Wed May 12 12:49:28 2010 -0700 (5 days ago)
Added the automated test plist file and modified .hgignore to be less aggressive.
Discovered that the new plist file for the automated test was not checked in because the .hgignore file was too aggressive. I changed the .hgignore to not ignore the Xcode directory and instead added a longer list of things to ignore.
Discovered that the new plist file for the automated test was not checked in because the .hgignore file was too aggressive. I changed the .hgignore to not ignore the Xcode directory and instead added a longer list of things to ignore.
The refactored tests were written in recognition that Mac and iPhone current working directories are usually not going to work. Resource directories are in bundles and write directories are restricted to certain areas. In theory, other platforms may have this problem too, hence the refactoring.
Also updated the Xcode iPhone project to use 3.2 as the Base SDK, but 3.1 as the Deployment SDK (for iPhone/iPad compatibility.)
Looks like there is a problem with the source code base. UIScreenMode and possibly a few other 3.2 only APIs have crept in. 3.2 is only available for iPad, and not iPhone/iPod Touch.
The Mac OS X project has a new target called testsdl which builds the automated test. I looked at using Xcode's native unit test support, but the impedance mismatch between the existing automated test structure and Apple's was more than I could handle.
As such, the testsdl application is a full blown proper OS X application, which means it is a self-contained .app bundle. This immediately revealed some problems from the automated test. The largest problem was the assumption about the current working directory and where to find resources. (I suspect Windows may have a similar problem depending on circumstance.) To open resources, the test was looking in directories relative to the SDL source directory, but this will not work well with self-contained .app bundles and Xcode which can place its built applications almost anywhere. And for real-world situations, this is pretty useless anyway.
So I modified SDL_RWFromFile to do special things on OS X. First, it will look for a file in the .app bundle. If not found, it will fallback and just call fopen as it used to do.
I also had to modify the automated test itself because it had a contrieved test which called fopen directly to do read from an existing FILE pointer. In addition, there was a write test. Since a .app bundle is likely going to be read-only, I added a special case for OS X to write to NSTemporaryDirectory.
I expect these changes should work for both Mac and iPhone OS (which includes iPad).
I will update the iPhone Xcode project next.
Finally, FYI, the X11 automated test seems to be failing. Below is my output.
Pending breakpoint 4 - "-[NSException raise]" resolved
Platform : All tests successful (2)
SDL_RWops : All tests successful (5)
Rect : All tests successful (1)
SDL_Surface : All tests successful (6)
Rendering with cocoa driver : All tests successful (3)
Assert Failed!
Blit output not the same.
Test Case 'Renderer x11'
Test Suite 'Rendering with x11 driver'
Last SDL error ''
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSGetSurfaceBounds
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorFailure: Set a breakpoint @ CGErrorBreakpoint() to catch errors as they are logged.
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSBindSurface: Invalid window 0xa150
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSBindSurface: Invalid window 0xa150
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSBindSurface: Invalid window 0xa150
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSGetWindowBounds
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSGetSurfaceBounds
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSBindSurface: Invalid window 0xa150
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSBindSurface: Invalid window 0xa150
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSBindSurface: Invalid window 0xa150
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSGetWindowBounds
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSGetSurfaceBounds
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSBindSurface: Invalid window 0xa150
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSBindSurface: Invalid window 0xa150
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSBindSurface: Invalid window 0xa150
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSGetWindowBounds
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSGetSurfaceBounds
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSBindSurface: Invalid window 0xa150
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSBindSurface: Invalid window 0xa150
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSBindSurface: Invalid window 0xa150
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSGetWindowBounds
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSGetSurfaceBounds
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSBindSurface: Invalid window 0xa150
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSBindSurface: Invalid window 0xa150
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSBindSurface: Invalid window 0xa150
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSGetWindowBounds
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSGetSurfaceBounds
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSBindSurface: Invalid window 0xa150
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSBindSurface: Invalid window 0xa150
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSBindSurface: Invalid window 0xa150
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSGetWindowBounds
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSGetSurfaceBounds
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSBindSurface: Invalid window 0xa150
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSBindSurface: Invalid window 0xa150
Sat May 8 00:30:34 iMacAL.local testsdl[71586] <Error>: kCGErrorIllegalArgument: CGSBindSurface: Invalid window 0xa150
Rendering with x11 driver : Failed 1 out of 4 testcases!
Rendering with dummy driver : All tests successful (3)
SDL_Audio : All tests successful (1)
Tests run with SDL 1.3.0 revision 1095906
System is running Mac OS X and is little endian
The iPad (and iPhoneOS 3.2) introduce both a larger screen and the ability to
enumerate and configure external displays. This hooks up SDL's multi-display
API to this new functionality, and removes the hardcoded iPhone resolution.
