspberry Pi
37b1f989	2018-11-25T22:13:09	metal: use a staging texture in SDL_UpdateTexture, to make sure it doesn't stomp texture data being drawn in a previous frame on the GPU.
872936a4	2018-11-21T23:46:37	metal: Fix an incorrect division.
4a58722b	2018-11-21T23:37:23	metal: SDL_RenderFillRects uses one draw call per 16k rectangles (within the given FillRects call), instead of one draw call per rectangle. Reduces CPU usage when drawing many rectangles.
5c5ba0e3	2018-11-19T21:35:59	Fixed bug 4394 - Crash in SDL_PumpEvents() after SDL_DestroyWindow() Cameron Gutman After updating to SDL 2.0.9, I got a user report that my app was crashing when closing a SDL_WINDOW_FULLSCREEN window to return to my Qt-based UI. It looks like the dead SDL window is getting a spurious updateLayer call which is causing SDL to dereference a null SDL_WindowData pointer. For some reason, this only happens when using SDL_WINDOW_FULLSCREEN and not windowed or SDL_WINDOW_FULLSCREEN_DESKTOP. I was also unsuccessful in my attempt to get a simple reproducer for this crash. The Session.cpp code is available https://github.com/moonlight-stream/moonlight-qt/blob/688c4a90d994aa23e7b0af3ffcbb8707886db780/app/streaming/session.cpp but I slightly modified it (adding a SDL_PumpEvents() call at 1179 to immediately trigger the issue, otherwise it happened when Qt next pumped the event loop). The crashing line is: NSMutableArray *contexts = data->nscontexts;
b73703b9	2018-11-19T21:17:00	Fixed bug 4391 - hid_enumerate() sometimes causes game to freeze for a few seconds Daniel Gibson Even though my game (dhewm3) doesn't use SDL_INIT_JOYSTICK, SDL_PumpEvent() calls SDL_JoystickUpdate() which ends up calling hid_enumerate() every three seconds, and sometimes on my Win7 box hid_enumerate() takes about 5 seconds, which causes the whole game to freeze for that time.
d3fa42b8	2018-11-18T19:28:20	os/2 bits for SDL_malloc.c -- from libffi
1a02403e	2018-11-18T11:50:20	libm: Watcom defines huge=__huge: undefine it to fix build using Watcom.
457e58c4	2018-11-17T16:24:52	opengles: Fixed compiler warnings.
9719f89d	2018-11-17T12:12:29	Back out change initializing renderer blend mode incorrectly.
fde7592a	2018-11-17T14:39:42	direct3d11: Fixed missing rendering of solid primitives. Fixes Bugzilla #4388.
782f1685	2018-11-17T14:37:51	Fixed a few compiler warnings.
29e15ce6	2018-11-17T00:58:45	The default draw blendmode is SDL_BLENDMODE_NONE
1f6bd951	2018-11-15T18:22:30	Emscripten: make CloseAudio actually close audio cf. https://bugzilla.libsdl.org/show_bug.cgi?id=4176
9262c0de	2018-11-14T22:38:58	software: fix blits with color mods that change during a command queue run.
2e348c1f	2018-11-14T13:37:22	Fixed bug 3193 - Dualshock 3's motion sensors overwrite analog stick maxxus The Dualshock 3's motion sensors don't seem to be reported by the call to EVIOCGBIT but they still send EV_ABS events. Because they're not reported by EVIOCGBIT they're not assigned a proper axis ids and the default of 0 is used, which is the valid id for the left analog sticks left/right axis.
b815ad56	2018-11-12T16:42:49	Fixed bug 4366 - Compile throws a warning on RPI (Raspbian Stretch) midwan When trying to compile on a Raspberry Pi 3, running Raspbian Stretch (fully updated), a warning appears: /home/pi/projects/SDL/src/test/SDL_test_memory.c: In function ?SDL_TrackAllocation?: /home/pi/projects/SDL/src/test/SDL_test_memory.c:112:109: warning: format ?%llx? expects argument of type ?long long unsigned int?, but argument 5 has type ?unw_word_t {aka unsigned int}? [-Wformat=] snprintf(entry->stack_names[stack_index], sizeof(entry->stack_names[stack_index]), "%s+0x%llx", sym, offset);
c525ff35	2018-11-10T20:56:23	cocoa: fix building with the macOS 10.7 SDK (thanks Riccardo!) Fixes bug #4368
5029d50e	2018-11-10T16:15:48	Add SDL_TouchDeviceType enum and SDL_GetTouchDeviceType(SDL_TouchID id). Touch device types include SDL_TOUCH_DEVICE_DIRECT (a touch screen with window-relative coordinates for touches), SDL_TOUCH_DEVICE_INDIRECT_ABSOLUTE (a trackpad-style device with absolute device coordinates), and SDL_TOUCH_DEVICE_INDIRECT_RELATIVE (a trackpad-style device with screen cursor-relative coordinates). Phone screens are an example of a direct device type. Mac trackpads are the indirect-absolute touch device type. The Apple TV remote is an indirect-relative touch device type.
9665a508	2018-11-06T16:57:07	Added Vulkan headers version 1.1.91 Downloaded from https://github.com/KhronosGroup/Vulkan-Headers
48917e0e	2018-11-07T01:08:35	wayland: fix resizing and fullscreen toggling For starters, we need to correctly respond to 0,0 configure after unsetting fullscreen. Also, turns out that there should be no drawing calls at all in between eglSwapBuffers and wl_egl_window_resize, as otherwise EGL can already allocate a wrongly sized buffer for a next frame, so handle those together.
5f980514	2018-11-04T21:08:40	wayland: ask xdg-decoration protocol extension to use server-side decorations if possible.
244b79e1	2018-11-04T14:31:56	metal: SDL_RenderReadPixels on macOS synchronizes the render target's texture data if it's managed, before reading from it.

