|
3df77ced
|
2017-01-06T00:40:22
|
|
Just roll back the entire portion of the commit from a8253d439914 which caused bug 3544 until we figure out what the right static analysis fix is.
|
|
748f4605
|
2017-01-06T03:38:14
|
|
audio: Add an assert to make sure non-streaming audio uses good buffer sizes.
|
|
|
356c2ead
|
2017-01-06T00:32:06
|
|
Fixed bug 3544 - Memory freeing bug in SDL_DestroyRenderer/SDL_DestroyTexture
felix
Here's a snippet of SDL_DestroyRenderer from hg revision 10746:7540ff5d0e0e:
SDL_Texture *texture = NULL;
SDL_Texture *nexttexture = NULL;
/* ... */
for (texture = renderer->textures; texture; texture = nexttexture) {
nexttexture = texture->next;
SDL_DestroyTexture(texture);
}
SDL_DestroyTexture removes the texture from the linked list pointed to by the renderer and ends up calling SDL_DestroyTextureInternal, which contains this:
if (texture->native) {
SDL_DestroyTexture(texture->native);
}
If it happens that texture->native is an alias of nexttexture two stack frames up, SDL_DestroyRenderer will end up trying to destroy an already freed texture. I've had this very situation happen in dosemu2.
Bug introduced in revision 10650:a8253d439914, which has a somewhat ironic description of "Fixed all known static analysis bugs"...
|
|
|
345c5989
|
2017-01-06T03:15:27
|
|
haiku: Patched to compile.
|
|
|
3443dc19
|
2017-01-05T23:53:46
|
|
Don't do any audio conversion if none is necessary
|
|
|
41be9756
|
2017-01-05T23:26:13
|
|
Fixed bug 3546 - SDL_EVDEV_is_console() uses type of wrong size when calling ioctl
Rob
When calling ioctl(fd, KDGKBTYPE, &type) in SDL_EVDEV_is_console(), we declare type as an 'int'. This should be a 'char'. The subsequent syscall, and kernel code, only writes the lower byte of the word.
See: http://lxr.free-electrons.com/source/drivers/tty/vt/vt_ioctl.c?v=4.4#L399
ucval = KB_101;
ret = put_user(ucval, (char __user *)arg);
I've observed intermittent behavior related to this, and I can force an error condition by using an int initialized to 0xFFFFFFFF. The resulting ioctl will set type to 0XFFFFFF02, and the conditional return in SDL_EVDEV_is_console() will fail.
Recommend changing to char, or masking off unused bits.
|
|
|
b3e8db80
|
2017-01-06T01:07:34
|
|
audio: rename fake_stream to work_buffer.
It's more than an alternative for when the OS can't provide a DMA buffer, now.
|
|
|
992124d4
|
2017-01-06T01:02:58
|
|
audio: Fixed SDL_AudioStreamGet() function parameters.
There was a draft of this where it did audio conversion into the final buffer,
if there was enough room available past what you asked for, but that interface
got removed, so the parameters didn't make sense (and we were using the
wrong one in any case, too!).
|
|
|
99fc1ef9
|
2017-01-06T00:56:29
|
|
naclaudio: Untested attempt to migrate to SDL_AudioStream.
|
|
|
115d0ce7
|
2017-01-06T00:50:01
|
|
haikuaudio: Untested attempt to get this working with SDL_AudioStream.
|
|
|
1a90c72d
|
2017-01-06T00:49:35
|
|
emscriptenaudio: don't get stuck in infinite loop if SDL_AudioStreamPut fails.
|
|
|
f07a1a5a
|
2017-01-05T21:31:02
|
|
emscriptenaudio: Reworked to use SDL_AudioStream.
|
|
|
3761b5f6
|
2017-01-05T20:11:19
|
|
Fixed a few compiler warnings.
|
|
|
4aa9e369
|
2017-01-05T19:45:57
|
|
Patched to compile on some compilers.
|
|
|
52130bde
|
2017-01-05T19:30:45
|
|
diskaudio: Use SDL_Log, not fprintf.
|
|
|
30178a9b
|
2017-01-05T19:29:38
|
|
audio: Added SDL_AudioStream. Non-power-of-two resampling now works!
