|
2e2b84fb
|
2015-05-26T16:31:11
|
|
Some setups need _GNU_SOURCE to make LLONG_MAX available (thanks, Ozkan!).
Fixes Bugzilla #2721.
|
|
|
a21672c8
|
2015-05-26T16:25:22
|
|
Fix fallback define for DECLSPEC for non-Windows platforms.
Looks like it was a copy/paste error?
GCC doesn't support visibility attributes until gcc4, so just make it blank.
Fixes Bugzilla #2720.
|
|
|
59f69f63
|
2015-05-26T16:14:25
|
|
Whoops, fix the static analysis fix.
|
|
|
d1980b93
|
2015-05-26T12:52:28
|
|
Mac: Fix compiler warning when building with a min target >= 10.6.
|
|
|
37f4eb53
|
2015-05-26T12:47:03
|
|
Darwin haptic: Fixed a static analysis warning if axes==0.
|
|
|
f99d6e1d
|
2015-05-26T12:03:51
|
|
Linux joystick: Look at entire axis namespace for controls (thanks, "spaz16"!).
This apparently has fallout: the PS4 (and maybe PS3?) controllers apparently
report some bogus axes, but it won't change the axes we currently expect, and
thus the game controller config string is still stable.
Fixes Bugzilla #2719.
|
|
52306459
|
2015-05-26T08:52:02
|
|
Fixed bug 2869 - Controllers connected on launch are reported twice.
Since all device detection/removal happens on the main thread now, post events inline with when the status changes occur.
Also fixed rare cases when joystick API functions could return data about removed joysticks when called with a device index.
|
|
|
80916e01
|
2015-05-26T11:38:04
|
|
Cocoa: Fixed relative mouse mode when app loses/regains focus (thanks, Eric!).
Fixes Bugzilla #2718.
|
|
|
d074a532
|
2015-05-26T11:32:06
|
|
Make dot easier to see in testrelative.
|
|
|
a2c28ec1
|
2015-05-26T11:13:51
|
|
Updated configure script.
|
|
|
903d0f73
|
2015-05-26T11:12:28
|
|
Removed -XCCLinker from MingW command line (thanks, Fredrik!).
Fixes Bugzilla #2707.
|
|
|
b11b3493
|
2015-05-26T11:08:30
|
|
Windows: don't beep on Alt-* key combos (Thanks, historic_bruno!).
Fixes Bugzilla 2669.
|
|
|
bb437f02
|
2015-05-26T11:01:19
|
|
Cocoa: report SDL_WINDOWEVENT_EXPOSED events to the app (thanks, David!).
Fixes Bugzilla #2644.
|
|
|
aba4d783
|
2015-05-26T10:25:15
|
|
Don't look for (and fail without) glGetIntegerv() until we need to.
Fixes Bugzilla #2615.
|
|
|
22704ac2
|
2015-05-26T09:55:41
|
|
Cocoa: don't fail outright if we see an unknown display format.
Just treat it as ARGB8888.
|
|
|
13487426
|
2015-05-26T06:32:19
|
|
Fixed bug 1392 - Debian patch: do not propagate -lpthread
|
|
|
2c4a6ea0
|
2015-05-26T06:27:46
|
|
Updated the copyright year to 2015
|
|
|
c133c8c8
|
2015-05-26T06:27:12
|
|
Script from Sylvain to automate updating the copyright year
|
|
|
d88b6738
|
2015-05-26T06:16:43
|
|
Fixed bug 2989 - Memory loss in clipboard_testClipboardTextFunctions
|
|
|
507157ab
|
2015-05-25T16:22:09
|
|
Wait for devices to finish initializing when inserted, before using them. Fixes hotplug issue with XBox 360 game controller.
|
|
|
b0c5e201
|
2015-05-25T14:52:41
|
|
Added support for Razer Serval Bluetooth mode
|
|
|
44719498
|
2015-05-23T09:33:20
|
|
Sign any subframeworks - not needed for SDL proper, but included for consistency with other projects
|
|
1968ff04
|
2015-05-22T22:34:08
|
|
Android: Fixed touch pressure being out of range.
According to the documentation of Android's MotionEvent, the getPressure() may
return values higher than 1 on some devices. To prevent passing such values into
SDL they are now corrected to 1 in Java before the JNI call (where it is assumed
to be correct).