3cfd907d

2019-01-11T15:33:02

Android: Audio thread is already setup for the JVM In 'src/thread/pthread/SDL_systhread.c' RunThread() calls first 'Android_JNI_SetupThread()'

9d82f4e9

2019-01-11T15:27:53

Android: use pthread_once for creating thread key 'mThreadKey'

9a98e5af

2019-01-11T14:50:43

Android: don't call Android_JNI_ThreadDestroyed() for Java SDLThread SDLThread is a Java Thread, it's not needed to call 'Detach' from the JVM. Clear mThreadKey, so that the pthread_create destructor is not called for this thread.

b44a7aea

2019-01-10T21:49:00

Android: fix prototype of Android_JNI_InitTouch

7a1d1bae

2019-01-10T21:40:57

Android: add name for Touch devices and simplification, from bug 3958

d23c2f07

2019-01-10T18:05:56

Fixed bug 3930 - Android, set thread priorities and names SDLActivity thread priority is unchanged, by default -10 (THREAD_PRIORITY_VIDEO). SDLAudio thread priority was -4 (SDL_SetThreadPriority was ignored) and is now -16 (THREAD_PRIORITY_AUDIO). SDLThread thread priority was 0 (THREAD_PRIORITY_DEFAULT) and is -4 (THREAD_PRIORITY_DISPLAY).

66fbfe1d

2019-01-10T15:43:07

Android: nativeQuit for SDLActivity thread - destroy Android_ActivityMutex - display any SDL error message that may have occured in this thread, since SDL_GetError() is thread specific, and user has no access to it.

dad81611

2019-01-10T15:35:46

Android: only send Quit event to SDLThread if it's not already terminated And it avoids reporting errors using Android_Pause/ResumeSem that are NULL.

8dd91550

2019-01-09T23:19:26

Android: prevent a dummy error message sending SDL_DISPLAYEVENT_ORIENTATION In the usual case, first call to onNativeOrientationChanged() is done before SDL has been initialised and would just set an error message "Video subsystem has not been initialized" without sending the event.