|
|
|
f12ab8f2
|
2017-01-05T19:12:20
|
|
audio: More effort to improve and simplify audio resamplers.
|
|
|
d024c724
|
2017-01-05T08:16:39
|
|
Fixed signedness issue when blitting on a big endian platform, as reported by capehill
For example, if sR is 0 and dR is 255, we will get -255*sA casted to an unsigned value. Basically results are quite large numbers instead of the expected 0-255 range.
|
|
|
e6e6613c
|
2017-01-05T02:53:29
|
|
Fixed build warning on Haiku
|
|
|
4938c505
|
2017-01-04T10:28:07
|
|
Added SDL_JoystickGetAxisInitialState() to get a joystick axis' initial value.
This is useful for controller mapping programs to determine an axis' zero state
|
|
|
99e10ef5
|
2017-01-04T07:06:48
|
|
Assume D-pad or thumbstick style axes are centered at 0
|
|
|
9d042052
|
2017-01-04T09:33:47
|
|
x11: deal with xrandr display size in millimeters being zero.
Xquartz on macOS reports a zero size, which leads to a division by zero here.
|
|
|
26f84d74
|
2017-01-04T06:21:17
|
|
Added mappings for several GameCube and SNES controllers
|
|
|
d2a01b6e
|
2017-01-04T06:19:56
|
|
Added the NEXT SNES Controller to the list of zero-centered joysticks
|
|
|
c7780497
|
2017-01-04T05:56:47
|
|
Increased joystick jitter tolerance for PS3 controllers
|
|
|
1ddff75c
|
2017-01-04T05:09:02
|
|
Some controllers have trouble getting out to 20000
|
|
|
082132a7
|
2017-01-03T23:39:28
|
|
Fixed binding the D-pad on some Super NES style controllers
Fixed a case where partial trigger pull could be bound to another button
There is a fundamental problem not resolved by this commit:
Some controllers have axes (triggers, pedals, etc.) that don't start at zero, but we're guaranteed that if we get a value that it's correct. For these controllers, the current code works, where we take the first value we get and use that as the zero point and generate axis motion starting from that point on.
Other controllers have digital axes (D-pad) that assume a zero starting point, and the first value we get is the min or max axis value when the D-pad is moved. For these controllers, the current code thinks that the zero point is the axis value after the D-pad motion and this doesn't work.
My hypothesis is that the first class of devices is more common and that we should solve for that, and add an exception to SDL_JoystickAxesCenteredAtZero() as needed for the second class of devices.
|
|
|
d3591800
|
2017-01-03T00:44:05
|
|
Fixed bug 3519 - SDL_GetDisplayMode fails to report mode.format when using Wayland backend
Ryan C. Gordon
Kristian says you can't do it with Wayland, and that going forward, it'll just handle whatever you throw at it anyhow.
https://twitter.com/hoegsberg/status/816148272402165761
So I say we mark it SDL_PIXELFORMAT_RGB888, which is what my X11 display currently reports, and leave it at that.
|
|
|
18d9b23c
|
2017-01-02T10:30:32
|
|
Fixed bug 3539 - SDL2, missing MIR LDFLAGS
Gianfranco
Hello, this is the failure I got
https://launchpadlibrarian.net/300679206/buildlog_ubuntu-zesty-amd64.libsdl2_2.0.5+dfsg1-2ubuntu1_BUILDING.txt.gz
LTLINK build/libSDL2.la
build/.libs/SDL_mirvideo.o: In function `MIR_InitDisplayFromOutput':
././src/video/mir/SDL_mirvideo.c:258: undefined reference to `mir_output_get_current_mode'
collect2: error: ld returned 1 exit status
|
|
|
67ed8943
|
2017-01-01T19:10:36
|
|
Fixed bug 3529 - SDL_EGL_UnloadLibrary is not called at all on SDL_Quit
kaisyu
In case of OpenGLES, the sequences of loading and unloading driver library should be like that:
SDL_Init
...
SDL_GL_LoadLibrary
SDL_EGL_LoadLibrary
...
SDL_Quit
...
SDL_GL_UnloadLibrary
SDL_EGL_UnloadLibrary
...
