|
577ef5e0
|
2014-11-02T10:38:29
|
|
WinRT: fixed crash when trying, and failing, to load the opengles2 renderer
The crash would occur when a WinRT app explicitly tried to create an
SDL_Renderer using the "opengles2" renderer (via SDL_HINT_RENDER_DRIVER), but
OpenGL ES 2 / ANGLE .dlls weren't packaged in the app.
|
|
|
18ecc34d
|
2014-11-02T09:02:01
|
|
WinRT: updated an OpenGL-related code-comment
|
|
|
0bcc254e
|
2014-11-02T08:47:05
|
|
WinRT: expanded OpenGL ES 2 support to enable recent updates to ANGLE/WinRT
ANGLE for WinRT has at least two versions:
- an older version, which supports Windows 8.0 and 8.1. This is currently
the "winrt" branch in MSOpenTech's ANGLE repository (at
https://github.com/msopentech/angle)
- a newer version, which drops support for Windows 8.0, but is under more
active development (via MSOpenTech's "future-dev" branch), and which was
recently merged into the ANGLE project's official "master" branch
(at https://chromium.googlesource.com/angle/angle)
Both versions are setup using slightly different APIs. SDL/WinRT will now
attempt to detect which version is being used, and configure it appropriately.
|
|
43ad9670
|
2014-10-30T13:44:31
|
|
Patched to compile on pre-C99 compilers.
|
|
3a916c6d
|
2014-10-29T22:55:40
|
|
X11: Fixed SDL_GetGlobalMouseState() to work better with multimonitor configs.
Fixes Bugzilla #2770.
Thanks to Epic Games for contributing this fix.
|
|
37720dd5
|
2014-10-29T22:44:35
|
|
Fixed gamecontroller database sorting script not sorting the whole database.
The script was originally written for an SDL_gamecontrollerdb.h which had groups
of mappings separated by single lines starting with a "#". This was changed some
time ago to #endif/#if pairs. Because the script assumed only a single #endif in
the whole file it no longer worked correctly and only processed the first group.
|
|
|
0c60f023
|
2014-10-29T22:41:01
|
|
Removed unnecessary NULL check.
|
|
|
2bdc5c9d
|
2014-10-29T20:35:21
|
|
Updated gamecontroller database sorting script to work with Python 3.
|
|
|
032b14e7
|
2014-10-29T20:29:32
|
|
Fixed SDL_AddHintCallback() crashing if no more memory available.
The return value of SDL_malloc() was not checked and NULL therefore not handled.
Also added setting of error message for the other SDL_malloc() in this function.
|
|
|
2a46a9df
|
2014-10-29T20:20:47
|
|
Fixed bug 2647 - Memory leak in SDL_AddHintCallback function - SDL_hints.c
Nitz
Variable entry going out of scope leaks the storage it points to, at:
/* Need to add a hint entry for this watcher */
hint = (SDL_Hint *)SDL_malloc(sizeof(*hint));
if (!hint) {
return;
}
Patch is attached.
|
|
|
7268eabb
|
2014-10-26T23:28:45
|
|
Added handling of NULL as input for SDL_GameControllerAddMapping().
|
|
|
33a2b58c
|
2014-10-26T23:22:53
|
|
Fixed SDL_GameControllerMappingForGUID() crashing if no more memory available.
The return value of SDL_malloc() was not checked and NULL therefore not handled.
NULL returned by SDL_GameControllerMapping()/SDL_GameControllerMappingForGUID()
now either means "no mapping" (as before) or "no memory" (just crashed before).
|
|
|
30d6cec1
|
2014-10-26T17:53:16
|
|
Added handling of NULL as input for SDL_GameControllerMapping().
For consistency with the similar functions getting SDL_GameController as input.
Also NULL is no SDL_GameController and therefore can not have a mapping anyway.
|
|
|
90387daf
|
2014-10-26T17:46:11
|
|
Removed two wrong documentation comments from gamecontroller source.
No replacement because correct documentation is already in SDL_gamecontroller.h.
|
|
|
0a053ed6
|
2014-10-26T17:44:00
|
|
Fixed wording in SDL_GameControllerAddMappingsFromRW() error message.
|
|
|
01f3498c
|
2014-10-26T12:33:10
|
|
WinRT: more keyboard code cleanups
- made keycode listings a bit easier to read
- listed VK_ keycode names, where appropriate
- removed a few pieces of dead + commented-out code
- applied a common 'WINRT_' prefix to internal function names
|
|
|
e057c7dd
|
2014-10-25T08:54:23
|
|
WinRT: Fixed bug whereby key-detection code could call SDL_Log
|
|
|
ed2fff6b
|
2014-10-25T08:50:41
|
|
WinRT: fallback to scancode key detection, if VirtualKey/VK detection fails
|
|
|
20b19aac
|
2014-10-25T08:43:51
|
|
WinRT: more keyboard code cleanups
- use SDL_arraysize where appropriate
- made long lines of debug-logging code span multiple lines, for easier reading
|
|
|
a1bac50d
|
2014-10-24T22:09:47
|
|
WinRT: Fixed bug whereby a right-shift key might get reported as left-shift
|
|
|
e9c79620
|
2014-10-24T21:25:21
|
|
WinRT: minor keyboard code cleanup
An internally-used hash table was replaced with a plain switch statement.
|
|
|
d5d3721b
|
2014-10-24T19:50:29
|
|
Added new key codes from Android 4.4 (API 20) and 5.0 (API 21).
|
|
a1dcebb2
|
2014-10-24T06:52:54
|
|
Remove unused variable.
|
|
|
04dbc492
|
2014-10-24T06:50:09
|
|
Merge patch for bug 2750. Thanks Sylvain!
|
|
|
5f193f0c
|
2014-10-20T22:19:09
|
|
Fixed setting of screen saver state crashing on some version of Android.
