Ozkan Sezer
Since changeset 11607:60cd425a2f14, I am getting the following
error upon quit. Running testsprite2, clicking the mouse, and
quiting it is enough to trigger it. This is on my old Fedora9
x86-Linux:
X Error of failed request: BadCursor (invalid Cursor parameter)
Major opcode of failed request: 2 (X_ChangeWindowAttributes)
Resource id in failed request: 0xb057340
Serial number of failed request: 905
Current serial number in output stream: 906
Reverting https://hg.libsdl.org/SDL/rev/60cd425a2f14 removes
the error.
The audioqueue thread needs to keep running, and processing the CFRunLoop
until the AudioQueue is disposed of, otherwise CoreAudio will hang waiting for
final data to feed the device.
At least, I think this is how it all works. It definitely fixes the bug here!
Since AudioQueueDispose() calls AudioQueueStop() internally, there's no need
for our thread to handle this, either, which is good because the AudioQueue
would be disposed by this point. So now the AudioQueue is disposed first, and
then our thread is joined, and everything works out okay.
Just in case, we mark the device "paused" before setting everything in motion,
so any further callbacks from CoreAudio will write silence and not fire the
app's audio callback again.
Fixes Bugzilla #3868.
Trent Gamblin
The documentation for SDL_TouchFingerEvent says that the x and y coordinates are normalised between 0-1. I've found that to be true on Windows, Android and iOS but on X11 they are in pixel coordinates. This patch fixes the issue. This was the cleanest way I could do it with what was available without changing things around a lot but you may know a better way.
(I thought padding size ranged from 5 frames to ~30 frames (based around
RESAMPLER_ZERO_CROSSINGS, which is 5), but it's actually between 512 and
several thousands (based on RESAMPLER_SAMPLES_PER_ZERO_CROSSING)). It gets
big fast when downsampling.
Manuel
I would like this small patch merged that adds support for my GreenAsia Inc. PSX to USB converter, so SDL_IsGameController() returns true when using this adaptor.
It's interesting because PSX/PS2 controllers connected using this model won't be detected as gamecontrollers otherwise, only as joysticks.
Sylvain
There are various YUV-RGB conversion coefficients, according to https://www.fourcc.org/fccyvrgb.php
I choose the first (from Video Demystified, with integer multiplication),
but the current SDL2 Dither functions use in fact the next one, which follows a specifications called CCIR 601.
Here's a patch to use the second ones and with previous warning corrections.
There are less multiplications involved because Chroma coefficient is 1.
Also, doing float multiplication is as efficient with vectorization.
In the end, the YUV decoding is faster: ~165 ms vs my previous 195 ms.
Moreover, if SDL2 is compiled with -march=native, then YUV decoding time drops to ~130ms, while older ones remains around ~220 ms.
For information, from jpeg-9 source code:
jpeg-9/jccolor.c
* YCbCr is defined per CCIR 601-1, except that Cb and Cr are
* normalized to the range 0..MAXJSAMPLE rather than -0.5 .. 0.5.
* The conversion equations to be implemented are therefore
* Y = 0.29900 * R + 0.58700 * G + 0.11400 * B
* Cb = -0.16874 * R - 0.33126 * G + 0.50000 * B + CENTERJSAMPLE
* Cr = 0.50000 * R - 0.41869 * G - 0.08131 * B + CENTERJSAMPLE
jpeg-9/jdcolor.c
* YCbCr is defined per CCIR 601-1, except that Cb and Cr are
* normalized to the range 0..MAXJSAMPLE rather than -0.5 .. 0.5.
* The conversion equations to be implemented are therefore
*
* R = Y + 1.40200 * Cr
* G = Y - 0.34414 * Cb - 0.71414 * Cr
* B = Y + 1.77200 * Cb
Sylvain
Few issues with YUV on SDL2 when using odd dimensions, and missing conversions from/back to YUV formats.
1) The big part is that SDL_ConvertPixels() does not convert to/from YUV in most cases. This now works with any format and also with odd dimensions,
by adding two internal functions SDL_ConvertPixels_YUV_to_ARGB8888 and SDL_ConvertPixels_ARGB8888_to_YUV (could it be XRGB888 ?).
The target format is hard coded to ARGB888 (which is the default in the internal of the software renderer).
In case of different YUV conversion, it will do an intermediate conversion to a ARGB8888 buffer.
SDL_ConvertPixels_YUV_to_ARGB8888 is somehow redundant with all the "Color*Dither*Mod*".
But it allows some completeness of SDL_ConvertPixels to handle all YUV format.
It also works with odd dimensions.
Moreover, I did some benchmark(SDL_ConvertPixel vs Color32DitherYV12Mod1X and Color32DitherYUY2Mod1X).
gcc-6.3 and clang-4.0. gcc performs better than clang. And, with gcc, SDL_ConvertPixels() performs better (20%) than the two C function Color32Dither*().
For instance, to convert 10 times a 3888x2592 image, it takes ~195 ms with SDL_ConvertPixels and ~235 ms with Color32Dither*().
Especially because of gcc vectorize feature that optimises all conversion loops (-ftree-loop-vectorize).
Nb: I put no image pitch for the YUV buffers. because it complexify a little bit the code and the API :
There would be some ambiguity when setting the pitch exactly to image width:
would it a be pitch of image width (for luma and chroma). or just contiguous data ? (could set pitch=0 for the later).
2) Small issues with odd dimensions:
If width "w" is odd, luma plane width is still "w" whereas chroma planes will be "(w + 1)/2". Almost the same for odd h.
Solution is to strategically substitute "w" by "(w+1)/2" at the good places ...
- In the repository, SDL_ConvertPixels() handles YUV only if yuv source format is exactly the same as YUV destination format.
It basically does a memcpy of pixels, but it's done incorrectly when width or height is odd (wrong size of chroma planes). This is fixed.
- SDL Renderers don't support odd width/height for YUV textures.
This is fixed for software, opengl, opengles2. (opengles 1 does not support it and fallback to software rendering).
This is *not* fixed for D3D and D3D11 ... (and others, psp ?)
Only *two* Dither function are fixed ... not sure if others are really used.
- This is not possible to create a NV12/NV12 texture with the software renderer, whereas other renderers allow it.
This is fixed, by using SDL_ConvertPixels underneath.
- It was not possible to SDL_UpdateTexture() of format NV12/NV21 with the software renderer. this is fixed.
Here's also two testcases:
- that do all combination of conversion.
- to test partial UpdateTexture
Aaron
As of 2.0.6, all of my games are failing with the following error:
process 31778: arguments to dbus_type_is_basic() were incorrect, assertion "dbus_type_is_valid (typecode) || typecode == DBUS_TYPE_INVALID" failed in file dbus-signature.c line 322.
This is normally a bug in some application using the D-Bus library.
D-Bus not built with -rdynamic so unable to print a backtrace
(patch by Ozkan Sezer)
Evgeny Kapun
Commit 490bb5b49f11 [1], which was a fix for bug #3790, introduced a new bug: now, calling SDL_FreeSurface(surface) deallocates surface->map even if there are other references to the surface. This is bad, because some functions (such as SDL_ConvertSurface) assume that surface->map is not NULL.
Robert Turner
SDL_windowsevents.c contains code to retrieve the x and y coordinate for a requested hit test. It does this as follows:
POINT winpoint = { (int) LOWORD(lParam), (int) HIWORD(lParam) };
LOWORD(lParam) does not correctly mask off high bits that are set if the point is on a second (or third, etc.) monitor. This effectively offsets the x-coordinate by a large value.
MSDN documentation suggests that LOWORD() and HIWORD() are the wrong macros for the task, instead suggesting we should be doing something like the following:
POINT winpoint = { GET_X_LPARAM(lParam), GET_Y_LPARAM(lParam) };
Testing this change on my Windows 10 machine with 2 monitors gives the correct results.
- Fixing rendering borderless window. Need to force windows to send a WM_NCCALCSIZE then return 0 for non-client area size.
- Adding WS_CAPTION | WS_SYSMENU | WS_MINIMIZEBOX to borderless windows, for reasons noted in comments.
- Fix SetupWindowData() setting SDL_WINDOW_BORDERLESS. This was being cleared at window creation, causing hanlding for the first WM_NCCALCSIZE message to fail
Previously, the padding was silence, which was a problem when streaming since
you would sample a little bit of this silence between each buffer.
We still need a means to get padding data for the right hand side, but this
patch makes the resampler output more correct.
