Workaround some Windows OpenGL drivers mishandling wglMakeCurrent(). You should be able to do wglMakeCurrent(NULL, NULL), according to MSDN, since the first parameter is ignored if the second is NULL, but this causes problems on some drivers, so make sure we always have a valid HDC in the first parameter (and exit early if we already have no current context in the NULL, NULL case).

Fix for bug 2001, verify that the AudioTrack is properly initialized ...and fail to initialize the audio backend if it's not!

Fixed bug 2000 - The framework INSTALL_PATH still uses @executable_path dak180 Since the MACOSX_DEPLOYMENT_TARGET is 10.5; INSTALL_PATH should use @rpath. See http://www.dribin.org/dave/blog/archives/2009/11/15/rpath/ for the details on why this is a good idea.

Added code signature step to Framework build process

Check for negative source rectangle in SDL_SoftBlit() (thanks John!)

Added support for the two new key codes from Android 4.3.

Fixed compiler warning with gcc

Fixed stack smashing crash when using the GCC compiled DLL with Visual Studio.

Workaround older libpulse that fails in pa_context_new() with a NULL appname. Newer libpulses do sane defaults if you pass a NULL there. In lieu of a proper way to specify this in SDL at the moment, this will do. Fixes Bugzilla #1119.

These should be __inline__, not inline.

Attempt to fix compiler warning. Fixes Bugzilla #1784. (or at least the only part of it that was a real bug.)

Added documentation saying the render API isn't for multi-threading.

Changed test program to be more compatible C.

Removed unused variable from test program.

Removed unused parameter in test program.

Corrected words in comments of test programs.

Fixed return value documentation for SDL_HapticUpdateEffect(). It was a cut-and-paste error from SDL_HapticNewEffect(), right above it.

Backed out hg changeset 7f26fd1df927; docs were wrong, not the code.

Added a FIXME discussion to last commit.

Cocoa: Make the next-highest window gain focus when a window is closing. (if the closed window wasn't the foreground, this is effectively a no-op.)

Fixed SDL_HapticOpened() returning -1 instead of 0. According to header file it should only return 0 or 1.

Removed C++ macro setup in internal header for Android port which is only C now.

Fixed SDL_HapticRumblePlay() maybe working because of SDL_HapticUpdateEffect().

Corrected SDL_HapticUpdateEffect() returning 0 instead of index of effect. According to documentation in header and wiki the index should be returned. This change may break existing programs which assume only 0 means a success.

Added example of using the software renderer and window surface API, contributed by Nitin Jain.

Fixed bug 1983 - SDL_thread.h broken under MinGW crosscompiling environment q66 after updating SDL2 to the latest RC I'm unable to build my game engine under mingw crosscompiling environment for Windows (32bit, hosted on freebsd 64bit). The error is in SDL_thread.h with messages like this: http://codepad.org/jEQXd3Yq The problem is, while _beginthreadex return type is correctly mapped to uintptr_t (as specified by Microsoft), the return type under mingw is unsigned long int (same size, different signature), resulting in the error above. The reason it didn't error before is this: http://codepad.org/8FAbKAxz You can see the _beginthreadex case is only used without HAVE_LIBC defined; at one point though, SDL_config_windows.h was changed like this: -/* Enabled for SDL 1.2 (binary compatibility) */ -#define HAVE_LIBC 1 +/* This is disabled by default to avoid C runtime dependencies and manifest requirements */ resulting in these errors.