Setting Window flags seems to affect Views and must be handled on UI thread.
|
|
79035b39
|
2014-10-20T10:10:39
|
|
Bug 2739 - [Android] No support for SDL_DisableScreenSaver by Martin Gerhardy
|
|
|
ec4dfdfc
|
2014-10-17T23:36:45
|
|
Fixed bug 2476 - Allow custom main() arguments
rettichschnidi
I would like to pass custom arguments from my Java code (subclass of SDLActivity) to the native SDL2 binary.
|
|
|
7cabaebe
|
2014-10-17T23:26:02
|
|
Fixed typo in source comment.
|
|
38f94e21
|
2014-10-16T09:54:13
|
|
Added a platform cleanup function for the Vivante video driver
|
|
|
5df970a6
|
2014-10-16T09:41:42
|
|
[udev] Fixes #2654, ID_INPUT_KEY devices are ignored.
Also added some references to udev code where these flags are set.
|
|
|
29eb6e9b
|
2014-10-15T18:33:43
|
|
[X11] Obey DISPLAY environment variable when selecting screen (Steaphan Greene)
Discussed in bug #2192, tested with Xvfb in dual head configuration.
|
|
|
1b341f58
|
2014-10-15T15:50:35
|
|
WinRT build fix
|
|
|
251ca855
|
2014-10-15T09:18:17
|
|
Expanded the iMX6 video driver into a general Vivante video driver that works across multiple SoCs
|
|
|
f7abb7c4
|
2014-10-15T09:14:09
|
|
Better error reporting for video drivers with less than full functionality.
|
|
|
3b70f65d
|
2014-10-15T09:13:36
|
|
Fixed bug where EGL function pointers wouldn't get saved in optimized gcc builds
|
|
|
348cf3ee
|
2014-10-15T09:11:15
|
|
Fixed spurious SDL error on shutdown if mouse capture isn't supported
|
|
|
89127c2d
|
2014-10-15T09:10:50
|
|
Fixed input device detection on systems that aren't running the udev daemon
|
|
|
639b39c4
|
2014-10-15T09:09:57
|
|
Fixed crash if the event subsystem didn't initialize properly
|
|
|
20a6c623
|
2014-10-14T09:53:46
|
|
WinRT: fixed bug whereby offscreen-rendered content could get improperly rotated
Attributes on the host device's rotation were getting applied to offscreen
textures in an invalid manner. This was causing some apps to look different,
depending on how the device was rotated.
|
|
|
72f703e3
|
2014-10-13T14:10:41
|
|
Fixed log typo
|
|
|
557d43e6
|
2014-10-13T10:46:27
|
|
Minor cleanups
|
|
|
47658057
|
2014-09-18T11:03:34
|
|
[Android] Better fix for #2480, pause/resume audio
|
|
|
5f9ea7ed
|
2014-09-17T14:49:36
|
|
Haptic: Deal with negative periodic magnitudes (thanks, Elias!).
A negative periodic magnitude doesn't exist in Windows' and MacOS' FF APIs
The periodic magnitude parameter of the SDL Haptic API is based on the Linux
FF API, so it means they are not directly compatible:
'dwMagnitude' is a 'DWORD', which is unsigned.
Fixes Bugzilla #2701.
|
|
|
4544343b
|
2014-09-17T11:41:12
|
|
[Android] Fixes #2480, music does not pause when process backgrounded
This modifies SDL_PauseAudio behavior to pause all audio devices instead of
just the default one (required on Android, at least for testmultiaudio on my
Nexus 4 which reported 2 audio devices).
It also changes SDL_PauseAudioDevice to retain the device lock from pause until
resume in order to save battery in mobile devices.
|
|
|
5f39ea89
|
2014-09-13T02:15:18
|
|
Fixed bug 2415 - Message Boxes aren't implemented on Android
Philipp Wiesemann
I attached a patch for an incomplete implementation of the messagebox parts.
It was not tested on lots of devices yet and features a very fragile workaround to block the calling SDL thread while the dialog is handled on Android's UI thread. Although it works for testmessage.c I assume there are lot of situations were it may fail (standby, device rotation and other changes). Also not all flags and colors are implemented.
On the other hand most uses of the messagebox are to show an error on start and fragility (or working at all) may not matter there.
|
|
|
2bfb3c3e
|
2014-09-12T15:09:33
|
|
Fixed using uninitialized display data in iMX6 initialization
|
|
24c86b55
|
2014-09-11T19:24:42
|
|
[X11] Reconcile logical keyboard state with physical state on FocusIn
since the window system doesn't do it for us like other platforms.
This prevents sticky keys and missed keys when going in and out
of focus, for example Alt would appear to stick if switching away
from an SDL app with Alt-Tab and had to be pressed again.
CR: Sam
|
|
7798b4d6
|
2014-09-09T17:26:16
|
|
video: waylandmouse: handle mmap error properly
Signed-off-by: Ryo Munakata <ryomnktml@gmail.com>
|