Anthony
This is what's making the software renderer crash with rotated destination rectangles of w or h = 0:
SDL_SetHint(SDL_HINT_RENDER_SCALE_QUALITY, "2");
This time it's using real math from a real whitepaper instead of my previous
amateur, fast-but-low-quality attempt. The new resampler does "bandlimited
interpolation," as described here: https://ccrma.stanford.edu/~jos/resample/
The output appears to sound cleaner, especially at high frequencies, and of
course works with non-power-of-two rate conversions.
There are some obvious optimizations to be done to this still, and there is
other fallout: this doesn't resample a buffer in-place, the 2-channels-Sint16
fast path is gone because this resampler does a _lot_ of floating point math.
There is a nasty hack to make it work with SDL_AudioCVT.
It's possible these issues are solvable, but they aren't solved as of yet.
Still, I hope this effort is slouching in the right direction.
This would cause playback problems in certain situations, such as on the
Raspberry Pi. The device that the wait was added for seems to not benefit from
it in modern times, and standard desktop Linux seems to do the right thing
when a USB device is unplugged now, without this patch.
Fixes Bugzilla #3599.
Andreas Falkenhahn
My app opens a 640x480 window. When I click on the window's maximize button, the window correctly fills the entire screen and loses its borders. But clicking on the restore button now doesn't restore the window to its original 640x480 size. Instead, the window size is identical to the screen size now. The only difference to the previous state is that the window now has borders again but it isn't restored to 640x480.
bastien.bouclet
The window is now resized to its specified size, but it moves to the top left corner of the screen. That is unexpected because neither the user nor the program moved it there. Test program attached (the same one as before).
Simon Hug
When RWops seeks with fseek or fseeko it uses the types long or off_t which can be 32 bits on some platforms. stdio_seek does not check if the 64-bit integer for the offset fits into a 32-bit integer. Offsets equal or larger than 2 GiB will have implementation-defined behavior and failure states would be very confusing to debug.
The attached patch adds range checking by using the macros from limits.h for long type and some bit shifting for off_t because POSIX couldn't be bothered to specify min and max macros.
It also defines HAVE_FSEEKI64 in SDL_config_windows.h so that the Windows function gets picked up automatically with the default config.
And there's an additional error message for when ftell fails.
Andreas Falkenhahn
When compiling SDL for the Raspberry Pi, I have to use the --host parameter to enable compilation of the native Raspberry Pi video driver, like so:
--host=arm-raspberry-linux-gnueabihf
It took me a while to figure out that this was necessary in order to have the native Raspberry Pi video driver compiled in. I think it would be better if there was an option like --enable-video-rpi that could be passed to configure and that would also show up when saying configure --help. Currently, it?s rather difficult to figure out that you have to use the --host parameter with arm-raspberry-linux-gnueabihf in order to get Raspberry Pi video support. It?s also somewhat inconsistent because most other video drivers can in fact be enabled/disabled through specific configure parameters but there is no such parameter for the native Raspberry Pi video driver.
Simon Hug
These are the remaining compiler warnings I see in the current tip cb049cae7c3c.
- SDL_test_log.c defines _CRT_SECURE_NO_WARNINGS without checking if it was already set.
- SDL_windowskeyboard.c converts integers to pointers without going over the (U)INT_PTR types. That bothers MSVC.
Now we try the new (hardware-specific) pathnames first, and if those fail to
load, we'll try the more generic names that earlier versions of Raspbian used.
Fixes Bugzilla #3800.
Mart?n Golini
I'm having a very slow initialization of the video subsystem that locks the window creation for about 500 ms ( tested in at least 4 different systems ). What i found is that X11_InitModes_XRandR is using XRRGetScreenResources, that explicitly ask to poll the hardware for changes. This is not really necessary since if the data is already available you can use XRRGetScreenResourcesCurrent.
I attached a tentative patch that fix this issue. With the patch there's no lock when the subsystem is initialized and the window creation is instant in my applications. The patch only uses XRRGetScreenResourcesCurrent in X11_InitModes_XRandR but it could be potentially used in X11_GetDisplayModes and X11_SetDisplayMode.
bastien.bouclet
When creating two surfaces and blitting them onto the other, SDL's internal reference counting fails, and one of the surfaces is not freed when calling SDL_FreeSurface.
Example code :
SDL_Surface *s1 = SDL_CreateRGBSurfaceWithFormat(0, 640, 480, 32, SDL_PIXELFORMAT_ARGB8888);
SDL_Surface *s2 = SDL_CreateRGBSurfaceWithFormat(0, 640, 480, 32, SDL_PIXELFORMAT_ARGB8888);
SDL_BlitSurface(s1, NULL, s2, NULL);
SDL_BlitSurface(s2, NULL, s1, NULL);
SDL_FreeSurface(s2);
SDL_FreeSurface(s1);
With this example, s1 is not freed after calling SDL_FreeSurface, its refcount attribute is still positive.
This also seems to fix the follow-up issue in bug #3719, whereby the initial fix caused the SDL window to move, after transitioning from fullscreen to windowed-mode
This is necessary because the Raspberry Pi is a strange beast, that believes
it has OpenGL support (through glX?) but generally has GLES2 support.
So when using the raspberry video target, we need to force this to default
to a GLES2 context, or by default SDL_CreateWindow() will fail, deep down
when it tries to load the proper GL library.
Fixes testsprite2 (and basically everything else that wasn't testgles2) when
run on a Raspberry Pi without a X server.
Please note that other targets might also need this filled in, the Raspberry
Pi is just the most prominent and readily-available System-On-A-Chip style
thing on my desk. :)
Romain Tisserand
Using KMS/DRM driver from WIP SDL2.0.6 on Linux/ARM SoC RockChip RK3328 (ARM Mali 450 MP2 GPU).
The current code is using GBM_BO_FORMAT_XRGB8888 as GBM buffer format specifier.
The Mali driver (it has been confirmed some other vendor implementations too) expects GBM_FORMAT_XRGB8888.
The Mesa implementation is actually handling both values as the same, but it's not implemented like this into every gbm.h vendor header.
https://github.com/ideak/mesa/blob/master/src/gbm/backends/dri/gbm_dri.c
So with stock SDL2 on my card (Mali vendor implementation), it does not work, eglCreateWindowSurface fails, and gbm_is_format_supported fails too (with the BO variant).
It runs fine with GBM_FORMAT_XRGB8888.
Here is a link of the gbm.h from Mali user-space driver :
https://github.com/rockchip-linux/libmali/blob/rockchip/include/gbm.h
This is necessary because we need to see if GLES compat extensions exist.
All of this code (including ShouldUseTextureFramebuffer()) should be
revisited after 2.0.6 ships; ideally we don't make throwaway contexts if
we can avoid it...but maybe we can't. I hear Vulkan is pretty cool.
Fixes Bugzilla #3725.
benjamin.feng
Probable underlying cause: https://bugzilla.libsdl.org/show_bug.cgi?id=3124#c5
"If you download and build the HID Calibrator sample you can see that these are totally legitimate HID devices (except for inverting the Y-axis of joysticks, which is contrary to the HID specification but does make them more compatible with games compiled expecting XBOX controllers)."
Colin Barrett
Using the pre-built x86 devel libs from here:
https://www.libsdl.org/release/SDL2-devel-2.0.5-VC.zip
If I have:
SDL_GL_SetAttribute(SDL_GL_CONTEXT_PROFILE_MASK, SDL_GL_CONTEXT_PROFILE_ES);
SDL_GL_SetAttribute(SDL_GL_CONTEXT_MAJOR_VERSION, 2);
SDL_GL_SetAttribute(SDL_GL_CONTEXT_MINOR_VERSION, 0);
SDL_GL_SetAttribute(SDL_GL_CONTEXT_FLAGS, SDL_GL_CONTEXT_DEBUG_FLAG);
and I'm using ANGLE/(a GL driver that doesn't provide an ES2 context) such that SDL_EGL_CreateContext is called by SDL_GL_CreateContext, I get the error "Could not create EGL context (context attributes are not supported)" and no context is created.
Looking at the code in SDL_EGL_CreateContext - if gl_config.flags is non-zero, it looks like the code in the section guarded with "#ifdef EGL_KHR_create_context" should be executed - but it apparently isn't.
Is it possible this section hasn't been compiled into the pre-built libraries? If I build SDL2.dll myself using the Visual C++ solution (VS2015 Community Update 3) then the call succeeds as I expect
bastien.bouclet
When exiting a "fullscreen space" on OS X, windows don't go to their defined "windowed mode size", but go back to their previous size.