Fixed variable scoping for Windows build

Fixed bug 1272 - Bogus numlock key up/down events being reported on MacOS X Vern Jensen The problem is that in certain situations I'm getting THREE keyUp/keyDown events when I push certain keys. In my event code I added: case SDL_KEYUP: printf("SDL KeyScanCode for KEYUP event: %d\n", event->key.keysym.scancode ); … and case SDL_KEYDOWN: printf("SDL KeyScanCode for KEYDOWN event: %d\n", event->key.keysym.scancode ); … The result of one test run where I push 2 keys and then release them is this: SDL KeyScanCode for KEYDOWN event: 92 // Pushed keypad 4 SDL KeyScanCode for KEYDOWN event: 83 // Pushed left shift SDL KeyScanCode for KEYUP event: 83 SDL KeyScanCode for KEYDOWN event: 225 SDL KeyScanCode for KEYUP event: 92 // Released keypad 4 SDL KeyScanCode for KEYDOWN event: 83 SDL KeyScanCode for KEYUP event: 83 SDL KeyScanCode for KEYUP event: 225 // Released left shift There *should* be only a total of 4 events above… 2 for each key being pushed, and 2 for each being released. But instead some bogus events for numlock being pushed/released are sent from SDL. These events did not occur. I did not push numlock. The value above for numlock is 83. Comments above show when I pushed each key. As you can see, when I push left shift, THREE events are instantly sent to my application, keyDown and then keyUp for numlock, and then the valid event for left shift (the key that was actually pushed). You could replace keypad 4 with pretty much any keyPad key and it'll still happen. You can also replace it with any arrow key and it'll happen. However, when trying it with normal letter keys on the main keyboard it didn't. It happens with other modifier keys too, not just left shift. The order in which the keys are pressed matter. For instance, if I do: 1) keypad 4 2) left shift 3) release left shift 4) release keypad 4 Then at step 2, I get the 3 events above (when there should be only one), but steps 3 and 4 work properly… I don't get extra keyUp/keyDown events for steps 3 or 4. Thereas if the order of steps 3 and 4 are reversed, I get the bogus extra events for numlock. Also, the problem can occur even when pushing just a single key by itself. If I push left shift, then keypad 4, then release left shift, then release keypad 4, then the following push of left shift will cause the bug. If I continue pushing and releasing left shift though, it won't happen again until I again involve keypad keys. --- Sam Lantinga According to the Apple documentation, NSNumericPadKeyMask is set for any arrow or numeric keypad event. Indeed this is what's happening. I verified that we get the correct events for the numlock key and the mod state gets set correcly, so it should be safe to remove this bogus code.

Fixed bug 1743 - CMake produces libraries with wrong filename/SONAME David Gow As discussed on the list, the autotools build backend uses libtool's "release" option, giving us the SONAME libSDL2-2.0.so.0, whereas CMake doesn't, giving us libSDL2.so.0 While libSDL2.so.0 has some small advantages (being simpler and matching the names on some other OSes better), many products have already been developed expecting libSDL2-2.0.so.0, which better matches SDL 1.2's SONAME. It seems clear, therefore, that most developers prefer this name. This patch emulates libtool's functionality, making libSDL2-2.0.so.0 the name of the shared library, while leaving libSDL2.a as the filename of the static library. Unlike with libtool, no libSDL2.so symlink is yet made. I also haven't tested this on anything but Linux, so it might break other platforms. :/

Improved error checking in the haptic system, preventing crashes in some cases. Edgar Simo 2012-05-06 02:33:39 EDT I recall that driver being buggy back in the day, but looking over the code there's a number of things being done wrong which I've fixed and it should now properly error out instead of crashing. Also make sure you initialize the haptic subsystem before using haptic commands (which I now more explicitly try to enforce).

Fixed bug 1919 - Window icon disappears as soon as a renderer is created Sebastian Setting a window icon works just fine until a renderer is added to the window. After adding the renderer the icon disappears. Reproduce by: - Take the example code from the wiki: http://wiki.libsdl.org/moin.fcg/SDL_SetWindowIcon - Add the following two lines after SDL_FreeSurface(surface); SDL_Delay(1000); SDL_Renderer* ren = SDL_CreateRenderer(window, -1, SDL_RENDERER_ACCELERATED); -compile and run You will see the window icon correctly at first. After the Delay the Icon will disappear.

Fixed bug 1526 - X11 - XUnmapWindow should not be called directly driedfruit In X11_HideWindow, we call XUnmapWindow to unmap. According to ICCCN2.0, this should never happen, and XWithdrawWindow should be called instead. http://www.tronche.com/gui/x/icccm/sec-4.html#s-4.1.4 "The reason for requiring the client to send a synthetic UnmapNotify event is to ensure that the window manager gets some notification of the client's desire to change state, ***even though the window may already be unmapped when the desire is expressed***." Additionally, this can be observed at http://cgit.freedesktop.org/xorg/lib/libX11/tree/src/Withdraw.c#n65 Failure to comply leads to "MapNotify" event never appearing on non-reparenting WMs after subsequent show-hide-show requests. (I'm currently observing this behavior, thus my bug report).