02f292eb

2019-01-09T22:49:49

Android: add some SetError for Android_SetWindowFullscreen First error could happen if Android_SetWindowFullscreen somehow gets called between SurfaceDestroyed() and SurfaceCreated() Second error should not happen has native_window validity is guaranteed. (It would happens previously with error -19)

68c0e69f

2019-01-09T22:41:52

Android: native_window validity is guaranteed between surfaceCreated and Destroyed It's currently still available after surfaceDestroyed(). And available (but invalid) between surfaceCreated() and surfaceChanged(). Which means ANativewindow_getWidth/Height/Format() fail in those cases. https://developer.android.com/reference/android/view/SurfaceHolder.html#getSurface()

1803944b

2019-01-09T15:18:41

Android: concurrency issue for Android_SetWindowFullscreen() It accesses data->native_window, which can be changed by onNativeSurfacedChanged(). Currently, Android_SetWindowFullscreen() may access data->native_window after it has been released, and before a new reference is acquired. (can be reproduced by adding some SDL_Delay() in onNativeSurfacedChanged and Android_SetWindowFullscreen() ).

cfe2924d

2019-01-07T11:35:31

Android: some robustness when quitting application from onDestroy() Make sure there is not pending Pause accumulated, so the the application doesn't remain paused and stucked in onDestroy(). Can be tested by adding: SDLActivity.nativePause(); SDLActivity.nativePause(); mSingleton.finish();

462e62e1

2019-01-06T20:25:54

Android: better fix for bug 3186. Run those commands from SDL thread.

9f23d181

2019-01-06T17:35:42

Android: allow multiple calls to nativeResume() Doesn't seem to happen manually, but symetrical to the pause handling. Can be tested by adding: mNextNativeState = SDLActivity.NativeState.PAUSED; handleNativeState(); mNextNativeState = SDLActivity.NativeState.RESUMED; handleNativeState(); mNextNativeState = SDLActivity.NativeState.PAUSED; handleNativeState(); mNextNativeState = SDLActivity.NativeState.RESUMED; handleNativeState(); Before, it ends in 'paused' state. Now, it ends in 'resumed' state.

e4f558a5

2019-01-05T22:46:52

Android: un-needed check of "isPausing" and minor typos

35722b64

2019-01-05T22:27:25

Android: fix wrong state after immediate sequence pause() / resume() / pause() It may happen to have the sequence pause()/resume()/pause(), before polling any events. Before, it ends in 'resumed' state because as the check is greedy. Now, always increase the Pause semaphore, and stop at each pause. It ends in 'paused' state. Related to bug 3250: set up a reconfiguration of SurfaceView holder. Turn the screen off manually before the app starts (repro rate is not 100%..)

d4c0f498

2019-01-04T22:09:38

Fixed bug 4255 - SDL_GetGlobalMouseState() returns incorrect Y on secondary display Julian Raschke I use an open Mac laptop with an additional external monitor. The coordinate spaces from SDL_GetGlobalMouseState() and SDL_GetWindowPosition() match on the primary display, but not on the secondary display. Cocoa window coordinates are vertically flipped in relation to the primary display: https://github.com/spurious/SDL-mirror/blob/release-2.0.8/src/video/cocoa/SDL_cocoawindow.m#L219-L222 However, Cocoa_GetGlobalMouseState inverts the cursor Y coordinate per-display: https://github.com/spurious/SDL-mirror/blob/release-2.0.8/src/video/cocoa/SDL_cocoamouse.m#L320-L323 Suggested fix: Replace the for-loop with this simpler calculation: *x = (int) cocoaLocation.x; *y = (int) (CGDisplayPixelsHigh(kCGDirectMainDisplay) - cocoaLocation.y);

5e13087b

2019-01-04T22:01:14

Updated copyright for 2019

2a885eb0

2019-01-04T23:39:27

Android: fixed immediate transition to pause and resume. "Pause" transition will add events: SDL_WINDOWEVENT_ENTER SDL_WINDOWEVENT_FOCUS_LOST SDL_WINDOWEVENT_MINIMIZED SDL_APP_WILL ENTER BACKGROUND SDL_APP_DID ENTER BACKGROUND "Resume" transition will add events: SDL_APP_WILL ENTER FOREGROUND SDL_APP_DID ENTER FOREGROUND SDL_WINDOWEVENT_FOCUS_GAINED SDL_WINDOWEVENT_RESTORED If Android application doesn't empty the event loop in between, it enters in "paused" state when SDL_WINDOWEVENT_RESTORED is fetched. See bug 3250 for pratical case.

