Mark Pizzolato
On Windows with Visual Studio, when building SDL as a static library using the x86 (32bit) mode, several intrinsic operations are implemented in code in SDL_stdlib.c.
One of these, _allshr() is not properly implemented and fails for some input. As a result, some operations on 64bit data elements (long long) don't always work.
I classified this bug as a blocker since things absolutely don't work when the affected code is invoked. The affected code is only invoked when SDL is compiled in x86 mode on Visual Studio when building a SDL as a static library. This build environment isn't common, and hence the bug hasn't been noticed previously.
I reopened#2537 and mentioned this problem and provided a fix. That fix is provided again here along with test code which could be added to some of the SDL test code. This test code verifies that the x86 intrinsic routines produce the same results as the native x64 instructions which these routines emulate under the Microsoft compiler. The point of the tests is to make sure that Visual Studio x86 code produces the same results as Visual Studio x64 code. Some of the arguments (or boundary conditions) may produce different results on other compiler environments, so the tests really shouldn't be run on all compilers. The test driver only actually exercised code when the compiler defines _MSC_VER, so the driver can generically be invoked without issue.
Martin Gerhardy
According to https://msdn.microsoft.com/de-de/library/2kzt1wy3%28v=vs.120%29.aspx when one is using /MT for msvc compilations the libcmt.lib is already linked to the binary. This lib includes the symbol that is now guarded (see attached patch) by the #ifndef _MT.
Fixes Bugzilla #2895.
His notes:
The following trivial changes make SDL2 tree (mostly) compatible with Visual
Studio 2005:
* SDL_stdlib.c: Similar to VS2010 and newer, VS2005 also generates memcpy(),
(it also generates memset(), see below), so propagate the #if condition to
cover VS2005.
* SDL_pixels.c (SDL_CalculateGammaRamp): VS2005 generates a memset() call for
gamma==0 case, so replace the if loop with SDL_memset().
* SDL_windowsvideo.h: Include msctf.h only with VS2008 and newer, otherwise
include SDL_msctf.h
* SDL_windowskeyboard.c: Adjust the #ifdefs so that SDL_msctf.h inclusion is
always recognized correctly.
This is a little macro magic to use malloc() directly instead of SDL_malloc(),
etc, so static analysis tests that know about the C runtime can function
properly, and understand that we are dealing with heap allocations, etc.
This changed our static analysis report from 5 outstanding bugs to 30.
5x as many bugs were hidden by SDL_malloc() not being recognized as malloc()
by the static analyzer!
SDL 2.x recently accepted patches to enable OpenGL ES 2 support via Google's ANGLE library. The thought is to try to eventually merge SDL/WinRT's OpenGL code with SDL-official's.
Ghassan Al-Mashareqa
The SDL_ceil function is implemented incorrectly when HAVE_CEIL is not defined (HAVE_LIBC not defined).
The following code:
double val = SDL_ceil(2.3);
printf("%g", val);
prints "2.0", as STD_ceil is defined as:
double
SDL_ceil(double x)
{
#ifdef HAVE_CEIL
return ceil(x);
#else
return (double)(int)((x)+0.5);
#endif /* HAVE_CEIL */
}
This functions is used in the SDL_BuildAudioResampleCVT function of the audio subsystem (SDL_audiocvt.c), and causes a bug in that function.
Having the SDL functions inline is causing build issues, and in the case of malloc(), etc. causing malloc/free mismatches, if the application build environment differs from the SDL build environment.
In the interest of safety and consistency, the functions will always be in the SDL library and will only be redirected to the C library there, if they are available.
See the following threads on the SDL mailing list for the gruesome details:
* SDL_stdinc.h inlines problematic when application not compiled in exact same feature environment
* Error compiling program against SDL2 with -std=c++11 g++ flag
Colin Barrett
I see this manifest itself (VS2012 x86) as:
"Run-Time Check Failure #0 - The value of ESP was not properly saved across a function call. This is usually a result of calling a function declared with one calling convention with a function pointer declared with a different calling convention."
in the first call to SDL_GetTicks in my application. The disassembly at the problem line is:
hires_now.QuadPart *= 1000;
00AD0792 push 0
00AD0794 push 3E8h
00AD0799 mov eax,dword ptr [ebp-10h]
00AD079C push eax
00AD079D mov ecx,dword ptr [hires_now]
00AD07A0 push ecx
00AD07A1 call _allmul (0AE7D40h)
00AD07A6 mov dword ptr [hires_now],eax
00AD07A9 mov dword ptr [ebp-10h],edx
Apparently _allmul should be popping the stack but isn't (other similar functions in SDL_stdlib.c - _alldiv and whatnot - DO pop the stack).
A 'ret 10h' at the end of _allmul appears to do the trick
All SDL_* functions are always available as real symbols, so you can always
link against them as a stable ABI. By default, however, all the things that
might have dithered down to macros in your application are now force-inlined,
to give you the same effect as before and theoretically better performance,
but still solve the classic macro problems.
Elsewhere, we provide real functions for these things that simply wrap the
inline functions, in case one needs to have a real function available.
Also: this exposed bugs: SDL_abs() does something different if you had the
macro vs the libc function, SDL_memcpy() returns a void* in the function
but not the macro, etc.