Fix cmake breakage (thanks Marcus) My recent change broke the sdlcheck.cmake by not escaping some quotes properly. Thanks to Marcus von Appen for pointing it out and providing the patch.

Consider the case where an attribute is set to EGL_DONT_CARE in eglChooseConfig SDL doesn't currently use EGL_DONT_CARE, but it doesn' hurt to future proof.

[Android] Bug 1827, select the best matching config provided by eglChooseConfig

Fixed crash and garbled output when converting from F32 to S16 audio.

Updated the copyright year for the test programs

Corrected return value to be of correct type.

Corrected internal documentation in source. The parameter silent is no longer optional.

Removed unreachable code.

Removed unused variables to fix warnings in test program.

Fix for recent GLX error bug Lee Salzman I messed up in the patch I sent you regarding gobbling up the GLX error codes signaled when trying to create a context. After reading the spec I realized those error codes are relative to a base error that needs to be queried when setting up the GLX extension... So I have made a patch that fixes it for a user I had who was still getting the bug with my old patch. Without this patch my previous one won't work, so it is recommended to merge this...

A little cleanup on the cleanup, just for consistency.

= {{0}} generates a memset() in MSVC, which isn't linked. Whoops!

Fix build errors from last change.

Fix some clang analyzer warnings. This fixes some analyzer warnings and a couple of minor memory leaks.

Mac: Codify SDK and OS requirements, and clean up. This #errors if you're using an SDK or deployment target that is less than 10.6 and 10.5, respectively, and cleans up uses of MAC_OS_X_VERSION_MIN_REQUIRED and MAC_OS_X_VERSION_MAX_ALLOWED according to those requirements.

Add a timeout to the clipboard code to avoid hangs when using synergy. When a timeout is detected we set the clipboard so that we own it so that future timeouts are avoided. Testing has confirmed that this timeout and setting only occurs when the clipboard contains text from my Windows machine. That is, if you copy from the Windows clipboard and then launch Dota and then repeatedly paste to a terminal window then the text will disappear when Dota hits the timeout. If you then select on the Linux side you can repeatedly paste. If you select again on Windows then the text will get cleared again. Note that Dota only looks at the clipboard when it has focus. CR: Saml

Added Direct3D shader to hardware accelerate YV12 and IYUV textures.

Finished updating testoverlay2 for the new API and added it to the Visual Studio 2010 projects

Fixed SDL printf output for 0x%.8x

Updated blend semantics so blending uses the following formula: dstRGB = (srcRGB * srcA) + (dstRGB * (1-srcA)) dstA = srcA + (dstA * (1-srcA)) This allows proper compositing semantics without requiring premultiplied alpha. Needs full unit test coverage and bug fixes!

Android: Removed all unnecessary dependencies on C++. C++ is a bit of a minefield on Android. Much functionality still doesn't work, and Android can't decide on which C++ standard library to use, so it provides 3 different ones, all of which are incompatible with each other. (It looks like clang is coming too which will add a new compiler and a 4th standard library.) As middleware, SDL might be distributed as a binary and intermixed with other projects already using C++. If C++ is intermixed in a bad way, bad things will happen. Removing dependencies on C++ will avoid this problem and downstream users won't have to worry/care.

Corrected internal documentation in source.

Fix bug 1494: add missing test projects to VS2010/VS2010 solution

Fixed building with cygwin Note that building with cygwin gcc results in a DLL that depends on cygwin1.dll

Removed not needed SDL_WINDOW_SHOWN from iOS example programs. See bug #1667.

Fixed bug 1813 - MouseMotion relative values do not respect renderer LogicalSize driedfruit A trivial issue, the xrel and yrel values of MouseMotion event struct are not adjusted to renderer logical size.

Fixed bug 1973 - test/Makefile.in fails on Mac OS X with X11 enabled... Ryan C. Gordon If you have the X11 SDK installed on Mac OS X, you'll build with X11 support, but the Makefile doesn't build anything but the Cocoa testnative code for Mac OS X, which is fine, but then testnative fails to link because testnative.h enables the X11 support based on the SDL config header. Building the testnativex11.c code on Mac OS X doesn't work because you need to explicitly link with "-L/usr/X11/lib -lX11" if this code is enabled, but you can't link with that by default because the X11 SDK might be missing. So this needs a little configure magic to do the right thing (or forcibly disabling X11 support on Mac OS X in testnative.h).

