Merge audio capture work back into the mainline.

Patched to compile.

alsa: Implemented hotplug support, cleaned up device names.

ios: patched to compile.

alsa: don't enumerate virtual devices, just physical hardware.

Added SDL_PrivateJoystickAdded() and SDL_PrivateJoystickRemoved() Updated the removal code to iterate over all joystick add messages instead of just the first one.

commit 1170112da3776fdb06425f62d57b63144c33dc51 Author: James Zipperer <james.zipperer@synapse.com> Date: Sun Aug 21 01:19:19 2016 -0700 bugfix for controller / joystick add / remove being in the event queue at the same time

Emscripten: Fixed crash if closing removed joystick (thanks, Sylvain!).

Android: Fixed crash if closing removed joystick (thanks, Sylvain!). Fixes fix for Bugzilla #3408.

WinRT: fixed bug where Win10 GameBar, when shown + hidden, might not restore a cursor's hidden state The repro steps were this: 1. run an sdl2 winrt/uwp app, on Win10, v10.0.10586.0 or higher 2. hide the cursor, via a call to SDL_ShowCursor(0) 3. make the Win10 game bar appear, by pressing the Windows + G hotkey 4. observe that the mouse cursor appears, in order to interact with the game bar (this is expected behavior) 5. make the Win10 game bar disappear, either by pressing the Windows + G hotkey again, or clicking somewhere in the app EXPECTED RESULT: cursor disappears, as game bar disappears ACTUAL RESULT: cursor didn't always disappear

Added link in header comment.

Emscripten: Fixed opening previously closed joystick.

Android: Fixed opening previously closed joystick (thanks, Sylvain!). Fixes Bugzilla #3408.

[ OSX ] enables Xbox One S Guide (System Main Menu) button detection. the one possible impact is that since button identifiers on OSX are just their sequence in the HID report descriptor we might change the button order, and any existing/saved mappings may get their order changed.

Add Xbox One controller GUIDs to the XInput filter in the DirectInput joystick driver. The Windows 10 Anniversary Update (1607) breaks the method uses that SDL uses to detect XInput devices. That is, on Windows 10 Anniversary Update, it is no longer possible to query RAWINPUT for HID devices, and check for "IG_" in the device name. Presumably, this will be fixed in the future. This patch works around the issue by adding the Xbox One controller series to the well-known device list. This skips the more expensive RAWINPUT check for those devices, and causes them to be detected as XInput devices once again.

Fixed two old identifiers in header comments.

PSP: Fixed memory leak if video init failed.

WinRT: Removed not needed SDL_zerop() after SDL_calloc(). SDL_calloc() already sets memory to zero.

Added brackets to function names in header comments so that doxygen links them.

testaudiocapture: ask for way more output samples. Fixes Emscripten builds on Chrome for Android.

emscripten: send fake mouse events for touches, like other targets do. (This really should be handled at the higher level and not in the individual targets, but this fixes the immediate bug.)

emscripten audio: check for an "undefined" object, remove some console.log().

audio: Cleaned up "extern AudioBootStrap" list.

audio: Renamed some internal driver symbols in various targets.

android: implement audio capture support.

directsound: Implemented audio capture support.

windows: directsound should also map audio device GUIDs to proper names. Moved this code from winmm into core so both can use it. DirectSound (at least on Win10) also returns truncated device names, even though it's handed in as a string pointer and not a static-sized buffer. :/

audio: Patched to compile on C89 compilers.

emscripten audio: fix timer on capture's silence callback.

directsound: recalculate audiospec size before creating secondary buffer. I think this was a bug before? Maybe I'm misunderstanding this, but it looks like it was working because we allocate room for 8 chunks...

emscripten audio: Added audio capture support.