However, according to my test results, the varible '_this->gl_config.driver_loaded' does not allow 'SDL_GL_UnloadLibrary' to call 'SDL_EGL_UnloadLibrary'.
|
|
|
0badbdde
|
2017-01-01T18:47:29
|
|
Run sed in binary mode so we don't have to worry about fixing line endings
|
|
|
9e6d729e
|
2017-01-01T18:36:35
|
|
Fixed typo in copyright update script
|
|
|
45b774e3
|
2017-01-01T18:33:28
|
|
Updated copyright for 2017
|
|
|
0f586d58
|
2017-01-01T18:32:57
|
|
Convert some files back into DOS line endings after updating copyright
|
|
|
49225f5f
|
2016-12-31T18:11:19
|
|
We don't currently support blitting to < 8 bpp surfaces, return an error instead of corrupting memory in that case.
|
|
|
2ba66d05
|
2016-12-31T16:21:55
|
|
Fixed bug 3535 - Misplaced comment #if/#endif closure comment
Coriiander
This notice is about a misplaced comment.
Often times when we use an #if #endif sequence, the #endif is followed by a comment to indicate what #if statement it belonged to. The SDL_xaudio2.c file contains a misplaced comment, as follows (I stripped the other comments):
#ifdef __GNUC__
# define SDL_XAUDIO2_HAS_SDK 1
#elif defined(__WINRT__)
# define SDL_XAUDIO2_HAS_SDK
#include "SDL_xaudio2.h"
#else
#if 0
#include <dxsdkver.h>
#if (!defined(_DXSDK_BUILD_MAJOR) || (_DXSDK_BUILD_MAJOR < 1284))
# pragma message("Your DirectX SDK is too old. Disabling XAudio2 support.")
#else
# define SDL_XAUDIO2_HAS_SDK 1
#endif
#endif
#endif /* 0 */
That final /* 0 */ should be moved one line up. Like this (I tabbed it out for you to make it more clear):
|
|
|
880842cf
|
2016-12-31T16:14:51
|
|
Fixed bug 3531 - internal SDL_vsnprintf implementation access memory outside given buffer ranges
Tristan
The internal SDL_vsnprintf implementation accesses memory outside buffer. The bug existed also inside the format (%) processing, which was fixed with Bug 3441.
But there is still an invalid access, if we do not have any format inside the source string and the destination string is shorter than the format string. You can use any string for this test, as long it is longer than the buffer.
Example:
va_list argList;
char buffer[4];
SDL_vsnprintf(buffer, sizeof(buffer), "Testing", argList);
The bug is located on the 'else' branch of the format char test:
while (*fmt) {
if (*fmt == '%') {
...
} else {
if (left > 1) {
*text = *fmt;
--left;
}
++fmt;
++text;
}
}
if (left > 0) {
*text = '\0';
}
As you can see that text is always incremented, even when left is already one. When then on the last lines, *text is assigned the NULL char, the pointer is located outside bounds.
|
|
|
7f2068da
|
2016-12-31T10:30:07
|
|
Fixed bug 3541 - DisplayIndex out of bounds in SDL_SetWindowPosition
Intellectual Kitty
In SDL_video.c, on line #1756, in SDL_SetWindowPosition (from today's distribution, 12-31-2016, https://hg.libsdl.org/SDL/shortlog/bf19e0c84483):
if (displayIndex > _this->num_displays) {
should be:
if (displayIndex >= _this->num_displays) {
|
|
e61daa72
|
2016-12-30T19:57:50
|
|
Fixed warning and missing animation delay in testoverlay2 program.
|
|
0cfa0aa1
|
2016-12-29T11:49:18
|
|
improved SDL_GetError() output generated by EGL code
This change attempts to report the EGL error codes generated by SDL's calls
into EGL, along with the name of the EGL function that failed.
|
|
|
c2b90f2d
|
2016-12-28T20:11:29
|
|
Fixed compiling of testgamecontroller program with C++.
|
|
|
de79828b
|
2016-12-28T20:11:12
|
|
Fixed warning about unused variable in controllermap program.
|
|
|
af263798
|
2016-12-28T20:10:48
|
|
Fixed crash if allocating memory for game controller failed.
|
|
|
2e2572a4
|
2016-12-27T23:48:43
|
|
Added SDL_ReserveSpaceInDataQueue() to make space without copying data.
|
|
|
8000e6d9
|
2016-12-27T09:59:36
|
|
Added documentation for the game controller axis values
|
|
|
21cb42d7
|
2016-12-27T09:51:58
|
|
Make sure we go all the way back (within the XBox controller dead zone) to prevent accidentally binding axes inverted
|
|
|
cb8685c0
|
2016-12-27T02:04:38
|
|
Fixed comment style
|
|
|
6d7da088
|
2016-12-27T01:39:07
|
|
Split controller axes into positive and negative sides so each can be bound independently.