Added X include path so SDL_syswm.h works correctly on systems with X11 in a non-standard directory.

Fixed cursor leak when quitting the mouse subsystem

Fixed bug 1976 - SDL2.m4 shares same variables as SDL.m4 so they cache values Edgar Simo So basically the issue is that: AC_PATH_PROG(SDL_CONFIG, sdl2-config, no, [$PATH]) in sdl2.m4 shares the same variable (SDL_CONFIG) as sdl.m4: AC_PATH_PROG(SDL_CONFIG, sdl-config, no, [$PATH]) What happens is if you run the AM_PATH_SDL2 and then AM_PATH_SDL to say fall back on SDL if SDL2 isn't found, the AM_PATH_PROG command gets cached (ac_cv_path_SDL_CONFIG) so that unless you unset this cache variable, the result for AM_PATH_SDL will be EXACTLY the same as for AM_PATH_SDL2. To fix I would recommend renaming SDL_CONFIG to SDL2_CONFIG in sdl2.m4, replacing all instances so it won't cache to the same variable.

Fixed bug 1979 - X11_SetGammaRamp doesn't check the return value of XStoreColors Sven Hesse Here on my laptop with an ATI Mobility Radeon 9600 PRO Turbo (RV350), the free radeon drivers and X.org 1.12.4, SDL_SetGamma() fails because it tries to use SDL_SetGammaRamp() if available. SDL_SetGammaRamp() uses XStoreColors(), and that call always fails with BadRequest (for whatever reason, I don't know). The SDL_Visual class check goes through, though. Now, XStoreColors failing is maybe another bug in SDL, or in Xorg, or in the radeon drivers or anywhere else, but I think SDL should still check that the XStoreColors() doesn't error as well.

Add support for the _NET_WM_BYPASS_COMPOSITOR hint This hints to window managers to unredirect the window to improve performance.

Cursor mask was uninitialized data, causing problems. Fixes Bugzilla #1461.

Include SDL_assert.h to fix broken build.

Don't allocate memory if we're just going to fail when checking parameters.

Don't try to clear errors in GL_ActivateRenderer() before we MakeCurrent. Otherwise, if we destroyed a different renderer, next time this one draws, it'll clear errors forever (GL_INVALID_OPERATION for having no current context, at least on Windows), hanging up the program in an infinite loop. Fixes Bugzilla #1775.

Fixed another compiler warning. This one is compliments of the FreeBSD buildbot finally having the X11 libraries installed. :)

One more compiler warning.

More compiler warning fixes.

Fixed some compiler warnings that Visual Studio reported.

Make XInput haptic code respect effect timeouts. This is really just a hack until this code expands to be a robust haptic mixer. (This is also untested, beyond compiling. Sorry!)

Changed documentation comment for SDL_CreateShapedWindow(). See bug #1667.

Replaced use of malloc()/free() with SDL_malloc()/SDL_free() in test program.

Removed not needed SDL_WINDOW_SHOWN from test programs. See bug #1667.

Replaced use of strcmp() with SDL_strcmp() in tests.

Removed not needed variable and work in standard library. Found by Cppcheck (that variable's value was never used).

Workaround crash bug in libXi <= 1.4.1 (thanks, Steve!). Fixes Bugzilla #1812.

Corrected comments in test library source.

Changed return value of internal function to shorten generic haptic source.

Fixed bug 1977 - D3D_UpdateClipRect() sets the wrong width for the clip rect Bithika Mookherjee SDL_RenderSetClipRect() calls into renderer->UpdateClipRect(renderer). I am not sure if UpdateClipRect() can point to a number of clip rect update functions, but on my platform it calls D3D_UpdateClipRect(). In that function, the rect to pass to IDirect3DDevice9_SetScissorRect() has it's right field set as: r.right = rect->w + rect->w; But actually, this should be: r.right = rect->x + rect->w;

Fixed wrong type being passed to eglGetConfigAttrib() (thanks, Alexander!). This wanted an EGLint*, not a VisualID* cast to an EGLint*. Without this, 64-bit X11 fails here, because the datatype sizes are different. Fixes Bugzilla 1686.