Steps to reproduce:
1. Create a windowed mode SDL window
2. Toggle it to fullscreen with the SDL_WINDOW_FULLSCREEN_DESKTOP flag
3. While in fullscreen, change the windowed mode size using SDL_SetWindowSize
4. Toggle the window back to windowed mode
Expected result:
- The window has the size specified during step 3.
Actual result:
- The window has the size specified when creating the window in step 1.
Attached is a minimal reproduction test case.
The attached test case works as expected on X11 and Windows.
Ozkan Sezer
fix windows build after revision 11382: commit 2026e42e377a renamed
_SDL_msctf_h to SDL_msctf_h_ . SDL_windowskeyboard.c relies on that
macro, so update it accordingly.
Ozkan Sezer
Since the Vulkan merge, building against a Mac OS X SDM older than
10.11 fails in SDL_cocoametalview.m because Metal.framework is not
present. There is no conditional compiling in SDL_cocoametalview.m
either, so --disable-video-vulkan doesn't help with anything. (The
configury doesn't check darwin for x86_64 either, but it's another
story.)
I cross-build against 10.8 SDK on linux using clang-3.4.2 and this
is a problem for me. Will this be fixed?
Simon Hug
This issue actually raises the question if this API change (requirement of initialized audio subsystem) is breaking backwards compatibility. I don't see the documentation saying it is needed in 2.0.5.
"Major changes, roughly in order of appearance:
- Use float math everywhere, instead of promoting to double and casting back
all the time.
- Conserve sound energy when downmixing any channel into two other channels.
- Add a QuadToStereo filter. (The previous technique of reusing StereoToMono
never worked, since it assumed an incorrect channel layout for 4.0.)
- Add a 71to51 filter. This removes just under half of the cases the previous
code would silently break in.
- Add a QuadTo51 filter. More silent breakage fixed.
- Add a 51to71 filter, removing another almost-half of the silently broken
cases.
- Add 8 to the list of values SDL_SupportedChannelCount will accept.
- Change SDL_BuildAudioCVT's channel-related logic to handle every case, and
to actually fail if it fails instead of silently corrupting sound data and/or
crashing down the road."
(Note that SDL doesn't otherwise support 7.1 audio yet, but hopefully it will
soon and the 7.1 converters are an important piece of that. --ryan.)
Fixes Bugzilla #3727.
David Brady
When I attempted to make a mapping file for Android gamepads, I quickly discovered that most of the ones that I have here show up as the same device (Broadcom Bluetooth HID), meaning that it was impossible to make mappings on Android, since every device looked the same.
This patch will check for the existence of the getDescriptor function added in Jelly Bean, and use it if it's there. The Android Dashboard says that the majority of Android phones should support this function, and doing it this way will not force us to bump up our API version.
Clayton Craft
The default path used by directfb for libGL is different than the default path used by x11 in SDL2:
./src/video/directfb/SDL_DirectFB_opengl.c:
path = "libGL.so";
./src/video/x11/SDL_x11opengl.c:
#define DEFAULT_OPENGL "libGL.so.1"
On at least one distro (Alpine Linux), libGL.so is not created (or more accurately the symlink to libGL.so.1 is not created). For consistency, the 'path' variable in SDL_DirectFB_opengl.c should patch the DEFAULT_OPENGL in SDL_x11opengl.c ("libGL.so.1")
Error message was:
[mvk-info] MoltenVK version 0.18.2. Vulkan version 1.0.51.
[***MoltenVK ERROR***] VK_ERROR_INITIALIZATION_FAILED: On-screen rendering requires a view that is backed by a layer of type CAMetalLayer.
2017-08-28 02:17:29.579 testvulkan[95627:1716939] ERROR: SDL_Vulkan_CreateSurface(): vkCreateMacOSSurfaceMVK failed: VK_ERROR_INITIALIZATION_FAILED
David Ludwig
I've created a new set of patches. I am happy to create more, if it would help.
One version only copies 'size'.
A second version copies both 'size' and 'silence'. When looking over the documentation for SDL_OpenAudio in SDL_audio.h, it mentioned that both 'size' and 'silence' were things that SDL_OpenAudio would calculate.
Regarding *both* patches, I did notice that SDL 1.2 appears to have always modified desired's size and silence fields. The SDL wiki, at https://wiki.libsdl.org/SDL_OpenAudio#Remarks , does note:
Carlos
We would like to add a switch (define) that allows us to compile Angle statically with SDL. That is, getting rid of the OpenGL DLL. Usually you need OpenGL to be loaded dynamically as DLL because implementation is provided by the system but no need with Angle.
Only 2 files need modification and it shouldn't affect current behaivor:
include/SDL_egl.h and src/video/SDL_egl.c, as in here
https://github.com/native-toolkit/sdl/pull/10/files
The flag name could be SDL_VIDEO_STATIC_ANGLE (instead of NATIVE_TOOLKIT_STATIC_ANGLE) as discussed here https://github.com/native-toolkit/sdl/pull/10
We have tested this with both Windows and UWP, using NME engine (https://github.com/haxenme/nme).
Releated issue: https://bugzilla.libsdl.org/show_bug.cgi?id=1820
Sylvain
Hi! here's a patch for that with two class loaded regarding API level.
Test both case : before API 11 and after.
I also remove now unused GetSystemServiceFromUIThread() and minor clean-up (haptic warning prototype).
This fixes a strange corner case (notes appended below), and should be
safe to do anyhow.
Fixes Bugzilla #3674.
"I did more tests.
It appears the bug only happens if there is
another window on the screen that has "always
on top" property. For me it is xawtv - it is
always opened in a screen corner. Closing
xawtv or removing "always on top" property
from it makes the problem to go away.
Plus, it doesn't appear like the buttons are
not delivered at all. It appears that instead
the button presses are delivered on some mouse
positions, but not delivered when you move the
mouse to other part of the window... So this is
really weird and is likely somewhere deep in the
Xorg.
Maybe somehow it happens that the cursor is
actually above the xawtv window, but, because
my app uses grab, it is not visible there, and
in that case the events are not delivered to
my app?
But with my patch the button events are
always delivered flawlessly, it seems.
Hmm, and that indeed seems to explain my problem:
if the mask is set properly and my app uses
grab, then, even if the mouse is above some
other window, the events would still be delivered
to the grabbing app, which is what actually wanted
because my app uses relative mouse mode, so it
doesn't know the pointer can cross some other window
(my app draws the pointer itself).
So my current theory is that my patch only enforces
the mouse grab, which otherwise can be tricked by
some other window preventing the button events
delivery (but motion events are still delivered
via xinput2, which makes it all look very obscure)."
This patch was originally written by Marc Di Luzio for glX and enhanced by
Maximilian Malek for WGL, etc. Thanks to both of you!
Fixes Bugzilla #3643.
Fixes Bugzilla #3735.
Pegasus Epsilon
With the system dialog font set to Arial or Tahoma or another variable-width font, everything works just as expected. When using a fixed-width font, like Courier or DejaVu Sans Mono, the text gets cut off. Example screenshots attached.
Martijn Courteaux
I implemented precise scrolling events. I have been through all the folders in /src/video/[platform] to implement where possible. This works on OS X, but I can't speak for others. Build farm will figure that out, I guess. I think this patch should introduce precise scrolling on OS X, Wayland, Mir, Windows, Android, Nacl, Windows RT.
The way I provide precise scrolling events is by adding two float fields to the SDL_MouseWheelScrollEvent datastructure, called "preciseX" and "preciseY". The old integer fields "x" and "y" are still present. The idea is that every platform specific code normalises the scroll amounts and forwards them to the SDL_SendMouseWheel function. It is this function that will now accumulate these (using a static variable, as I have seen how it was implemented in the Windows specific code) and once we hit a unit size, set the traditional integer "x" and "y" fields.
I believe this is pretty solid way of doing it, although I'm not the expert here.
There is also a fix in the patch for a typo recently introduced, that might need to be taken away by the time anybody merges this in. There is also a file in Nacl which I have stripped a horrible amount of trailing whitespaces. (Leave that part out if you want).
manuel.montezelo
Original bug report (note that it was against 2.0.0, it might have been fixed in between): http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=733015
--------------------------------------------------------
Package: libsdl2-2.0-0
Version: 2.0.0+dfsg1-3
Severity: normal
Tags: patch
I have occasional crashes here caused by the X11 backend of SDL2. It seems to
be caused by the X11_Pending function trying to add a high number (> 1024)
file descriptor to a fd_set before doing a select on it to avoid busy waiting
on X11 events. This causes a buffer overflow because the file descriptor is
larger (or equal) than the limit FD_SETSIZE.