cf122344

2019-01-04T23:11:21

Android: make Android_PumpEvents() more readable No behavior change in this commit.

ca184ac3

2019-01-03T23:22:50

Android: concurrency issue with egl_surface EGL_BAD_SURFACE - (bug 4142) Occurs when application goes to background: - Java activity is destroying SurfaceView holder and "egl_surface" (in onNativeSurfaceDestroyed()) - While native thread is in Android_GLES_SwapWindow(), prepared to call SDL_EGL_SwapBuffers() The error is "call to eglSwapBuffers failed, reporting an error of EGL_BAD_SURFACE" It an be reproduced easily by adding a SDL_Delay(100) at the begining of SDL_EGL_SwapBuffers(), and putting the application into background.

2e19343d

2019-01-03T20:18:29

Android: use Mutex instead of Semphore for bug 4142

23478642

2019-01-03T16:22:33

Android: prevent the error message from SDL_EGL_CreateSurface() to be masked.

cc8f1136

2019-01-03T14:18:06

Fixed bug 4142 - Concurrency issues in Android backend Use a semaphore to prevent concurrency issues between Java Activity and Native thread code, when using 'Android_Window'. (Eg. Java sending Touch events, while native code is destroying the main SDL_Window. )

d11f7615

2019-01-03T13:38:33

Android: minor preparation for bug 4142 (concurrency issues)

5dc25fef

2019-01-03T13:14:16

Android: preparation bug 4142, reduce usage of global variable Android_Window

a95f91bc

2019-01-02T18:06:33

Fixed bug 3250 - Wrong backbuffer pixel format on Android, keep getting RGB_565 Use the egl format to reconfigure java SurfaceView holder format. If there is a change, it triggers a surfaceDestroyed/Created/Change sequence.

03b0e1de

2018-12-30T22:44:25

Android: on rare occasion, prevent Android_JNI_GetNativeWindow() from crashing If Java getNativeSurface() returns null, then ANativeWindow_fromSurface() would crash().

1e22fc15

2018-12-30T15:41:28

Android: fixed comments and spaces

2a412eb9

2018-12-30T15:39:37

Fixed bug 3186 - Android SW keyboard not restored when app becomes foreground.

9cdd0dd9

2018-12-29T17:59:34

PSP renderer: use colors from 'draw' union (very likely, but un-tested)

ebd9efb3

2018-12-29T16:37:44

opengles 1: same fix as in bug #4433

87b76369

2018-12-29T16:34:50

opengles 1: use color from 'draw' union in SetDrawState()

dc344360

2018-12-21T20:53:31

render: Fix internal state getting out of sync when destroying a texture that was just rendered and then creating a new one, in the GL and GLES2 backends. Fixes bug #4433.

0a705901

2018-12-19T18:10:02

render: Prefer the Metal renderer over OpenGL. This is the best option for macOS and iOS, the only platforms with Metal. Pre-Metal versions of these platforms will fall back to OpenGL (ES), as appropriate. Huge thanks to Alexander Szpakowski, who worked incredibly hard to get the Metal renderer to such a high-quality state!

c7db6ade

2018-12-19T18:27:21

metal: Implement SDL_LockTexture for YUV formats.

ce8c716a

2018-12-18T14:23:05

metal: Implement SDL_LockTexture for non-YUV textures.

c0c8f2d7

2018-12-16T11:15:21

Gesture: remove warnings when ENABLE_DOLLAR is undefined.

13869f19

2018-12-16T01:03:17

cocoa: Implement OpenGL swap interval support with CVDisplayLink. Not only does this fix macOS 10.14 ("Mojave")'s broken NSOpenGLCPSwapInterval support, it also lets us implement "adaptive vsync" on macOS! CVDisplayLink is supported back to macOS 10.4 ("Tiger"), so we just use it universally without version checks and dump NSOpenGLCPSwapInterval, Mojave or not.

1ed60219

2018-12-15T16:21:24

Fixed bug 4426 - allows re-creation of software renderer Switching between renderers "software -> opengl -> opengles2 -> software" fails. "opengl -> opengles2" calls SDL_RecreateWindow() and frees "window->surface" without marking it as "surface_invalid".