Using this a D-Pad can be mapped to a thumbstick and vice versa.
Also added support for inverted axes, improving trigger binding support
|
|
|
7c316366
|
2016-12-26T23:02:14
|
|
x11: Don't loop forever if the X server refuses a pointer grab.
|
|
|
76f2ae3c
|
2016-12-26T22:58:58
|
|
Don't use pushd/popd in autogen.sh; Ubuntu's /bin/sh doesn't support it.
|
|
|
b4e069e7
|
2016-12-26T02:12:21
|
|
Fixed bug 3517 - Compiler warnings with gcc -Wstrict-prototypes
felix
Compiling even a simple SDL2 'hello world' program with gcc -Wstrict-prototypes (GCC 6.2.1) results in warnings like:
/usr/include/SDL2/SDL_gamecontroller.h:143:1: attention : function declaration isn't a prototype [-Wstrict-prototypes]
extern DECLSPEC int SDLCALL SDL_GameControllerNumMappings();
^~~~~~
It seems there is a missing 'void' between the parentheses.
|
|
|
9492492d
|
2016-12-26T01:56:52
|
|
Fixed bug 3516 - fix build on illumos
Sylvain
trivial patch to fix the build on illumos
-Werror=declaration-after-statement
https://gist.github.com/wiedi/15b71456667f7aa2a7f8815663723bb3
|
|
fd85f574
|
2016-12-23T22:49:37
|
|
Mac: back out commit 3e9b2ae41adf. It causes significant overhead on many GPUs.
|
|
|
d719374c
|
2016-12-23T22:08:18
|
|
Mac: Fix over-saturated colors on P3 displays (e.g. the 2016 MBPs).
|
|
|
3e928451
|
2016-12-23T20:36:24
|
|
Windows: Fixed compile error.
|
|
|
7c60a638
|
2016-12-23T20:36:12
|
|
Corrected header file guard comments.
|
|
|
4fc0fe1f
|
2016-12-23T02:23:44
|
|
Removed debug print statements
|
|
|
ad26769c
|
2016-12-22T18:43:00
|
|
Fixed compile errors on various platforms
|
|
|
ca019dad
|
2016-12-22T17:33:45
|
|
Fixed issue where the throttle and other axes on racing wheels don't start at zero and show up as immediate input when the wheel is turned for the first time. Wait until they are actually moved before generating input from them.
|
|
|
b2f6c4c1
|
2016-12-19T11:15:53
|
|
Fixed bus error when converting 16-bit to float for non-integral-multiple sample rates
|
|
|
366c77a9
|
2016-12-18T20:17:33
|
|
audio: fixed one more incorrectly-hardcoded value in the resamplers.
|
|
|
eda74fda
|
2016-12-18T13:05:14
|
|
tvOS: Expose remote swipe gestures as arrow key presses (thanks oviano!)
|
|
|
787a54c8
|
2016-12-18T12:28:28
|
|
iOS bug #3377: work around bugs in some third party iOS libraries (e.g. Google admob) where they assume the optional UIApplicationDelegate ?window? property always exists and will crash if it doesn?t.
|
|
|
f956df23
|
2016-12-17T16:15:24
|
|
audio: fixed arbitrary upsampling (thanks, Sylvain!).
This was a leftover of simplifying the resamplers down from autogenerated
code; I forgot to make something that the generator hardcoded into something
variable.
Fixes Bugzilla #3507.
|
|
|
b515b34d
|
2016-12-16T22:58:32
|
|
Fixed hotplug with more than one device in testjoystick program.
|
|
|
0a3f9d0c
|
2016-12-16T22:58:16
|
|
Fixed warning about unused variable in controllermap program.
|
|
|
0c5e7a10
|
2016-12-15T14:27:22
|
|
Fixed handling joysticks that send multiple events for a single control, e.g. both a button and axis event for a trigger.