Added a FIXME.

testaudiocapture: open capture device to same spec as output device. ...since our resampler is still terrible (sorry!).

audio: Only allocate fake_stream if we're using the standard audio threads.

audio: simplifed check for internal callback. Easier to check when it's NULL instead of a list of known internal functions.

dsp: Implemented audio capture support.

winmm: Try to get full device names from the Windows Registry.

winmm: Added a FIXME for truncated device names.

winmm: Implemented audio capture support.

disk audio: Make default i/o delay match what device is meant to be running at.

disk audio: Implemented "capture" support, cleaned up some things.

testaudiocapture: use capture device buffer queueing, for better test coverage.

audio: Implemented buffer queueing for capture devices (SDL_DequeueAudio()).

testaudiocapture: don't use fullscreen for the window.

audio: SDL_ClearQueuedAudio() should free everything but two packets. Otherwise, if you had a massive, one-time queue buildup, the memory from that remains allocated until you close the device. Also, if you are just using a reasonable amount of space, this would previously cause you to reallocate it over and over instead of keeping a little bit of memory around.

nas: initial shot at audio capture support. Compiles, but not tested.

audio: Changed OnlyHasDefaultInputDevice to OnlyHasDefaultCaptureDevice.

audio: changed some SDL_memset() calls to SDL_zero(), other minor corrections.

audio: Removed internal SDL_audiomem.h and macros. I think this was important for SDL 1.2 because some targets needed special device memory for DMA buffers or locked memory buffers for use in hardware interrupts or something, but since it just defines to SDL_malloc and SDL_free now, I took it out for clarity's sake.

audio: Clean up some CloseDevice() interface details. - It's now always called if device->hidden isn't NULL, even if OpenDevice() failed halfway through. This lets implementation code not have to clean up itself on every possible failure point; just return an error and SDL will handle it for you. - Implementations can assume this->hidden != NULL and not check for it. - implementations don't have to set this->hidden = NULL when done, because the caller is always about to free(this). - Don't reset other fields that are in a block of memory about to be free()'d. - Implementations all now free things like internal mix buffers last, after closing devices and such, to guarantee they definitely aren't in use anymore at the point of deallocation.

arts: backed out audio capture support. Turns out that libartsc isn't thread-safe, so if we run a capture and playback device at the same time, it often crashes in arts's internal event loop. We could throw mutexes around the read/write calls, but these are meant to block, so one device could cause serious latency and stutter in the other. Since this audio target isn't in high-demand (Ubuntu hasn't offered a libartsc package for years), I'm just backing out the capture support. If someone needs it, they can pull it out of the revision history.

dummy audio: Patched to compile.

dummy audio: Implemented dummy audio capture support. :)

bsdaudio: first shot at audio capture support! (untested, uncompiled...for now.)

bsdaudio: this appears to be using the wrong variable...? (We probably never noticed because this is meant to block until it fully writes a buffer, and would only trigger an issue if we had a short write that wasn't otherwise an error condition.)

arts: Patched to (maybe) compile.

arts: implemented audio capture support. (completely untested! Not even compiled!!).

alsa: capture devices don't need a mixbuf allocated.

testaudiocapture: made test app interactive. (hold down mouse/finger to record, then it plays back what it heard. Repeat.)

coreaudio: Implemented audio capture for iOS.

testaudiocapture: Make a simple green/red window when recording/playing.

alsa: Implemented audio capture support!

alsa: Cleaned up the 5.1 swizzle mess a little. Shouldn't this be something ALSA handles for us with channel maps, so we can just delete this code?

audio: a little more robustness in the capture device's thread.

pulseaudio: Implemented audio capture support!

audio: implemented higher level infrastructure for running capture devices.

audio: Made some SDL_AudioDevice fields atomic. This makes sure they're properly communicated to the audio threads.

testaudiocapture: Let specific devices be opened.

BUGS.txt: Bugzilla supports SSL; changed link to https:// ...

audio: changed some internal ints to be SDL_bools.

nacl: unlock audio thread mutex when done with it.

audio: Make SDL_AudioDevice::shutdown an atomic value. Just to make sure this get communicated to the audio thread properly.

audio: Implemented capture support for Mac OS X CoreAudio. I don't know what iOS wants yet, so this code might work there, too...?

audio: Initial bits to enable audio capture support.