Attached is a possible workaround patch.
Please also keep in mind that fd_set are also used in following files which
may have similar problems.
src/audio/bsd/SDL_bsdaudio.c
src/audio/paudio/SDL_paudio.c
src/audio/qsa/SDL_qsa_audio.c
src/audio/sun/SDL_sunaudio.c
src/joystick/linux/SDL_sysjoystick.c
--------------------------------------------------------
On Tuesday 24 December 2013 00:43:13 Sven Eckelmann wrote:
> I have occasional crashes here caused by the X11 backend of SDL2. It seems
> to be caused by the X11_Pending function trying to add a high number (>
> 1024) file descriptor to a fd_set before doing a select on it to avoid busy
> waiting on X11 events. This causes a buffer overflow because the file
> descriptor is larger (or equal) than the limit FD_SETSIZE.
I personally experienced this problem while hacking on the python bindings
package for SDL2 [1] (while doing make runtest). But it easier to reproduce in
a smaller, synthetic testcase.
Martin Gerhardy
just for easier debugging issues in the own code...
SDL_CreateRenderer should maybe also use this macro
Ryan C. Gordon
I'll go one better: it should have an SDL_assert().
Leonardo
Structure SDL_gestureTouch gets reallocated for every new added gesture but its never freed.
Proposed patch add the function SDL_GestureQuit() that takes care of doing that and gets called when TouchQuit is called.
Gabriel Jacobo
Thanks for the patch. I think it needs a bit of extra work though, looking at the code in SDL_gesture.c , I see that SDL_numGestureTouches only goes up, I think the right fix here involves adding SDL_GestureDelTouch (hooked into SDL_DelTouch) as well as SDL_GestureQuit (as you posted in your patch).
Rainer Deyke
I've written a small patch that adds a small SDL_DuplicateSurface function to SDL. I've written the function as part of a larger (as yet unfinished) patch, but I think this function is useful enough that it merits inclusion in SDL on its own.
Alvin
I'm interested in this bug as well. I have experienced it when trying to embed an SDL_Window into a FLTK application. To do this, I create a FLTK window (window inside a window - think video player) and then use SDL_CreateWindowFrom() on the inner most window's Xlib Window*. After which, I create a renderer.
In my situation I am using the FLTK GUI toolkit.
What I have experienced is that the SDL_CreateRender() will recreate the window in order to properly setup OpenGL capability. As part of this process, the window is hidden and a call is executed that waits indefinitely for an acknowledgement that the window was indeed unmapped. This is where my program hangs.
Please correct me if I am wrong, but should SDL2 not make Xlib calls that effect the Xlib Window in this situation (e.g. When SDL_CreateWindowFrom() is used)? The toolkit being used typically assumes responsibility and, I presume, tracks all Xlib Windows it creates.
On line src/video/SDL_video.c:1372 the comment associated with setting SDL_WINDOW_FOREIGN reads:
/* Can't destroy and re-create foreign windows, hrm */
Since I do not know the reason for hiding the window in the first place, the attached patch simply does not wait for a response when X11_XWithdrawWindow() and X11_XMapRaised() are issued by X11_HideWindow() and X11_ShowWindow(), respectively. I presume that the GUI toolkit (GTK, FLTK, etc.) has or will consume the acknowledging event as it is managing the Xlib Window (or it thinks it is).
I have tested the patch against hg 5c645d037de2 and I have successfully tested:
* Embedding the SDL_Window inside a FLTK application.
* Calling SDL_SetWindowSize() when FLTK resizes the window (e.g. dragging cursor on the edge of the window).
* Filling the renderer's default target blue and drawing a red fill square at the centre (exciting, I know!)
* Calling SDL_Quit() when the application terminates
I do not receive any Xlib erorr messages (BadWindow, etc.) in any of those situations.
UX-admin
I am compiling with the Sun Studio 12 u2 compiler. There are multiple issues with the build, but this particular issue appears to be that it is illegal to declare a union of a struct of floats and a float. While GCC 4.8.1 does not flag this as an error, Sun Studio is much more standards compliant and strict, halting further compilation with an error.
afwlehmann
Sorry for re-opening, but it turns out that the current interval is indeed not updated. I've just checked the source code of the 2.0.3 release again:
163 if (current->canceled) {
164 interval = 0;
165 } else {
166 interval = current->callback(current->interval, current->param);
167 }
168
169 if (interval > 0) {
170 /* Reschedule this timer */
171 current->interval = interval; // <-- this line is missing
172 current->scheduled = tick + interval;
173 SDL_AddTimerInternal(data, current);
174 } else {
According to the documentation: "The callback function is passed the current timer interval and the user supplied parameter from the SDL_AddTimer() call and returns the next timer interval. If the returned value from the callback is 0, the timer is canceled."
If I understand the text correctly, then the current interval should in fact be updated according to the returned value. Otherwise there would be a discrepancy between the next time for which the timer is actually re-scheduled and the value that's passed to the callback once the timer fires again.
This could be fixed by adding line #171.
Ozkan Sezer
The attached patch removes SDLCALL attribute from SDL_BlitFunc() funcptr.
As far as I can see, *SDL_BlitFunc() is completely internal to SDL with
no specific calling convention requirements. The actual functions assigned
to SDL_BlitFunc seem to not have any calling conventions specified. So,
easy solution is simply removing the strict calling convention from the
type.
Sylvain
Here's a patch.
It tries to get the hint first. Resizable will allow any orientation. Otherwise it uses width/height window.
setOrientation method is splitted in static and non-static, so that it can be overloaded in a user subclass.
Some artefact observed :
surfaceChanged() can be called twice at the beginning. When the phone starts in portrait and run a landscape application.
Clayton Craft
linux_input module is disabled by default, despite the comments in source code that it is otherwise:
src/video/directfb/SDL_DirectFB_video.c:
devdata->use_linux_input = readBoolEnv(DFBENV_USE_LINUX_INPUT, 0); /* default: on */
src/video/directfb/SDL_DirectFB_video.h:
#define DFBENV_USE_LINUX_INPUT "SDL_DIRECTFB_LINUX_INPUT" /* Default: on */
When using the directfb driver, the linux_input module is suppressed unless the SDL app is started with "SDL_DIRECTFB_LINUX_INPUT=1" set in the environment. I recall seeing at one point that the directfb folks recommended using linux_input over the other input drivers, but I am having trouble locating this recommendation. In any case, I believe that this should really be defaulted to 'on' since it's vastly superior to the other dfb input drivers!
Jimb Esser
Note: This is using DirectInput, I have to disable XInput as that causes all but the first 4 controllers to be completely ignored by SDL (I can find no way to reconcile XInput devices with DirectInput devices, otherwise I would make a patch that accepts the fifth and later controllers with DirectInput...). XInput does not seem to have the problem below, only DirectInput.
I plug in 3 identical wireless Xbox 360 controllers, call them J1, J2, J3. Direct Input shows them as having GUIDs G1, G2, G3. I unplug J1, then J2 and J3 show up as having GUIDs G1 and G2! Not so "unique"... I start my SDL app when just J2 and J3 are plugged in, and open J2 and J3. Then I plug in a new controller, SDL sees that now G3 exists, assigns that a new SDL joystick instance ID, which I request to be opened, but G3 at this point is J3, which I already had opened! So I end up with two instances of J3 opened, and none of J1. "Re-"opening G1 would get the actual handle to the newly attached controller, but there's no current way to know this. This is clearly a bug or poor design in DirectInput or my wireless receiver drivers, but is a showstopping bug for my 8-20 player games (as soon as any one controller runs out of battery or goes to sleep and gets turned back on, suddenly things are busted requiring a restart (or, at least, a reinitialization of all controllers - the game can't go on)).
The solution I found is to use HID paths instead of GUIDs to uniquely identify joysticks. GUIDs are still needed to open a controller, however I have added code to re-find the GUIDs for all joysticks whenever a new joystick is attached or removed. This does now require opening of all joysticks (instead of just enumerating them), though if your app, like mine, is opening all of them anyway so that any can press a button to join, that doesn't change much (although perhaps they joysticks should be kept open in this case, instead of closed and re-opened). If your app only ever opens one joystick, this will do more work at startup than it did previously.
Jonas Kulla
This eliminates the need to manually compile in SDL_main_android.c.