18296927

2018-12-15T09:57:05

emscripten: SDL_PrivateJoystickAdded() wants an instance id, not device index.

e5476c65

2018-12-15T14:50:12

Fixed bug 4425 - promote to alpha format, palette surface with alpha values. SDL_CreateTextureFromSurface() forgets to choose a texture format with alpha for surfaces that have palettes with alpha values.

bd08d72d

2018-12-11T20:04:10

Fixed building with the 10.10 SDK

98829d37

2018-12-11T19:50:58

Handle both "Sony Interactive Intertainment" and "Sony Computer Entertainment" when ignoring motion sensors

1c9595b1

2018-12-08T11:06:40

Fixed bug 4415 - SDL menu bar is nonstandard on Mac foo.null I'm on macOS 10.14 and I think I'm using or around SDL 2.0.9. This is about the menu bar that SDL sets up which looks like: <App Name> <Window> <View> 1. View menu never proceeds after the Window menu in any Mac application (it is always before). 2. For SDL, the only purpose of the View menu is for a single fullscreen menu item, which is not justifiable enough to reserve space for a menu. The View menu should thus be removed, and the full screen menu item should be added at the end inside of Window's menu. See built in apps like Dictionary, Chess, App Store (on 10.14) that do this. 3. SDL should add a "Close" menu item to the Window's submenu, and it should be the first item. Its key equivalent should map to command w. Without this, you cannot close the game window via this shortcut, and you cannot close the app's About window via this shortcut. 4. Apps typically use "Enter Full Screen" or "Exit Full Screen" depending on context, not "Toggle Full Screen" which is less user friendly -- I personally care about this point the least.

70ce0f2e

2018-12-07T12:02:08

Added the hint SDL_HINT_GAMECONTROLLERCONFIG_FILE to specify a file to load at initialization containing SDL game controller mappings

0b6e24f7

2018-12-06T14:26:18

Linux Haptic: Fix periodic.magnitude value

898644d1

2018-12-06T09:09:05

Made it more clear that the values being compared are floats

f9192ab8

2018-12-06T10:39:33

wayland: Do not try to lock on an invalid pointer This happens if you try to lock the pointer and (caps & WL_SEAT_CAPABILITY_POINTER) is false Leading to input->pointer being NULL which ends up bringing the wayland client down (at lease on weston)

7468d1e0

2018-12-06T15:46:40

Fix warnings detected on Android build

39ec1699

2018-12-06T10:24:44

opengles2: fix prototype of glDeleteBuffers

252dc85e

2018-12-06T09:22:00

Fix warnings detected on Android build

a7563bcd

2018-12-05T19:03:15

joystick: Removed unused variable.

c7b71371

2018-12-05T17:53:38

Whoops, forgot to add a new source file. :/

6ed76ae1

2018-12-05T14:46:03

Fixed the ROCCAT Tyon mouse showing up as a joystick on Windows

fbead635

2018-12-05T16:55:59

joystick: Added controller config for IMS Passenger Control Unit Devices.

2878d4f8

2018-12-05T16:53:15

egl: Don't force X11 support when testing for EGL. Fixes building Wayland support on embedded systems without X11. (TODO: maybe move the EGL test out of the X11 tests at some point, too.)

1689e9f9

2018-12-05T16:51:22

linux: Move SDL_LinuxSetThreadPriority() elsewhere to fix build. Fixes Bugzilla #4393.

bd3ee07c

2018-12-05T16:49:38

wayland: Send SDL_TOUCH_MOUSEID mouse events for touches.

6259a726

2018-12-05T16:13:12

Warnings: fix a documentation warning and missing prototypes

db320e46

2018-12-04T14:21:29

Fixed the PS4 motion controls showing up as a separate game controller on Linux

09b46204

2018-12-04T16:50:31

Windows: NEON detection and intrinsic includes on Visual Studio Visual Studio doesn't define __ARM_ARCH nor _ARM_NEON, but _M_ARM and _M_ARM64, so SDL_HasNEON() was bypassed. PF_ARM_NEON_INSTRUCTIONS_AVAILABLE doesn't see to be defined (but still works when defined as 19).

aea7e56a

2018-12-04T12:34:45

android: use __ARM_NEON instead of __ARM_NEON__ to include <arm_neon.h> Only __ARM_NEON is defined with Android NDK and arm64-v8a Tested on ndk-r18, ndk-r13 and also Xcode. (Visual Studio needs a different fix). Fixes Bugzilla #4409.

cca9d24c

2018-12-03T09:26:05

direct3d: be more aggressive about resetting invalidated cached state. Fixes Bugzilla #4402.