Fixed information in touch README and changed identifiers so doxygen links them.

Android: Updated a comment.

Android: Replaced deprecated AbsoluteLayout with RelativeLayout. Fixes Bugzilla #2211.

Linux: Fixed memory leak in haptic implementation (thanks, Martin!). Fixes Bugzilla #3238.

Pandora: Fixed memory leak in video implementation.

Pandora: Fixed compile warnings in video implementation (thanks, Bombo!). Fixes Bugzilla #2734.

ALSA: Fixed adding wrong status to error message (thanks, romain145!). Fixes Bugzilla #3364.

Pandora: Fixed SDL version in debug output.

Pandora: Updated SDL version in Makefile.

Pandora: Fixed referring to no more existing atomic source file in Makefile.

Updated comments in video implementations.

Pandora: Fixed compile error in video implementation (thanks, Bombo!). Partially fixes Bugzilla #2734.

Android: Fixed androidbuild.sh not working on OS X (thanks, Steffen!). Fixes Bugzilla #2429.

Android: Fixed missing export of a JNI function (thanks, Andrii!). Fixes Bugzilla #3393.

Android: Fixed compile warning about a missing prototype (thanks, Sylvain!). Fixes Bugzilla #3351.

Fixed compile warnings about evaluating undefined identifiers. Partially fixes Bugzilla #3351.

Tests: Would be wise to compile this

Tests: Somehow tabs leaked in...

Tests: Add a manual test for a custom cursor (Taken from the API docs)

Tests: Mir needs the window to swap at lease 1 frame for the cursor to show. So render in testwm2

MIR: Fix system cursors. Been broken since custom cursor support was added!

Android: Added new key codes from API 24.

iOS: Fix an OpenGL ES error when rotating the device if MSAA is used (thanks Cole Campbell!) Fixes bug #3378.

Fixed three source comments in tests.

Added a simple test case for SDL_sscanf() to tests. It fails on platforms where SDL's custom implementation is used. Relates to Bugzilla #3341.

Fixed freeing strings created by strdup() with SDL_free(). This only worked on platforms where SDL_free() wraps free().

Mac: Fixed crash if SDL_GetDisplayDPI() got NULL pointers for output parameters.

Fixed compile warnings in test library about formats strings not being literals. Partially fixes Bugzilla #3375.

Mac: Updated file name in README.

Fixed Bug 3147 - Windows: Crash when resizing Window since hg 333216331863 Thanks for the fix, Gab!

Mac: Updated availability macro in build script.

MIR: Add better error handling for a failed connection

MIR: Support relative mouse mode

Windows: Fix other window positions/sizes being changed when exiting exclusive fullscreen (bug #3315, thanks Simon Hug!)

SDL 1.2 Windows: Fix other window positions/sizes when exiting fullscreen (bug #3315, thanks Simon Hug!)

Mac: Fix a crash when SDL is compiled with SDL_MAC_NO_SANDBOX enabled, by increasing the stack size of the mouse tap thread back to OS X' default of 512 KB.

Removed Mac OS 10.5 support (bug #3137). Also fixed a warning about deprecated Carbon code when using SDL_audio (bug #3127, thanks Dominik!)

Updated documentation of SDL_IsScreenSaverEnabled(). The screensaver is deactivated by default since SDL 2.0.2.

Windows: Fixed missing error message if SDL_GetDisplayDPI() failed.

X11: Fixed typos in error messages and source comments.

X11: Fixed missing error message if SDL_GetDisplayDPI() failed.

WinRT: workaround a possible Windows bug, whereby hiding cursors, disables mouse-moved events This workaround, unfortunately, requires that apps directly link to a set of Win32-style cursor resource files (that contain a transparent cursor image). Copies of suitable resource files are in src/core/winrt/, and should be included directly in an app's MSVC project. A rough explanation of this workaround/hack, and why it's needed (and seemingly can't be done through programmatic means), is in this change's code.