Instead, add "-lSDL2main -Wl,-u,SDL_main_dummy" when linking.
I don't know how the nkd-build process works, but unless it was
for some reason linking libSDL2main.a it should be unaffected.
Alexey
Seems to be a missing functionality. I want to set an icon from RC file. I cant pass MAKEINTRESOURCE(X) string to SDL_RegisterApp() cause string returned by MAKEINTRESOURCE string is not actually a string and SDL_strlen will crash. Moreover LoadImage seems to be loading wrong icon size. LoadIcon seems to be fine.
Edward Rudd
Device: Logitech Rumble Gamepad F510 in Xinput mode.
Upon opening the joystick the values of the axes are queried via PollAllValues are not actually set on the device all the time.
This can easily be seen in the testjoystick or testgamecontroller test programs,as the testjoystick shows all axes in the center until one 'tickles' the triggers., and the testgamecontroller will show the triggers as 'on' until on 'tickles' the triggers.
Upon further research the culprit is the SDL_HINT_JOYSTICK_ALLOW_BACKGROUND_EVENTS hint. In the default value events are ignored until there is an active window, Thus in cases where the joystick system is initialized and controllers opened before the initial window is created & focuses, the initial values will be incorrect.
Here is my current workaround in the game I'm working on porting..
SDL_SetHint(SDL_HINT_JOYSTICK_ALLOW_BACKGROUND_EVENTS, "1");
SDL_GameController* gamepad = SDL_GameControllerOpen(index);
SDL_SetHint(SDL_HINT_JOYSTICK_ALLOW_BACKGROUND_EVENTS, "0");
Edmund Horner
When a 16-bit "565 format" surface has a colour key set, it will blit with correct transparency. If, however, it has its colour key set then is converted to a 32-bit ARGB format surface, the colour key in the converted image will not necessarily be the same pixel value as the transparent pixels. It may not blit correctly, because the colour key does not match the right pixels.
In my case, with an image using 0xB54A for transparency, the colour key was converted to 180,170,82; but the corresponding pixels (with the same original value) were converted to 180,169,82. Blitting the converted image did not use transparency where expected.
I have attached a test case. The bug has been replicated on both x86_64 Linux (SDL 2.0.2), and 32-bit MS C++ 2010 on Windows (SDL 2.0.0).
malferit
Hello, I began a little program with SDL2 on Linux in C, and when I call SDL_Init(SDL_INIT_VIDEO) I get an error and this is printed in the console:
XDM authorization key matches an existing client!
I searched through Internet, and found that some people suggest to run 'xhost +' or to specify this in /etc/X11/xdm/xdm-config:
DisplayManager*authName: MIT-MAGIC-COOKIE-1
I don't think an end user needs to know that...
But what bothered me is that first I started this little program in Pascal using the Freepascal compiler and it works. In freepascal you only use some thin header bindings in Pascal and then it links with the dynamic SDL library, so I don't understood why it worked with Freepascal and not in C.
I run ldd to the two generated applications:
Application in C:
linux-gate.so.1 (0xffffe000)
libSDL2-2.0.so.0 => /usr/lib/libSDL2-2.0.so.0 (0xb76ac000)
libpthread.so.0 => /lib/libpthread.so.0 (0xb766e000)
libc.so.6 => /lib/libc.so.6 (0xb74e2000)
libm.so.6 => /lib/libm.so.6 (0xb74a0000)
libdl.so.2 => /lib/libdl.so.2 (0xb749a000)
librt.so.1 => /lib/librt.so.1 (0xb7491000)
/lib/ld-linux.so.2 (0xb77b3000)
Application compiled with Freepascal:
linux-gate.so.1 (0xffffe000)
libSDL2-2.0.so.0 => /usr/lib/libSDL2-2.0.so.0 (0xb762a000)
libX11.so.6 => /usr/lib/libX11.so.6 (0xb74f3000)
libc.so.6 => /lib/libc.so.6 (0xb7367000)
libm.so.6 => /lib/libm.so.6 (0xb7325000)
libdl.so.2 => /lib/libdl.so.2 (0xb731f000)
libpthread.so.0 => /lib/libpthread.so.0 (0xb7305000)
librt.so.1 => /lib/librt.so.1 (0xb72fc000)
libxcb.so.1 => /usr/lib/libxcb.so.1 (0xb72dc000)
libXau.so.6 => /usr/lib/libXau.so.6 (0xb72d9000)
libXdmcp.so.6 => /usr/lib/libXdmcp.so.6 (0xb72d3000)
/lib/ld-linux.so.2 (0xb7755000)
It seems that Freepascal is linking with libX11, libxcb, libXau and libXdmcp .
Linking my C application with libxcb solved the problem (linking with libXau and/or libXdmcp without libxcb didn't work). Linking with X11 links all the other libraries and works as well.
So I fill this bug report mainly to let you know about this. I don't know if it is a problem that can be solved on the libSDL side or not, but at least I hope it will help.
Hi, some tests:
1. Disabled XDM. Login in console and running 'startx'. The program works without having to link with X11.
2. Enabled XDM. Added 'DisplayManager*authName: MIT-MAGIC-COOKIE-1' to /etc/X11/xdm/xdm-config.The program works without having to link with X11.
3. Enabled XDM without 'DisplayManager*authName: MIT-MAGIC-COOKIE-1' in /etc/X11/xdm/xdm-config . I get the authentication error unless I link with X11.
Coriiander
There's a slight mistake in the function "GetWindowStyle" found in file "SDL_windowswindow.c".
When a window is marked to be resizable, the resizable style is being added regardless of whether the window has a border or not. While for some arcane, hidden semantics this can be ok, it's still inconsistent in this case.
Witek Jachimczyk
I'm using SDL to develop a video viewer for MATLAB. The window is scrambled while using thightVNC with its default mode of RGB656.
SDL does not correctly recognize the pixel mode.
I found a solution for this problem. The solution involves modifying
SDL/src/video/SDL_pixels.c
Adding the following "if statement" under case 16: of SDL_MasksToPixelFormatEnum resolves the issue:
if (Rmask == 0x003F &&
Gmask == 0x07C0 &&
Bmask == 0xF800 &&
Amask == 0x0000) {
return SDL_PIXELFORMAT_RGB565;
}
I hope that this helps someone. I took me a while to figure it out.
Yann Dirson
When SDL_GL_GetProcAddress returns in error, the cause of the error is overwritten
in GL_GL_GetAttribute, reporting to the user "Failed getting OpenGL glGetString entry point", whereas the original "OpenGL library not loaded" never makes it
to the user.
Pushed a fix to:
f94cb13708
Note that the "OpenGL library not loaded" error looks like no root cause either,
and I'm still puzzled by the code path used: I'm forcing opengles2 renderer on
the x11 video driver on a rpi2, as in https://bugzilla.libsdl.org/3169, and although I now know that I must force the use of the RPI video driver instead
of the x11 one, I suspect even more accurate info can be given to user.
Daniel Gibson
AZERTY keyboard layouts (which are the default layouts in France and Belgium) don't have the number keys (1, 2, ..., 9, 0) in the first row of keys, but ?, &, ?", ', (, -, ?_, ??), = (with small differences between the France and Belgian variants). Numbers are reached via shift.
On Linux and OSX, SDL seems to use the corresponding ISO 8859-1 codes (231 for ?232 for ?tc) as SDL_Keycode (but no SDK_* constants exists for those values!), while on Windows SDL seems to map those keys to SDLK_1, SDLK_2 etc, like you'd get on QWERTY.
I don't know how other platforms behave.
So we have two problems:
1. At least on Linux and OSX invalid/undefined SDL_Keycodes are returned
2. Different platforms behave differently (Windows vs Linux/OSX)
It's unclear what behavior is desired: Should SDL_* constants for those keys be introduced (and Windows behavior changed accordingly)?
Or should all platforms behave like Windows here and use the existing SDLK_1, ..., SDLK_0 keycodes?
This bug on the mailing list:
https://forums.libsdl.org/viewtopic.php?t=11555 (my post about Linux/Windows)
https://forums.libsdl.org/viewtopic.php?t=11573 (Tim Walters discovered the same problem on OSX about 1.5 weeks later).
Sylvain
Let's you have a SDL_Surface that has ColorKey, but no Alpha Modulation.
When this surface is duplicated with SDL_ConvertSurface function, the result has ColorKey and Alpha Modulation (BLEND, and Opaque 255).
