Simon Hug
There's a chance that an audio conversion from many channels to a few can use more than 9 audio filters. SDL_AudioCVT has 10 SDL_AudioFilter pointers of which one has to be the terminating NULL pointer. The SDL code has no checks for this limit. If it overflows there can be stack or heap corruption or a call to 0xa.
Attached patch adds a function that checks for this limit and throws an error if it is reached. Also adds some documentation.
Test parameters that trigger this issue:
AUDIO_U16MSB with 224 channels at 46359 Hz
V
AUDIO_S16MSB with 6 channels at 27463 Hz
The fuzzer program I uploaded in bug 3667 has more of them.
This only affects Wayland and DirectFB, as a Unix system generally has X11
support. Other platforms also have different sizes for the C union in
question, but are likely the only target for that platform, etc.
Apps that might run on Wayland or DirectFB will need to be compiled against
new headers from an official 2.0.6 release, or be prepared to force the x11
target, or not use SDL_GetWindowWMInfo().
Fixes Bugzilla #3428.
It's easier for Visual Studio users that want this information to turn it on
or live without it, than it is to explain why every debugger that isn't Visual
Studio crashes out here. Eventually SetThreadDescription() will be the thing
everyone uses anyhow.
Fixes Bugzilla #3645.
(and several others).
Failing to check if a key was known to be pressed by SDL was causing
SDL_SendKeyboardKey to send duplicate key pressed events with the repeat
property set to true.
Fixes Bugzilla #3637.
The message sent upon the window being activated or deactivated, to trigger
the call to SDL_SetKeyboardFocus was missing a mandatory parameter. So
keyboard focus was never properly set.
Fixes Bugzilla #3658.
We don't fill buffers just to throw them away during shutdown now, we let the
AudioQueue free its own buffers during disposal (which fixes possible warnings
getting printed to stderr by CoreAudio), and we stop the queue after running
any queued audio during shutdown, which prevents dropping the end of the
audio playback if you opened the device with an enormous sample buffer.
Fixes Bugzilla #3555.
We need more than two buffers to flip between if they are small, or CoreAudio
won't make any sound; apparently it needs X milliseconds of audio queued when
it needs to play more or it drops any queued buffers. We are currently
guessing 50 milliseconds as a minimum, but there's probably a more proper
way to get the minimum time period from the system.
Fixes Bugzilla #3656.
"In particular, only one VkSurfaceKHR can exist at a time for a given window. Similarly, a native window cannot be used by both a VkSurfaceKHR and EGLSurface simultaneously"
CR: SamL
If an Emscripten app is in relative mouse mode and the user presses Escape
(or whatever is appropriate), then the pointer lock is broken by the browser.
This keeps track of those losses, and next time the user presses a mouse
button down on the canvas, if the app is still meant to be in relative mouse
mode, we will attempt to regrab the pointer.
This makes it much more seamless for things like first-person shooters, and
the app doesn't need any manual intervention.
Samuel Hopkins
Just confirming that the patch from Andreas (attachment 1715 [details]) works for me under SDL 2.0.3 with xmonad.
Stas Sergeev
Confirming that the patch in this ticket fixes the full-screen switching for dosemu2 on ubuntu-16.04. Note that I am not using xmonad, so this bug appears to be generic.
This gracefully recovers when a device format is changed, and will switch
to the new default device if the current one is unplugged, etc.
This does not handle when a new default device is added; it only notices
if the current default goes away. That will be fixed by implementing the
stubbed-out MMNotificationClient_OnDefaultDeviceChanged() function.
"ISO C90 forbids mixed declarations and code [-Werror=declaration-after-statement]"
Moving some variable declarations to the top of Android_SetScreenResolution()
* alsa hotplug thread is low priority
* give a chance for other threads to catch up when audio playback is not progressing
* use nonblocking for alsa audio capture
There is a bug with SDL hanging when an audio capture USB device is removed, because poll never returns
We will throw away the data anyhow, but some apps depend on the callback
firing to make progress; testmultiaudio.c, if nothing else, is an example
of this.