X11: Fixed SDL_GL_GetSwapInterval() returning -1 if interval is unknown. It should return 0 as a safe default if the interval can not be determined.

EGL: Fixed SDL_GL_GetSwapInterval() returning -1 on error. It should return 0 as a safe default on errors. -1 is returned for late swaps.

Updated documentation of SDL_CreateWindow() and SDL_CreateWindowFrom().

iOS: Added brackets to function names in README so that doxygen links them.

Fixed missing error message if SDL_GetDisplayDPI() is unsupported.

Fixed crash in shape test program if memory allocation failed.

Fixed error return values in filesystem test program.

WinRT: allow on-screen keyboard to be shown via SDL APIs, Win10/UWP only

Fixed memory leak in game controller test program.

Updated documentation of SDL_HINT_THREAD_STACK_SIZE in header file.

Added missing links at documentation of SDL_FreeCursor() in header file.

Cocoa mouse code: Replaced NSPointInRect with NSMouseInRect (thanks Eric Wasylishen!) Fixes bug #3312.

Mac: replaced a deprecated CGSetLocalEventsSuppressionInterval call that I missed in commit f0645cd8fd39

Android: Updated name of README file.

WinRT: another README tweak

WinRT: README tweaks

Wayland: Removed not needed including and setting of errno. One internal function was setting errno on error but it was not read afterwards.

Fixed markdown formatting in porting README for doxygen output. The underscores and the newlines for the list were lost before.

Added missing error return in test program.

PSP: Fixed compile error.

WinRT: build/link fix

Patched to compile on Haiku...?

windows: created threads' stack sizes should be reserved, not committed.

Patched to compile on various platforms.

threads: Move SDL's own thread creation to a new internal API. This allows us to set an explicit stack size (overriding the system default and the global hint an app might have set), and remove all the macro salsa for dealing with _beginthreadex and such, as internal threads always set those to NULL anyhow. I've taken some guesses on reasonable (and tiny!) stack sizes for our internal threads, but some of these might turn out to be too small in practice and need an increase. Most of them are simple functions, though.

threads: Handle SDL_HINT_THREAD_STACK_SIZE at top level, implement elsewhere.

WinRT: bug hack-fix - gamepad detection was failing on Xbox One Win10's 'GamepadAdded' event seems to need to have something registered with it in order for Xinput-based gamepad detection to work. This 'fix' simply causes a dummy event-handler to be added for this event, in case an app wants to use gamepads on Xbox One (most likely).

iOS: Fixed SDL_GL_CreateContext crashing instead of returning null when a GLES3 context is requested on iOS 6 and older.

alsa: Make device enumeration robust against weird results with NULL strings.

iOS: Also do the Dictation crash workaround before executing the animation callback.

Add a new hint SDL_HINT_MAC_MOUSE_FOCUS_CLICKTHROUGH, which allows mouse click events to occur when clicking to focus a window in Mac OS X. Fixes bug #3300.

iOS: Workaround for a crash after Dictation is used.

iOS: Fixed compiling demos on C89 compilers.

PSP: Fixed returning success from unsupported SDL_RenderReadPixels(). This also fixed the missing error message.

PSP: Fixed returning success for unsupported SDL_SetTextureColorMod(). Partially fixes Bugzilla #3298.

PSP: Fixed missing error message for unsupported SDL_CreateWindowFrom().

Updated info about required Ubuntu version for Mir support in README.

Wayland: Fixed missing error message if creating a custom cursor failed. SDL_GetError() returned no error message because it was written to stderr only.

Removed not needed SDL_WINDOW_SHOWN from chessboard test program.

iOS: Removed not needed SDL_WINDOW_SHOWN from rectangles demo.

iOS: Fixed keyboard demo ignoring quit requests and window closing. This also fixed a warning about an unused variable.

iOS: Replaced #import with #include in two demos. This extension to C is not required here and made the demos less portable.

iOS: Restored older version of rectangles demo. The demo did not draw rectangles like described in the demo README.

iOS: Updated SDL version in demo README.