Yann Leprince 2010-03-31 11:07:53 PDT
Please add a #serial line as below to sdl.m4 and increase the serial number
with each revision of this file. This allows using aclocal --install, thereby
enabling automatic updating of sdl.m4 in SDL-based packages that distribute it.
# serial 1
The complete documentation and rationale for #serial can be found in the
Automake documentation:
<http://www.gnu.org/software/automake/manual/html_node/Serials.html>.
Tatu Kilappa 2010-02-11 12:13:20 PST
When compiling with -Wconversion, gcc complains about a cast in SDL_endian.h
that might change the result as we are casting from an int into an Uint16. This
is of course not visible unless we are on a non-x86 platform where the
assembler is not available.
While it's not really an error, the warning is really annoying. To fix, change
SDL_endian.h line 87 to:
return(Uint16)((x<<8)|(x>>8));
Thank you.
* Switched event type to enum (int32)
* Switched polling by mask to polling by type range
* Added SDL_RegisterEvents() to allow dynamic user event registration
* Spread events out to allow inserting new related events without breaking binary compatibility
* Added padding to event structures so they're the same size regardless of 32-bit compiler structure packing settings
* Split SDL_HasEvent() to SDL_HasEvent() for a single event and SDL_HasEvents() for a range of events
* Added SDL_GetEventState() as a shortcut for SDL_EventState(X, SDL_QUERY)
* Added SDL_FlushEvent() and SDL_FlushEvents() to clear events from the event queue
Ozkan Sezer 2010-02-06 12:31:06 PST
Hi:
Here are some small fixes for compiling SDL against mingw-w64.
(see http://mingw-w64.sourceforge.net/ . Despite the name, it
supports both win32 and win64.)
src/audio/windx5/directx.h and src/video/windx5/directx.h (both
SDL-1.2 and SDL-1.3.) I get compilation errors about some union
not having a member named u1 and alike, because of other system
headers being included before this one and them already defining
DUMMYUNIONNAME and stuff. This header probably assumes that those
stuff are defined in windef.h, but mingw-w64 headers define them
in _mingw.h. Easily fixed by moving NONAMELESSUNION definition to
the top of the file.
src/thread/win32/SDL_systhread.c (both SDL-1.2 and SDL-1.3.) :
The __GNUC__ case for pfnSDL_CurrentBeginThread is 32-bit centric
because _beginthreadex returns uintptr_t, not unsigned long which
is 32 bits in win64. Changing the return type to uintptr_t fixes
it.
video/SDL_blit.h (and configure.in) (SDL-1.3-only) : MinGW-w64
uses msvcrt version of _aligned_malloc and _aligned_free and
they are defined in intrin.h (similar to VC). Adding proper
ifdefs fixes it. (Notes about macros to check: __MINGW32__ is
defined for both mingw.org and for mingw-w64 for both win32 and
win64, __MINGW64__ is only defined for _WIN64, so __MINGW64__
can't be used to detect mingw-w64: including _mingw.h and then
checking for __MINGW64_VERSION_MAJOR does the trick.)
SDL_win32video.h (SDL-1.3-only) : Tweaked the VINWER definition
and location in order to avoid multiple redefinition warnings.
Hope these are useful. Thanks.
Kalle Olavi Niemitalo 2010-02-28 09:15:50 PST
It seems the SDLK_LMETA and SDLK_RMETA constants have been removed from SDL
1.3. I grepped for them in the SDL source tree and these were the only hits:
./include/SDL_compat.h:230:#define SDLK_LSUPER SDLK_LMETA
./include/SDL_compat.h:231:#define SDLK_RSUPER SDLK_RMETA
./src/video/bwindow/SDL_BWin.h:194: keymap[0x66] = SDLK_LMETA;
./src/video/bwindow/SDL_BWin.h:195: keymap[0x67] = SDLK_RMETA;
I don't know how compatible SDL 1.3 is supposed to be with applications
designed for SDL 1.2. However, as you can see, SDL itself is still trying to
use the removed constants, and that is clearly a bug.
Because SDL_compat.h defines KMOD_LMETA as KMOD_LGUI, I suppose it should also
define SDLK_LMETA as SDLK_LGUI, and SDLK_RMETA likewise.
Reset actuators and set autocenter might not actually be supported on all
implementations. We'll just disable error checking since they aren't
critical to neither opening the joystick nor the haptic subsystem.