Tested with the 8Bitdo NES30 Pro on Linux
|
|
|
b4ea63ec
|
2016-12-14T06:25:09
|
|
Fixed crash if there are multiple joysticks closed during the joystick update loop
|
|
|
f50a0400
|
2016-12-13T00:22:42
|
|
windows: add whitespace to fix macro preprocessing issue (thanks, Sven!).
Apparently without a space here, "fastbuild -cache" breaks.
|
|
|
85d8a79f
|
2016-12-12T09:19:48
|
|
Fixed updated return value for SDL_GL_SwapWindow()
Ozkan Sezer
http://hg.libsdl.org/SDL/rev/464a2676d8ab seems to have
forgotten removing the return from SDL_dynapi_procs.h, and this patch
does that. Without it, MSVC warns:
c:\sdl2\src\dynapi\SDL_dynapi_procs.h(598) : warning C4098:
'SDL_GL_SwapWindow_DEFAULT' : 'void' function returning a value
c:\sdl2\src\dynapi\SDL_dynapi_procs.h(598) : warning C4098:
'SDL_GL_SwapWindow' : 'void' function returning a value
|
|
|
e81bf12b
|
2016-12-12T09:18:42
|
|
Fixed edid parsing code for older gcc compilers
Ozkan Sezer
This adds the name 'ad' to two unnamed unions in edid.h
and adjusts edid-parse.c for it. Nameless unions are not supported in
ancient gcc, which I happened to use on one of my ancient setups.
|
|
|
8414c3d4
|
2016-12-11T12:01:44
|
|
Fixed ABI, don't change the return type of SDL_GL_SwapWindow()
|
|
|
6211668e
|
2016-12-11T12:01:01
|
|
Fixed creating a renderer on the dummy driver on Mac OS X
|
|
|
cb5a0b0f
|
2016-12-11T11:45:33
|
|
Fixed crash when creating a dummy window on Mac OS X
|
|
|
d6bcec8f
|
2016-12-10T15:23:17
|
|
WinRT: build fixes
These fixes are lumped into two categories:
1. add new file, SDL_dataqueue.c, to UWP/WinRT build-inputs (via MSVC project
files)
2. implement a temporary, hack-fix for a build error in SDL_xinputjoystick.c.
Win32's Raw Input APIs are, unfortunately, not available for use in UWP/WinRT
APIs. There does appear to be a replacement API, available in the
Windows.Devices.HumanInterfaceDevice namespace.
This fix should be sufficient to get SDL compiling again, without affecting
Win32 builds, however using the UWP/WinRT API (in UWP/WinRT builds) would
almost certainly be better (for UWP/WinRT builds).
TODO: research Windows.Devices.HumanInterfaceDevice, and use that if and as
appropriate.
|
|
|
454d9cb9
|
2016-12-09T05:19:31
|
|
PP_OK isn't available in the NaCl build environment on buildbot. It's defined as 0
|
|
|
97d05b0d
|
2016-12-09T05:12:27
|
|
Fixed a bunch of SwapWindow calls that needed their return value updated
|
|
|
7a39681e
|
2016-12-09T05:04:18
|
|
Fixed X11 OpenGL ES build
|
|
|
fed8cbcd
|
2016-12-09T05:00:35
|
|
Fixed build for EGL platforms
|
|
|
b936a4c3
|
2016-12-09T04:57:54
|
|
Added support for the XiaoMi Game Controller
|
|
|
70aa2a57
|
2016-12-09T04:17:10
|
|
Only print out the controller mappings if we're not going to test a controller
|
|
|
524bf3c2
|
2016-12-09T01:47:43
|
|
Fixed bug 3513 - SDL_GL_SwapWindow does not return error status
Return an error code from SDL_GL_SwapWindow(), like the other SDL APIs.
|
|
|
3b18c796
|
2016-12-09T01:29:52
|
|
Fixed bug 3512 - Memory leak of SDL_Joystick axes_zero array
Benjamin Harris
Found with valgrind and confirmed in the 2.0.5 source code.