Fixed comment in gesture source.

1.2 stdlib: include SDL_config.h so we can see if HAVE_QSORT is actually set.

gcc-fat.sh: SDL 1.2 won't build with the Mac OS X 10.11 SDK with < 10.7 compat. Several deprecated APIs it uses are flat-out removed from the SDK now. We already had preprocessor checks around them, so gcc-fat.sh just needs its minimum OS compatibility bumped up to 10.7. Leaving the deprecated code in place, however, in case someone wants to target an ancient OS X by compiling with an older SDK.

SDL_qsort in SDL 1.2 needs to use the system includes.

Mac: avoid calling CGSetLocalEventsSuppressionInterval, it was deprecated in OS X 10.6.

Windows CMake: don't try to link dxguid, it's unavailable when using Visual Studio's XP-compatible toolset.

Wayland: Fixed crash if memory mapping failed while creating a custom cursor.

Wayland: Fixed storing a theme cursor which is never used in video data.

Quick fix for qsort off-by-one error.

Pick up new display mode information after a mode change (Windows only right now).

Quick fix for qsort off-by-one error.

Emscripten: Deactivated custom cursor support because it created system cursors.

Wayland: Fixed drawing created cursors without transparency.

Wayland: Fixed showing created cursors incorrectly before the first redraw.

WinRT: Removed dead code and fixed memory leak if allocation for driver failed.

Fixed compiling IME test program with HAVE_SDL_TTF on C89 compilers.

Removed unnecessary include statement in test program.

A simple test program for SDL_qsort().

Another commit to test revision control.

Minor tweak to a README to test Mercurial server.

Fixed infinite timeout in SDL_WaitEventTimeout() - thanks Константин Гончарик

Allow mappings to be added before GameControllerInit

alsa: Implemented basic device detection. This is kind of nasty, because ALSA reports dozens of "devices" that aren't really things you'd ever want, or things that should be listed this way, but the default path still works as before, and it at least allows these devices to be available to apps. This does not handle hotplugging yet. You get a device list at init time and that's it.

x11: Fix a few more XMoveWindow() calls to adjust for border size. Also, fix my inability to do basic math ('+' should have been '-').

x11: Deal with window borders better. - Cache the _NET_FRAME_EXTENTS data locally, so we don't have to query the X server for them (instead, we update our cached data when PropertyNotify events alert us to a change). - Use our cached extents for X11_GetWindowBordersSize(), so it's a fast call. - Window position was meant to refer to the client area, not the window decorations, so adjust appropriately when getting/setting the position.

Raspberry: Fixed crash if memory allocation for cursor failed.

Fixed compile warnings about unused variables in IME test program.

Wayland: Fixed fault in event handling which might have caused a crash someday. Found by Cppcheck.

Wayland: Fixed crash if allocating memory for cursor failed. Also added missing error message if first allocation failed.

Replaced strlen() with SDL_strlen() in IME test program.

wayland: Add wl_proxy_marshal_constructor_versioned sym wl_proxy_marshal_constructor_versioned was introduce in wayland-client 1.10.

Mir: Do not use opengl to find the valid pixel format if we dont use opengl

Mir: Replaced memcpy() with SDL_memcpy().

Mir: Fixed comment at conditional compilation macro.

Fixed doxygen warnings about markdown formatting.

Mir: Fixed crash if allocating memory for cursor failed.

Regenerated configure script with Mir changes.

Fix API/ABI breakage in Mir 0.13/0.14.

Windows: Just use WaitForSingleObjectEx() everywhere. (It's supported on WinXP, no reason to have an #ifdef here...I think.)