Capture also will now fire the callback in these conditions, offering nothing
but silence.
Apps can check SDL_GetAudioDeviceStatus() or listen for the
SDL_AUDIODEVICEREMOVED event if they want to gracefully deal with
an opened audio device that has been unexpectedly lost.
This is just enough to get you through a file that just used the extended
header for float or int data. It doesn't handle all the other things that
you expect from this header, like 24-bit samples inside a 32-bit container
or speaker masks.
This should remain binary compatible with Windows XP, as we dynamically
load anything we need and fall back to DirectSound/WinMM/XAudio2 if not
available.
Like other C runtimes, it should probably produce the string "(null)".
This bug probably only affected Windows, as most platforms use their standard
C runtime's snprintf().
Walter van Niftrik
We have found that since SDL 2.0.5 the audio callback thread is created with a very small stack size. In our application this is leading to stack overflows.
We believe there is a bug at http://hg.libsdl.org/SDL/file/391fd532f79e/src/audio/SDL_audio.c#l1132, where the is_internal_thread flag appears to be inverted.
Volumetric
In X11 the SDL error "Unknown touch device" can occur after which the application stops recognizing touch events. For a kiosk-type application this results in a hang as far as the user is concerned. This is reproducible on HP Z220/Z230/Z240 workstations by swapping USB cables for a while and it also occurs with no physical changes, probably due to USB device power management. A workaround is to make SDL re-enumerate the touch devices like it does at startup. A patch is attached.
Matthew
Its possible to set SDL_CaptureMouse() so you continue receiving mouse input while the mouse is outside your window. This works however There is then a gap where no messages send, which is when the mouse is hovering the title bar and the window edges.
X11 seemed to be confused by the broad definition, so WEIGHT_NAME,
SLANT and SETWIDTH_NAME were defined, thus fixing the font lookup
on some systems (tested on Mageia 6 with X11 1.19.1).
Fixes bug 3571.
Tom Seddon
GL_ActivateRenderer may call GL_UpdateViewport, which leaves the GL_PROJECTION matrix selected. But after GL_ResetState, the GL_MODELVIEW matrix is selected, suggesting that's the intended default state.
It seems at least like these should be consistent. Presumably GL_UpdateViewport should be doing a glMatrixMode(GL_MODELVIEW) before it finishes.
Also updates the naming of these Xbox Wireless Controllers connected via USB (and thus the third-party Xbox Controller Driver) to match.
The Xbox Wireless Controller entries are now listed, in order, via USB, bia Bluetooh (with older firmware) and via Bluetooth (with firmware 3.1.1221.0).
This is a bleeding edge API, added to Windows 10 Anniversary Edition (build
1607, specifically).
https://msdn.microsoft.com/en-us/library/windows/desktop/mt774976(v=vs.85).aspx
Nothing supports this yet, including WinDbg, Visual Studio, minidumps, etc,
so we still need to also use the RaiseException hack. But presumably tools
will use this API as a more robust and universal way to get thread names
sooner or later, so we'll start broadcasting to it now.
Firmware revision 3.1.1221.0 changes the mapping of the Xbox One S
controller in Bluetooth mode. Aside from changing the layout of
other buttons, this revision also changes the triggers to act as
Accelerator and Brake axes from the simulation controls page.
The Darwin sysjoystick code didn't previously map anything at these
axes, making it impossible to detect input on these two buttons.
This defaults to the internal SDL resampler, since that's the likely default
without a system-wide install of libsamplerate, but those that need more can
tweak this.
This currently favors libsamplerate over the fast path (quality over speed),
but I'm not sure that's the correct approach, as there may be surprising
changes in performance metrics depending on what packages are available on
a user's system. That being said, currently, the only thing with access to
SDL_AudioStream is an SDL audio device's thread, and it might be mostly idle
otherwise, so maybe this is generally good.