I think SDL_ConvertSurface should strictly keeps the input format.
example
=======
SDL_Surface *input; // ... Set up a surface with ColorKey and no AlphaMod
SDL_Surface *output = SDL_ConvertSurface(input, input->format, input->flags);
// "output" surface has a ColorKey but *also* AlphaMod (BLEND, and Opaque 255).
Simon Hug
The bug is in the GL_ResetState and GLES_ResetState functions which get called after a new GL context is created. These functions set the cached current color to transparent black, but the GL specification says the initial color is opaque white.
The attached patch changes the values to 0xffffffff to reflect the initial state of the current color. Should the ResetState functions get called anywhere else in the future, this probably has to call the GL functions itself to ensure that the colors match.
Adam M.
The keysym.mod field does not reflect the state of the modified keys when processing key down events for the modifier keys themselves. The documentation says that it returns the current key modifiers, but they are not current for key down events involving modifier keys. I interpret "current" to mean "equal to SDL_GetModState() at the instant the event is processed/enqueued".
For example, if you depress the Shift key you get a key down event with .mod == 0. However, .mod should not be 0 because a shift key is down. If you then release the Shift key, you get a key up event with .mod == 0. Neither event reports the modifier key.
If you press Shift and then A, .mod is incorrect (== 0) when Shift is pressed, but is correct later when A is pressed (== KMOD_LSHIFT).
You might say this behavior is deliberate, i.e. keysym.mod is the value /before/ the event, not the current value as documented, but that explanation is incorrect because only key down events behave that way. Key up events correctly give the current value, not the value before the event.
Not only is it inconsistent with itself, I think it makes keyboard processing harder.
The problem is near line 740 in SDL_keyboard.c:
if (SDL_KEYDOWN == type) {
modstate = keyboard->modstate; // SHOULD THIS BE MOVED DOWN?
switch (keycode) {
case SDLK_NUMLOCKCLEAR:
keyboard->modstate ^= KMOD_NUM;
break;
case SDLK_CAPSLOCK:
keyboard->modstate ^= KMOD_CAPS;
break;
default:
keyboard->modstate |= modifier;
break;
}
} else {
keyboard->modstate &= ~modifier;
modstate = keyboard->modstate;
}
In the key down path, modstate (and thus keysym.mod) ends up being the modifier state /before/ the event, but in the key up path modstate ends up being the modifier state /after/ the event. Personally I think the "modstate = keyboard->modstate" line should just be moved after the entire if/else statement, so that keysym.mod always reflects the current state.
owen
I removed all the static variables from SDLActivity.java
Updated all the SDL_android.c jni calls as well
I added a new function to SDL_android.c/ h
void Android_JNI_SeparateEventsHint(const char* c);
This is called by SDL_androidtouch.c so that this TU doesn't need to call any JNI functions.
Philipp Wiesemann
There is another problem with the current implementation which maybe should be fixed first (to prevent some work). It was written as if it would get the number of a button from the Java side but actually it gets the state of all buttons. That is why it should not work if more than one button is pressed at once.
Ian Abbott
I just spotted what I think is a bug in "src/thread/pthread/SDL_sysmutex.c" in the SDL_TryLockMutex function when FAKE_RECURSIVE_MUTEX is defined (for an implementation of Pthreads with no recursive mutex support). It calls pthread_mutex_lock instead of pthread_mutex_trylock, so it will block until the mutex is available instead of returning SDL_MUTEX_TIMEDOUT if it cannot lock the mutex immediately.
Coriiander
Here is a minor correction for a non-breaking mistake in SDL_setenv for __WIN32__ platform. See below for details.
FILE:
"SDL/src/stdlib/SDL_getenv.c"
FUNCTION: (__WIN32__ platform)
int SDL_setenv(const char *name, const char *value, int overwrite)
CODE:
if (!overwrite) {
char ch = 0;
const size_t len = GetEnvironmentVariableA(name, &ch, sizeof (ch));
if (len > 0) {
return 0; /* asked not to overwrite existing value. */
}
}
WHAT'S WRONG:
The 3th argument to GetEnvironmentVariable (being DWORD nSize) must be the number of characters, not the number of bytes. SDL currently passes "the size of 1 char", rather "1". While it is non-breaking (1=1 after all), it is incorrect. Furthermore there is no need to specify the 2nd and 3th arguments at all.
CORRECTION 1: (corrected argument_
if (!overwrite) {
char ch = 0;
const size_t len = GetEnvironmentVariableA(name, &ch, 1);
if (len > 0) {
return 0; /* asked not to overwrite existing value. */
}
}
CORRECTION 2: (stripped of unneeded code)
if (!overwrite) {
if (GetEnvironmentVariableA(name, NULL, 0) > 0) {
return 0; /* asked not to overwrite existing value. */
}
}
Juha Niemim?
On AmigaOS 4 platform with Newlib 'C' library, there is a problem with failing fseeko64. This seemed to be caused by using fopen instead of fopen64.
Littlefighter19
When trying to mirror something on the PSP, I've stumbled upon the problem,
that using SDL_RenderCopyEx with SDL_FLIP_HORIZONTAL flips the image vertically, vise-versa SDL_FLIP_VERTICAL flips the image horizontally.
Proposed patch would be swapping the check in line 944 with the one in line 948 in SDL_render_psp.c
romain.lacroix
For the windows implementation of SDL_ShowMessageBox() : ./src/video/windows/SDL_windowsmessagebox.c:345 WIN_ShowMessageBox()
The implementation in 2.0.4 uses "button index" for parameter "id" of function AddDialogButton().
It then expects the value provided in param wParam of function MessageBoxDialogProc() to be a valid index of a button.
It uses this value to index in the array of buttons when DialogBoxIndirect() returns (line 474 : *buttonid = buttons[which].buttonid;)
However, when dismissing this box with Escape, the return value of DialogBoxIndirect will be SDL_MESSAGEBOX_BUTTON_ESCAPEKEY_DEFAULT (=2) which is not always a valid index of array buttons.
When the array buttons has a length less or equal than 2, the memory access is invalid; I can see that the value written to *buttonId is uninitialized memory (random value).
The fix I propose : use value "buttonid" (field of button) for parameter "id" of AddDialogButton(), then copy return value of DialogBoxIndirect() in *buttonid. This way, we will not use an out-of-bounds index in array buttons.
Simon Hug
The RGBA_FROM_PIXEL macro in src/video/blit.h [1] is not designed to work with more than 8 bits per channel and the ARGB2101010 format makes it read outside of the array bounds causing access violations. This can happen during blitting with the BlitNtoNPixelAlpha and SDL_Blit_Slow functions.
When SDL_InitFormat tries to calculate the loss of the channels [2], the Uint8 will wrap around and it will end up at 254 for the 10-bit channels. Clearly way over the 9 entries of the SDL_expand_byte array. (Not that a signed integer would help.) Then the macro tries to access the lookup table with the channel value which could be up to 1023. If the previous indirection didn't cause an access violation this one will.
I guess it's not worth modifying this macro for a format that only a few will use. It will only make the other blitters slower. I don't have good ideas to solve this issue.
Attached is a test case that does three blits. A copy one that work and the two that use the functions mentioned above.
[1] https://hg.libsdl.org/SDL/file/cd1994d4f3c6/src/video/SDL_blit.h#l303
[2] https://hg.libsdl.org/SDL/file/cd1994d4f3c6/src/video/SDL_pixels.c#l540
Daniel
SDL_RenderReadPixels with SDL_RENDERER_SOFTWARE reads pixels from wrong coordinates.
SW_RenderReadPixels adjusts the rect coordinates according to the viewport. But since this is already done by SDL_RenderReadPixels, the final rect has x2 bigger X and Y.
Fabian Greffrath
we use SDL_GetPrefPath() in Chocolate Doom to get a reasonable directory to save and restore config files and savegames:
https://github.com/chocolate-doom/chocolate-doom/blob/sdl2-branch/src/m_config.c#L2162
However, since there is no "organization" behind Chocolate Doom and there is really only one "product" called Chocolate Doom, we pass an empty string for the org parameter and the package string for app.
This leads to two consecutive slashes in the path returned by SDL_GetPrefPath() like this:
/home/user/.local/share//chocolate-doom/
While this is harmless, it sure looks bad.
I believe that it should be possible to either pass a NULL pointer for the org parameter or at least have the function detect an empty string as a means to express "there is no origanization, just a single product". The generation of the path string to be returned by the function will have to get adapted accordingly.
