makes the SDL_main code shorter
Also added a generic SDL_RunApp() implementation for platforms that
don't really need it.
Some platforms (that use SDL_main but haven't been ported yet) are
still missing, but are added in the following commits.
and move the #undef main and #define main SDL_main to the start/end of
SDL_main_impl.h instead of repeating it in every platform implementation
Thanks to SDL_N3DSRunApp we don't need the #include <3ds.h> in
SDL_main_impl.h - that caused conflicts with testthread.c, because both
have (different) ThreadFunc typedefs.
As the implementation requires C++, the user will have to include
SDL_main.h in a C++ source file (that needs to be compiled with /ZW).
It's ok to keep the standard main() implementation in plain C and use
an otherwise empty C++ source file for the SDL_main implementation part,
if both source files #include <SDL3/SDL_main.h>
Including SDL_main.h in a C source file will print a message at
compilation (when building for WinRT or possibly other not yet implemented
platforms that require C++ for main), to remind the user of also
including it in a .cpp source file. This message/warning can be disabled
with #define SDL_MAIN_NOIMPL before including SDL_main.h in the C file.
When including it in a .cpp file, there will be a compiler error with
helpful message if it's not compiled with /ZW
For this I renamend _SDL_MAIN_NOIMPL to SDL_MAIN_NOIMPL, because now it's
not for internal use only anymore, but also useful for users (that want
their main() function in a different file than the SDL_main implementation)
Add a project for the testdraw2.c test to the WinRT solution to at least
get some minimal testing on WinRT.
I won't add all tests because it's a lot of manual clicking per test,
but this should be better than nothing :)
(remaining platforms will follow)
SDL_main.h is *not* included by SDL.h anymore, users are supposed to
include it directly now, usually only in the file they implement main() in.
If they need the header elsewhere or don't want SDL_main to implement
main() (but only call SDL_SetMainReady() or whatever), they
can #define SDL_MAIN_HANDLED first, same as before.
For SDL-internal usage, I added _SDL_MAIN_NOIMPL, which *also* skips the
implementation and `#define main SDL_main`, but still defines
SDL_MAIN_AVAILABLE and SDL_MAIN_NEEDED in SDL_main.h, as before.
To make the implementaion in the header shorter and avoid including windows.h,
I moved most of the Win32 SDL_main code into SDL3.dll via SDL_Win32RunApp(),
so the header-only part is just the different main functions calling
SDL_Win32RunApp(SDL_main, NULL)
Note that I changed changed the return value and type of OutOfMemory()
to return -1 instead of FALSE, so main() (or WinMain() or whatever)
returns -1 instead of 0 in case of an out-of-memory error
Compared to original Win32 SDL_main, I tweaked the part of the
implementation in SDL_main_impl.h a bit to avoid linker warnings
and conflicts with stuff from windows.h:
- replaced windows.h with own define of WINAPI
and typedef-ing HINSTANCE and LPSTR.
This prevents conflicts between all the generically-named #defines and
types in windows.h and user code (like DrawState in some SDL tests)
- only using one of main() or wmain() gets rid of a MSVC linker error
("warning LNK4067: ambiguous entry point")
If this still causes problems, we might try getting rid of wmain(),
seemed to me like MSVC can use regular main() in UNICODE mode as well
- simplified the UNICODE logic for that - while this is not exactly
equivalent to the old, it should make sense and Works For Me
This simplifies some things, clarifies some things, and also allows
for the possibility of RWops that offer non-blocking i/o (although
none of the current built-in ones do, intentionally, we could add this
later if we choose, or people could provide things like network socket
RWops implementations now, etc.
Fixes#6729.
The annotations have been added to SDL_mutex.h and have been made public so applications can enable this for their own code.
Clang assumes that locking and unlocking can't fail, but SDL has the concept of a NULL mutex, so the mutex functions have been changed not to report errors if a mutex hasn't been initialized. We do have mutexes that might be accessed when they are NULL, notably in the event system, so this is an important change.
This commit cleans up a bunch of rare race conditions in the joystick and game controller code so now everything should be completely protected by the joystick lock.
To test this, change the compiler to "clang -Wthread-safety -Werror=thread-safety -DSDL_THREAD_SAFETY_ANALYSIS"
If you care about timestamps you'll also want to catch all of the sensor events instead of just polling the current state. For example, Nintendo Switch controllers generate 3 sensor events with distinct values for each polling interval.
If a program built against one version of SDL is run in an
environment where there is an earlier version of the SDL .so library
installed, the result varies depending on platform configuration; in
the best case, it won't start at all, at worst it aborts in the
middle of the user doing "something important" (systems implementing
lazy symbol resolution). verdefs on the other hand are always checked
on startup.
The dependency information present in programs and shared libraries
is not only of value to the dynamic linker, but also to a
distribution's package management. If the dynamic linker is able to
tell that a program is not runnable per the above, a package manager
is able to come to the same conclusion — and block the installation
of a nonfunctional program+library ensemble.
Because there are a lot more symbols than there are libraries (I am
going to throw in "10^4 to 1 or worse"), package managers generally
do not evaluate symbols, but only e.g. the SONAME, NEEDED and VERNEED
fields/blocks. Because the SONAME is the same between two SDL
versions like 2.0.24, and 2.0.26, everything rests on having verdefs.
This patch proposes the addition of verdefs.