One-line fix in SDL_JoystickClose?
|
|
|
a5250171
|
2016-12-08T10:13:45
|
|
Protect the game controller API the same way the joystick API is protected from multi-threaded access
|
|
|
1b08f0c1
|
2016-12-07T11:41:20
|
|
Added the guide button for the NVIDIA Shield Controller
|
|
|
91491824
|
2016-12-07T11:02:02
|
|
Removed unneeded warning when running from Visual Studio
|
|
|
acabb5d5
|
2016-12-06T14:22:28
|
|
Added support for XBox One controllers on Mac OS X, using the driver at: https://github.com/360Controller/360Controller/releases
|
|
|
09ed5cc0
|
2016-12-06T13:33:02
|
|
xcode: Make sure SDL_dataqueue sources are included in all targets.
|
|
|
6cb2f5bb
|
2016-12-06T12:30:31
|
|
xcode: updated macOS and iOS project files with SDL_dataqueue sources.
|
|
|
c0231875
|
2016-12-06T12:23:17
|
|
audio: Fixed compiler warnings.
|
|
|
68d7be39
|
2016-12-06T00:40:09
|
|
Fixed bug 3508 - variably modified ?SDL_dummy_size? at file scope in test/testatomic.c
Ciro Santilli
GCC 6, Ubuntu 16.10, cd test; ./configure; make
/bin/sh config.status Makefile
config.status: creating Makefile
gcc -o loopwave loopwave.c -g -O2 -D_REENTRANT -I/usr/include/SDL2 -DHAVE_OPENGLES2 -DHAVE_OPENGL -DHAVE_SDL_TTF -g -lSDL2_test -lSDL2
gcc -o testatomic testatomic.c -g -O2 -D_REENTRANT -I/usr/include/SDL2 -DHAVE_OPENGLES2 -DHAVE_OPENGL -DHAVE_SDL_TTF -g -lSDL2_test -lSDL2
In file included from /usr/include/SDL2/SDL_main.h:25:0,
from /usr/include/SDL2/SDL.h:32,
from testatomic.c:14:
/usr/include/SDL2/SDL_stdinc.h:261:20: error: variably modified ?SDL_dummy_size? at file scope
typedef int SDL_dummy_ ## name[(x) * 2 - 1]
^
testatomic.c:106:1: note: in expansion of macro ?SDL_COMPILE_TIME_ASSERT?
SDL_COMPILE_TIME_ASSERT(size, CountTo>0); /* check for rollover */
^~~~~~~~~~~~~~~~~~~~~~~
Makefile:114: recipe for target 'testatomic' failed
make: *** [testatomic] Error 1
If I remove the line SDL_COMPILE_TIME_ASSERT(size, CountTo>0); /* check for rollover */ it works, lazy to figure out the best way to do this.
|
|
|
a0e003ee
|
2016-12-06T02:23:54
|
|
Refactored the audio queueing code to a generic SDL_DataQueue interface.
This is not a public API (at the moment), but we will be needing this for
other internal things soon.
|
|
|
8b960d4e
|
2016-12-06T02:20:58
|
|
Added SDL_VARIABLE_LENGTH_ARRAY so this #ifdef is localized to one place.
|
|
|
af7f30de
|
2016-12-05T21:28:27
|
|
Fixed SDL always being built because of missing headers in the project.
|
|
|
54188c8b
|
2016-12-03T09:59:43
|
|
Fixed crash at startup
|
|
|
264138cc
|
2016-12-03T08:39:21
|
|
Added support for the Hori Fighting Commander 4
|
|
|
e7efcfba
|
2016-12-03T00:40:13
|
|
Added Linux mapping for the DragonRise Inc. Generic USB Joystick
|
|
|
16074e67
|
2016-12-02T22:18:05
|
|
Fixed compile warning on Visual Studio
|
|
|
4eda58ba
|
2016-12-02T21:01:13
|
|
Added SDL_HINT_MOUSE_NORMAL_SPEED_SCALE and SDL_HINT_MOUSE_RELATIVE_SPEED_SCALE to scale the speed of the mouse.
This currently doesn't affect absolute motion, which would need to be implemented on each windowing system so the cursor matches the reported mouse coordinates.
|
|
|
26f05ecb
|
2016-12-02T02:25:12
|
|
Fixed missing prototypes on Android, patch from Sylvain
|