Eric Wasylishen
Alt-Up/Down/Left/Right switches between displays using SDL_WINDOWPOS_CENTERED_DISPLAY
Shift-Up/Down/Left/Right shifts the window by 100px
Simon Hug
Some code in SDL loads libraries with SDL_LoadObject to get more information or use newer APIs. SDL_LoadObject may fail, set an error message and SDL will continue with some fallback code. Since SDL will overwrite the error or exit the function with a return value that indicates success, the error form SDL_LoadObject for the optional stuff might as well be cleared right away.
Eric Wasylishen 2017-07-26 18:42:58 UTC
I set up an (admittedly exotic) 3-monitor setup, and when I enter fullscreen-desktop on the middle display (#2), the SDL window is off center. (covers half of monitor #2 and most of monitor #3).
The displays are arranged from left to right:
Display #1 (main): 2880x1800, 200% scaling
Display #2: 1920x1200, 150% scaling
Display #3: 1920x1080, 100% scaling
SDL display bounds:
INFO: Bounds: 1440x900 at 0,0
INFO: Bounds: 1281x801 at 1921,0 (these are incorrect)
INFO: Bounds: 1920x1080 at 4800,0
Correct bounds reported by calling EnumDisplayMonitors and printing the LPRECT param of the callback:
1440x900 at (0, 0)
1280x800 at (2880, 0)
1920x1080 at (4800, 0)
It seems like you need 3 displays to reproduce this, and the left two need DPI scaling, and the 3rd display needs to have a different scale factor than the others.
Related: https://bugzilla.libsdl.org/show_bug.cgi?id=3709
SDL: current hg (11235:6a587b9e0ec8)
Windows 10, Version 10.0.15063 Build 15063
Tested with testdraw2 and testgl2, and pressing alt+enter to enter fullscreen desktop.
This patch reworks SDL_windowsmodes.c to use EnumDisplayMonitors instead of EnumDisplayDevices, so we always have an HMONITOR for each SDL display.
With access to an HMONITOR, we can get the monitor bounds in virtual screen coordinates the proper way, by calling GetMonitorInfo. (whereas the original code was doing some calculations - e.g. "data->DeviceMode.dmPosition.x * data->ScaleX" - to try to get virtual screen coordinates. These worked in simple cases, but failed in more complex cases like this bug)
The one potential problem with my patch is, the ChangeDisplaySettingsEx docs say that you're supposed to get the display name from EnumDisplayDevices, but I'm getting the display name from GetMonitorInfo now.
Simon Hug
KMSDRM_VideoInit allocates and frees some connectors and encoders but doesn't set the pointer to NULL after freeing. The cleanup code at the end may free one of those garbage pointer should an error happen in the initialization.
Simon Hug
When WIN_WindowProc processes the WM_TOUCH message, it doesn't check if the touch functions have been properly loaded and may call a NULL pointer. It's probably an extremely rare case, but here's a patch that adds some checks anyway.
Eric wing
Hi, I think I found a bug when using SDL_WINDOW_ALLOW_HIGHDPI with SDL_RenderSetLogicalSize on iOS. I use SDL_RenderSetLogicalSize for all my stuff. I just tried turning on SDL_WINDOW_ALLOW_HIGHDPI on iOS and suddenly all my touch/mouse positions are really broken/far-off-the-mark.
I actually don't have a real retina device (still) so I'm seeing this using the iOS simulator with a 6plus template.
Attached is a simple test program that can reproduce the problem. It uses RenderSetLogicalSize and draws some moving happy faces (to show the boundaries/space of the LogicalSize and that it is working correctly for that part).
When you click/touch, it will draw one more happy face where your button point is.
If you comment out SDL_WINDOW_ALLOW_HIGHDPI, everything works as expected. But if you compile with it in, the mouse coordinates seem really far off the mark. (Face appears far up and to the left.)
Alex Szpakowski on the mailing list suggests the problem is
"I believe this is a bug in SDL_Render?s platform-agnostic mouse coordinate scaling code. It assumes the units of the mouse coordinates are always in pixels, which isn?t the case where high-DPI is involved (regardless of whether iOS is used) ? they?re actually in ?DPI independent? coordinates (which matches the window size, but not the renderer output size)."
Additionally, if this is correct, the Mac under Retina is also probably affected too and "as well as any other platform SDL adds high-dpi support for in the future".
Levi Bard
In some environments, xrandr modes initialization can fail even though xrandr support is present and of a sufficient version.
(The one I encountered was an AWS instance running a virtual display)
The attached patch allows SDL to keep trying other methods if xrandr modes initialization fails (still subject to SDL_VIDEO_X11_REQUIRE_XRANDR).
Manuel
The attached patch adds support for KMS/DRM context graphics.
It builds with no problem on X86_64 GNU/Linux systems, provided the needed libraries are present, and on ARM GNU/Linux systems that have KMS/DRM support and a GLES2 implementation.
Tested on Raspberry Pi: KMS/DRM is what the Raspberry Pi will use as default in the near future, once the propietary DispmanX API by Broadcom is overtaken by open graphics stack, it's possible to boot current Raspbian system in KMS mode by adding "dtoverlay=vc4-kms-v3d" to config.txt on Raspbian's boot partition.
X86 systems use KMS right away in every current GNU/Linux system.
Simple build instructions:
$./autogen.sh
$./configure --enable-video-kmsdrm
$make
Now the clipboard isn't lost if you destroy a specific SDL_Window, as it
works on other platforms. You will still lose the clipboard data on
SDL_Quit() or process termination, but that's X11 for you; run a
Clipboard Manager daemon.
Fixes Bugzilla #3222.
Fixes Bugzilla #3718.
Eric Wasylishen
I think I found a better fix.
The problem with https://hg.libsdl.org/SDL/rev/ebdc0738b1b5 is setting the styleMask to 0 clears the NSWindowStyleMaskFullScreen bit, which then confuses Cocoa later when you try to leave fullscreen. Instead I'm just clearing the NSWindowStyleMaskResizable bit, although SetWindowStyle(window, NSWindowStyleMaskFullScreen); seems to also work.
Eric Wasylishen
Unfortunately this commit seems to have broken exiting desktop-fullscreen.
- Launch testgl2.
- Press alt+enter to go fullscreen-desktop
- Press alt+enter again. The spinning cube will freeze, and the window stays fullscreen desktop.
Simon Hug
SDL_GL_GetAttribute doesn't check if a video driver has been initialized and will access the SDL_VideoDevice pointer, which is NULL at that point.
I think all of the attributes require an initialized driver, so a simple NULL check should fix it. Patch is attached.
Jason Wyatt
After hiding the window, SDL_WINDOW_HIDDEN/SDL_WINDOW_SHOWN flags on a window are correctly updated. However on the next SDL_PumpEvents, they are set incorrectly.
This appears to be because X11_GetNetWMState does not check whether the _NET_WM_STATE property exists (it shouldn't on unmapped windows, see https://specifications.freedesktop.org/wm-spec/wm-spec-1.3.html#idm140130317598336). This results in an empty list of atoms for the state, which would imply that the window is not hidden.
(Seen on Fedora 24, Gnome)
--
Dan Ginsburg
More details on my proposed patch: I am on Kubuntu 16.04.2. I ran into this same bug, but with Jason's patch I found that actualType != None was true so the SDL_WINDOW_HIDDEN would still not be set. My fix instead is to explicitly check for whether the window is unmapped rather than relying on the returned values in XGetWindowProperty.
felix
The functions in src/render/SDL_yuv_mmx.c contain the following inline assembly snippet:
/* tap dance to workaround the inability to use %%ebx at will... */
/* move one thing to the stack... */
"pushl $0\n" /* save a slot on the stack. */
"pushl %%ebx\n" /* save %%ebx. */
"movl %0, %%ebx\n" /* put the thing in ebx. */
"movl %%ebx,4(%%esp)\n" /* put the thing in the stack slot. */
"popl %%ebx\n" /* get back %%ebx (the PIC register). */
Here's how it ended up in a binary on my old laptop:
0xb5c17dbd <ColorRGBDitherYV12MMX1X+93>: push $0x0
0xb5c17dbf <ColorRGBDitherYV12MMX1X+95>: push %ebx
0xb5c17dc0 <ColorRGBDitherYV12MMX1X+96>: mov 0xc(%esp),%ebx
0xb5c17dc4 <ColorRGBDitherYV12MMX1X+100>: mov %ebx,0x4(%esp)
0xb5c17dc8 <ColorRGBDitherYV12MMX1X+104>: pop %ebx
Apparently the compiler, oblivious to the fact that the assembly snippet manipulates the %esp register, decided to refer to the operand via that same register instead of via %ebp (I believe -fomit-frame-pointer enables this). This causes %ebx to be loaded with the wrong value, which later leads to a null pointer dereference.