Turns out that iterating from 0 to channels-1 was a serious performance hit!
These cases now tend to match or beat the original audio resampler's speed!
This allows us to avoid an extra copy, allocate less memory and reduce cache
pressure. On the downside: we have to do a lot of tapdancing to resample the
buffer in reverse when the output is growing.
It's expensive and (hopefully) unnecessary. If this becomes an overflow
problem, we could multiply both values by 0.5f before adding them, but let's
see if we can get by without the extra multiplication first.
We never seem to overflow the source buffer now; this might have been a
leftover from a bug that was covered by Vitaly's fixes?
Removing this conditional makes the resampler 10-20% faster. Left an
assert in there for debug builds, in case this still happens.
white.magic
The logic which decides if a device enumerated via the direct input system in the function EnumJoysticksCallback in SDL_dinputjoystick.c is processed is discarding valid joystick devices due to the assumption that devices of the type DI8DEVTYPE_SUPPLEMENTAL are not valid devices.
This change was added with 2.0.4 with this commit http://hg.libsdl.org/SDL/rev/1b9d40126645 that is linked to this bug report https://bugzilla.libsdl.org/show_bug.cgi?id=2460 which indicates that in that case devices of the type DI8DEVTYPE_SUPPLEMENTAL were not desirable as they caused a singular device to emit multiple "device added" events.
Since then there appear to have been a few fixes to handle devices that fall into various other classes in the following two commits:
http://hg.libsdl.org/SDL/rev/10ffb4787d7a and http://hg.libsdl.org/SDL/rev/6a2bbac05728
Two devices I have reports of failing to be listed when the DI8DEVTYPE_SUPPLEMENTAL type is excluded are ECS Gametric Throttle and Thrustmaster MFD Cougar.
Sam Lantinga
I verified that the OUYA controller shows up as a single device with this change, so I've reverted the change to ignore supplemental devices, leaving framework in place to easily add devices that we want to ignore.
Removed some needless things ("len / sizeof (Uint8)"), and made sure the
int32 -> float code uses doubles to avoid working with large integer values
in a 32-bit float.
Mark Callow
The attached patch does the following for the X11 and Windows platforms, the only ones where SDL attempts to use context_create_es_profile:
- Adds SDL_HINT_OPENGL_ES_DRIVER by which the application can
say to use the OpenGL ES driver & EGL rather than the Open GL
driver. (For bug #2570)
- Adds code to {WIN,X11}_GL_InitExtensions to determine the maximum
OpenGL ES version supported by the OpenGL driver (for bug #3145)
- Modifies the test that determines whether to use the OpenGL
driver or the real OpenGL ES driver to take into account the
hint, the requested and supported ES version and whether ES 1.X
is being requested. (For bug #2570 & bug #3145)
- Enables the testgles2 test for __WINDOWS__ and __LINUX__ and adds
the test to the VisualC projects.
With the fix in place I have run testdraw2, testgl and testgles2 without any issues and have run my own apps that use OpenGL, OpenGL ES 3 and OpenGL ES 1.1.
Lukasz Biel
Tried to compile SDL2 using newest version of VS.
Got:
SDL_audiocvt.obj : error LNK2019: unresolved external symbol memcpy referenced in function SDL_ResampleCVT
1>E:\Users\dotPo\Lib\SDL\VisualC\x64\Release\SDL2.dll : fatal error LNK1120: 1 unresolved externals
whole compilation process: http://pastebin.com/eWDAvBce
Steps to reproduce:
clone http://hg.libsdl.org/SDL using tortoise hg,
open SDL\VisualC\SDL.sln,
when promted if should retarget solution click ok,
select release x64 build type,
Build/Build Solution
attempt 2, using Visual Studio cmake support:
open folder SDL\
select release x64 build type,
run CMake\Build CMakeLists.txt
build fails
When switched to debug build type, buils succeeds in both cases.