Windows: let threads be named in the debugger. We now only raise the magic exception that names the thread when IsDebuggerPresent() returns true. In such a case, Visual Studio will catch the exception, set the thread name, and let the debugged process continue normally. If the debugger isn't running, we don't raise an exception at all. Setting the name is a debugger trick; if the debugger isn't running, the name won't be set if attached later in any case, so this doesn't lose functionality. This lets this code work without assembly code, on win32 and win64, and across various compilers. The only "gotcha" is that if you have something attached that looks like a debugger but doesn't respect this magic exception trick, the process will likely crash, but that's probably a deficiency of the attached program. Fixes Bugzilla #2089.

1.2 stdlib: Updated qsort with licensing resolved (thanks, Gareth!).

stdlib: Restored previous qsort() implementation; the licensing is resolved. Thanks to Gareth McCaughan for changing his code to the zlib license on our behalf!

x11: better fix for the previous commit's fullscreen vs maximized issue.

x11: Don't mess with fullscreen vs maximized window state on unmapped windows.

Android: Restored Philipp's joystick change, lost in the previous merge. This patch, specifically: https://hg.libsdl.org/SDL/rev/2455bf5d1866

Merge Android C89 fixes from Eric Wing.

Android: C89 cleanup to avoid warnings/errors since the default gcc mode on Android is still pre-C99.

x11/wayland/mir: Make the dynamic loading macro salsa a little less messy.

Added support for the Mad Catz FightStick TE S+ PS3

Added support for the Mad Catz FightStick TE S+ PS4

Added support for the Akishop Customs PS360+ v1.66 and MadCatz PC USB Wired Stick 8818

Added support for several Mad Catz arcade sticks

Allow using the game controller API with arcade sticks and other XInput devices

Android: Changed an internal joystick function to return count instead of id. The returned value is currently not used by the caller. The instance id would also not be needed on Java side and providing it just complicated the function. Partially fixes Bugzilla #3234.

x11: Patched to compile with DEBUG_XEVENTS on C89 compilers.

x11: Removed an assert. This assert triggers when run under XMonad. It's safe to pass a zero here anyhow, as this will still work "well enough" and the original problem--GNOME printing a warning message--is still fixed because GNOME's window manager gives us a chance to grab a non-zero user-time value before this code is run.

Another attempt to fix Windows build.

Patched to compile on Visual Studio.

Replaced SDL_qsort with public domain code from PDCLib: http://pdclib.e43.eu/

Windows: make sure SDL_TicksInit has been called before calling Sleep(ms) in SDL_Delay. This ensures the Windows system timer resolution is properly set before Sleep is called.

PSP: Deactivated dynamic API. Thanks to Bugzilla #3240 for pointing this out.

PSP: Added missing object for generic TLS to Makefile. Thanks to the SDL mailing list (and later Bugzilla #3240) for pointing this out.

Fixed spaces in license comments.

x11: Updated imKStoUCS.* to latest from x.org. This was a version from XFree86 before now. :) Although not much has changed.

iOS: Implemented clipboard support.

David Carlier implemented SDL_GetBasePath() for OpenBSD

Return the full number of events from SDL_PeepEvents() if NULL is passed in with SDL_PEEKEVENT

x11: Fix AltGr generating an invalid keycode (bug #3244). Thanks Thomas!

PSP: Removed inlined copy of SDL_CalculateAudioSpec()'s functionality.

Android: Fixed ignoring a return value while handling motion events. This should not have caused problems yet because the return value was identical.

Fixed creating fullscreen windows on Steam Link

Expose the EGL display and window for Vivante SDL windows

Fixed compile warnings about type conversion. Found by buildbot.

Android: Fixed finishing Activity on some devices if right mouse button pressed. Partially fixes Bugzilla #3227.

Fixed two typos in documentation.

Android: Added mouse initialization to reset state. If the app is launched again then the shared object may be reused (on Android).

Android: Fixed a comment.

Android: Added mapping of mouse forward button and mouse back button.

Removed dead code (caught by Clang's static analyzer).

CMake: only set "-O3 -g" defaults if CMAKE_BUILD_TYPE wasn't set at all.