Currently SDL only sends SDL_FINGERMOTION events if the touch state (position or
pressure) changed. By correcting pressure down to 1 some events may get dropped
in the rare case that only the pressure was changed but was out of range and the
position did not change.
|
|
|
4f00dda9
|
2015-05-21T21:27:53
|
|
Removed redundant NULL check in test program.
|
|
|
b00214c3
|
2015-05-21T21:25:32
|
|
Changed clean-up order in test program.
|
|
|
39d27a63
|
2015-05-21T21:25:14
|
|
Fixed undefined key access in test program.
|
|
|
63653814
|
2015-05-20T16:28:21
|
|
Added game controller support for the Razer Serval
|
|
|
2c4ad51d
|
2015-05-18T21:21:14
|
|
Added missing loop cancel for Emscripten in test programs.
|
|
|
ac10a2f2
|
2015-05-18T21:17:13
|
|
Fixed compiling test program with Emscripten.
|
|
|
1ed86d22
|
2015-05-18T21:12:16
|
|
Fixed handling only one event per frame in test program.
|
|
fcee3e73
|
2015-05-16T17:35:36
|
|
Code style cleanup in the GLES and GLES2 render backends.
|
|
|
240a3fbf
|
2015-05-16T16:55:56
|
|
iOS: Added support for SDL_DisableScreenSaver and SDL_EnableScreenSaver.
|
|
|
f7ac0201
|
2015-05-16T21:15:59
|
|
Android: Replaced logging tag strings with constant.
|
|
|
a54d038b
|
2015-05-16T21:15:27
|
|
Emscripten: Fixed wrong source comment and updated web link.
|
|
|
d57b9735
|
2015-05-16T12:05:42
|
|
Mac: Use CoreFoundation headers instead of Carbon headers, in GetPowerInfo code.
|
|
afc97cbd
|
2015-05-14T14:40:56
|
|
Fix duplicate raw mouse events with XInput2
Make XGrabPointer calls in X11_SetWindowGrab and X11_CaptureMouse
consistent by passing False to owner_mask along with proper event_mask.
|
|
|
1c6ea0f2
|
2015-05-13T22:39:32
|
|
Added a userdata parameter to SDL_SetWindowsMessageHook()
|
|
|
b2be9253
|
2015-05-13T22:39:27
|
|
Fixed Mac and Linux builds
|
|
|
7de242e7
|
2015-05-13T22:39:20
|
|
Added SDL_SetWindowsMessageHook() to facilitate full IME support on Windows
|
|
|
094c4a16
|
2015-05-13T22:37:26
|
|
Added generic xinput fallback for XBox compatible controllers on Linux
|
|
|
f1a7c00d
|
2015-05-11T21:03:36
|
|
Refactored SDL_EGL_CreateContext: It now supports context flags and OpenGL ES 3+ contexts, and its behavior more closely matches the GLX and WGL context creation code.
Improved the code style consistency of SDL_egl.c.
Fixes bugzilla #2865.
|
|
|
5919a859
|
2015-05-09T22:42:23
|
|
Android: Fixed empty parameter list in signatures of internal functions.
|
|
|
6e4e9ceb
|
2015-05-08T21:53:02
|
|
Fixed SDL_TouchFingerEvent documentation in header file.
|
|
|
8f4bcc72
|
2015-05-06T21:11:06
|
|
Android: Replaced spaces with tab in Android.mk file.
|
|
|
3f51758e
|
2015-05-06T21:10:48
|
|
Emscripten: Fixed touch coordinates not being normalized.
|
|
|
5715097a
|
2015-05-06T21:09:33
|
|
Fixed implicit function declaration in test program.
|
|
|
ac27b511
|
2015-05-06T12:54:51
|
|
Fixed SDL_GL_GetAttribute queries for framebuffer component sizes in Core Profile OpenGL contexts.
Fixes bugzilla #2060.
|
|
|
c4fe8c80
|
2015-05-06T12:42:14
|
|
Fixed building the iOS Demo files in debug mode
|
|
|
4fc40266
|
2015-05-05T19:01:55
|
|
Replaced all remaining uses of NSAutoreleasePool with @autoreleasepool blocks (bugzilla #2680.)
|
|
|
6c20b682
|
2015-05-05T16:24:05
|
|
Fixed a warning when SDL_syswm.h is included in code compiled for iOS with clang.
|
|
|
d1372bb9
|
2015-05-05T16:20:11
|
|
Fixed the window offset on iOS when resuming an app with a borderless or fullscreen window that has the on-screen keyboard visible.
|
|
|
d603bb30
|
2015-05-05T16:16:10
|
|
Fixed a crash on iOS when none of the orientations in Info.plist match the SDL window's actual orientation.