939bf1c4

2018-12-03T02:06:17

render: fix some static analysis warnings.

33f78eb1

2018-12-03T01:58:23

direct3d: Make sure streaming textures update before being used for drawing. Fixes Bugzilla #4402.

b744108a

2018-12-02T21:57:33

Patched to compile on C89 compilers.

3c936150

2018-12-02T20:55:57

direct3d: Release and NULL out vertex buffers on reset. Otherwise they are irretrievably lost on window resize, etc, which makes rendering freeze and other disasters. Fixes Bugzilla #4358.

65a7c98c

2018-12-02T02:33:06

directfb: Updated render backend to new internal API. Totally untested, beyond it now compiles again. Probably needs some fixes. Fixes Bugzilla #4405.

583f61b2

2018-12-01T16:31:56

kmsdrm: Check for resources when validating KMSDRM device in check_modesetting. Fixes Bugzilla #4403.

b3ac87d5

2018-12-01T13:09:00

kmsdrm: uninitialized KMSDRM fixes * Search for valid encoder in connector when it doesn't have an encoder ID set. * Search for valid CRTC in resources and encoder's possible CRTCs when encoder doesn't have one set. * Select default mode if CRTC doesn't have a valid one. * Pick current_mode's W/H/Refresh Rate basing on current and valid CRTC mode, not the saved one.

3323b355

2018-12-01T12:19:11

android: use cpufeatures to support SDL_HasNEON() (thanks, Sylvain!). Fixes Bugzilla #4406.

fac40f8e

2018-12-01T11:14:20

Patched to compile on Linux with threads enabled. (whoops!)

c4bc59a5

2018-12-01T10:36:26

Patched to compile on Linux with --disable-threads. Fixes Bugzilla #4393.

f205f3a8

2018-11-27T15:10:26

Added support for the Razer Raiju Mobile

a95291c9

2018-11-27T11:16:04

Fixed the hotspot for cursors on Raspberry Pi

aea48357

2018-11-27T10:20:29

Fixed bug changing cursors on Raspberry Pi

37b1f989

2018-11-25T22:13:09

metal: use a staging texture in SDL_UpdateTexture, to make sure it doesn't stomp texture data being drawn in a previous frame on the GPU.

872936a4

2018-11-21T23:46:37

metal: Fix an incorrect division.

4a58722b

2018-11-21T23:37:23

metal: SDL_RenderFillRects uses one draw call per 16k rectangles (within the given FillRects call), instead of one draw call per rectangle. Reduces CPU usage when drawing many rectangles.

5c5ba0e3

2018-11-19T21:35:59

Fixed bug 4394 - Crash in SDL_PumpEvents() after SDL_DestroyWindow() Cameron Gutman After updating to SDL 2.0.9, I got a user report that my app was crashing when closing a SDL_WINDOW_FULLSCREEN window to return to my Qt-based UI. It looks like the dead SDL window is getting a spurious updateLayer call which is causing SDL to dereference a null SDL_WindowData pointer. For some reason, this only happens when using SDL_WINDOW_FULLSCREEN and not windowed or SDL_WINDOW_FULLSCREEN_DESKTOP. I was also unsuccessful in my attempt to get a simple reproducer for this crash. The Session.cpp code is available https://github.com/moonlight-stream/moonlight-qt/blob/688c4a90d994aa23e7b0af3ffcbb8707886db780/app/streaming/session.cpp but I slightly modified it (adding a SDL_PumpEvents() call at 1179 to immediately trigger the issue, otherwise it happened when Qt next pumped the event loop). The crashing line is: NSMutableArray *contexts = data->nscontexts;

b73703b9

2018-11-19T21:17:00

Fixed bug 4391 - hid_enumerate() sometimes causes game to freeze for a few seconds Daniel Gibson Even though my game (dhewm3) doesn't use SDL_INIT_JOYSTICK, SDL_PumpEvent() calls SDL_JoystickUpdate() which ends up calling hid_enumerate() every three seconds, and sometimes on my Win7 box hid_enumerate() takes about 5 seconds, which causes the whole game to freeze for that time.

d3fa42b8

2018-11-18T19:28:20

os/2 bits for SDL_malloc.c -- from libffi

1a02403e

2018-11-18T11:50:20

libm: Watcom defines huge=__huge: undefine it to fix build using Watcom.