Added finger events to SDLTest_PrintEvent().

Mac: Handle SDL_CreateWindow with SDL_WINDOW_MINIMZED. This fixes bug #1446. You can now create a window with SDL_CreateWindow(..., SDL_WINDOW_MINIMIZED), and not have it immediately restore itself. It also changes SDL_RaiseWindow() to be a no-op on minimized or hidden windows, which is how it behaves on Windows.

Made PND_setwindowgrab() a no-op. It was a cut/paste of PND_destroywindow(). This looks like the rest of the code doesn't deal with windows, and probably just deals with a single fullscreen GLES context, like a console would, so making setwindowgrab a no-op makes sense in this case; it's already "grabbed." Fixes Bugzilla #1850.

Changed strdup() to SDL_strdup().

Mac: Bring back FS windows when appropriate This automatically restores FS windows when the application is made active (Cmd-Tab, you click the Dock icon, or you launch the .app again).

Mac: CGMakePoint -> CGPointMake in previous change.

Mac: Fix SDL_WarpMouseInWindow in fullscreen. If you switched from a window to fullscreen, your SDL_WarpMouseInWindow calls would be offset by the x and y coordinates of the original window.

Turn the system mouse cursor back on before VideoQuit(). This is good policy, so it doesn't have a chance to leave it hidden on targets that wouldn't necessarily reset it by default, but it also fixes a crash if you try to use a message box after SDL_Quit() is called. Fixes Bugzilla #1969.

Minor indentation clean up.

Make winmm and directsound audio targets robust against unsupported formats. It now tries to make sure the hardware can support a given format, and if it can't, it carries on to the next best format instead of failing completely.

Fix #1445: Use xcrun to find CpMac This should make the DMG building step more reliable on all current and future Xcodes, by using xcrun to find the path to CpMac.

Fix #1667: Docs about SDL_CreateWindow and flags. Updates the docs to say you can use SDL_WINDOW_HIDDEN, instead of (the ignored) SDL_WINDOW_SHOWN.

Re-enabled other messagebox tests, moved the SDL_Init() call as late as possible to show the circumstances where it's needed.

Fixed bug 1970 - Cocoa message boxes ignore parent window requests Ryan C. Gordon Cocoa_ShowMessageBox() ignores the "window" field of SDL_MessageBoxData, which means you can't assign a parent window to a message box. This is particularly egregious on Mac OS X, because it'll actually make the NSAlert visually part of the parent window instead of just concerning itself with window focus.

Added testing of messagebox with a parent window

Fixed code example for SDL_GetWindowWMInfo()

Removed obsolete assertion code

Make sure a window is valid for a subsystem before using it in a messagebox

If the video system has been initialized, only use that message box system. Don't pass a window from one video driver to the messagebox function of another video driver. This makes bad things happen. :)

Fixed compiler warnings in test program by using wrapped functions.

Changed use of isspace() to SDL_isspace().

Implement float32 support for winmm and directsound targets (Thanks, John!). Fixes Bugzilla #1657.

Fixed compiler warning in testtimer.c

Protect SDL_PauseAudio*() with the audio callback lock. Otherwise, you can pause audio and still have the callback running, or run one more time. This makes sure the callback is definitely stopped by the time you return from SDL_PauseAudio().

Fixed off-by-one error in SDL_AudioQuit() (thanks, Rainer!). Audio Devices IDs are offset by one. Fixes Bugzilla #1971.

Removed unused internal function. Found by Cppcheck.

Fixed compiler warnings in test program by using wrapped functions.

Fixed compiler warnings in test programs by adding return statements.

Fixed compiler warnings in test programs by adding includes directives.

Changed some for-loops to be a bit more clear.

Don't copy structs to stack in cmpmodes(), use const pointers instead. (and return 0 immediately if the pointers are the same.)

Don't use SDL_memcmp() on modes to check if they've already been added. They might have different driverdata values and thus not match. Fixes Bugzilla #1407.

Explicitly check for Xext.h in the CMake project, fail if missing. Added to match configure script change in hg changeset 8f118396264b.

Fixed compiler warnings in test programs.

Added missing fclose() in test program. Found by Cppcheck.

Removed unused variable and not needed assignment in test program. Found by Cppcheck.

Fixed printf() usage in test program. Found by Cppcheck.