VS 2017 is still beta.
It causes audio pops if you're converting in chunks (and needs to
allocate/initialize/free on each convert). We'll either adjust this interface
when we break ABI for 2.1 to make this usable, or publish the SDL_AudioStream
API for those that want a streaming solution.
In the meantime, the "simple" resampler produces "good enough" audio without
pops and doesn't have to be initialized, so that'll do for now on the
SDL_AudioCVT interface.
K_NORMTAB, K_SHIFTTAB, K_ALTTAB, K_ALTSHIFTTAB
In the normal case we'll load all the keymaps from the kernel, but this reduces the size of the SDL library for the fallback case when we can't get to the tty.
Mark Logan 2015-08-24 15:57:50 UTC
In SDL_windowsopengles.c, WIN_GLES_SetSwapInterval is as follows:
WIN_GLES_SetSwapInterval(_THIS, int interval)
{
/* FIXME: This should call SDL_EGL_SetSwapInterval, but ANGLE has a bug that prevents this
* from working if we do (the window contents freeze and don't swap properly). So, we ignore
* the request for now.
*/
SDL_Log("WARNING: Ignoring SDL_GL_SetSwapInterval call due to ANGLE bug");
return 0;
}
With a recent version of ANGLE (early July) calling SDL_EGL_SetSwapInterval with a D3D11 backend appears to work just fine. I am working on testing this with D3D9.
--
Alex Szpakowski
I found the bug, it was fixed in 2013. https://bugs.chromium.org/p/angleproject/issues/detail?id=481
In my opinion it should be safe to unconditionally use SetSwapInterval now. Anyone who encounters the bug should update their ANGLE to a version less than 3 years old, especially since they'd be using a SDL version that's 3+ years newer than their ANGLE version.
Rob
I've ran into an issue where I successfully receive SDL_KEY[UP,DOWN] events but not SDL_TEXTINPUT or SDL_TEXTEDITING. In my case the code in SDL_EVDEV_do_text_input() is returning early (on error) prior to calling SDL_SendKeyboardText(). I'm running on the RaspberryPi 3, without X11.
In SDL_EVDEV_do_text_input() there is a condition to check keysyms with a type value below 0xf0, then subtract 0xf0 from type. Without understanding the purpose of this code, I disabled it, recompiled, and I'm getting correct SDL_TEXTINPUT events. I'm going to guess that my hack/fix is going to be problematic in some other environment, but after some initial testing it looks like everything is running fine in my setup.
Richard Russell
Resuming from a suspended state results in a black screen. This only happens when using GLES 1.1 (GLES 2 resumes correctly) and when the render target has been changed using SDL_SetRenderTarget. This problem is new in 2.0.4.
The attached test case demonstrates the issue.
Sylvain Becker has apparently found a fix as follows:
"In the opengles leaf function (in 'src/render/opengles/SDL_render_gles.c'), it appears there is a call to 'GLES_ActivateRenderer' in 'GLES_SetRenderTarget', which is not present in opengles2. When commenting out this 'GLES_ActivateRenderer', it seems to resume fine".
This appears to fix the testcase perfectly, but I don't know whether it could have any undesirable side-effects.
tvc
I believe this patch should fix it, instead of looping through all the tty's and seemingly selecting the wrong one and corrupting the console I've just made SDL open /dev/tty which is the console attached to the current process anyway.
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"...
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.
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!).
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.
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.
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.
Ryan C. Gordon
Philipp Wiesemann
Sam Lantinga
Alex Szpakowski
Olli Kallioinen
Bastien Bouclet
Charlie Birks
Patrick Monaghan
Joshua Granick
Juha Kuikka
Drew Bliss
James Legg
Brandon Schaefer
David Ludwig
R?mi Verschelde
Jessica Stokes
Misty De Meo