457e58c4

2018-11-17T16:24:52

opengles: Fixed compiler warnings.

9719f89d

2018-11-17T12:12:29

Back out change initializing renderer blend mode incorrectly.

fde7592a

2018-11-17T14:39:42

direct3d11: Fixed missing rendering of solid primitives. Fixes Bugzilla #4388.

782f1685

2018-11-17T14:37:51

Fixed a few compiler warnings.

29e15ce6

2018-11-17T00:58:45

The default draw blendmode is SDL_BLENDMODE_NONE

1f6bd951

2018-11-15T18:22:30

Emscripten: make CloseAudio actually close audio cf. https://bugzilla.libsdl.org/show_bug.cgi?id=4176

9262c0de

2018-11-14T22:38:58

software: fix blits with color mods that change during a command queue run.

2e348c1f

2018-11-14T13:37:22

Fixed bug 3193 - Dualshock 3's motion sensors overwrite analog stick maxxus The Dualshock 3's motion sensors don't seem to be reported by the call to EVIOCGBIT but they still send EV_ABS events. Because they're not reported by EVIOCGBIT they're not assigned a proper axis ids and the default of 0 is used, which is the valid id for the left analog sticks left/right axis.

b815ad56

2018-11-12T16:42:49

Fixed bug 4366 - Compile throws a warning on RPI (Raspbian Stretch) midwan When trying to compile on a Raspberry Pi 3, running Raspbian Stretch (fully updated), a warning appears: /home/pi/projects/SDL/src/test/SDL_test_memory.c: In function ?SDL_TrackAllocation?: /home/pi/projects/SDL/src/test/SDL_test_memory.c:112:109: warning: format ?%llx? expects argument of type ?long long unsigned int?, but argument 5 has type ?unw_word_t {aka unsigned int}? [-Wformat=] snprintf(entry->stack_names[stack_index], sizeof(entry->stack_names[stack_index]), "%s+0x%llx", sym, offset);

c525ff35

2018-11-10T20:56:23

cocoa: fix building with the macOS 10.7 SDK (thanks Riccardo!) Fixes bug #4368

5029d50e

2018-11-10T16:15:48

Add SDL_TouchDeviceType enum and SDL_GetTouchDeviceType(SDL_TouchID id). Touch device types include SDL_TOUCH_DEVICE_DIRECT (a touch screen with window-relative coordinates for touches), SDL_TOUCH_DEVICE_INDIRECT_ABSOLUTE (a trackpad-style device with absolute device coordinates), and SDL_TOUCH_DEVICE_INDIRECT_RELATIVE (a trackpad-style device with screen cursor-relative coordinates). Phone screens are an example of a direct device type. Mac trackpads are the indirect-absolute touch device type. The Apple TV remote is an indirect-relative touch device type.

9665a508

2018-11-06T16:57:07

Added Vulkan headers version 1.1.91 Downloaded from https://github.com/KhronosGroup/Vulkan-Headers

48917e0e

2018-11-07T01:08:35

wayland: fix resizing and fullscreen toggling For starters, we need to correctly respond to 0,0 configure after unsetting fullscreen. Also, turns out that there should be no drawing calls at all in between eglSwapBuffers and wl_egl_window_resize, as otherwise EGL can already allocate a wrongly sized buffer for a next frame, so handle those together.

5f980514

2018-11-04T21:08:40

wayland: ask xdg-decoration protocol extension to use server-side decorations if possible.

244b79e1

2018-11-04T14:31:56

metal: SDL_RenderReadPixels on macOS synchronizes the render target's texture data if it's managed, before reading from it.

kc3-lang/SDL/src

src

Log