Fixed bug 1953 - Crash at memcpy X11_DispatchEvent(_THIS) Function Nitz In Function X11_DispatchEvent(_THIS), case SelectionNotify : static void X11_DispatchEvent(_THIS) { // Some Code case SelectionNotify: { //Some Code SDL_bool expect_lf = SDL_FALSE; char *start = NULL; // Initialised with NULL char *scan = (char*)p.data; char *fn; char *uri; int length = 0; while (p.count--) { if (!expect_lf) { if (*scan==0x0D) { expect_lf = SDL_TRUE; } else if(start == NULL) { start = scan; length = 0; } length++; } else { if (*scan==0x0A && length>0) { uri = malloc(length--); memcpy(uri, start, length); // Problem is Here, start is still NULL if control comes to else statement without initialising the start pointer, which is wrong uri[length] = 0; fn = X11_URIToLocal(uri); if (fn) SDL_SendDropFile(fn); free(uri); } expect_lf = SDL_FALSE; start = NULL; } scan++; } } As shown above how start pointer remains NULL, Patch for this issue would be: if (*scan==0x0D) { expect_lf = SDL_TRUE; } if(start == NULL) { start = scan; length = 0; } Just replace else if statement with if.

Added src/thread/windows/SDL_systls.c to CMakeLists.txt (thanks, Charles!). Fixes Bugzilla #1955.

Added PowerPC and ARM versions of the memory barrier functions.

Added release/acquire memory barriers to the atomic API * Added a destructor to clean up TLS memory at thread shutdown * Refactored the TLS code to have platform independent code and a small platform dependent core with a fallback to generic code if platform dependent functions fail. * Fixed recursion issues with SDL_GetErrBuf()

Fixed Haiku build

Implemented an API for thread-local storage: SDL_TLSCreate(), SDL_TLSSet(), SDL_TLSGet()

SDL_GL_MakeCurrent: Only no-op redundant calls on *same* thread. This ensures that we only no-op redundant calls if they're made on the same thread as the last MakeCurrent with those arguments. This should maek SDL behave better with multithreaded environments.

Mac: Remove dead FULLSCREEN_TOGGLEABLE code. This code was written almost 2 years ago, and the flag hasn't been changed since. Cleaning up the code by removing the conditional blocks, so that they behave the same way they have for the past two years. FULLSCREEN_TOGGLEABLE used to cause us to use -[NSOpenGLContext setFullScreen] and a pixel format with NSOpenGLPFAFullScreen.

Removing video/uikit/*.c from configure's iOS sources Unexplicable computer sciences phenomenon: Instead of returning an empty set, *.c in an folder with no .c files produces the "*.c" string to be added as a source. I'm sorry future generations, we are doing the best we can :)

Backout hg changset 898992405fa7; lots of things still use SDL_types.h. :/ Will remove this again at some point in the future, though.

Made the SDL_memset4() comment even clearer so we don't accidentally replace it with memset() in the future.

Backed out commit 898992405fa7 because memset() does a byte fill and SDL_memset4() does a uint32 fill and this change breaks SDL_FillRect()

Adds Input Focus handling on Android to improve pausing/resuming behavior Ref: http://android-developers.blogspot.com/2011/11/making-android-games-that-play-nice.html

Removed deprecated SDL_types.h header. Fixes Bugzilla #1945.

Replaced SDL_memset4() implementation with a call to SDL_memset(). The implementation was slower than the C runtime on Mac OS X, Linux, and Windows...quite a bit slower when using the C fallback instead of the inline asm, too. Fixes Bugzilla #1755.

change fsaa argument for testgl to accept a # of samples for FSAA

Disable OSS and BSD audio targets on OpenBSD.

Make sure that srcdir has a full pathname so source indexing works.

Fixed bug 1943 - Wrong handling of legacy 32bpp BMP files Kang Seonghoon While BMP format supports alpha channel, it is enabled only when the header is at least 56 bytes long (BITMAPV3INFOHEADER and later). For very common 40-byte-long header (BITMAPINFOHEADER) 32bpp format should be interpreted as BGRX format, but currently SDL interprets them as BGRA format and causes a significant compatibility problem as many 32bpp files use a padding byte of 0 ("transparent" in BGRA interpretation). --- I fixed this by checking to see if the alpha channel is all 0, and if so, setting it opaque.

sndio dynamic loading fix.

Backed out changeset c8b16b3a3c9b

Fixed bug 1943 - Wrong handling of legacy 32bpp BMP files Kang Seonghoon While BMP format supports alpha channel, it is enabled only when the header is at least 56 bytes long (BITMAPV3INFOHEADER and later). For very common 40-byte-long header (BITMAPINFOHEADER) 32bpp format should be interpreted as BGRX format, but currently SDL interprets them as BGRA format and causes a significant compatibility problem as many 32bpp files use a padding byte of 0 ("transparent" in BGRA interpretation).

Added automated test to validate conversion between all supported formats.

Added surface conversion support for ARGB2101010 formats

Fixed converting ARGB2101010 surfaces to 8-bit surfaces It's not like this is very useful, but it fixes a crash in this case.

Updated configure with the sndio audio backend

Corrected name of constant in header file comment.

Changed include directive to standard header.

Fixed SDL_RWread() returning -1 as unsigned instead of 0 if error on Android. Found by Cppcheck (pointed out check of unsigned < 0 which was left in place).

Added an SDL2 OpenBSD sndio(7) audio target.