Fixes bug #2967.
|
|
|
74d83ead
|
2015-05-04T21:47:40
|
|
Fixed bug 2976 - Fix RGBA<->RGBA blit that was broken with the optimization from Bug 11
id.zeta
The optimization from Bug 11 added a code branch on cases where the source RGB masks match the destination RGB masks and a optimized blit function Blit4to4MaskAlpha that always overrides the source alpha info would be chosen. Unfortunately, the branch also errorneously took over the RGBA<->RGBA blitting cases where the source alpha info should be copied, while they would instead get overriden in Blit4to4MaskAlpha.
The attached patch fixes that by handling the RGBA<->RGBA cases correctly in that branch with the original BlitNtoNCopyAlpha as well as uses an optimized Blit4to4CopyAlpha along the same vein.
|
|
|
589c46dd
|
2015-05-01T01:20:28
|
|
X11: send keypress events before textinput events.
|
|
|
0045e2e6
|
2015-05-01T01:19:00
|
|
checkkeys: report SDL_TEXTEDITING events.
|
|
|
42326278
|
2015-05-01T01:12:48
|
|
checkkeys: Readded the KEYUP event test.
|
|
|
9979bab2
|
2015-04-30T21:45:29
|
|
Android: Deactivated debug log messages on joystick device events.
|
|
|
93bd476d
|
2015-04-26T20:46:07
|
|
Moved code signature step to after the framework build step is complete, and don't hardcode the codesign identity
|
|
|
931f6e4e
|
2015-04-26T20:21:06
|
|
Turn off code signing by default
Code signature can be added after build with the following command line:
codesign --force --sign 76BB5ACAC44CA5EFA5F879434D157B81DA842CFB SDL2.framework/Versions/A
|
|
|
43692a00
|
2015-04-26T15:47:40
|
|
Updated project and added code signing for release builds
|
|
d9d1a1b9
|
2015-04-26T13:53:46
|
|
X11: Use our own cut-buffer for intermediate clipboard storage.
XA_CUTBUFFER0 is not defined for holding UTF8 strings.
|
|
|
5c09bf73
|
2015-04-25T20:49:26
|
|
Only check for Linux-specific input APIs on Linux targets (thanks, Marcus!).
This is only for the configure script. The CMake project files already make
this Linux-exclusive.
Fixes Bugzilla #2659.
|
|
|
75a3ad1d
|
2015-04-22T20:25:19
|
|
CMake: Slightly better fix for installation target (thanks, Anthony!).
Fixes Bugzilla #2474.
|
|
|
034933d7
|
2015-04-22T21:43:22
|
|
Fixed typos in header file documentation comments.
|
|
|
71c65b15
|
2015-04-20T20:03:40
|
|
Fixed SDL_GameControllerOpen() and SDL_JoystickOpen() documentation in header.
|
|
|
83c088a0
|
2015-04-20T20:03:26
|
|
Fixed unsupported doxygen tag in header file.
|
|
|
69f6f646
|
2015-04-20T13:43:24
|
|
Cleaned up the macro salsa in the Windows timer code.
- Removed USE_GETTICKCOUNT code; it's never used now.
- Reduced the number of preprocessor checks for WinRT.
- Renamed timeSetPeriod() so it doesn't look like a Win32 API call.
|
|
|
b72938c8
|
2015-04-20T12:22:44
|
|
Windows: Always set the system timer resolution to 1ms by default.
An existing hint lets apps that don't need the timer resolution changed avoid
this, to save battery, etc, but this fixes several problems in timing, audio
callbacks not firing fast enough, etc.
Fixes Bugzilla #2944.
|
|
e5f9c347
|
2015-04-08T12:14:36
|
|
Remove trailing spaces in Android source code.
|
|
eff61ee3
|
2014-06-05T11:55:37
|
|
Add an entry for X11 "/" key for Brazilian keyboard.
SDL2 reports the following message when we type the "/" on br-abnt2 keyboards:
The key you just pressed is not recognized by SDL. \
To help get this fixed, please report this to the SDL mailing list \
<sdl@libsdl.org> X11 KeyCode 97 (89), X11 KeySym 0x2F (slash).
That's because the corresponding entry in the scancodes table is
marked with value SDL_SCANCODE_UNKNOWN.
This commit fixes that adding the value SDL_SCANCODE_SLASH for this entry.
|