Elis?e Maurer
The attached minimal program sets the SDL_HINT_MOUSE_RELATIVE_MODE_WARP to 1, enables relative mouse mode then logs all SDL_MOUSEMOTION xrel values as they happen.
When moving the mouse exclusively to the right:
* On a Windows 10 installation before Fall Creators update (for instance, Version 10.0.15063 Build 15063), only positive values are reported, as expected
* On a Windows 10 installation after Fall Creators update (for instance, Version 10.0.16299 Update 16299), a mix of positive and negative values are reported.
3 different people have reproduced this bug and have confirmed it started to happen after the Fall Creators update was installed. It happens with SDL 2.0.7 as well as latest default branch as of today.
It seems like some obscure (maybe unintended) Windows behavior change? Haven't been able to pin it down more yet.
(To force-upgrade a Windows installation to the Fall Creators update, you can use the update assistant at https://www.microsoft.com/en-us/software-download/windows10)
Eric Wasylishen
Broken GetCursorPos / SetCursorPos based games on Win 10 fall creators are not limited to SDL.. I just tested winquake.exe (original 1997 exe) and it now has "jumps" in the mouse input if you try to look around in a circle. It uses GetCursorPos/SetCursorPos by default. Switching WinQuake to use directinput (-dinput flag) seems to get rid of the jumps.
Daniel Gibson
A friend tested on Win10 1607 (which is before the Fall Creators Update) and the the bug doesn't occur there, so the regression that SetCursorPos() doesn't reliably generate mouse events was indeed introduced with that update.
I even reproduced it in a minimal WinAPI-only application (https://gist.github.com/DanielGibson/b5b033c67b9137f0280af9fc53352c68), the weird thing is that if you don't do anything each "frame" (i.e. the mainloop only polls the events and does nothing else), there are a lot of mouse events with the coordinates you passed to SetCursorPos(), but when sleeping for 10ms in each iteration of the mainloop, those events basically don't happen anymore. Which is bad, because in games the each iteration of the mainloop usually takes 16ms..
I have a patch now that I find acceptable.
It checks for the windows version with RtlGetVersion() (https://msdn.microsoft.com/en-us/library/windows/hardware/ff561910.aspx) and only if it's >= Win10 build 16299, enables the workaround.
All code is in video/windows/SDL_windowsevents.c
and the workaround is, that for each WM_MOUSEMOVE event, "if(isWin10FCUorNewer && mouseID != SDL_TOUCH_MOUSEID && mouse->relative_mode_warp)", an addition mouse move event is generated with the coordinates of the center of the screen
(SDL_SendMouseMotion(data->window, mouseID, 0, center_x, center_y);) - which is exactly what would happen if windows generated those reliably itself.
This will cause SDL_PrivateSendMouseMotion() to set mouse->last_x = center_x; and mouse->last_y = center_y; so the next mouse relative mouse event will be calculated correctly.
If Microsoft ever fixes this bug, the IsWin10FCUorNewer() function would have to
be adjusted to also check for a maximum version, so the workaround is then disabled again.
Andrey
Seems latest google angle library successfully built & tested under macOS'es.
https://github.com/google/angle
We need to use GLES2 to implement true cross-platform code.
tomwardio
Proposed patch loads eglCreatePbufferSurface in same manner as other 1.1 functors. This allows custom video drivers to create pbuffer surfaces.
Manuel Alfayate Corchete
This fixes a problem with KMSDRM on some graphics hardware where only bigger cursor sizes are supported, such as current Intel gfx. (The kernel-side driver is what limits this: had to look for failing IOCTLs...)
That caused SDL_SetCursor() to fail silently, and we were left with a missing cursor without further explanation.
With this patch, different "standard" sizes are tried and a bigger one is used (with an intermediate and clean buffer only used to write the new cursor to the BO where it will live after) if we get, let's say, 16x16 which is pretty common but our hardware does not support that.
C Snover
SDL_AddDisplayMode returns an SDL_bool corresponding to whether or not the given display mode was added or not. It will return SDL_FALSE if a matching display mode already exists in the display's list of display modes, which causes ownership of the mode driverdata to remain with the caller. Some video drivers ignore the return value of SDL_AddDisplayMode, so leak the driverdata memory when SDL_AddDisplayMode returns SDL_FALSE.
raist66676
Here is the bug in latest SDL 2.0.8 development repo. It is obvious and simple to fix by correcting typos on six lines of code.
In src/video/SDL_yuv.c on lines 217, 249, 280, 321, 353, and 384 the wrong conversion functions are called for SDL_PIXELFORMAT_ABGR8888 and SDL_PIXELFORMAT_BGR888. Instead of ABGR functions, BGRA functions are called. These are typos.
New functions get and set the YUV colorspace conversion mode:
SDL_SetYUVConversionMode()
SDL_GetYUVConversionMode()
SDL_GetYUVConversionModeForResolution()
SDL_ConvertPixels() converts between all supported RGB and YUV formats, with SSE acceleration for converting from planar YUV formats (YV12, NV12, etc) to common RGB/RGBA formats.
Added a new test program, testyuv, to verify correctness and speed of YUV conversion functionality.
Alex Szpakowski <slime73@gmail.com> 2017-07-12 21:28 -0300
macOS: Expose more display modes on retina screens. Fixes an issue found in BZFlag.
http://hg.libsdl.org/SDL/rev/cfb3ddf796c3
Alex Szpakowski <slime73@gmail.com> 2017-07-12 21:32 -0300
Fix a potential crash in macOS 10.7 and earlier.
http://hg.libsdl.org/SDL/rev/4941c8867075
tomwardio
Remove static int vm_error and vm_event, use local variables instead.
This fixes unused variable errors when compiling with SDL_VIDEO_DRIVER_X11_XINERAMA undefined.
src/video/x11/SDL_x11modes.c:505:22: error: unused variable 'vm_error' [-Werror,-Wunused-variable]
src/video/x11/SDL_x11modes.c:505:12: error: unused variable 'vm_event' [-Werror,-Wunused-variable]
tomwardio
Add support to be able to set EGL_SURFACE_TYPE bits when creating an EGL config. This is usefule when wanting to create pixel buffer surfaces in custom video drivers.
Changing the background color causes the titlebar to blend against it on
modern macOS releases, making all SDL windows look wrong by default. This was
set to make the window not flash white before a GL context is ready, but we
can accomplish this in our window's view's drawRect implementation, too.
Manuel
I noticed that, at least on Intel GPU hardware, passing SDL_RENDERER_PRESENTVSYNC would result on a static console instead of the program graphics.
That was due to the fact that calling drmModePageFlip() only works if we have previously set up CRTC to one of the GBM buffers with a drmModeSetCrtc() call, so now it's done and things work as expected.
The KMSDRM_GLES_SetupCrtc() call is done only one time, only when needed (when egl_swapinterval is not 0: when it's 0, there's no need for it because we flip by calling drmModePageFlip() anyway).
The place where KMSDRM_GLES_SetupCrtc() call is done may look strange, but it's right: it needs EGL completely ready because it needs to call eglSwapBuffers() internally to work (see more comments about it in the code).
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.
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
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.