Recent GCC can use the %ebx register normally: <https://gcc.gnu.org/bugzilla/show_bug.cgi?id=47602#c16>. There is even an explicit constraint "b" for allocating it.
Holger Schemel
Summary: This patch adds support for key events for the "rewind" and "fast forward" media keys on the Amazon Fire TV remote control.
How to reproduce the problem: Run Android build of SDL2 application on the Amazon Fire TV (tested with "stick" version) and log key events.
Expected behaviour: Every key pressed on the Fire TV remote control should result in a corresponding key event (pressed/released).
Observed behaviour: Of the bottom row of buttons on the Fire TV remote control, only the "play/pause" (middle) button generates a key event, while the "rewind" (left) and "fast forward" (right) buttons to not generate any event at all.
The attached patch adds support for these two missing buttons/keys.
Note 1: Some missing definitions were added for the already existing key codes SDL_SCANCODE_APP1 and SDL_SCANCODE_APP2 (to keep up the correct order of enumerations / array positions when adding the two new key codes).
Note 2: Definitions in "scancodes_linux.h" and "scancodes_xfree86.h" (to also add support for these keys on other platforms) were added without testing. However, I was unable to find corresponding definitions for these two media keys for Windows and Mac OS X.
Note 3: I have also updated the (broken) link to the USB usage page standard PDF document (comment in "include/SDL_scancode.h").
kdrakehp
The attached patch adds capture support to the sndio backend.
The patch also allows the `OpenDevice' function to accept arbitrary device names.
Bogomancer
On X11, windows created using the shaped window API appear distorted unless the width of the shape surface is divisible by 8.
Steps to reproduce:
1) Use your favorite image editor to resize one of the images in test/shapes/ to a width that's not a multiple of 8.
2) Compile and run test/testshape.c on the image you edited.
3) The shaped window will appear twisted and distorted.
It appears the bug was not caught sooner because all the test images are either 640 or 256 pixels wide.
I tracked down the bug to SDL_CalculateShapeBitmap() in SDL_shape.c. The shape surface is reduced to a 1-bit-per-pixel mask, but the original code doesn't take into account that X11 apparently wants each scanline to begin on a new byte.
Ozkan Sezer
(In reply to Ryan C. Gordon from comment #9)
> I've put this patch in as https://hg.libsdl.org/SDL/rev/7213ae46e870 ...can
> you verify this works on the latest MinGW?
>
> Thanks,
> --ryan.
This patch is wrong: the structure in question has nothing to do with any
gcc version in use. I suggest reverting this adding a conigury check for
it, instead. Something like the following should do it: (configure needs
regenerating.)
Mark Callow
SDL_ShowMessageBox calls SDL_CaptureMouse which, in the UIKit driver, raises a ?That operation is not supported? error, overwriting the SDL error that an application may be trying to report.
This is because UIKit SDL_CaptureMouse returns SDL_Unsupported() which ends up calling SDL_SetError() which has the following code:
/* If we are in debug mode, print out an error message */
SDL_LogDebug(SDL_LOG_CATEGORY_ERROR, "%s", SDL_GetError());
The SDL_GetError call here overwrites the static buffer?..
Although an application can avoid this by using SDL_GetErrorMsg(char* errstr, int maxlen) to avoid the static buffer, SDL should be fixed.
The fix is simple. In SDL_SetError change
SDL_LogDebug(SDL_LOG_CATEGORY_ERROR, "%s", SDL_GetError());
to
SDL_LogDebug(SDL_LOG_CATEGORY_ERROR, "%s", error);
where error is the pointer to the buffer where it assembled the message.
Amruth Raj
- My app runs in full screen to play video(I use SDL_WINDOW_FULLSCREEN_DESKTOP)
- Cmd-tab to go out of full screen to another app
- Cmd-tab again to get back to my app
- Press left mouse button at one of the edges of the screen, don't release yet.
After this point the main thread is stuck until I release the left mouse button and hence video rendering doesn't happen anymore.
On debugging more, I see that thread 0 is stuck as shown below with sendEvent processing left mouse down. It comes out only after it receives a left mouse up. There are some frames below which show NSWindowResizing, but my window flag doesn't have SDL_WINDOW_RESIZABLE set.
Thread 0:: CrBrowserMain Dispatch queue: com.apple.main-thread
0 libsystem_kernel.dylib 0x00007fffbe13d34a mach_msg_trap + 10
1 libsystem_kernel.dylib 0x00007fffbe13c797 mach_msg + 55
2 com.apple.CoreFoundation 0x00007fffa889d434 __CFRunLoopServiceMachPort + 212
3 com.apple.CoreFoundation 0x00007fffa889c8c1 __CFRunLoopRun + 1361
4 com.apple.CoreFoundation 0x00007fffa889c114 CFRunLoopRunSpecific + 420
5 com.apple.HIToolbox 0x00007fffa7dfdebc RunCurrentEventLoopInMode + 240
6 com.apple.HIToolbox 0x00007fffa7dfdcf1 ReceiveNextEventCommon + 432
7 com.apple.HIToolbox 0x00007fffa7dfdb26 _BlockUntilNextEventMatchingListInModeWithFilter + 71
8 com.apple.AppKit 0x00007fffa6396a54 _DPSNextEvent + 1120
9 com.apple.AppKit 0x00007fffa6b127ee -[NSApplication(NSEvent) _nextEventMatchingEventMask:untilDate:inMode:dequeue:] + 2796
10 com.apple.AppKit 0x00007fffa66f568d +[NSWindow(NSWindowResizing) _mouseHysteresisCheck:withExpiration:andDistance:finalMouseLocation:] + 525
11 com.apple.AppKit 0x00007fffa65eedb5 -[NSWindow(NSWindowResizing) _hitTestWithHysteresisCheck:forEvent:allowWindowDragging:] + 394
12 com.apple.AppKit 0x00007fffa6c8f0db -[NSWindow(NSEventRouting) _handleMouseDownEvent:isDelayedEvent:] + 1873
13 com.apple.AppKit 0x00007fffa6c8ca6c -[NSWindow(NSEventRouting) _reallySendEvent:isDelayedEvent:] + 1942
14 com.apple.AppKit 0x00007fffa6c8bf0a -[NSWindow(NSEventRouting) sendEvent:] + 541
15 org.libsdl.SDL2 0x000000010d46d74a -[SDLWindow sendEvent:] + 90
16 com.apple.AppKit 0x00007fffa6b10681 -[NSApplication(NSEvent) sendEvent:] + 1145
17 org.libsdl.SDL2 0x000000010d46532b -[SDLApplication sendEvent:] + 139
18 org.libsdl.SDL2 0x000000010d466b2f Cocoa_PumpEvents + 495
19 org.libsdl.SDL2 0x000000010d44c1d5 SDL_PumpEvents_REAL + 53
20 org.libsdl.SDL2 0x000000010d44c2f5 SDL_WaitEventTimeout_REAL + 53
21 org.libsdl.SDL2 0x000000010d44c2b7 SDL_PollEvent_REAL + 23
22 org.libsdl.SDL2 0x000000010d51bb24 SDL_PollEvent + 36
23 libTest.dylib 0x000000010cf3e0e8 SDLEventProcessor::processEvents(int) + 568
24 Test 0x000000010cde6bba BrowserApp::RunAppMessageLoop(BAInstData*, CefStringBase, CefStringBase) + 810
25 Test 0x000000010ce04bbc main + 17980
26 libdyld.dylib 0x00007fffbe016235 start + 1
I further noticed that while entering full screen in SDL_cocoawindow.m NSResizableWindowMask is set. If I clear it inside windowDidEnterFullScreen, then, the issue doesn't repro.
This is discussed at https://discourse.libsdl.org/t/main-thread-gets-stuck-on-left-mouse-down/22753/3 and thanks to Eric for the pointers.
The Xlib documentation demands that 32-bit values here be passed in a long,
even when long itself isn't a 32-bit value. Otherwise libx11 might read
memory incorrectly.
Fixes Bugzilla #3692.
Sam Lantinga
Ryan C. Gordon
Ozkan Sezer
Ethan Lee
Brandon Schaefer
Alex Szpakowski
Mark Callow
Patrice Mandin
David Ludwig
Ozkan Sezer
ouned
Philipp Wiesemann